Sasanian Archaeology: Settlements, Environment and Material Culture 9781803274188, 9781803274195, 1803274182

Table of contents :
Cover
Title Page
Copyright page
Contents Page
General Introduction
St John Simpson
PART ONE: Sites, Settlements and Landscapes
Part One: Sites, Settlements and Landscapes - Introduction
St John Simpson
Parthian and Sasanian Site and Settlement Patterns on the Deh Luran Plain, Khuzistan Province, South-West Iran
J.A. Neely
The Sasanian Colonisation of the Mughan Steppe, Ardebil Province, North-West Iran
Jason Ur1 and Karim Alizadeh2
The Army and Urbanism at the Sasanian Empire’s Northern Frontiers:Fieldwork on the Linear Barriers, Fortresses and Cities at the Margins of the Gorgan Plain (Results of a Joint Project Between the Iranian Cultural Heritage, Handicraft and Tourism Organisa
Hamid Omrani Rekavandi (HO), Eberhard W. Sauer (EWS), Tony J. Wilkinson (TW),† Ghorban Ali Abbasi (GA), Roger Ainslie (RA), Francesco Caputo (FC), Mohammad Ershadi (ME), Morteza Fattahi (MF), Nikolaus Galiatsatos (NG), Kristen Hopper (KH), Julian Jansen V
A Temple Among the Fields: Preliminary Reconstruction of an Agricultural Landscape in the Serakhs Oasis, Southern Turkmenistan
Barbara Kaim
Excavations at Kirpichli depe in Dehistan
A.E. Atagarryev
The First Season of Excavations at the Sasanian Site of Tole Qaleh Seyfabad (TQS), South-west Iran: Preliminary Report
Parsa Ghasemi,1 Reza Noruzi2 and Azizallah Rezaei3
Sasanian and Early Islamic Settlement Patterns North of the Persian Gulf
Ahmadali Asadi,1 Seyed Mehdi Mousavi Kouhpar,2 Javad Neyestani,2 Alireza Hojabri Nobari3
Bushehr, Dashtestan and Siraf: the Transformation of the Maritime Trade Network in the Middle Persian Gulf
Seth, M. N. Priestman
PART TWO: Environment and Resources
Charcoal and Wood in Fuel Residues in Sasanian and Early Islamic Merv: Wood Resources and Environmental Implications
Rowena Gale
Exploitation of Wood at Kush (Ras Al-Khaimah, UAE) in Sasanian Times: Results of the Charcoal Analysis
Margareta Tengberg
Two Sasanian Archaeofaunas from the Hamrin Basin, Eastern Iraq
F.G. Fedele
Fracture Patterns Amongst Sasanian Bones at Merv
Ian Smith
PART THREE: Material Culture
Part Three: Material Culture - Introduction
St John Simpson
Sasanian Pottery: Archaeological Evidence for Production, Circulation and Diachronic Change
St John Simpson
Imitations and Innovations in Kushano-Sasanian Pottery
V.A. Zavyalov†
Parthian, Sasanian and Islamic Pottery and Occupation of the Deh Luran Plain, South-west Iran
David V. Hill
Magic and Inter-Communal Relations in Sasanian Iraq: the Case of the Incantation Bowls
Ali G. Scotten
The Dating of Sasanian Glass in the Light of Excavated Finds from Bahrain
S.F. Andersen
Reconstruction of the Process of Cutting of Two Sasanian Glass Bowls in the Okayama Orient Museum and the British Museum
Tami Ishida
Sasanian Swords and Scabbards in the British Museum: A Scientific Analysis
J. Lang1 and St J. Simpson2
General Conclusion
St John Simpson
Bibliography
Index
Contents
General Introduction
St John Simpson
Part One: Sites, Settlements and Landscapes - Introduction
St John Simpson
St John Simpson
Parthian and Sasanian Site and Settlement Patterns
J.A. Neely
The Sasanian Colonisation of the Mughan Steppe
Jason Ur and Karim Alizadeh
The Army and Urbanism
Hamid Omrani Rekavandi
A Temple Among the Fields
Barbara Kaim
Excavations at Kirpichli depe in Dehistan
A.E. Atagarryev
The First Season of Excavations at the Sasanian Site
Parsa Ghasemi
Sasanian and Early Islamic Settlement Patterns
Ahmadali Asadi
Bushehr, Dashtestan and Siraf
Seth, M. N. Priestman
Part Two: Environment and Resources - Introduction
St John Simpson
St John Simpson
Charcoal and Wood in Fuel Residues
Rowena Gale
Exploitation of Wood at Kush
Margareta Tengberg
Two Sasanian Archaeofaunas from the Hamrin Basin
F.G. Fedele
Fracture Patterns Amongst Sasanian Bones at Merv
Ian Smith
St John Simpson
Part Three: Material Culture - Introduction
St John Simpson
Sasanian Pottery
St John Simpson
Imitations and Innovations in Kushano-Sasanian Pottery
V.A. Zavyalov
Parthian, Sasanian and Islamic Pottery and Occupation
David V. Hill
Magic and Inter-communal Relations in Sasanian Iraq
Ali G. Scotten
The Dating of Sasanian Glass
S.F. Andersen
Reconstruction of the Process of Cutting of Two Sasanian Glass Bowls
Tami Ishida
Sasanian Swords and Scabbards in the British Museum
J. Lang and St J. Simpson
General Conclusion
St John Simpson
Bibliography
Index

Citation preview

Sasanian Archaeology Settlements, Environment and Material Culture

Edited by

St John Simpson

Sasanian Archaeology Settlements, Environment and Material Culture

Edited by

St John Simpson

Archaeopress Archaeology

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

ISBN 978-1-80327-418-8 ISBN 978-1-80327-419-5 (e-Pdf) © the individual authors and Archaeopress 2022

Cover: Archaeological excavations and environmental sampling underway in Erk-Kala, the Sasanian citadel of ancient Merv, modern Turkmenistan (photograph: St J. Simpson, 1992). The first sherd of Sasanian pottery to be published is from Layard’s 19th century excavations on the mound of Kuyunjik at Nineveh (British Museum, N.1806)

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

Contents General Introduction..................................................................................................................................................................1 St John Simpson

PART ONE: Sites, Settlements and Landscapes Part One: Sites, Settlements and Landscapes - Introduction.........................................................................................21 St John Simpson Parthian and Sasanian Site and Settlement Patterns on the Deh Luran Plain, Khuzistan Province, South-West Iran..........................................................................................................................................................................47 J.A. Neely The Sasanian Colonisation of the Mughan Steppe, Ardebil Province, North-West Iran.........................................80 Jason Ur and Karim Alizadeh The Army and Urbanism at the Sasanian Empire’s Northern Frontiers:Fieldwork on the Linear Barriers, Fortresses and Cities at the Margins of the Gorgan Plain (Results of a Joint Project Between the Iranian Cultural Heritage, Handicraft and Tourism Organisation and the Universities of Edinburgh and Durham)...........................................................................................................................................................90 Hamid Omrani Rekavandi (HO), Eberhard W. Sauer (EWS), Tony J. Wilkinson (TW),† Ghorban Ali Abbasi (GA), Roger Ainslie (RA), Francesco Caputo (FC), Mohammad Ershadi (ME), Morteza Fattahi (MF), Nikolaus Galiatsatos (NG), Kristen Hopper (KH), Julian Jansen Van Rensburg (JJ), Eve Macdonald (EM), Majid Mahmoudi (MM), Amin Nazifi (AN), Jebrael Nokandeh (JN), Chris Oatley (CO), Seth Priestman (SP), James Ratcliffe (JR), Kourosh Roustaei (KR), Esmail Safari Tamak (EST), Bardia Shabani (BS) and Lucian Steven Usher-WilsoN (SU) A Temple Among the Fields: Preliminary Reconstruction of an Agricultural Landscape in the Serakhs Oasis, Southern Turkmenistan.............................................................................................................................................103 Barbara Kaim Excavations at Kirpichli depe in Dehistan.........................................................................................................................109 A.E. Atagarryev The First Season of Excavations at the Sasanian Site of Tole Qaleh Seyfabad (TQS), South-west Iran: Preliminary Report..................................................................................................................................................................127 Parsa Ghasemi, Reza Noruzi and Azizallah Rezaei Sasanian and Early Islamic Settlement Patterns North of the Persian Gulf............................................................141 Ahmadali Asadi, Seyed Mehdi Mousavi Kouhpar, Javad Neyestani and Alireza Hojabri Nobari Bushehr, Dashtestan and Siraf: the Transformation of the Maritime Trade Network in the Middle Persian Gulf...............................................................................................................................................................................153 Seth M. N. Priestman

PART TWO: Environment and Resources Part Two: Environment and Resources - Introduction������������������������������������������������������������������������������������� 178 St John Simpson Charcoal and Wood in Fuel Residues in Sasanian and Early Islamic Merv: Wood Resources and Environmental Implications.................................................................................................................................................192 Rowena Gale Exploitation of Wood at Kush (Ras Al-Khaimah, UAE) in Sasanian Times: Results of the Charcoal Analysis......212 Margareta Tengberg Two Sasanian Archaeofaunas from the Hamrin Basin, Eastern Iraq..........................................................................224 F.G. Fedele Fracture Patterns Amongst Sasanian Bones at Merv.....................................................................................................232 Ian Smith i

PART THREE: Material Culture Part Three: Material Culture - Introduction.....................................................................................................................239 St John Simpson Sasanian Pottery: Archaeological Evidence for Production, Circulation and Diachronic Change.....................255 St John Simpson Imitations and Innovations in Kushano-Sasanian Pottery...........................................................................................375 V.A. Zavyalov† Parthian, Sasanian and Islamic Pottery and Occupation of the Deh Luran Plain, South-west Iran...................389 David V. Hill Magic and Inter-Communal Relations in Sasanian Iraq: the Case of the Incantation Bowls...............................396 Ali G. Scotten The Dating of Sasanian Glass in the Light of Excavated Finds from Bahrain...........................................................420 S.F. Andersen Reconstruction of the Process of Cutting of Two Sasanian Glass Bowls in the Okayama Orient Museum and the British Museum.........................................................................................................................................................431 Tami Ishida Sasanian Swords and Scabbards in the British Museum: A Scientific Analysis.......................................................445 J. Lang and St J. Simpson General Conclusion..................................................................................................................................................................474 St John Simpson Bibliography..............................................................................................................................................................................477 Index............................................................................................................................................................................................523

ii

General Introduction St John Simpson Department of the Middle East, The British Museum, London, WC1B 3DG.

‘Our knowledge of the pre-Islamic past of Iran can be advanced only by new discoveries, and archaeology is far more important in this land with few written sources than in more favoured regions of the Mediterranean’. Frye 1976: 93.

The Sasanian empire was one of the great powers of Late Antiquity (Figure 1). Founded in AD 224 by Ardashir I (AD 224–242), an Iranian noble whose family came from the city of Istakhr in Fars province of southern Iran, he adopted the previous Parthian capital of Ctesiphon in southern Iraq as his own. He established a powerful family dynasty and an empire which thrived for four centuries. Political continuity was strengthened by closer integration of the administration of the provinces of Iran, Mesopotamia, and adjacent territories. Rapid military successes by his successor Shapur I (AD 242– 272) triggered crisis in the Roman empire, followed by drastic economic reforms and re-organisation of the Roman army to recreate a temporary status quo until a century later when Shapur II (AD 309–379)

C

AU

B LAC K S EA

aggressively expanded his empire westwards into upper Mesopotamia. The Sasanian threat remained constant and, under Khusrau II (AD 590–628), armies surged across the eastern Roman empire again but now holding onto, and consolidating, their gains for up to a generation. However, brilliant counter-offensives by Heraclius, political implosion in Ctesiphon, and increasingly bold Arab probes along the southern frontiers of both great powers finally led to the collapse of the Sasanian empire during the second quarter of the 7th century. Sasanian studies have been traditionally dominated by Iranologists, art historians, numismatists and/or scholars most familiar with Classical, Armenian, Jewish

CENTRAL ASIA C

AS

U

S

Gurgan wall

Merv

C ASPIAN

rT

igr

MEDITER R A NEA N S EA

Jerusalem

IRAN

is

Ri ve

r Eu phr

ate

s

Ctesiphon Veh Ardashir Susa Babylon

Hira

ARABIA

A F G H A N I S TA N

Masjid-i Suleiman Istakhr Naqsh-i Rustam Bishapur Qasr-i Abu Nasr PE RS Firuzabad I A Rishahr N GU

d

ve

Antioch

us

In

TURKEY

Bandiyan

er

Dara Nisibis R i

SEA Tabriz Takht-i Suleiman Rayy Nineveh Taq-i Bustan

Ri v

Constantinople

LF

RE D

N

SE

INDIAN OCEAN

A

0

300 miles

0

500 kilometres

Approximate extent of The Sasanian Empire

Figure 1. Map showing the extent of the Sasanian empire at its height (drawing: P. Goodhead)

Sasanian Archaeology (Archaeopress 2022): 1–19

INDIA

Sasanian Archaeology

Figure 2. Monumental Sasanian relief at Naqsh-i Rustam showing Hormizd II (photograph: author, 2000)

or later Arab sources. Academic research began as an antiquarian numismatic pursuit with correlations of coin portraits with well known rulers, and thence to create a typology which could be adapted to the identification of individuals on rock reliefs or silver vessels. In the absence of many surviving primary historical sources, numismatic, glyptic, art-historical and architectural studies have defined the field, and discussions of rock reliefs, stuccoes, silver and silks, cut glass, coins, glyptic and bullae have shaped opinions on the development and impact of Sasanian art and material culture.1

‘Silk Roads’ trade in luxury commodities and studies of Sasanian silver coinage have clouded appreciation of the importance of internal trade and everyday transactions. There is increasing interest in these topics, how the empire really functioned and the need for serious historiographic analyses has been reiterated by several scholars. This collection of essays builds on some of these questions and offers an approach which is based almost entirely on archaeological sources. These have attracted much less attention among non-archaeologists and syntheses of the data, either by region or category, are rare.3 The first attempts to identify archaeological data of this period began with antiquarian travellers passing through the region in the early 19th century. Several observed the eroded remains of the Gorgan wall and its interval forts, the large fortress at Gomish Tappeh and remains of old canals, but most concurred in the opinion that these were built by Alexander, referred to locally as Iskander.4 Many others drew attention to Sasanian rock reliefs at Bishapur, Naqsh-i Rustam, Naqsh-i Rajab and another, subsequently erased, at Rayy and speculated on their identifications (Figure 2). A distinctive class of decorated pottery typical of northern Iraq was first recognised as being Sasanian by Claudius James Rich when he found sherds of it on the surface of the important yet never re-investigated site of Eski Kifri

These studies have in turn created the picture of a rich courtly culture with a strong Iranian character moderated in part by Western artistic influences. However, the archaeological evidence ‘points to a considerable regional heterogeneity of material culture within the Sasanian empire, although the historical sources indicate a tendency towards political centralism right from the beginning of the state’.2 The corresponding consumable culture of the masses is much less understood, as is the economy which not only sustained and supported the Sasanian empire for four centuries but also underpinned its consistent military superiority over its western rivals and effective defence against successive powerful threats along its eastern frontier. Moreover, the glamorous concept of 1  Harper et al. 1978; Kröger 1982; Gyselen 1989; Overlaet (ed.) 1993; Whitehouse 2005; Demange (ed.) 2006; Harper 2006. 2  Huff 1987: 302.

3  4 

2

Huff 1987; Mousavi and Daryaee 2012. Sauer et al. 2013: 4–14, 630–47.

St John Simpson: General Introduction

Figure 3. The first sherd of Sasanian pottery to be published is from Layard’s 19th century excavations on the mound of Kuyunjik at Nineveh (British Museum, N.1806)

in 1820.5 Further fragments were recognised by almost all of the subsequent 19th century excavators of the Assyrian palaces on the mound of Kuyunjik at Nineveh and retained with other diagnostic sherds, glassware, metalwork and coins recovered from the extensive and deeply stratified late period occupation at the site (Figure 3).6 Throughout the 19th century, many British political agents, envoys and soldiers also commented on and/or made their own investigations of evidence for Zoroastrian funerary practices found in the form of reused jar or carved stone ossuaries across the Bushehr peninsula (Figure 4).7 During the 1920s and early 1930s scholars were presented with a steadily increasing body of archaeological data from this period, firstly from Iraq as excavations there revealed elaborate stuccoes and distinctive types of pottery and glass at Kish,8 Ctesiphon,9 Nineveh10 and Nuzi11 (Figure 5).

Figure 4. Re-used Sasanian torpedo jar containing excarnated human remains and found at Bushehr in the 19th century (British Museum, 91952)

year further stuccoes were found decorating a pillared building south of Damghan, investigated as part of the University Museum Philadelphia / Pennsylvania Museum of Art joint project at nearby Tepe Hissar (Figure 6),12 and during the extensive clearance of the late Sasanian citadel of a small town, possibly that of old Shiraz, at Qasr-i Abu Nasr (Figure 7).13 The longterm French excavations at Susa revealed an area of housing in Level IV which the excavator dated to the mid-4th century on the grounds that this equated to the

The abolition of the French monopoly in October 1927 and the ratification of a new Antiquities Law on 3rd November 1930 were followed by a scramble of archaeological activity across Iran. In the following Rich 1836: vol. I, 20; see Simpson 2013a. Simpson 1996a; 2005a; 2005b. 7  Simpson 2007a; Simpson and Molleson 2014; Simpson 2019c. 8  Langdon 1931a; 1931b; 1932; Langdon and Harden 1934; see Moorey 1978: 122–46. 9  Reuther 1929; Kühnel 1933; Schmidt 1934; see Kröger 1982. 10  Thompson and Mallowan 1933; see Simpson 1996a; 2005a. 11  Ehrich 1939; see Potts 1996. 5  6 

12  13 

3

Pope 1932; Kimball 1937. Frye (ed.) 1973; Whitcomb 1985.

Sasanian Archaeology

Figure 5. Sasanian stuccoes shortly after their discovery at Kish (after Langdon 1931b: fig. 3)

Figure 6. The remains of the columned hall of a small residence near Tepe Hissar with in situ stuccoes (after Pope 1932: fig. 6)

remains and cleared part of the palace complex at the eastern end, revealing mosaics and part of an enigmatic partly sunken structure commented on by many European travellers since 1809 and interpreted by the excavator as a fire-temple (Figure 8).16

occupation destroyed by Shapur II in AD 341, and the city went into decline thereafter until the late Sasanian period;14 however, many of the early reports of Sasanian structures, tombs and pottery are incorrectly dated and most of the recovered finds poorly stratified.15 Between 1935 and 1941, the French expedition also excavated at Bishapur where they created the first plans of the

During the late 1950s a trickle, rising to a flood, of Sasanian antiquities and modern imitations began to

Ghirshman 1952. e.g., de Mecquenem 1931; see Boucharlat 1987; Boucharlat and Haerinck 2011. 14  15 

16 

4

Ghirshman et al. 1956; 1971; see Frye 1976.

St John Simpson: General Introduction

Figure 7. View looking across the summit of the citadel of Qasr-i Abu Nasr with the bastion overlooking the entrance at the rear and modern Shiraz behind (photograph: author, 2000)

Figure 8. View of the so-called fire temple or temple of Anahita at Bishapur (photograph: author, 2000)

cut glass bowls and a much smaller number of closed forms, and most major public and private collections with Iranian interests – from America to Japan – soon acquired examples of these (Figure 10).18 The Gilan region of north-west Iran was said to be the source

enter the art market following a combination of largescale licensed commercial and illegal excavations in Iran. Most notably, these included large numbers of silverwares and a small number of rock crystal objects (Figure 9).17 There were also hundreds of Sasanian 17 

e.g., Grabar et al. 1967; Muscarella 2000: 203–205.

18 

5

e.g., Fukai 1973; 1977; Whitehouse 2005; Goldstein et al. 2005.

Sasanian Archaeology of these and earlier objects, although this is unlikely to be the only provenance and many silverwares are of dubious authenticity. A small number of helmets, long swords in highly decorated scabbards (Figure 11), and plain or fluted vessels cast from high-tin bronze also appeared on the market,19 and were reported to come from the Dailaman district of Gilan where one Iranian claimed to have seen children playing with ancient swords during his visit ‘to inspect my timber business’ in 1951.20 There were strong vested interests in the commercial value of such discoveries and those archaeologists who ventured into this region faced tough challenges. An Iranian archaeological expedition excavating at Marlik in 1961/62 not only witnessed illegal digging around the site before, during and after their work, but on one occasion had their tent slashed at night,21 and the Japanese expedition to Dailaman also regularly encountered systematic looting at sites through the region.22 However, both made discoveries that support reports that this region was one of the sources for Sasanian objects circulating on the art market, and several Sasanian dipper jugs found in the disturbed upper layers at Marlik reflect its proximity to a Sasanian fortress.23 In the meantime, the Japanese mission investigated reports of numerous cemeteries having been recently commercially excavated in the Halimehjan valley: the reported finds included iron swords, glassware, stamp seals, silver plates, gold ornaments and carnelian beads which are consistent with a Sasanian date,24 and a more recent evaluation of these looted sites confirms a large number date to the Partho-Sasanian period (Figure 12).25

Figure 9. Late Sasanian silver bowl, said to have been found in Iran (British Museum 133033)

Throughout the 1960s and 1970s, archaeological fieldwork projects not only had a research agenda but also served as part of a larger package of national promotion alongside the government’s ambitious redevelopment and restoration programmes. Largescale irrigation projects in south-west Iran were accompanied by extensive archaeological surveys, notably of Khuzestan in 1960/61,26 the Deh Luran plain in 1968/69,27 and the Susiana plain in 1973.28 Moreover, research into the Sasanian period was given a high public profile within Iran as it marked one of the high points, alongside that of the Achaemenids, in 19  e.g., Grancsay 1963; Overlaet 1982; 2004; Melikian-Chirvani 1974; Moorey 1976; Mahboubian 1997: 275, cats 354a–b; see Lang and Simpson, this volume. 20  Mahboubian 1997: 27. 21  Negahban 1996: vol. I, 5–11, vol. II, col. pl. VIIIA. 22  Fukai 1968: 4; Ohtsu, Nokandeh, Yamauchi and Adachi (eds) 2006: fig. 22. 23  Negahban 1996: vol. I, 227–28, vol. II, fig. 22, pls 106–107, nos 542– 46; see Simpson, this volume 316, fig. 63. 24  Fukai and Matsutani 1977; 1980. 25  Simpson 2015a. 26  Adams 1962a. 27  Neely 1970; 1974. 28  Wenke 1975; 1975/76.

Figure 10. Late Sasanian cut glass vase, said to have been found near Amlash (British Museum 135712)

6

St John Simpson: General Introduction at Firuzabad (Gur) he cleared part of the Sasanian palace (Figure 16), while at Bishapur a 250 m stretch of the 9 m thick fortifications and the foundations of a triumphal gateway were excavated (Figure 17).32 In 1957 a group from the Fars army concluded several years of work in the Mudan cave at Bishapur by resurrecting the fallen upper part of the carved statue of Shapur I on reinforced concrete legs, with fragments displayed nearby, enlarging the mouth of the cave enlarged and laying out paths (Figure 18). Between 1959 and 1978 the German Archaeological Institute conducted careful excavations and consolidation of the late Sasanian fire-temple sanctuary of Adhur-Gushnasp at Takht-i Sulaiman (Figure 19);33 following a preliminary survey of the Firuzabad plain in 1972,34 the same institute investigated Ardashir I’s palace at Qaleh-i Dukhtar near Firuzabad (Figure 20).35 Between 1968 and 1978, excavations at Haftavan tepe, south-west of Lake Urmia, revealed a number of Sasanian graves on the summit of the mound and cutting an earlier Sasanian phase of occupation when the mound had been crowned by a circular fortification wall with repeating horseshaped shaped towers projecting from the exterior (Figure 21).36 The appearance of this fort resembles a second discovered on the summit of the steep-sided main mound of Tureng tepe in the Gorgan plain during French excavations there between 1967 and 1975 (Figure 22).37 Meanwhile, between 1970 and 1977, the late M.Y. Kiani was exploring the sites on the Gorgan plain, including the remains of the Gorgan wall to the north: he proposed a Parthian date for this during the reign of Mithradates II (123–87 BC) based on a combination of brick sizes, pottery, and architectural parallels and a misconception over the height of the Caspian sealevel in antiquity, although he admitted the presence of Sasanian occupation and illustrated a late Sasanian faceted glass cylinder vessel from Trench F at Fort 12.38 The British Institute of Persian Studies supported a photographic survey of some of the major rock reliefs,39 and an eagerly anticipated Sasanian origin for the early Islamic Persian Gulf port of Siraf was matched by the discovery of a fortified complex beneath the massive early 9th century Congregational Mosque (Figure 23).40 The excavator, the late David Whitehouse, was quoted in The Iran Tribune (April 1970) as saying that ‘one might speculate that we have here the remains of a port which served that great Sassanian city of Firoozabad’ and further evidence for Sasanian involvement in maritime

Figure 11. Late Sasanian crucible steel sword with decorated silver scabbard, said to have been found in Gilan (British Museum 135739 / 135747)

the sequence of pre-Islamic Iranian monarchies.29 As a result, during the military parades held at Persepolis during the Shah’s controversial celebration of the 2,500th Anniversary of the Founding of the Persian Empire in October 1971, recreations of Achaemenid and Sasanian army uniforms constituted 38% of the total (Figure 13).30 The site below the Achaemenid royal tombs at Naqsh-i Rustam was partly levelled and formal paths laid in preparation for the associated events, revealing two minor Sasanian rock reliefs on the lower cliff face in the process (Figure 14),31 although construction of an access road cut through the massive Sasanian mudbrick fortifications enclosing the site (Figure 15). Ali Akbar Sarfaraz was among those archaeologists at this time who embraced the new nationalist ideology:

Sarfaraz 1969; 1970; Yasi 1977. Naumann and Naumann 1976; Naumann, Huff and Schneider 1975; Huff 2011a: 102–105. 34  Huff 1974. 35  Huff 1976; 1978. 36  Burney 1968; 1972. 37  Boucharlat and Lecomte 1987. 38  Kiani 1982a; 1982b. 39  Herrmann 1976. 40  Whitehouse 1971a; 2009. 32  33 

Abdi 2001: 67. Ministry of Information 1971. 31  Roaf 1974. 29  30 

7

Sasanian Archaeology

Figure 12. Map of archaeological sites located on the 2001–2006 Iran/Japan archaeological survey of the western Sefid Rud, with Partho-Sasanian sites highlighted as solid dots (after Ohtsu, Nokandeh, Yamauchi and Adachi (eds) 2006: fig. 22)

trade was argued by Andrew Williamson.41 He delivered a lecture on this subject at the British Academy on 21st June 1972, which was attended by Reza Shayegan, the managing director of the National Shipping Lines: the full text owas enthusiastically published shortly afterwards in The Times and Kayhan International. This promotion of the military and economic aspects of the port of Siraf was timely given the Shah’s obsession to develop Iran as a major sovereign military power and rival of Iraq, and play a larger role in Middle Eastern politics, develop more Persian Gulf ports and air-bases and safeguard oil exports through the Persian Gulf and Indian Ocean.42 41  42 

During these decades many bridges, weirs, and causeways across Iran were attributed a Sasanian date,43 ruined chahar taq structures interpreted as Zoroastrian fire temples,44 free-standing rock-cut monuments described as open-air Zoroastrian fire-altars,45 and a circular moated rampart at Tal-i Khandagh near Sar Mashhad interpreted as an exposure dakhma founded by the powerful Zoroastrian priest Kirdir in the 3rd century.46 Alongside these reports, the remains of a large Nestorian monastery on Kharg island was extensively excavated by the French expedition and a preliminary 202, 252–53, 261, 362. 43  e.g., Stein 1940: 15–16, 48, 71–74, 171–74. 44  e.g., Vanden Berghe 1961. 45  e.g., Stronach 1966; Trümpelmann 1973. 46  Trümpelmann 1984: 317–18.

Williamson 1972; Whitehouse and Williamson 1973. Simpson 2019a; cf. Alam 1991: 103–104, 153, 155–56, 176, 180, 200–

8

St John Simpson: General Introduction

Figure 14. View of a small Sasanian relief discovered at Naqsh-i Rustam (photograph: author, 2000)

Figure 13. Costume for a Sasanian commander in chief, designed for the 2,500th Anniversary of the Founding of the Persian Empire parades held at Persepolis in October 1971 (Ministry of Information 1971)

Figure 15. View of the unexcavated Sasanian fortifications around the complex at Naqsh-i Rustam (photograph: author, 2000)

9

Sasanian Archaeology

Figure 16. View of the Sasanian palace at Firuzabad/Gur (photograph: author, 2000)

report published which interpreted it as evidence for a 6th century Christian presence.47 This dating has been regularly repeated in more recent years, as further discoveries of churches and monasteries along the Persian Gulf and in the Iraqi Western Desert have been cited as evidence for the rise of the Nestorian Church during the late Sasanian period.48 However, despite these numerous discoveries and reports, the literature is filled with assumptions and unreliable or circular dating. Frye criticised the situation in typical style in the opening words of a paper given at a conference of Iranologists in Munich: ‘During eight years of intermittent residence in Shiraz, I have made many trips to sites and monuments of Fars province. From the beginning of my acquaintance with many structures, I have raised the question: just how do isolated monuments receive a designation as Achaemenid, Sasanian or Islamic when there are no written sources which tell us what the remains were and their dates? Sometimes the identification and dating seemed to have been passed on throughout the years, decades or even centuries, from a chance remark of a traveller rather than from a detailed study on the premises. It is not my intention here to answer any questions regarding architecture, or such technicalities which Ghirshman 1960; Steve 2003. Bowman 1974; Gropp 1991; Lerner 1992; Hellyer 2001; Elders 2003; Hauser 2007a; Horn and Hunter 2012. 47 

Figure 17. View of the exterior of the excavated Sasanian fortifications at Bishapur (photograph: author, 2000)

48 

10

St John Simpson: General Introduction

Figure 18. Real postcard photographic print showing visitors admiring the restored statue of Shapur I inside the Mudan cave (EPH-ME 8154, presented by Dr G. Herrmann)

Figure 19. View of the lake and the late Sasanian fire-temple sanctuary of Adhur-Gushnasp at Takht-i Sulaiman (photograph: author, 2004)

11

Sasanian Archaeology

Figure 20. General view of Ardashir I’s fort at Qaleh-i Dukhtar near Firuzubad (photograph: Paul Gotch, 1966)

Figure 21. Plan of the excavated section of a Sasanian fort on the summit of Haftavan tepe in north-west Iran (after Burney 1973: fig. 10)

12

St John Simpson: General Introduction

Figure 22. View of the main mound crowned by the Sasanian fort at Tureng tepe in north-east Iran (photograph: author, 2004)

Figure 23. Plan of the so-called Sasanian fort and enclosure at Siraf (after Whitehouse 1972: fig. 3)

13

Sasanian Archaeology

Figure 24. View of the exterior of the so-called Sasanian palace at Sarvistan (photograph: author, 2000)

are the province of the archaeologist, for I am quite unqualified to even use a sextant in the field much less to analyze squinches or arches. Rather I wish to raise questions which I believe need answer and perhaps thus stimulate an interest in the historical remains of Fars province, the homeland of the Achaemenids and the Sasanians. In many cases, how can we distinguish between the original construction of a building or a dam, repairs made later, or extensions or radical changes in the plan or execution of the structure? One observation I have made which has impressed me over the years is the continuity in methods and types of construction in Iran throughout the centuries. Such statements that smooth and fine stone work on an isolated, ruined structure, must be Achaemenid because the Sasanians or the Iranians of Islamic times never did, or could not execute, comparable works, in my opinion, should be discarded’.49 The small vaulted building at Sarvistan is a good instance of this (Figure 24). This structure was first recorded by European travellers in the mid-19th century, and as early as 1910 was interpreted as the remains of a minor Sasanian palace on the basis of a literary reference by al-Tabari (d. AD 923) to the construction of a palace near Fīrūzabad by Mihr Narseh, the vizier of Varahran V (AD 421–438). A number of later scholars have questioned this interpretation and dating, including Bier who argued for a considerably later – mid-8th to mid-10th 49 

Figure 25. Aerial view of the same building taken in 1936 (after Schmidt 1940: pl. 21)

Frye 1979: 335–37.

14

St John Simpson: General Introduction

Figure 26. Aerial view of the Taq-i Kisra (British Museum/Middle East archives)

been residential during the reign of Kavad (AD 488–496, 499–531), Khusrau I (AD 531–579) constructed a palace here with the assistance of ‘Roman expertise’ and ‘Greek marble, building experts, and craftsmen skilled in ceilings’ which were supplied by Justinian,55 and another description refers to Khusrau commissioning a palace mural commemorating his sack of Antioch in AD 540.56 A royal aviary, game reserve, stud, treasury and church are also attested in written sources, a bath-house has been partially excavated a short distance due west of the Taq-i Kisra, and the foundations and platform of a large building originally decorated with stuccoes located immediately to the south at the spot known today as Tell ad-Dhabai.57 Pottery and coins found in the vicinity of the iwan suggest a late Sasanian date of construction but piecemeal planning and construction of the surrounding complex was suggested by the early excavators and may explain some of the discrepancies between later Arab authors.

century – date on the basis of architectural features, and suggested that it may have been a Zoroastrian fire sanctuary rather than a palace.50 Stein commented on how the building ‘now rises in splendid isolation, but the amount of broken pottery, much of it glazed green ware, strewn around it over a considerable area shows that habitations of a humbler sort must have stood near by’.51 A number of walls in this area are visible on aerial photographs taken by Schmidt in March 1936, leading Whitcomb to reiterate that it was part of an urban complex rather than an isolated monument (Figure 25).52 These outlying remains have since been largely destroyed by agricultural works, but soundings excavated around the main building in 2002 suggest a possibly late Sasanian foundation although the most significant phase of occupation was indeed in the early Islamic period.53 The Taq-i Kisra near Ctesiphon provides another good example (Figure 26). This monument corresponds to the ‘great iwan’ of Arab Conquest historians such as Yaqubi which he specifies as being in the royal city of Aspanabr (or Aspanbor) and which lay the equivalent of a mile [1.6 km] south of Ctesiphon.54 This area is said to have

Within highland Iran, the ‘fire-bowls’ have now been proven to simply be the remains of astodans which have lost their lids,58 many of the chahar taq structures demonstrated to be much later imamzadehs,59 and

Bier 1986. Stein 1936: 179. 52  Whitcomb and Sumner 1999: 215–16. 53  Askari Chāverdi 2010; 2012. 54  el Ali 1968/69: 426.

Theophylact Simocatta History, V.6.10. Shahid 1995: 235–36. 57  Kröger 1982: 18–37, pls 3–5. 58  Huff 1998; 2004. 59  e.g., Huff 1974: 157.

50 

55 

51 

56 

15

Sasanian Archaeology

Figure 27. View of the excavated early medieval monastery at Sir Bani Yas (photograph: Emma Thompson, 2009)

the site of Tal-i Khandagh (described above) is more likely to be a fortified site.60 The historically attested importance of the Persian Church is unchallenged but its architectural footprint is now nebulous as the monasteries on Kharg island and Sir Bani Yas, as well as the parochial churches at al-Qusur on Failaka and Akkaz island, have been convincingly re-dated to between the late 7th and early 9th centuries.61 Moreover, they have been shown to be part of a wider development of coenibitic monasticism and proselytisation on the desert margins of Arabia following a period when evidence for permanent settlement is remarkably scarce (Figure 27).62

The shift to the west in the political power-base is also reflected in the increased number of building projects in western and north-west Iran. In 1964, the Italian archaeological expedition to Iraq began investigations in the Seleucia/Ctesiphon area. Directed first by Gullini, then by Invernizzi, these were the first excavations to be conducted at the political centre of the Sasanian empire since 1931. Among their important results was clarification of the urban topography, confirmation that the ‘round city’ next to Seleucia was the remains of Veh Ardashir rather than Ctesiphon, and the exposure of large areas of vernacular architecture and workshops within the curving fortifications (Figure 28).64

Beyond the modern Iranian borders other important discoveries were being made during this period. Extensive archaeological surface surveys begun in Iraq in 1956 demonstrated large-scale Sasanian water engineering works across the Mesopotamian alluvial plains, leading Adams to conclude that this was ‘the apogee of ancient developments on the central Euphrates floodplain’.63 One implication of these results was that the Sasanian state had invested sufficiently heavily in the region to indicate that they regarded it as an integral part of its empire and, although the ceremonial heart of the dynasty lay on the plateau, Mesopotamia was a major economic and cultural centre.

Archaeology abroad is always at the mercy of foreign politics. The Islamic Revolution in Iran in 1979 was followed by the long Iran–Iraq war. Although local Iranian archaeologists persevered under difficult conditions, all foreign archaeological projects in that country stopped for many years. Many of these archaeologists moved across the Persian Gulf to work on earlier periods in eastern Arabia, particularly the United Arab Emirates, where as early as 1972 Beatrice de Cardi had published evidence for a small Sasanian settlement on an islet near the tip of the Musandam peninsula which she interpreted as the remains of a ‘a military outpost to watch movements of shipping in the

Ghasemi 2012a. Kennet 2007; Carter 2008; Simpson 2018a. 62  Payne 2011. 63  Adams 1981: 200. 60  61 

Gullini 1966; Invernizzi 1976; Gullini (ed.) 1985; Simpson 1997a; 2015b.

64 

16

St John Simpson: General Introduction

Figure 28. Plan showing the relationship of some of the sites forming the Ctesiphon conurbation (reconstruction by the author)

narrow Strait of Hormuz’.65 Ironically, the site has been levelled since as part of construction associated with a small naval station.

for glass production, and the range of cut, moulded and plain glass vessels illustrated from excavations at sites in Azerbaijan, eastern Georgia and Armenia suggest that this was a major centre for Sasanian glass in parallel with Mesopotamia.67 It is no coincidence that both regions border the eastern Roman empire and cross-cultural influences are therefore likely to have stimulated fashions and industries in both regions. All of the Sasanian finds from the Caucasus have been found at sites within the territory of the kingdoms of Iberia and Albania which became satellite kingdoms in

Although less well appreciated in the context of Sasanian studies, important discoveries have been made in the last few decades in the southern Caucasus and Central Asia. Within the former region, Sasanian silverwares have been found at Shemakha in Azerbaijan, and Mukuzani and Mtskheta in eastern Georgia.66 Excavations in the old city of Dvin have apparently produced evidence 65  66 

de Cardi 1972: 308; 1975: 30–34. Khalilov 1985; Soltes (ed.) 1999: 201, 204–205, cats 98, 104.

67 

17

Simpson 2015a.

Sasanian Archaeology

Figure 29. Oblique aerial view of the ancient city-site of Merv: the Sasanian and earlier city is on the right and its medieval successor is on the left (Google Earth, 2009)

the 3rd century and were finally absorbed as provinces of the Sasanian empire in the 5th and 6th centuries. This region offers a much more likely place of production for the complete glass vessels reportedly found in graves in the Gilan region of north-west Iran for this was, to adapt Adams’ catchphrase for the Diyala, ‘the land behind the Caucasus’ in this period, as it was before.68

excavations carried out in parts of the city-site of Merv led to significant discoveries of monumental and domestic architecture, most notably a Buddhist monastery founded in the 4th century (Figure 29).69 More recent analyses of the previously excavated coins have led to many re-attributions, entailing a redating of the associated stratigraphy in many cases,70 and the employment of more careful excavation and the recovery of environmental remains for the first time have added considerable new information about the development of the city and the surrounding oasis during this period.71

On Iran’s interface with Central Asia, evidence for Sasanian activity north of the Kopet Dagh was regularly encountered by Soviet archaeologists during their surveys and excavations in southern Turkmenistan. The strongly Marxist-influenced ideology resulted in this period being interpreted as a feudal society following the so-called ‘period of slave-owning society’ which was equated with the Parthian period. The frequent occurrence of low-denomination coins was used to help secure the dating and important

The interpretation of archaeological data is as subject to the paradigms of the day as the analysis of historical sources but, as this short review attempts to show, this alternative set of evidence offers an important avenue Pugachenkova and Usmanova 1995. Loginov and Nikitin 1993. 71  Simpson 2008a; 2008b; 2014b. 69 

68 

70 

Simpson 2014a; 2015a; Shikaku 2013: 357; 2019.

18

St John Simpson: General Introduction day Turkmenistan), Gorgan (Iran) and Kush (United Arab Emirates), also offer new avenues for exploring survey and excavation data for this period.76

for examining the achievements and development of Sasanian society, its economy and its material culture. Mapping these in detail against historical developments remains a great challenge, and both greater control of the dating and more cautious interpretation of how the archaeological evidence reflects specific historical events, as opposed to cultural-economic developments, is required in future. The need for syntheses of the data, as well as more critical editions, has been called for by historians and archaeologists alike in recent years.72

The approaches taken by the contributors to the present work vary necessarily according to the data they are concerned with, but in order to place them within a wider research framework, this monograph is divided into three parts in order to address some of the key archaeological issues. It begins with a selection of essays dealing with sites and landscapes in different regions of the Sasanian empire. This is followed by two further sections, one dealing with aspects of the agricultural economy and the other on how studies of surviving material culture help inform our ideas of craft and industry. Each section is preceded by a short introduction outlining some of the most important relevant textual and archaeological background to these broad topics. There are many areas which could be included but which have not. This is not an attempt to create a systematic overview, let alone the ‘inventory of all archaeological sites [which] should be established as a first measure toward the critical assessment of the archaeological data’ that one writer has proposed.77 It is intended instead to explore some fruitful avenues of academic discussion and includes research by different international scholars, some well established and others at early stages of their careers. It should go without saying that some statements and conclusions may prove to be wrong, but without publication and discussion there is intellectual sterility.

The development of the present volume has had a gradual evolution, rather than a sudden birth, and has taken much longer to publish than intended, but the publication still seems to be the right thing to do at the present time. It was inspired by the results of several specialist workshops held at the British Museum in connection with post-excavation analyses arising from the author’s excavations at Merv, as well as related research on the Sasanian collections, and was developed to include the present set of contributions. The editor is very grateful both to the contributors for submitting their papers and for their huge patience while this volume was assembled, and to the reviewers for their comments on the contents. This publication follows several international exhibition on Sasanian luxury arts,73 two published conferences devoted to Sasanian history and archaeology,74 and the first formal catalogues of Sasanian glass in public or private collections.75 The results of major research projects at Merv (present-

In memoriam: V.A. Zavyalov You were one of the finest archaeologists to work in Central Asia, and one of my closest and longest friends and colleagues, with almost 30 years of work together in Central Asia, Russia and England. I value your experience, scholarship, humour and company to the end, remember that everything must be logical, and dedicate this work to your memory.

e.g., Shayegan 2003; Morony 2008; Bakhos and Shayegan 2010; Mousavi and Daryaee 2012. 73  Demange (ed.) 2006. 74  Sarkhosh Curtis and Stewart (eds) 2008; Kennet and Luft (eds) 2008. 75  Whitehouse 2005; Goldstein et al. 2005. 72 

76  77 

19

Sauer et al. 2013; Simpson et al. forthcoming a; forthcoming b. Shayegan 2003: 368.

PART ONE:

Sites, Settlements and Landscapes St John Simpson Department of the Middle East, The British Museum, London, WC1B 3DG.

‘Population may have peaked on the Susiana Plain during the Parthian Period, but it was imperial Sasanian investment that radically transformed settlement and subsistence in the area. Kilometers of qanats, large new canals, and massive stone weirs were constructed to ensure and expand cultivation and to supply new urban centers. Much design work on the hydrological system was apparently done by captured Roman engineers. Extensive agricultural investment was also made in the Deh Luran area where population reached unprecedented levels’. Johnson 1987: 291.

Sasanian Archaeology (Archaeopress 2022): 20–46

St John Simpson: Part One: Sites, Settlements and Landscapes - Introduction Introduction At its greatest extent, the Sasanian empire extended from the plains of northern Syria to the central Asian steppe, and from the mountains of the Caucasus to the shores of the Indian Ocean. Intimate knowledge of these hugely varied landscapes helped shape a range of responses in terms of settlement, economy, and military defence. Extended travel times from one end of the empire to another must have prevented direct meddling of the state in everyday affairs but equally must have enabled administrators, whether civilian or military, to take efficient control of the regions and resources under their purview. The postings of these men, and the sharing of technical expertise and support by an efficient bureaucracy, must have been critical factors in the development and maintenance of the empire. The scale and quality of archaeological evidence which might be used to test this varies greatly. Most of the evidence relating to urbanisation and rural settlement patterns comes from Mesopotamia and south-west Iran, and this therefore forms the basis of the following section. Urban landscapes ‘Shapur left the territories of the Romans, taking with him prisoners whom he settled in the countries of Iraq, Susiana, Persia and in the towns his father had founded. He also founded three towns and gave them names derived from his own name. One was in the land of Maysan and was called Sod Shapur [Sadh-Shapur]. The second in Persia which is still called Shapur today [i.e., Bishapur]. He rebuilt Gundeshapur which had fallen into disrepair and named it Anti-Shapur [Antioch-Shapur], a word half-Greek and half-Persian, meaning: ‘You are Shapur’s equal’. He constructed a third town on the banks of the Tigris called Marw Habor which is ‘Akbora and its environs [Buzurg Sabur]. These towns were populated by his prisoners who were provided with lands and home to till’.1

Figure 1. View of the main structures in the centre of the city of Ardashir-Khurrah (photograph: author, 2000)

with a circular inner area which included a high tower (tarbal), chahar taq (Takht-i Neshin), and palace, and an outer area divided by radiating streets into 20 sectors.2 Brief excavations by the Cultural Heritage Organisation of Iran in 2005 exposed wall and floor-paintings near the tarbal although few details have yet been published.3 The city had four main gates and was fortified with high ramparts and a water-filled ditch fed by a canal (Figure 1), and the extension of the axes of the streets as radiating tracks extending across the surrounding plain underlines the symbiotic, as well as functional, relationship between the city and its hinterland.4

This statement in the medieval Nestorian Chronicle of Se’ert is one of several sources describing such effects of Sasanian military campaigns in the eastern Roman empire. They underline the fact that boosting the urban economy was central to Sasanian economic planning, and this is well illustrated by the archaeological evidence.

After his defeat of Artabanus V and adoption of the Parthian capital at Ctesiphon as his own, Ardashir founded another – more or less circular – city on the opposite bank of the river Tigris and next to the old Seleucid capital of Seleucia ad Tigrim. This new foundation was laid out around and on top of a small older settlement known in Aramic as Coche and given

In about AD 220, Ardashir Papakan, founder of the Sasanian dynasty, founded a circular city in his homeland of the Firuzabad plain and named it ArdashirKhurrah (‘divine glory of Ardashir’). This was carefully laid out as a perfect circle measuring 1.85 km across, 1 

Huff 1972; 1974: 157–59. Mousavi and Daryaee 2012: 1080. 4  Huff 1974: 159. 2  3 

Lieu 1986: 478.

21

Sasanian Archaeology the new name of Veh Ardashir (‘City of Ardashir’), later Arabicised as Bahusīr.5 The old core presumably held at least some of the major public buildings, although the limited excavations at Tell Baruda only revealed vernacular architecture.6 The city covered an area of about 700 hectares and was enclosed within a curving mudbrick fortification wall with regular repeating towers along the exterior (Figure 2). The Italian excavations in the southern portion suggest an initial phase characterised by wells, drains, and bread ovens, followed by a progressive increase in settled density from the 4th to the latter half of the 5th century.7 This change is marked by increasing subdivision of private properties and a gradual encroachment onto the adjacent streets. This sequence suggests the initial maintenance of well-prescribed property boundaries in an initial phase of comparatively low population density, followed by necessity to adapt with the inability to build elsewhere. The two excavated residential blocks are thought to largely consist of business shops and houses in the southern part (initially termed Area 1) and workshops and private houses in the northern part (Area 2). The business shops were typically situated along the street frontages. The mixture of functions within a given quarter resembles the traditional concept of a Middle Eastern suq, rather than a classical definition of segregated functions in different insulae, but marks another stage in a long-term development noted within late Parthian levels at Seleucia and also at Dura. A third, almost circular, city exists at Darabgird in Fars although the walls may date to the 8th century as the mid-10th century author Hamza al-Isfahani states that the city was originally triangular but was enclosed within the present walls by the Umayyad governor of Fars, Hajjaj ibn Yusuf;8 the surface pottery suggests continuity of occupation as late as the 12th century (Figure 3).9 These circular (or almost circular) cities were the exception rather than the norm and they resemble glorified versions of local market centres with a convergence of tracks from all directions seen, for instance, on aerial photographs of Hatra.10 Although a circular layout has military advantages, and is more economic in the use of building materials, it creates challenges for fitting the architecture within and it is unsurprising that most other cities were laid out on rectilinear plans. Some of these were continuations of much older centres although their occasional renaming was probably accompanied by infrastructure schemes.

Figure 2. Plan of the excavated portions of the southern quarters of Veh Ardashir (after Simpson 2015b: 10)

Ctesiphon had been founded as a Parthian capital on the left bank of the Tigris, opposite Seleucia, and, by the Sasanian period at least, surrounded by fired brick walls and a moat.11 Located a km north of the late Sasanian royal city of Aspanabr, the site consists of a large mounded area known today as al-Ma’aridh or Tulul Bawi and straddling the mouth of an old canal off-take of the Diyala (Figure 4). This confirms a representation on the arch of Septimius Severus in Rome which indicates that it was divided in two by a canal.12 Most of the site is now covered by housing and the only archaeological investigations that have taken place there were carried out in 1928/29 and 1931/32,

el Ali 1968/69: 433–34. Venco 1973/74; 1977. 7  Cavallero 1966; Negro Ponzi 1966; Venco 1968/69; Venco and Negro Ponzi 1985. 8  Huff 1996. 9  Morgan 2003: 333. 10  Bradford 1957: pl. 24. 5  6 

11  12 

22

Gregory of Nazianzus Or., 5.10 P.G. 36.676 B/C. Gullini 1966: 33.

St John Simpson: Part One: Sites, Settlements and Landscapes - Introduction

Figure 3. Aerial view of the round city at Dārābgird (Google Earth, 2010)

river Tigris.17 It was formally named Veh-as-AntiokKhusrau (‘City Better than Antioch [has] Khusrau [built this]’) and is said to have had public baths and a hippodrome; the street plan was modelled on that of Antioch and freedom of worship and burial was given to the Christian inhabitants. Known colloquially as Rumagan (‘Town of the Greeks’) and later Arabicised as al-Rumiyya, it probably corresponds with the site still known as Bustan Kisra (‘Gardens of Khusrau’) which is partly destroyed by the Tigris but the surviving portion still surrounded by the eroded remains of high rectilinear mudbrick walls.18

revealing a series of large late Sasanian residences near the southern edge.13 Merv was another old city which was occupied throughout the Sasanian period and the essentially rectangular layout of its Seleucid foundation remained unchanged as the positions of the main gates effectively fossilised the orientation of the major arteries (Figure 5).14 In AD 266, Shapur I (AD 242–272) founded a new city in Fars and named it Bay-Shapur (‘Excellent [or] Beautiful [city] of Shapur’). Better known today as Bishapur, it measures 1500 x 1000 m across, bounded by fortifications which survive on three sides (the fourth was destroyed by the Shapur river),15 and apparently divided by major streets into 200 m square blocks with smaller intervening streets (Figure 6).16 After his sack of Antioch in AD 540, Khusrau I founded another city which was said by al-Dinawari to be located one parasang [4–5 km] south of Ctesiphon and implied by al-Tabari to be on the same, i.e. eastern, side of the

Shapur I also founded another city in Susiana called Jundi Shapur (‘Military camp of Shapur’). This measured 3.1 km in length, 1.5 km across, is bounded by what el Ali 1968/69: 431. Reuther 1938: 574; Canepa (2018: 67) was influenced by the biased account of John of Ephesus to say that the fortifications, which he misleadingly describes as ‘an earth wall’, were ‘intended to keep the city’s thirty thousand residents interned rather than to defend it’. He also regards it as having been a labour camp akin ‘to a human zoo’ (Canepa 2018: 68). The alternative site identifications made by Adams (1965: 177) and Fiey (1967) are either too far removed or are confused with a second site with the same name.

17  18 

Kröger 1982: 80–136, figs 41, 47, 49. Simpson 2008a: 71. 15  contra Talbot Rice 1935. 16  See Salles 1939/42: 93, pl. XVII; Amiri and Genito et al. 2013. 13  14 

23

Sasanian Archaeology

Figure 4. Aerial view of the site of Ctesiphon (Google Earth, 2016)

Figure 5. View of the ancient city-site of Merv with the citadel of Erk-Kala and the north wall on the right (photograph: author, 1992)

24

St John Simpson: Part One: Sites, Settlements and Landscapes - Introduction

Figure 6. Aerial view of the city-site of Bishapur (Google Earth, 2010)

structures, and the whereabouts and layout of Sasanian structures could not be determined from the surface as these lay below plain level. The second concerned its water supply. Soundings in the bed of the Siah Mansour, a watercourse immediately west of the city, revealed a row of ashlar masonry piers which must have originally supported a bridge, whereas an adjacent tunnel was part of an elaborate inverted siphon which delivered water from a canal off-take from the Dez river, passing under the Siah Mansour and into the city (Figures 10–11).23 The authors suggested that ‘the large grid traceable within the city on the air photographs consist not so much of a rectangular arrangement of streets as a network of such tunnels’.24 This feature explains how the engineers were able to deliver large and regular amounts of fresh water to the population, and these may have been standard features of Sasanian urban planning as they recur as features within forts and the city-site of Dasht Qal‘eh in the Gorgan plain.25 Despite damage inflicted during the Iran–Iraq war

is now a low gravel bank on three sides and contains low scattered mounds within (Figures 7–8). The site was described by Ghirshman as having ‘practically disappeared beneath the plough, but the plan was based on a great rectangle, curiously reminiscent of a Roman military camp. The same arrangement was noticed by the writer at Iwan-i Karkha near Susa, where Shapur II is said to have settled his western prisoners’ (Figure 9).19 Ghirshman published a vertical air photograph of the latter city,20 carried out a single season of excavations in 1950 and revealed parts of two monumental buildings, one with remains of figural wall-paintings, although few details were published and their precise date unclear.21 In 1963 a topographic survey was carried out at Jundi Shapur and four small soundings excavated, but the results were regarded as ‘uniformly unpromising’.22 However, two important observations were made. One was that the prominent mounds in the central third of the walled area were composed of re-deposited fill to support early Islamic Ghirshman 1978: 320. Ghirshman 1978: 179. 21  Ghirshman 1952: 11–12. 22  Adams and Hansen 1972: 301. 19 

See Sauer et al. 2013: fig. 12: 13. Adams and Hansen 1972: 301. 25  Sauer et al. 2013: 235–36, 312–18, 387.

20 

23  24 

25

Sasanian Archaeology

Figure 7. Aerial view of the city-site at Jundi Shapur (Google Earth, 2013)

Figure 8. Schematic plan of the remains at Jundi Shapur (after Adams and Hansen 1968: fig. 1)

26

St John Simpson: Part One: Sites, Settlements and Landscapes - Introduction

Figure 9. Aerial view of the city-site at Iwan-i Karkheh (Google Earth, 2012)

al-Qusair being either elements in an outer line of defense or posts designed to protect routes coming from the south’.26

and through agriculture and construction, these two city-sites offer great potential in understanding the development of major Sasanian cities.

Although apparently still not surveyed, the site does indeed appear to be another example of this form of Sasanian urban planning (Figure 12), and many more surprises are likely to lie in store.

These are not the only two such cities of this distinctive plan. Following his survey of the Eridu–Ur region of southern Iraq in 1966, Henry Wright commented that: ‘Visible on the air photographs, which I did not see until after I returned to Baghdad, about eight kilometres west of the survey border is a roughly rectangular walled town, covering 55 hectares and divided into three parts. It is similar in plan to Jundi Shahpur and Iwan-i Karkheh in south-western Iran (Adams 1961, figs. 5, 7) and is probably a Parthian or Sasanian center. Perhaps when the present levee of the Euphrates is examined in detail we will find that EP-65 is one of a series of fortified settlements linked to this larger town, with installations such as

Other Sasanian urban centres are attested archaeologically from Iran but almost all remain to be investigated archaeologically. On the Damghan plain, the extensive but heavily water-eroded ruins of the early Parthian capital of Hecatompylos at Shahr-i Qumis continued to be occupied throughout the Sasanian period according to Trinkaus.27 This 26  27 

27

Wright 1981: 335. Trinkaus 1986: 30.

Sasanian Archaeology

Figure 10. Plan of the inverted siphon supplying fresh water to the city of Jundi Shapur (after Adams and Hansen 1968: fig. 2)

supports Hansman’s earlier conclusion that ‘a recent and detailed study ... of aerial photographs covering all the Damghan plain produced no other likely site for Sasanian Kōmish’.28 The fact that Parthian, Sasanian and medieval pottery is concentrated in different areas suggests that the centre of occupation here shifted occasionally, explaining the lack of deep stratigraphy and its currently deflated condition.29 Observation of satellite imagery suggests that the site is laid out on a grid orientated north-west/south-east, there are substantial unexcavated architectural remains in the north-west part, and much of the gully erosion has a very regular pattern suggesting that it is following the lines of least resistance offered by the main streets. During the same period, a wall appears to have been constructed around the town of Damghan itself, located

some 30 km to the north-east.30 On the Gorgan plain, the city-site of Dasht Qal‘eh covers an area of about 3 km2, therefore exceeding that of Bishapur or Gur, and geophysical survey, followed by test excavation, proved that at least the eastern portion was divided by avenues of brick piers.31 The unexcavated ruins of an important Sasanian site, ‘possibly the position of a considerable military camp’, have been noted near the medieval site of Ghubayra, in the Bardsir valley some 70 km south of Kirman, and perhaps this was the source for the Pahlavi-inscribed pottery jar noted by the excavators of Ghubayra itself.32 Likewise, on the Marv Dasht plain, graves, a pottery kiln, and a stray personal seal found during excavations of the much earlier site of Tal-i Malyan may have been associated with the nearby Adle 1993. Sauer et al. 2013: 382–422. 32  Bivar et al. 2000: 3, 177, pl. 431a. 30 

Hansman 1968: 138–39. 29  See Gye et al. 1969: 34–35; Hansman and Stronach 1970a: 34. 28 

31 

28

St John Simpson: Part One: Sites, Settlements and Landscapes - Introduction

Figure 11. View of the inverted siphon and the piers of the overlying bridge at Jundi Shapur (photograph: author, 2003)

Figure 12. Unsurveyed rectangular anomaly 30 km south-west of Samawah (Google Earth, 2012)

29

Sasanian Archaeology

Figure 13. Aerial view of the city-site at Istakhr (Google Earth, 2012)

historically attested remains at Rayy and Nishapur remain to be convincingly identified on the ground.37

(unexcavated) Sasanian settlement site of Tal-i Baiza A, which al-Tabari refers to as the seat of a governor of the province of Istakhr during the early part of the reign of Ardashir I.33 The main population centre on the Marv Dasht plain was Istakhr, and Whitcomb has suggested a plausible difference between the Sasanian (and earlier) and subsequent Islamic parts of the city-site based on different alignments of the streets visible on air photographs;34 the trial excavations by the University of Chicago in 1932 and 1934 are essentially unpublished but, in any case, were focused on the latter area of the site (Figure 13). The date of foundation of the modern town of Shiraz remains ambiguous and the nearby site of Qasr-i Abu Nasr probably represent its Sasanian forerunner (Figure 14). Excavations there between 1932 and 1935 focused on the prominent citadel, parts of the fortifications and the raised western portion (now destroyed by modern housing), leaving the intervening urban core almost entirely unexplored.35 In other cases, such as Isfahan, the remains of the Sasanian city attested as a mint lie buried below the modern city,36 and the equivalent

Within the western portion of the Sasanian empire, corresponding largely to present-day Iraq, the results of older excavations also give some glimpses into the continuity of life within some of the old Mesopotamian urban centres. A few, such as Ur, had long since been abandoned as changes in the water-courses left some of these ancient sites stranded and waterless. Uruk is traditionally interpreted in the same manner,38 but in 1975 a 200 hectare extension of the city measuring 1600 x 1250 m across was discovered immediately to the south-east: it was dated to the late Sasanian period through the presence of torpedo jars, finger-trailed and honeycomb wares, and by comparison of the excavated and surface pottery with material from level III at Tell Abu Sarifa.39 This implies a deliberate refoundation of the city on an adjacent spot and probably straddling a new canal, although the exact date of this is uncertain. There is also extensive evidence for late and post-Sasanian occupation at Nippur: Gibson refers to this as

Miri 2009: 9; see Alden 1978; Balcer 1978; Alden et al. 2005: 45–46, fig. 4c; Simpson, this volume: 355. 34  Whitcomb 1979. 35  Whitcomb 1985. 36  Cereti 2004; Jung et al. 2016: 11; CAIS 2010. 33 

Keall 1979; Rante and Collinet et al. 2013. e.g., Crüsemann, van Ess, Hilgert and Salje (eds) 2013. 39  Finster and Schmidt 1976: 164–66; Finster 1983; see Simpson, this volume: 273-76. 37  38 

30

St John Simpson: Part One: Sites, Settlements and Landscapes - Introduction

Figure 14. Plan of the latest phases on the citadel at Qasr-i Abu Nasr, marking the excavated wall footings with mudbrick walls marked in solid black (after Whitcomb 1985: fig. 30)

The excavators found incantation bowls in situ in household contexts in the so-called ‘Scribal Quarter’ mounds, as well as in deflated surface layers there and elsewhere at the site;41 some of the post-Level I ‘late graves’ excavated in the same area probably date to this period too.42 Incantation bowls and surface

‘one of its peaks of occupation. The major part of the settlement seems to have been on the West Mound. A string of small tells along the western edge of the city may mark the bed of the most important canal of the time. The old bed of the Euphrates, through the middle of the site, was presumably still used as a canal, however … The mound called Ishan al-Sahra, south-east of Nippur, was predominantly a Sasanian settlement’.40 40 

Montgomery 1913: 13–14; McCown and Haines 1967: pls 163–67; McCown, Haines and Biggs 1978: pls 76–77.1–3; Gibson et al. 1975: 43–44, 117, figs 38.1–4, 89.3–5; Kaufman 1975; Hunter 1994; Faraj and Moriggi 2005: 71–77, pl. I. 42  McCown and Haines 1967: 73, 118–27. 41 

Gibson 1992: 52.

31

Sasanian Archaeology pottery also point to similar extensive late and postSasanian occupation at Kutha (Tell Ibrahim)43 and areas of Babylon, including Merkes:44 all three of these centres are named in the Babylonian Talmud, and Babylon is mentioned in the Arab Conquest accounts on account of a strategic fort commanding a crossing of the Euphrates here, most probably to be identified with the square fortress raised on a platform at the site known locally as Babil (Figure 15).45 The nearby site of Kish offers a slightly different picture with the remains of several elite complexes (described as ‘palaces’ in the modern literature), some originally decorated with painted stuccoes and which partly resemble the plan of similar-sized complexes at Chal Tarkhan and Damghan, and situated close to a sprawling area of residential occupation (Figure 16).46 Moving northwards, the 2nd millennium BC site of Nuzi, near Kirkuk, was re-occupied at this period, with a cemetery located on the elevated early mound in the centre of an extensive settlement which appears to date between the 5th and early 7th centuries judging by the associated pottery (Figure 17).47 Finally, the site of Nineveh provides clear evidence for continuity of occupation on the former Assyrian citadel of Kuyunjik, extending from the Seleucid to the medieval period and including evidence for a post-Sasanian church.48 Unfortunately, in none of these cases, is there evidence for how the settlements were organised, and modern excavations of this final period of occupation at any one of these old Mesopotamian urban centres is long overdue.

commanded by a vast network of new canals. The construction of these canals seems to adumbrate a new and radically different outlook. They imply the engineering competence to plan, and the political and financial power to execute, a fundamental reshaping of the landscape and its water resources in the endeavour further to enlarge and stabilize the revenues of the state’.49 This observation of sites in the Diyala basin in east central Iraq is mirrored across many other regions of the Sasanian empire. Cities need feeding and they cannot exist without efficient transport systems and supply networks. The immediate hinterlands of most highland Iranian cities were relatively small and confined to their immediately surrounding intermontane valley or plain, and the symbiosis of city and countryside is encapsulated by the layout of Gur where the radiating layout of the urban streets extends outwards into the field alignments beyond the fortifications.50 However, valleys connect and each micro-region is linked with the next. There is currently rather limited evidence for Sasanian rural settlement patterns across Iran but this is probably a misleading impression. The Damghan plain was intensively surveyed in 1976/77 by K.M. Trinkaus and the results suggest that there was much more extensive settlement in this period, compared to the preceding Parthian period, that there was an increase in the number of open sites but these were more widely dispersed, and during the late Sasanian period settlement appears to have moved to the fringes of the plain.51 This change in settlement pattern suggests a transformation in the management of the primary alluvial zone and consistent with land reforms. Juridical references to the foundation and maintenance of qanats prove that this form of agriculture and watersupply was important in some regions, particularly arid areas where measures to capture episodic run-off or to tap springs have formed the basis of agricultural activity for millennia.52 Surveys in the Kur river basin point to a significant increase in settlement along the main routes and irrigation agriculture during this period.53 Moreover, within the Behbehan plain, located midway between the plateau and Susiana, a survey carried out in 1970 indicated that the densest periods of settlement were in the 5th millennium BC and Sasanian-Islamic periods.54 It is likely that regions such as the Gorgan plain were extensively settled although attention there so far has focused either on earlier sites or the region’s military and hydraulic infrastructure during the

The success or otherwise of Sasanian urban planning can be gauged by how long these cities were occupied for. In many cases, there is insufficient evidence to indicate how long they continued after the Islamic Conquest but others, such as Bishapur, Istakhr, Merv and some of the older Mesopotamian towns, certainly thrived for several centuries thereafter and were only gradually abandoned as their populations drifted to nearby centres of power and patronage. It is important to recognise that this continuity of population must have had a strong effect on the transmission of knowledge and culture, as discussed in greater depth below. Rural and maritime landscapes ‘enough remains to indicate that virtually the entire cultivable area was brought under the plow and Reade 1986: 112–13. Koldewey 1914: 248, 251–52, figs 162, 168; see Hunter 2000a. 45  Schmidt 1941: 821–22, fig. 17; Wetzel, Schmidt and Mallwitz 1957: 24–25, pl. 13; Gibson 1972: 149 = Kish Survey Site 127; see Morony 1982: 26. 46  Harden 1934; Gibson 1972: 77–78, 113; Moorey 1978: 122–46; see below. 47  Simpson 2013a; see Ehrich 1939. 48  Simpson 1996a; 2005b. 43  44 

Adams 1965: 82–83. Huff 1974: 159. 51  Trinkaus 1981; 1983; 1984; 1986. 52  De Menasce 1966; Wilkinson et al. 2012: 168; see Moghaddam and Miri 2007: 50. 53  Hartnell 2014. 54  Dittmann 1984: 103–16, map 1. 49  50 

32

St John Simpson: Part One: Sites, Settlements and Landscapes - Introduction

Figure 15. Plan of the excavated Partho-Sasanian fortress at Babil (after Schmidt 1941: fig. 17)

Sasanian period.55 Elsewhere on the plateau, the wider settlement context of important sanctuaries such as Ganzak (Takht-i Suleiman), or forts such as Haftavan Tepe in the Urmia basin and that on the summit of Tepe Yahya (Level IA) in the Soghun valley of south-east Iran, remain uncertain.

Mesopotamia and Susiana in the south-west to Gorgan and Margiana in the north-east. Efficient and intensive agriculture, coupled with waterborne transport and communication in the case of lower Mesopotamia, were keys to the success of the urban landscape. One of the greatest contributions of archaeology to the understanding of the organisation of the Sasanian economy lie in the results of surface surveys carried out in the Mesopotamian lowlands between 1953 and 1975. These prove that there was massive investment in the regional hydraulic infrastructure designed to sustain intensive year-round agriculture and institute

The more extensive lowland plains and deltaic fans offered greater opportunities for massive development, and this is clear at both ends of the empire, from 55 

Sauer et al. 2013.

33

Sasanian Archaeology

Figure 16. Plan showing the relative locations of the excavated Sasanian buildings and other areas of occupation at Kish (after Moorey 1978: fig. J)

Figure 17. Plan of the site of Nuzi, the areas excavated indicated by the tone and the position of the early Sasanian cemetery on the north-west summit of the highest mound in the centre (after Starr 1937: vol. II, plans 2, 39)

34

St John Simpson: Part One: Sites, Settlements and Landscapes - Introduction in the alluvial plains downstream: as the middle Tigris possesses a steeper gradient than further south, canals drawn off at this point maintained a relatively higher level of water downstream and enabled secondary canals to divert supplies for irrigation. The first of these begins opposite modern Baiji, north of Tekrit, and stretches 62 km to join the Nahr al-Rasasi (Qatul al-a’la al-Kisrawi). A square palace measuring 250 m across with a central pool lies next to the junction and may be late Sasanian in date, and; the Sasanian town of Karkh Fairuz lies nearby (Figure 19).61 Further south, the Nahr al-Rasasi was joined by further off-takes, the Nahr al-Qaim and Nahr al-‘Ibra. Near the beginning of the first of these lies the remains of a Sasanian solid block tower, the Burj al-Qaim, which may represent a commemorative monument or fire signal station (Figure 20);62 Ammianus refers to such a construction at the Euphrates mouth of the Nahar Malcha (‘Royal Canal’), which he describes as ‘a tower of considerable height … like the Pharos’,63 and another, ‘from which the new moon used to be announced’, is described in the Babylonian Talmud as located at Biram near presentday Falluja.64 To the east, the Nahr al-Rasasi cuts across the southern end of the river Adheim and diverted its waterload into the Diyala and Nahrawan systems south of Baquba. Further evidence for the new degree of control placed on the natural drainage is indicated by the construction of a barrage and regulator on the Adheim at the point where it broke through the Jabal Adheim (Figure 21).65 The same feature was used at the point where the Diyala river breaks through the Jabal Hamrin.66 The purpose of this regulator was mainly to control the quantity of water entering the Qatul al-Kisrawi and Nahrawan canal systems during spring peak flow but also to deflect water into canal off-takes on the right bank of the Adheim. It might be added that the fan-like distribution of these suggests that a high density of rural settlement is to be expected here when the area is properly surveyed, and would explain references to Arab pastoralists being cleared from this section of valley in favour of arable agriculture.67 In addition to placing greater control over water supply, these canals must have facilitated transport both by water and along their banks (Figure 22).

an improved and closely integrated water transport system. The first important archaeological survey within the wide area covered by the Sasanian empire was begun in the Diyala region of east-central Iraq in 1936/37. Initiated by the Sumerologist Thorkild Jacobsen, it rapidly demonstrated that the earliest periods of settlement were heavily masked by alluvial silt, whereas the most significant late period of occupation corresponded to the Sasanian period and this was characterised by large-scale water-engineering projects involving canals and regulators.56 The results were developed and published in greater detail by Adams, and the principle of archaeological survey and use of diagnostic pottery types (so-called ‘type fossils’) to date sites was extended by him and others across large areas of central and southern Iraq and the Susiana plain of south-west Iran between 1956 and 1975.57 Adams later admitted that ‘in the northernmost part of the plain the exceptionally numerous and massive sites of the later periods were for the most part not recorded’ on the 1956/57 survey.58 Increasing attention was paid on later surveys to these late periods and a slightly refined chronology developed: thus Gibson’s more intensive survey of some 2000 sq km around Kish contradicts the results of Adams’ Akkad Survey (although they were published together in 1972) and suggests a Sasanian, rather than Parthian, date for an increase in both settlement density and settlement size.59 Irrespective of their individual inconsistencies, these different surveys concur in demonstrating that the extent of Sasanian engineering works was huge: not only were the Tigris and Euphrates systems combined for the first time, but the Greater and Lesser Zab and Adheim rivers were also controlled and diverted into canal off-takes. The beginning of this integrated water management system therefore lies considerably further north than previously recognised (Figure 18). A series of canal off-takes above and below the Sasanian town of Mahoza dhe Ariwan (present Tell Mahuz) on the left bank of the Lesser Zab were used to divert flow via the ‘Abbassi and Fil canals into the Adheim river system. The purpose of this appears to have been to increase water supply south of the Jabal Hamrin rather than for local irrigation purposes.60 Moreover, additional largescale water engineering works were undertaken along the middle Tigris with the cutting of a series of canal off-takes through Pleistocene (or older) conglomerate terraces on the left bank of the river Tigris. The primary purpose of these feeders was to increase water supplies

The region on the opposite, right, bank of the Tigris appears to have been dominated by pastoralist tribes and, during the 3rd century, members of the Qudā‘a tribe are said to have spent their winters here until they were later moved to the Hira region.68 A century later, Northedge 1987. Herzfeld 1924; 1948: 7. 63  Ammianus, History 24.2.7. 64  Oppenheimer 1983: 96–97. 65  Jones 1857: 121–22 and facing illustrations; Sulaiman 2011: 385–87, pl. XXX. 66  Adams 1965: 76. 67  See Adams 1965: 78, fig. 7. 68  Morony 1984: 215. 61  62 

Jacobsen 1954; Jacobsen and Adams 1958. Adams 1962a; 1965; 1972; 1981; 2006; Adams and Nissen 1972; Gibson 1972; Wenke 1975/76. 58  Adams 1981: 206. 59  Gibson 1972: 51–52; see Adams 1965: 70; 1972: 187–88. 60  Ionides 1938; see also Reade 1978: 172–75, figs 1–2. 56  57 

35

Sasanian Archaeology

Figure 18. Plan showing the reconstructed extent of Sasanian sites and canals in central and southern Iraq, the extent of the published archaeological surveys being indicated by the toned areas (after Simpson 2015b: 19)

and mixed flocks of sheep and goats.69 A desire to control the movement of such tribes from the south Jazira into the irrigated zone further south may have

the Iyad ibn Nizar occupied the area but were killed, resettled and/or driven over the Roman frontier during the reign of Shapur II; further re-settlement of this tribe occurred at Tekrit under the reign of Khusrau I when they are described as possessing herds of camels

69 

36

Morony 1984: 216.

St John Simpson: Part One: Sites, Settlements and Landscapes - Introduction

Figure 19. View of the fortifications around the Sasanian town at Karkh Fairuz near Samarra (photograph: A. Northedge, 1983)

motivated the possible early Sasanian construction of a long wall known today as el-Mutabbaq, which stretches for at least 10 km into the steppe from the right bank of the Tigris opposite Samarra (Figure 23). This was constructed as a solid casemate with packed gravel infilling and a mudbrick face on both sides, regular projecting semi-circular towers on the outer, western, face and a 20–30 m wide ditch along the front (Figure 24).70 In terms of construction it resembles the Gorgan and Tammisheh Walls, as well as the walls around Gur and Iwan-i Karkheh, but with the addition of towers.71 Traces of an equivalent construction known as Umm Rus have been noted in the middle Euphrates area and this may be the same as the ‘half-destroyed traces of walls’ reported by Ammianus in the mid-4th century: ‘these in early times had a wide extent, it was said, and protected Assyria from hostile inroads’.72 The Diyala basin fell within the late Sasanian administrative province of Khusrau Shadh Qubadh or Shadh Hormizd and was administered from Ctesiphon. The Qatul al-Kisrawi/Nahrawan canal system drawing water from the middle Tigris and Adheim was a major physical feature which linked shorter stretches of earlier canal in the lower Diyala basin. Joined by further offtakes from the upper Diyala, the aim of this trunk canal was to increase the quantity, reliability and regularity of water supplies in the lower Diyala and adjoining reaches of the Tigris. The latter were in turn linked by canals to the plains further south. The effect was to

Figure 20. View of the Sasanian block tower at Burj al-Qaim (photograph: A. Northedge, 1983)

Reade 1964. Nokandeh et al., this volume. 72  Ammianus, History 24.2.6. 70  71 

37

Sasanian Archaeology

Figure 21. Plan of a Sasanian regulator at Adheim (after Jones 1857: fig. 10)

Figure 22. Modern canal near al-Mada’in in southern Iraq (photograph: author, 2016)

38

St John Simpson: Part One: Sites, Settlements and Landscapes - Introduction

Figure 23. Aerial view of the Sasanian long wall known as el-Mutabbaq (Google Earth, 2013)

at Qalat Sefid, immediately south of Mandali, and Tell al-‘Aqar near Badra.75 It seems likely that this area sustained a high population in Sasanian and early Islamic times with hill-slope runoff being utilised for natural flow irrigation.76 This would not only match the situation documented in greater detail for the Deh Luran plain further east,77 but also explain the high density of Sasanian sites recorded on Iranian surveys of the intervening Mehran plain.78 Nevertheless, the close similarity in the plan of a late Sasanian courtyard house excavated by the Italian archaeological expedition to Iraq at Tell Mahmud in the Hamrin basin79 with those mapped on the surface in the Deh Luran reinforces the impression that these may even have been part of the same phase of rural development,80 and the analysis of the zooarchaeological remains from Tell Mahmud gives an important insight into the equivalent environmental remains .81

sustain a further increase in agriculture (attaining the maximum possible in the Diyala), settlement density and size of urban centres. The Sasanian period thus sees a continuation of trends, beginning in the Seleucid and Parthian periods, which were associated with the development of agriculture and infrastructure in the immediate hinterland of the new Tigris capitals of Seleucia and then Ctesiphon, and reflects the growing importance of this route into western Iran. A characteristic of Sasanian settlement here, and further south, was a tendency for low sprawling ribbon development alongside the waterways but these also doubled as arterial routes and potentially defensible water features. During the Sasanian period, Ctesiphon increasingly resembled a conurbation of cities, both old and new, explaining the respective Aramaic and Arabic terms of Mahoze and al-Mada’in (‘the cities’).73 East of the Diyala, the Mandali–Badra area was briefly surveyed in 1966.74 A Sasanian ‘monumental building’ is suspected at Tell Kusakh, the largest of a group of mounds clustered along a wadi leading off the Ab-i Naft, and Partho-Sasanian occupation is also reported 73  74 

Oates 1966: 54, fig. 2; Hrouda 1973. See Dowson 1921: 20. 77  Neely, this volume. 78  Zeidi 2012; see Nokandeh 2010. 79  Invernizzi 1980: fig. B. 80  Neely, this volume. 81  Fedele, this volume. 75  76 

Simpson 1997a; Negro Ponzi 2005a; Hauser 2007b. Oates 1968.

39

Sasanian Archaeology

Figure 24. View of the unexcavated Sasanian long wall at el-Mutabbaq (photograph: H. Abdulrasool)

from the city of az-Zandaward (possibly the 230 hectare site of Tell Jidr) and the latter from Kaskar (opposite the Islamic conquest foundation of Wasit). The region further south was the province of Ard Maysan, formed from southern Characene and Mesene, and centred on the cities of Astarabadh Ardashir (former Charax Spasinu, modern Naisan) and Bahman Ardashir (Pratta, modern Maghlub), both situated on the eastern or Madhar branch of the Tigris and largely submerged under the alluvium.84 Under the terms of the late Sasanian administrative quadripartite division of the empire, Ard Maysan was part of the so-called ‘Quadrant of the South’ which was centred on Fars and included the Persian Gulf, and this suggests that its economic orientation was re-orientated accordingly. The fact that two other cities are attested here, namely Apamea near Fam al-Silh and Shadh Shapur (founded by Shapur II), as well as the port at Ubulla (previously known as Apologos, later replaced by Old Basra), indicates that Ard Maysan was a flourishing province.85

South of the so-called Hit–Samarra line between the Tigris and Euphrates, these rivers finally break free of their incised valleys and enter the alluvial plains of lower Mesopotamia. The rivers deposit most of their silt load at this point as the velocity of the water decreases accordingly. One estimate suggests that, prior to the heavy damming of both rivers in recent decades, the Tigris alone was responsible for dumping three million tons of silt a day during the peak discharge months of April and May.82 The bed of the Euphrates is some 9 m higher than that of the Tigris between Baghdad and Ramadi – where the rivers come within 40 km of each other – and for this reason canals cut across from the Euphrates enable gravity flow irrigation across the northern part of the alluvium. The Mesopotamian alluvial plains correspond partly to the early Sasanian province of Asoristan (later Arab Sawad), but were divided in late Sasanian times into the provinces of Ard Babil, Ard Kaskar, and Ard Maysan.83 The first of these was eventually divided into four smaller provinces: Peroz-Shapur (after the city of the same name on the middle Euphrates, modern Falluja), Veh Ardashir (stretching from that city opposite Ctesiphon to Kutha), Veh Kavad (further south along the Babylon branch of the Euphrates) and az-Zawabi (on the Tigris near Nu’maniyya). To the south, Ard Kaskar was created as a province from the northern portion of the former kingdom of Characene and stretched from Veh Kavad/az-Zawabi in the northwest and west to the western channel of the Tigris. It was divided into two districts based on the Shatt anNil (to the west) and Tigris, the former administered 82  83 

Systematic archaeological surveys only cover the equivalent of one of these three provinces, namely Ard Babil, and even this has neither been fully surveyed nor covered to the same level of intensity. Nevertheless, the results clearly demonstrate that this region was covered by a complex and fully integrated network of canals during this period. A series of canals traversed the northern portion and connected the Euphrates with the Tigris. Near the site of Tell Nihar (possibly Talmudic Nehardea), a trunk canal ran south-east from the Euphrates down to Tell Jidr. This precursor of the later Shatt an-Nil was paralleled by two further canals,

Buringh 1957. Morony 1982.

84  85 

40

See Moon et al. 2016. Simpson 2019b.

St John Simpson: Part One: Sites, Settlements and Landscapes - Introduction been unwalled, judging by the amorphous shapes of the mounds, and this probably contributed to dynamic shifting settlement, as has been noted at Zibliyat where the 9th century pottery was noted closest to a prominent tower where dessicated reeds surviving within the brickwork enabled radiocarbon dating to c. AD 880.91 The same type of settlement organisation applies to some of the much older towns, such as Babylon, Borsippa, Kutha, Nippur and Warka (see above). The continuity and conservatism of population at these might also explain the frequency of Babylonian magic bowls at many of these sites.92 Contrasting with these settlements are a series of fortified new towns which appear to have been founded as part of the agrarian re-organisation. These include the sites of Tell ad-Dhiba’i (Medina), Ruqbat al-Mada’in (Figures 26–27), and Jidr. Adams and Nissen suggest these functioned as military strongholds,93 but their central locations and proximity to major canal junctions suggest they were also centres of local administration and industry, and the walls themselves may have been to denote status as well as defence. However, some other sites may have had a more explicitly military or police function and are characterised by being raised on artificial platforms (Nippur Survey site 716, Warka Survey sites 164, 170), whereas others appear to have functioned as watchtowers (Tell Shahal, Ishan al-Turmma).

located 6–7 km to the east and west respectively, which ran for some 85 km between Ishan al-Jihariz (a 72 hectare site) and Jidr. The site of Zibliyat (160 hectares) lies near the northern end of the easterly of these two canals, whereas Jidr itself is described as: ‘a very large and long-lived ancient town, whose full importance has yet to be recognized ... Nearest the center is a steeply elevated citadel (?) 200 m. in diameter. Numerous yellowish baked bricks (34 x 34 x 7 cm) here bear a stamped impression of trifurcating wavy lines, for both of which a Sassanian date seems probable. Immediately south-west of the citadel is a square mound, 200 m along each side and with corners oriented to cardinal directions, that rises almost to [the] same height. Here there are bricks of same dimensions (but without stamp), and much mortar’.86 The location of these major urban centres is clearly closely related to the design of this portion of the canal system. Upstream of Ishan al-Jihariz, a fan of further canal off-takes implies the existence of major weirs and regulators. Additional traces of waterworks with fired brick construction have been noted at several other sites.87 Adams concluded that ‘the countryside for a long distance to the south is divided into north–south strips by a fairly regular network of parallel, unusually straight and hence carefully laid out branch canals. At intervals these are intersected by other canals, either at right angles or along diagonals, whose uniformity of orientation again indicates that they were laid out according to a large-scale, carefully surveyed plan. The effect was to open up a very large new area for cultivation by dividing it into polygons of varying size, principally rectangles, trapezoids, and triangles of from as little as 20 or 30 to 1,000 or more hectares’.88

The extent of ancient permanent settlement east of the present western course of the Tigris (which effectively marks the eastern limit of modern archaeological surveys) is uncertain and, for the Sasanian period at least, has provoked some controversy as it has been regarded as being part of the marshes or ‘Great Swamp’.94 The extremely shallow gradient of the alluvium at this point renders it very susceptible to seasonal or permanent flooding if the river or canal banks are breached, particularly when combined with raised sea-level and/or temporary tectonic downfaulting. Prior to the large-scale draining at the end of the last century, the marshes offered rich fish, bird and reed resources, the narrow strips of raised land alongside the waterways were well suited to the cultivation of date palms, salt-tolerant barley and rice, and there was good grazing potential. Ironically, rather than representing a region of poor agricultural resources, the marshes therefore offered an unusually high diversity of minor eco-zones which were capable of sustaining small secluded communities, as well as larger population centres on the main river channels. However, the Arab accounts suggest that this was not always the case. Ibn Rustah states that ‘before Islam’ ships from India could travel as far as Ctesiphon along

The largest number of settlements across this zone belonged to village-sized settlements of ten hectares or less. However, the frequency and exceptionally large size of urban centres resulted in these villages totalling only some 42% of the total settled area according to Adams.89 Only one small settlement of this period has been excavated within this region, namely Tell Abu Sarifa, where excavations in Levels II and III revealed the remains of regularly aligned mudbrick walls which were dated to the 6th century and earlier, mainly on the basis of two incantation bowls found in Level III (Figure 25).90 The majority of these villages appear to have Adams and Nissen 1972: 219. Adams 1981: 256, 258, 273, 279, 290 = Nippur Survey sites 612, 686, 1121, 1280, 1572. 88  Adams 1981: 210. 89  Adams 1981. 90  Adams 1970. 86 

Burleigh 1980. Scotten, this volume. 93  Adams and Nissen 1972: 62. 94  e.g., Adams 1981: 205; see Hritz, Pournelle and Smith 2012; Verkinderen 2015.

87 

91  92 

41

Sasanian Archaeology the present eastern course of the Tigris but, according to Yaqut, this channel was blocked in the 5th century by Varahran V (AD 421–439) as part of a project to add water supply to the western Tigris channel. This is consistent with development of the province of Kaskar and suggests that the water was required for new field systems on freshly drained land east of Warka. AlBaladhuri states that there had been flooding of this area but the land was restored by Kavad I (AD 488–496, 499–531), until the area became flooded again in c. AD 627–629 – doubtless following lack of resources after the wars of Khusrau II and the political instability which followed – and the area turned to marsh. Conclusion A mixture of continuity and change runs through the Sasanian legacy to their Islamic successors. In some regions there appears to have been a collapse in settlement but how far these reflect economic, military or other causes is unclear, and a combination is likely. In some cases the appearance may also be misleading as the dating of the sites in question relies on good control of the pottery types present and, unless these are specified and/or quantified, questions remain over the precision of their original dating. Despite the length and scale of previous archaeological investigations, large areas of Iraq still remain almost entirely unsurveyed. For present purposes, these include the entire region between the Hamrin basin and Nineveh, the levees along the Tigris and Euphrates and the large region corresponding to the Sasanian province of Ard Kaskar. Moreover, the quality of survey for Sasanian sites north of Kish is very weak, and the methodology of survey which was based on vehicular transects led to an under-representation of low-lying sites, whether small or large. Nevertheless, the results suggest that the destructive effects of Julian’s AD 363 campaign have been over-emphasised,95 as has the extent of the marshes which at one point were even considered to extend as far north as Kufa/Najaf and Hilla.96 Moreover, the surveys also contradict a suggestion that ‘the urbanisation project in southwest Iran and Mesopotamia brought about a decline in rural settlement and migration to the cities’.97 In this case urban and rural growth went hand in hand as each relied on and stimulated the other, and Adams correctly concluded that this period ‘was the apogee of the ancient settlement and irrigation record in every respect’.98 There is similar evidence from Khuzestan,99 the Susiana

Figure 25. General plan of Tell Abu Sarifa (1) and the architectural remains excavated in Levels II–VI (2–6) (after Adams 1970: figs 1, 3)

plain100 and the Deh Luran,101 although the dating of the associated sites requires more precision. Moreover, it was during this period that Bushehr and its immediate hinterland also saw intensive settlement,102 and the discovery of large number of ossuaries containing secondary burials proves that a large proportion of the population were Zoroastrian.103 The results of more recent surveys in the Kur river basin of northern Fars and small valleys north of the Persian Gulf between Bushehr and Siraf suggest a similar intensification in

See Matthews 1989. Le Strange 1905: 40–43, map II; Obermeyer 1929: 97; Adams 1981: 180, 205. 97  Mousavi and Daryaee 2012: 1078. 98  Adams 1981: 252. 99  Adams 1962a; 1962b; Eqbal 1976: 116, table 20, fig. 49; Walstra, Heyvaert and Verkinderen 2010; cf. Moghaddam and Miri 2007: 48–52. 95  96 

Wenke 1975/76; Moghaddam and Miri 2003: 103–105. Neely 1974; this volume. 102  Carter et al. 2006. 103  Simpson and Molleson 2014; Simpson 2019c. 100  101 

42

St John Simpson: Part One: Sites, Settlements and Landscapes - Introduction

Figure 26. Plan of Ruqbat al-Mada’in with architecture visible on the surface mapped in three areas at the northern end (after Finster and Schmidt 1976: 153–57, figs 80–84), but note that the shape and position of the towers is incorrect.

the Abbasid period,105 although Priestman is more cautious,106 and Asadi et al. argue that local agricultural production was of greater importance along the northern Persian Gulf.107 Moreover, on the southern side of the Persian Gulf, a re-assessment of earlier surveys combined with more recent work shows that the extent of permanent settlement in eastern Arabia during this period is considerably less than previously assumed.108 However, the unexcavated Sasanian ports of Rev Ardashir (modern Rishahr) and Ubulla probably

those regions during the Sasanian period: they also offer evidence for a significant investment in irrigation and industrialised agriculture in the former region as well as the exploitation of more marginal lands to the south.104 The importance of maritime trade to the Sasanian economy and development of the Persian Gulf has been emphasised in many studies yet its scale and organisation are still unclear. Daryaee and Morony follow Whitehouse and Williamson in arguing that it was highly developed and similar to that known from

104 

Daryaeee 2003; 2010: 405–406; Morony 2004: 185–88; see Whitehouse and Williamson 1973. 106  Priestman, this volume. 107  Asadi et al., this volume. 108  Kennet 2007. 105 

Hartnell 2014; Asadi et al., this volume.

43

Sasanian Archaeology directions and some of the commodities of trade.113 The construction of a tower at Kush, along with other finds which indicate a militarised presence there in the late Sasanian period, may be interpreted along similar lines. This was the largest settlement in the northern Emirates at this period, situated next to a lagoon and dominating a fertile hinterland where date cultivation played an important economic role, alongside yearround fishing and the herding of sheep and goat.114 The intensification of agriculture had several significant effects on the Sasanian economy. It increased the population-carrying capacity of the land, as well as the volume and regularity of product from agro-industries such as textiles and leather-working, ensured greater supplies of fuel for pyro-industries such as metal, glass and pottery, and generated increased revenue through taxation of crop yields. Fertile regions therefore must have been regarded by the state and private individuals alike as highly desirable and very important assets. It is not surprising to see investment being made in these regions, and not simply through water engineering works. The extent of integrated planning is becoming clearer not only in Mesopotamia, but also both sides of the present Iran/Azerbaijan border,115 the Gorgan plain of north-east Iran,116 and the Merv oasis.117 These are all highly fertile regions. Although the increasing use of mechanisation in the agriculture is rapidly transforming the archaeological landscape into a zone of destruction, it is clear from the number, density and size of mounded sites that the Gorgan plain had a very lengthy and prosperous pattern of settlement beginning in the 6th millennium BC.118 From the early 19th century onwards, it was also noted that there was a very substantial linear feature running across this region between the Kopet Dagh and the Caspian sea. Its dating attracted many different opinions, some relying on hear-say and others citing evidence of Caspian sealevel changes, pottery and historical sources, to argue for dates ranging between the 5th or 4th century BC (when it was attributed either to the Achaemenids or Alexander the Great respectively) and the 6th century (when it was assigned to one of the Sasanian rulers). The results of the more recent joint Iranian-British project to study this wall using a wide variety of techniques have conclusively shown that it was constructed within a short period of time, probably during the second or third quarter of the 5th century although possibly slightly later, and must reflect an urgent imperial Sasanian response to the very real threat posed by the

Figure 27. Aerial view of Ruqbat al-Mada’in showing substantial areas of near-surface housing separated by streets with projecting horseshoe-shaped interval towers along the hollow curtain wall and signs of extensive looting (Google Earth, 2018)

served in tandem as gateways to and from Fars and Mesopotamia, thus the forerunners of Bushehr and Basra respectively. They must also have been magnets for investment, and Carter follows Williamson in suggesting that Rishahr and the neighbouring Iranian coastline took over, and benefited from, the bulk of the Persian Gulf pearling industry at this period.109 The excavation of a port such as this might completely change our understanding of Sasanian maritime trade, just as the recent decades of research in the Red Sea have shown that it was considerably more dynamic in Late Antiquity than previously recognised.110 Underwater discoveries of torpedo jars and military equipment off the Iranian coast between Gonaveh and Bushehr indicate that ships were sailing the region,111 and finds of Mesopotamian, south-east Iranian and Indian pottery, imported Sasanian glass and south or south-east Asian glass beads at the site of Kush in present-day Ras al-Khaimah (United Arab Emirates),112 or Sasanian torpedo jars at the pepper-exporting centre of Pattanam in southern India, illustrate

Tomber 2007. Simpson et al. forthcoming b. Alizadeh 2014; Ur and Alizadeh, this volume. 116  Sauer et al. 2013. 117  Simpson 2014b. 118  Sauer et al. 2013: 93. 113  114 

Carter 2012: 23–29; see Williamson 1972: 106. Tomber 2008; Sidebotham 2011. 111  Tofighian, Nadooshan and Mousavi 2011. 112  Kennet 2004. 109 

115 

110 

44

St John Simpson: Part One: Sites, Settlements and Landscapes - Introduction Susiana and in Assuristan and in every other nation where our own and our father’s and our forefather’s foundations were’.122

Hephthalites. A summary of the results is presented below.119 They necessitate a complete re-assessment of the organisation of the Sasanian army and imply that it must have been a professional body, rather than a feudal levy as often assumed, and suggest a closer relationship between military and non-military planners than previously considered.

These victories were a wake-up call for Rome and drastic measures put in place in order to generate more money to overhaul their eastern frontier. The first step was the confiscation of long-established urban revenues, such as local tolls, taxes and endowment proceeds, followed by debasement of the coinage and price fixing to compensate for the massive inflation resulting, and finally new regular taxes on economic production.123 Although a military status quo was re-established at first, the successes were relatively short-lived as a more aggressive policy was adopted by Shapur II (AD 309–379) with further campaigns in upper Mesopotamia. According to Ammianus, virtually the entire population of Singara was taken into captivity after a short siege in AD 359/60, ‘transported to the remotest parts of Persia’, and soon afterwards Shapur crossed the Tigris and sacked the frontier city at Bezabde:124 the hagiographical source known as the Testimony of the Captives refers to the deportation of some 9,000 of its inhabitants to Khuzestan.125

Additionally, the implications of the surveys are that the Sasanian state succeeded in transforming large areas of the natural landscape into highly productive sources of food and revenue, and the results of this must have had profound social as well as economic effects. Ensuring this not only meant guaranteeing large and regular supplies of water but also manpower to work the land. As stated at the outset of this section, Classical, Syriac and later Persian authors repeatedly refer to the transplanting of substantial urban populations from Roman Syria into lower Mesopotamia and Iran following campaigns by Shapur I. Although details are vague and partly contradictory, the gist is clear. People were regarded as valuable assets and the movement of populations in this manner was not restricted to Roman deportees as a later source, the Nihāyatu l’arab, refers to 12,000 Iranians from Istakhr and 4,000 from Isfahan being sent to Nisibis after its capture by Shapur II in AD 363 (see below). Another late memory of this was captured in the 8th or early 9th century Pahlavi text known as the Shahrestānīhā-ye Ērānshahr [Provincial Capitals of Iran]. This lists a large number of cities in what it claims to be the ‘land of the Iranians’, although the later date of the text means it adds places more familiar to an Abbasid readership and their alleged founders are often mythical or hear-say characters.120

These captured cities must have been repopulated with communities brought from elsewhere, and this explains the passage in the Nihāyatu l’arab described above.126 Singara also had a strong Iranian garrison which is said to have returned to Iran at the time of the Arab Conquest.127 This movement of Iranians into western frontier zones must have been intended to help bolster Iranian identity and allegiance in these newly conquered cities. Moreover, this process was not limited to urban population transfers as Kavad is said to have settled people in villages in different parts of Iraq.128 Shayegan views the wars as being ‘over the control of trade with the East’ and less through ideology than ‘the necessity to protect the economic interests of the urban clientele’.129 Classical and Syriac sources imply that the westward population transfers which followed Sasanian victories were intended as acts of revenge and involved great hardship, but the reality seems to be more complex. These populations were valuable human resources who included builders, craftsmen and simple manpower, all taxable and able to contribute to imperial projects. Trade clauses and border infringements may have been casus belli but the economic incentives of these operations must have been very tempting for a state which enjoyed

The Sasanian use of mass deportation of people from one region to another follows a long history in the Near East, and was recognised for its economic benefits as well as its symbolic value. In his three great Roman campaigns between AD 240 and 260, Shapur I took a large number of captives, and during his second campaign alone he sacked as many as 37 cities, including Antioch and Dura. In about 260 he defeated the emperor Valerian and his own Res Gestae states that ‘we took him prisoner with our own hands as well as the other commanders of the army, the Praetorian Prefect, senators and officials. All these we took prisoner and deported to Persis’.121 The inscription continues with a statement that they were not only moved deep into Iran but were distributed across old and new cities in different regions: ‘We led away into captivity men from the Empire of the Romans, non-Iranians, and settled them into our Empire of Iranians, in Persia, in Parthia, in

Gk., ll.34–35, p. 315. Heather 2005: 64–65. 124  Ammianus, History 20.6.7. 125  Lieu 1986: 496. 126  Browne 1900: 221. 127  Hill 1971: 86; see Morony 1982: 9. 128  Morony 1976: 41. 129  Shayegan 2003: 373. 122  123 

Omrani Rekavandi et al., this volume. Marquart 1931; Daryaee 2002. 121  Gk., ll.24–26, p. 313. 119  120 

45

Sasanian Archaeology military supremacy in the field. Moreover, these were not the only sources of labour available. Corvée work by local peasants, cheap immigrants and slaves was also likely. One wonders where the deportees of Sasanian campaigns were placed if these policies were to be maintained: were they moved deeper into Iran, were they dissipated across numerous regions, and how many might have been put to hard work on canal building or mining? Deportees were effectively enslaved, and history tells us that slaves are generally not treated very kindly and are given tasks that no-one else is prepared to do voluntarily.

might be regarded as Late Antique.130 As Heather has observed, frontiers are zones of contact rather than simply strict lines of demarcation.131 Alizadeh has argued, on the basis of archaeological work in northwest Iran and neighbouring Azerbaijan, that the borderlands ‘were as important as heartlands for the maintenance of internal order’,132 and this statement is supported by archaeological evidence from other frontier regions. Moreover, successful frontiers not only require good soldiers, but also rely on effective logistics ranging from the supply of regular and large quantities of food and fodder to materials, both in raw and finished form. The second and third sections in this volume address the question of what we know about the Sasanian agricultural economy, followed by some of the equivalent evidence for manufacturing crafts and industries.

The westward shift of the Sasanian frontier in the 4th century brought it into deeper contact with eastern Roman culture and thus directly propelled it into the modern debate as to how far the Sasanian empire

Morony 2008. Heather 2009: 73. 132  Alizadeh 2014: 111. 130  131 

46

Parthian and Sasanian Site and Settlement Patterns on the Deh Luran Plain, Khuzistan Province, South-West Iran J.A. Neely Department of Anthropology, University of Texas, Austin, Texas 78712, USA.

Introduction The purpose of this paper is threefold: (1) present an interim report on the Partho-Sasanian occupations of the Deh Luran plain, (2) compare the settlement patterns of the Deh Luran plain for the Partho-Sasanian periods with those published for other nearby regions of the Middle East, and (3) discuss the possible role(s) of the Deh Luran plain in relation to the larger economic and socio-political centres in the greater Mesopotamian sphere during these periods.1 The settlement pattern data presented in this paper derived from a comprehensive archaeological programme conducted on the Deh Luran plain of southwestern Iran, but curtailed since the Iranian Revolution of 1979. Sponsored by Rice University and funded by the National Science Foundation, this programme was under the general direction of Dr Frank Hole. The majority of the data were collected during a 1968/69 reconnaissance of the plain,2 although some information also comes from the 1963/64 fieldwork.3 This paper will relate the results of a second phase of analysis to consider the Partho-Sasanian occupations of the plain in a more detailed exposition. While the basic information and associations may be adjudged accurate, the temporal placement of some of the sites involved is tentative. The trends brought to light in this secondary analysis will most likely stand as presented, but the support of specialists is being sought to assess the results of our analyses of the unglazed pottery. It is possible that changes will have to be made in a few cases of site period assignment, as well as mean site size and population estimates. We are currently working

The following individuals have graciously provided their time, labour, and information to the betterment of this study: Elizabeth Carter, John V. Cotter, William E. Doolittle, Lynn Berry Fredlund, McGuire Gibson, John Hansman, David V. Hill, Frank Hole, Vance T. Holliday, Pierre de Miroschedji, St John Simpson, Jason A. Ur, the late Tony J. Wilkinson, and Henry T. Wright. It is with profound gratitude that I acknowledge their contributions. 2  Neely 1969; 1970; 1974; Neely and Wright 1994; Wright and Neely (eds) 2010. 3  Hole, Flannery and Neely 1969.

to present a final report on these periods in the near future.4 Environmental setting The environment of the plain has been recognised as playing an important role in the cultural development and settlement distributions through time. The details concerning the environment of the Deh Luran plain have been published in several venues.5 Only the most salient characteristics of the environment pertaining to the economy and settlement pattern study are reviewed here. The Deh Luran plain (Figure 1) is located in south-west Iran near the border with Iraq, some 300 km north of the Persian Gulf and 550 km south-west of Tehran. Note that this is about 200 km south-east of the Diyala region of Iraq and 125 km north-west of the Upper Khuzistan (Susiana) plain, areas to be referred to later in the paper in relation to the findings of the intensive archaeological surveys conducted by Adams and Wenke.6 The Deh Luran plain lies within the semi-arid steppe of the Zagros foothills biotic province at an elevation of about 150 to 300 m above sea level.7 The summer months are dry and hot, and high mean temperatures of over 50°C are quite common. Winter temperatures seldom fall below freezing. The annual precipitation of 250 to 350 mm is highly variable and not equally distributed throughout the year. In the winter, when the vast majority of the precipitation occurs, the alluvial plain is transformed in places into meadows of various grasses and wild flowers. The area is not a uniform or homogeneous environmental zone. In consultation with the project’s geographers, botanist and zoologist, at least four micro-environmental zones have been

1 

Sasanian Archaeology (Archaeopress 2022): 47–79

Neely and Wright in preparation; another version of the present paper was published by the author as Neely 2016. 5  Hole 1987; Hole, Flannery and Neely 1969; Kirkby 1977; Kirkby and Kirkby 1969; Neely 1974; Neely and Wright 1994; Wright and Neely (eds) 2010. 6  Adams 1962a; 1965; Wenke 1975/76; 1987. 7  See Hatt’s [1959] ‘Assyrian Steppe’. 4 

Sasanian Archaeology

Figure 1. The location of the Deh Luran plain in relation to the Diyala plain and the Upper Khuzistan (Susiana) plain

defined.8 These are based on the present-day situation,9 as well as data derived from studies pertaining to the early periods of occupation of the plain (Figure 2).10 The four zones are:

• • • •

rocky piedmont zone riverine zone alluvial plain zone shallow, salty marsh zone

The rocky piedmont (Figure 3) forms the northern portion of the survey area, north of the improved road passing south-east-to-north-west through the Deh Luran plain from the city of Dezful. The piedmont

Coe and Flannery 1964. 9  Hole, Flannery and Neely 1969; Kirkby 1977; Kirkby and Kirkby 1969. 10  Hole, Flannery and Neely 1969. 8 

48

J.A. Neely: Parthian and Sasanian Site and Settlement Patterns

Figure 2. The Deh Luran plain, showing the approximate boundaries of the four micro-environmental zones

channels in the foothills and mountains north of Deh Luran. Upstream, beyond the plain, the river channels are deep and narrow and are characterised by nearly vertical banks with heavy rock content. As they enter the plain from the north, the channels are over 10 m below the alluvium but become shallower to the south. The terraced flood plains lying between the main river channel and the alluvial plain are most extensive to the south and there the terraces are densely vegetated with grasses and low forests of Tamarisk (Tamarix), wild Licorice (Glycyrrhiza) and Poplar (Populus).

risAes in a fairly steep slope to the Kuh-i Siah range of the Zagros mountains. This micro-environment is rocky and highly dissected by erosional channels. A number of small springs, as well as tar and naphtha seeps, effloresce at and near the juncture of the piedmont with the Kuh-i Siah range. Compared with the alluvial plain, the piedmont is characterised by a substantial increase of perennial grasses as well as small trees and shrubs. This sub-zone served the ancient inhabitants of the plain for the grazing of sheep and goats, as it does today, but was also where some dry-farming activities (discussed below) were conducted and gristmills were constructed.11

The alluvial plain lies immediately south of the rocky piedmont zone and forms the largest of the four microenvironments (Figures 5–7). The improved road passing south-east to north-west through the plain from Dezful, the closest urban centre some 125 km to the east-south-east, generally forms the boundary between these two sub-zones. In the Deh Luran area, this road is believed to generally overlie the royal Achaemenid road that was constructed from Shush (Susa) to Sardis

The riverine micro-environment consists of the channels of the Mehmeh and Dawairij rivers and their flood plains (Figure 4). The rivers, especially the Mehmeh, carry brackish waters charged with gypsum and other salts from geological strata cut by their 11 

Neely 2011.

49

Sasanian Archaeology

Figure 3. Photographic overview of the rocky piedmont zone, looking north-east from DL-2 toward the Kuh-i Siah range of the Zagros mountains

Figure 4. Photograph of the Mehmeh river, illustrating the riverine zone. Note the depth of the river channel, looking north, with Tepe Farukhabad (DL-32) in the centre background

and the Aegean Sea. The alluvial plain is essentially a flat alluvial surface with small amounts of scattered, low vegetation. Natural topographic features consist of depressions or ‘sink holes’, the erosional cuts of the two main perennial river systems, a few smaller springfed watercourses, and numerous small intermittent drainages. Manmade features consist of canal and qanat

systems, and nearly all forms rising above the level of the plain’s surface. The shallow salty marsh (Figure 8) extends as an L-shaped zone divided into two segments. Lying predominantly in the west-central portion of the plain, the marsh area extends south-east until it takes 50

J.A. Neely: Parthian and Sasanian Site and Settlement Patterns

Figure 5. Looking south-east from DL-2, the southern edge of the rocky piedmont zone (fore- and middle ground) may be seen to blend into the alluvial plain zone (middle- and background): the two arrows point to tepe/tell sites situated near the northern edge of the plain

Figure 6. A view looking north-west across the alluvial plain, with the Kuh-i Siah range of the Zagros mountains in the background; the late afternoon October lighting clearly delineates large and small tepe/ tell sites

51

Sasanian Archaeology

Figure 7. A view looking south-west across the alluvial plain: Tepe Musiyan (DL-20), the largest site on the alluvial plain is in the background; note the several generations of canals clearly visible in the fore- and middle ground

a more easterly trend. The second, smaller segment, which forms the easternmost extreme of the L-shaped zone, lies across the Dawairij river and spans at least 15 km. The zone ranges from about one to three km in width and is over 35 km in total length. Characterised by saline soils and sparse low vegetation of highly saltresistant shrubs, the zone is dissected by numerous small erosional cuts. While marshy during the winter rainy season, this micro-environment is essentially dry and salt-encrusted during the summer. This area apparently served, as it does today to a limited extent, as a drainage area for excess irrigation waters.

70% of the plain and the discovery of an estimated 80% of the visible sites and features in each of the microenvironmental zones defined.13 Beyond the discovery of surface pottery, excavation would have been needed to determine the nature of occupations at most large sites. However, the smaller sites and features of the Parthian, Sasanian, and early Islamic periods were conspicuous on the ground surface. Not only could they be located with relative ease, but also with a minimum of trowelling it was often possible to determine and accurately map the various components of the sites and even the internal features of many of the houses (Figures 9–14). The information pertaining to the smaller occupations, which are often not visible or overlooked, may be considered a major contribution of this survey.

In summary, it is important to emphasise three environmental characteristics that affected the economy and settlement patterns: (1) the extreme aridity of the area, (2) the micro-environmental zones of the plain, and (3) the presence of springs and the two rivers available to supply waters for domestic and irrigation purposes.

This ease of location and mapping was probably the result of a number of factors: the relatively recent date of these features, the nature and location of their construction, and the fact that the project geographers determined that the plain has undergone a relatively long period in which little alluviation has taken place.14 Information has been derived in regard to trends in the site (community) and settlement patterns, as well as the water management and irrigation systems

Presentation of data A total of 331 sites were recorded during the archaeological survey of the Deh Luran plain.12 While the entire area of the nearly 1,000 sq km plain was not surveyed, the use of a ‘zone’ and ‘band’ sampling technique resulted in survey coverage of approximately 12 

13 

Neely and Wright 1994: appendix maps 1B, 1D.

14 

52

Neely 1969: 9–11; Neely and Wright 1994: 9. Kirkby 1977; Kirkby and Kirkby 1969: 2–3.

J.A. Neely: Parthian and Sasanian Site and Settlement Patterns

Figure 8. Photograph of a portion of the shallow salty marsh, looking north. This photograph was taken during the month of October just after light rain: traces of mineral salts appear as a thin white crust in the upper middle ground

Figure 9. Site DL-10: photograph of the remains of the Sasanian period structure foundation as discovered by the survey, looking south-west (see Fig. 10)

53

Sasanian Archaeology

Figure 10. Site DL-10: a small irrigation agriculture homestead on the alluvial plain constructed during the Sasanian period: the depression to the north of the structure may have resulted from the excavation of clay as building material, and may have subsequently been used as a small reservoir for domestic water storage.

Figure 11. Site DL-250: an irrigation agriculture compound on the alluvial plain constructed during the Parthian period

54

J.A. Neely: Parthian and Sasanian Site and Settlement Patterns

Figure 12. Site DL-12: founded during the Parthian period, this large site is situated on the north-west part of the alluvial plain

55

Sasanian Archaeology

Figure 13. Site DL-241: an irrigation agriculture homestead on the alluvial plain constructed during the Sasanian period. The semi-circular feature abutting the wall of one of the larger rooms may be a hearth and the rectangular feature in the floor of the easternmost room may be a ceramic coffin. This site was situated atop a mound site dating to the Khazineh and Mehmeh phases (5000-4600 BC: Neely and Wright 1994: 140–41)

utilised during these late occupations of the plain. Before discussing the site and settlement patterns, it is appropriate to present some overview data to place the information into perspective.

canals). In addition, one early road, three springs (two water and one tar/naphtha) and four isolated dry-laid stone field walls were recorded. Of the 263 habitation sites, 123 (47 %) were found to have pottery dating to late occupational components representing periods from the Parthian into medieval Islamic times, and dating from c. 210 BC to the post-10th century. Of these 123 late sites, 75 sites (61%) had two or more occupation components, and 109 (89%) had pottery dating to the Parthian and/or Sasanian periods. 30 (28%) of the 109

Overview of the data Of the 331 sites recorded, 263 (79%) were some form of habitation while 60 were water management features/ systems (15 qanats, 34 canals, and 11 qanats with 56

J.A. Neely: Parthian and Sasanian Site and Settlement Patterns

Figure 14. Site DL-241: photograph of the remains of the Sasanian period structure foundation as discovered by the survey, looking south-south-east (see Figure 13)

unglazed pottery. As will be noted in Table 1, the Optically Stimulated Luminescence dating of a small number of pottery samples,17 and the discovery of a limited number of coins have augmented and made more precise the dating of a few sites.

sites have both Parthian and Sasanian pottery, while only one site (0.01%) had only Parthian pottery and 78 sites (72%) had only Sasanian pottery. Of the 109 sites, 23 sites (21%) appear either as re-occupation or continuity of occupation of earlier sites. Six (26%) of the 23 sites had Parthian and earlier components and 17 (74%) of the 23 sites had Sasanian and later components. Of the 109 Parthian and Sasanian sites, only three (0.03%) were found with only glazed pottery, 35 sites (32%) had both glazed and unglazed (i.e., plain and surface manipulated) pottery, and 71 sites (65%) were found to have only unglazed (i.e., plain and surface manipulated) pottery. There was a great variety in the number and density of pottery found at these late sites, ranging from none to many hundreds of sherds. In general, as one would suspect, the number and density of pottery fragments increased as the site size increased. It was also noted that the number and density of sherds tended to be greater on the later sites.

Unfortunately, the glazed pottery from only 57 (56%) of the 101 sites with glazed pottery could be classified precisely enough to determine the period or periods of occupation (Table 1). This low level of success was due primarily to the presence of only blue-glazed pottery at 44 sites that could not be assigned to a time period more specific than ‘Sasanian/Islamic’. The fact that we could identify diagnostic pottery from 57 sites was, in fact, somewhat of a surprise considering the rural nature of the Deh Luran plain and the small size of most sites.18 Our inability to date the 44 sites was also affected by the lack of sherds present at some and the small size of some sherds. Of the 57 sites with diagnostic glazed pottery, 24 (42%) were multi-component sites containing glazed pottery diagnostic of from two to all five of the late periods (i.e., Parthian and later) identified in Hill’s pottery analysis.19

Table 1 presents a listing of the late sites on the Deh Luran plain considered in this paper. Those sites listed with diagnostic glazed pottery have at present the more precisely dated components.15 We are currently in the process of reclassifying the unglazed (i.e., plainware and surface manipulated) pottery.16 Therefore, for this paper, the dating must remain tentative for the components of those sites listed as containing only

31 (25%) of the 123 sites occupied during these periods had pottery dating to the Parthian period, 108 sites (88%) had pottery dating to the Sasanian and 7th century Islamic periods, 67 (55%) contained pottery Hill 2006. Neely 2011. 19  Hill 2006. 17 

Hill 2006. 16  Neely and Wright in preparation. 15 

18 

57

Sasanian Archaeology Table 1. Chronological placement of late site occupation components on the Deh Luran plain Parthian Sasanian and Eighth and Tenth Post - Tenth Site or 7th Century Ninth Century Century Century Notes and Independent Dating Method (s) Number Earlier Islamic Islamic Islamic Islamic DL-1 DL-2

B, W, X, Ω P, X, Ω

A, ES, C, L, W, PM, SG, Ω

Ω

Ω

LP, Ω

UG-B, Ω

DL-3

Ω

SG, Ω

DL-4

Ω

DL-5

X, Ω

Ω

DL-6

Ω

SG, Ω

Ω

Ω

Ω

Ω

LS, Ω

P, Ω

ES, SG, Ω

LS, Ω

DL-9

Ω

DL-10 DL-12 DL-13

W, Ω

Ω

Habitation site associated with DL-5. UG-B

Ω

DL-16

Ω

ES, Ω

DL-17

P, Ω

A, Ω

DL-18

P B, Ω

DL-24 DL-27

X, Ω

P, Ω

ES, C, WM, PM, Ω

Ω

Ω

Ω

ES, Ω

DL-28 DL-32

Ω Ω

DL-33

Reoccupied early site (Neely and Wright 1994:3133).

Reoccupied early site (Neely et al. In Preparation). Six small mounds or platforms.

One mound with a single, multi-roomed foundation. Large site with many structures visible. Qanat and Canal.

Reoccupied early site (Neely and Wright 1994:4447). Three small mounds bordering a small old canal. L-shaped platform, house foundations, and canals.

Ω

Reoccupied early site (Neely and Wright 1994:4853). Reoccupied early site (Neely and Wright 1994:5767).

LP, LS, Ω

Reoccupied early site (Neely and Wright 1994:7881). Reoccupied early site (Neely and Wright 1994:8384). Reoccupied early site (Neely and Wright 1994:8588).

Ω

Reoccupied early site (Neely and Wright 1994:9799). Cohen 1981.

P, X, Ω

Reoccupied early site (Neely and Wright 1994:99101).

X, Ω

DL-34

B, X, Ω

DL-35

X, Ω

Canal - Gristmill System. Drop-Tower Mill = 680+150 AD (OxL-1349). Isolated Artifact = 1750+100 AD (OxL-1348). NW to SE trending Canal Segment.

Ω

DL-15

DL-20

Probably the largest site on the Deh Luran Plain during these periods. Total area: 113 ha.

Habitation site associated with DL-5.

DL-7 DL-8

Cave – no visible architecture.

Reoccupied early site (Neely and Wright 1994:102-103). 490+180 AD (OxL-1350).

A, Ω

Ω

B, W, X, Ω X, Ω

A, ES, Ω

Ω

DL-36

W, Ω

P, X, Ω

L, SG, Ω

LP, LS, Ω UG-B, Ω

DL-38

W, Ω

P, X, Ω

A, ES, WM, SG, Ω

Large site with many mounds. 830+250 AD (OxL1352). 1290+150 AD (OxL-1351).

LP, LS, Ω Ω

Large site, two platforms, 8 mounds, and canals.

DL-39

Ω

DL-40

Ω

Ω

DL-41

Ω

Ω

DL-43

Ω

Ω

Reoccupied early site (Neely et al. In Preparation). Coins - Parthian Period.

Ω

Large site, similar to DL-38.

A walled town, many mounds, and associated canals. Large site, similar to DL-38.

Reoccupied early site (Neely and Wright 1994:104).

58

J.A. Neely: Parthian and Sasanian Site and Settlement Patterns

Parthian Sasanian and Eighth and Tenth Post - Tenth Site or 7th Century Ninth Century Century Century Notes and Independent Dating Method (s) Number Earlier Islamic Islamic Islamic Islamic DL-44

B

Three small mounds overlaid by modern blacktent camp.

P, Ω

DL-45

Ω

Ω

DL-48

Ω

Ω

DL-49

Ω

Ω

DL-52

Ω

ES, Ω

DL-51

One large and one small mound overlaid by modern black-tent camp.

Small platform with surrounding canal, similar to DL-43. Five small mounds along a canal.

A 25 by 50 m platform with 3 mounds atop.

Three platforms, two with mounds and plazas. Area of site: 3.2 ha.

LP, LS, Ω

DL-53 DL-54

Small mound, probably part of DL-52 or 54.

Ω

DL-55 DL-56

Ω Ω

B, W, Ω

DL-57 DL-58

B, W

DL-62 DL-63

DL-59

Reoccupied early site (Neely and Wright 1994:105-109).

Mound (22 by 30 m), little pottery, much lithics.

P, X, Ω

SG, Ω

Ω

Ω

P, Ω

Large site, many mounds (one 5 m high), walls, columns, canals.

Ω

Six small mounds in two clusters, house foundations visible.

One 90 by 160 m mound bordered by canals.

X, Ω

ES, SG, Ω

B, W

P, Ω

C, ES, Ω

LP, Ω

Ω

Ω

SG, Ω

Ω

DL-71

ES, Ω

Ω

DL-72

SG, Ω

Ω

DL-73

SG, Ω

Ω

A, ES, WM, SG, Ω

LP, LS, Ω UG-B, Ω

DL-68

DL-74

B, W, Ω

P

P, Ω

DL-75

Ω

DL-85

Ω

Ω

DL-88

Ω

Ω

DL-89

Ω

Ω

DL-90

P, Ω

DL-100

P, Ω

Ω

DL-104

Ω

Ω

DL-105

Ω

Ω

DL-108

Ω

Ω

Ω

Ω

DL-111 DL-112 DL-113

W, X, Ω

Large platform with 14 mounds atop.

Reoccupied early site (Neely and Wright 1994:112-115).

Large site, 78 small & 7 medium-size mounds.

Twelve by 30 m mound next to canals.

Reoccupied early site (Neely and Wright 1994:116-117).

Fifty-seven by 62 m platform, 7 mounds, foundations, walls.

Two mounds (24 by 32 and 10 by 33 m). Site area: 1.1 ha. Large site, large platform, many mounds. Sites DL-72, 73, 74 all one site? 1230+330 AD (OxL1353).

Platform (215 m circumf.) with two mounds atop. Reoccupied early site (Neely and Wright 1994:119-123). Small site next to a canal.

Reoccupied early site (Neely and Wright 1994:123).

Reoccupied early site (Neely and Wright 1994:123).

Small site between two secondary canals. Reoccupied early site (Neely and Wright 1994:124-130). Reoccupied early site (Neely and Wright 1994:124-130). Small site between two old canals.

Platform (75 by 90 m) supporting 3 mounds. Site area: 2.9ha.

One mound (45 by 60 m). May be part of DL-104, 105, 108, 111 complex.

Ω Ω

Ω

59

Small site located between primary and secondary canals.

Sasanian Archaeology

Parthian Sasanian and Eighth and Tenth Post - Tenth Site or 7th Century Ninth Century Century Century Notes and Independent Dating Method (s) Number Earlier Islamic Islamic Islamic Islamic DL-132

DL-138

W, Ω

Ω

Ω

DL-150

Ω

DL-171

Ω

DL-172

Ω

DL-194

Ω

Ω

DL-195

Ω Ω

Ω

DL-220

Ω

DL-222

X, Ω

DL-223

Ω

DL-224

Ω

DL-225

B

Dry-farming hamlet, 2 farm-houses with compounds, many terraces and check dams.

Site (area: 1.2 ha) with small mounds and foundations, lies between canals.

Site with one mound (9 by 25 m) and several foundations. Reoccupied early site (Neely and Wright 1994:138-139).

SG, Ω

Mound (25 by 32 m) with 7-room farmhouse and compound (19 by 25 m) atop.

Circular (Diam: 9.5 m) and rectangular (3 by 5.5 m) stone foundations among many terraces..

Ω GLAZED TILE

DL-227

Ω

DL-228

Ω

DL-230

Ω

ES, Ω

DL-231

Ω

Ω

DL-232

Ω

DL-233

Ω

DL-235

B

Three small mounds, one foundation (Total circumference: 400 m).

Small mound, 2 foundations, with terraces and check dams.

Two small (4 by 7 m) mounds near drainage with checkdams. LS, Ω

Ω

Ω

Ω

DL-240

Ω

DL-241

Ω Ω

DL-243

P, Ω

Platform (70 by 70 m) bordered to N and S by alignments of one-meter square rock-piles.

Mound (7 by 9 m) with terraces and rock-piles.

Platform (9 by 12 m) and foundation (5 by 10 m) with terraces and check dams.

Reoccupied early site (Neely and Wright 1994:139).

Reoccupied early site (Neely and Wright 1994:140-141). ES, Ω

Platform (22 by 30 m) and house foundations. Site circumference: 300 m.

LS

Ω B

Ω

Three+ room farmhouse (5 by 14 m) with compound.

Small platform with 3-room foundation & compound atop.

ES, Ω

Ω

Mound (40 by 50 m) with foundation atop.

Many dry-farming terraces and check dams (Area: 5 ha).

Ω

DL-239

Three mounds, one with 4-room farmhouse with compound atop.

Mound (25 by 25m) of boulders in mortar, on high terrace above Mehmeh River.

ES, Ω

DL-236

DL-244

Dry-farming hamlet in mouth of canyon, 3 farmhouses with compounds.

ES, WM, PM, Ω

Ω

DL-242

Piedmont site, with small qanat –irrigated (?) fields.

Small dry-farming hamlet with terraces and check dams.

DL-226

DL-238

Piedmont dry-farming and herding site.

Structure (15 by 18m), north of many dryfarming terraces. Probably a part of DL-172.

Ω

DL-219

Ω

Small farming site near edge of alluvium.

Ω

Platform (5 by 10 m) with house foundation atop. PM, Ω

Complex of 12 small mounds. Site circumference: 585 m.

Ω

60

J.A. Neely: Parthian and Sasanian Site and Settlement Patterns

Parthian Sasanian and Eighth and Tenth Post - Tenth Site or 7th Century Ninth Century Century Century Notes and Independent Dating Method (s) Number Earlier Islamic Islamic Islamic Islamic DL-247 DL-248

Reoccupied early site (Neely and Wright 1994:146-147).

Ω

DL-250

B, Ω

Ω

DL-251

B

Ω

DL-252

W

Ω

C, Ω

DL-253

Ω

ES, Ω

DL-254

Ω

ES, L, Ω

DL-255

Ω

DL-256

B

Ω

DL-262

P, Ω

DL-263

Ω

DL-264

Ω

DL-272

P, Ω

DL-274

Ω

DL-278

Ω

DL-275

Ω P

DL-286

P, Ω

Seven small platforms (2 are L-shaped). Site area: 1.44 ha. Farm hamlet, 3 structures with compounds and field terraces.

LP, LS, Ω

Platform (40 by 85 m) with a L-shaped & oval mound atop.

Mound (15 by 24 m), between qanat shaft branching.

Small mound (8 by 20 m), near canal, overlying qanat.

Two small mounds with one foundation eroding out.

Reoccupied early site (Neely and Wright 1994:147-148).

Mound (8 by 23 m), much chipped stone.

Two small mounds. Site circumference: 350 m.

Small farming site next to an old canal.

Small farming site next to an old canal. South of DL-271.

Ω

Small farming site. Old canals and fields visible.

Ω

Small farming site. Old canals and fields visible.

Small farming site between two large canals.

Small farming site between two large canals.

Ω

Small farming site between two small canals.

Reoccupied early site (Neely and Wright 1994:149-155).

Ω

DL-290 DL-291

Eight small house mounds along an old canal.

Ω

DL-271

DL-281

Platform with 2-room house and compound (18 by 25 m) atop.

LS

Ω

DL-257

DL-282

Reoccupied early site (Neely and Wright 1994:142-145).

ES, Ω

Reoccupied early site (Neely and Wright 1994:155).

A, Ω Ω

Ω

DL-294

Ω

DL-298

P, Ω

DL-303

One platform (30 by 70 m), near a canal branching.

Ω

DL-305

Ω

DL-307

Ω

DL-310

Ω

DL-311

Ω

Four-room house foundation with compound. Total size: 10 by 13 m.

Ω ES, Ω

DL-304

Platform (28 by 35 m) with mound (15 by 35 m) atop.

Ω

At least 9 stone-founded semi-circular structures. Site circumference: 350 m.

Ω

House foundation with compound (25 by 40 m). Site circumference: 500 m.

Three-room house foundation. Total size: 4 by 8 m.

A, Ω

Two house foundations (3 by 7 m & 4 by 5 m) with 2 field walls (55m & 1,350 m).

Two multi-room house foundations and compounds. Total site area: 0.72 ha.

A 6-room structure (4 by 14 m), much chipped stone.

61

Sasanian Archaeology

Parthian Sasanian and Eighth and Tenth Post - Tenth Site or 7th Century Ninth Century Century Century Notes and Independent Dating Method (s) Number Earlier Islamic Islamic Islamic Islamic DL-312

Reoccupied early site (Neely and Wright 1994:160-161).

Ω

SYMBOL KEY: A: APPLIQUE DECORATION. LP: PAINTED DESIGNS / LINES / PSEUDO-KUFIC ON GLAZED BACKGROUND. B: BLACK GLAZE. L: LUSTER. W: WHITE GLAZE. WM: WHITE MAJOLICA. P: PEA-GREEN / OLIVE GLAZE. PM: PRESS MOLDED. B/B: BLUE GLAZE ONE SURFACE / BLACK OBVERSE. SG: SGRAFFIATO. LP: LATE PAINTED. UG-B: UNDERGLAZE BLUE. ES: EARLY SPLASH (WHITE OR YELLOW - BASE GLAZE). X: DIAGNOSTIC RIM (APPLIED TO SASANIAN AND EARLIER CERAMICS). LS: LATE SPLASH (OLIVE / PEA-GREEN - BASE GLAZE). Ω: UNGLAZED PLAIN AND/OR SURFACE TEXTURED CERAMICS C: COBALT. NOTE: Rows showing no symbols, “blank rows”, represent sites with one or more undiagnostic turquoise-colored glazed sherds, and as such could represent any time between the Late Parthian and the tenth century Islamic periods. This table is modified from Hill (2006:Table B.10).

belonging to the 8th and 9th centuries, 27 (22%) were characterised by 10th century Islamic pottery, and only 8 sites (7%) were found to have post-10th century Islamic pottery.

plain, are linear in nature owing to the sites being located along canals coming from the Mehmeh and Dawairij rivers. This linear pattern had been long established in these areas before the Parthian period. Neely and Wright documented these linear site distributions and contemporary associated canals on the Deh Luran plain dating as early as the Chogha Mami Transitional phase (c. 5400–5200 BC).22 Qanat systems may have been introduced to the plain to augment already existing canals supplying domestic and irrigation waters perhaps as early as 2350 BC. This possible early use of qanat technology is based on the apparent association of a qanat/canal system with a large Early Dynastic I–III phase site (DL-34) and surrounding fields which would have lacked a water supply if the qanat/canal system had not been present.23

The Parthian period (c. 210 BC–AD 225) The Parthian period is represented by at least 31 sites on the Deh Luran plain. Eight (26%) of these sites, two in the north-west, one in the south-west, and five in the south-east, were continually occupied or re-occupied sites which had been founded in earlier periods. Ten (32%) of the Parthian sites were found to have only glazed pottery, eight (26%) had only unglazed pottery, while 13 (42%) had both glazed and unglazed sherds on their surfaces. Site and settlement patterns

The six sites not falling into the two main clusters are loosely aligned from north to south down the centre of the plain. Sites DL-2 and DL-172, at the northern extreme of the piedmont, were located on natural drainages, and probably had water resources from mountain runoff and springs. Three other sites were located along a drainage/canal carrying waters from the Ab-i Garm spring.24 The sixth site, DL-32,25 is a reoccupied site which probably received water from the Mehmeh river. Site DL-2 was evidently founded during this period and apparently becomes the largest late occupation on the plain, although it may not have attained that status until the Sasanian period.

The site and settlement pattern appears to reflect the changes noted by Wenke for the same period on the Susiana plain,20 and by Adams for the Diyala region of Iraq.21 This change was from large mounded sites (tells or tepes) that were frequently walled, to growing numbers of unwalled sites composed of many small buildings, often laid out apparently without regard to any overall settlement plan. This type of site increased in size and number during the Sasanian period. As may be seen on the site distribution map for this period (Figure 15), there are two main clusters of sites as well as a few scattered sites on the plain. The main clusters, in the north-west and south-east parts of the 20  21 

Neely and Wright 1994. Neely and Wright 1994: 190; Wright and Neely (eds) 2010. 24  Neely and Wright 1994. 25  Cohen 1981; Wright 1981. 22  23 

Wenke 1981: 313; 1987: 255. Adams 1965: 73.

62

J.A. Neely: Parthian and Sasanian Site and Settlement Patterns

Figure 15. The Deh Luran plain, with the distribution of dated sites for the Parthian period (c. 210 BC– AD 225): owing to the small scale of this map, the site numbers have not been indicated

ranged from 0.008 hectares to 18 hectares, with a mean of c. 2.1 hectares. Small sites were in the majority and often comprised one well-defined structure (e.g., see Figure 11). However, several of the larger sites were also found with well-defined multiple structures (e.g., see Figure 12). It is expected that the maximum range and mean of the dwelling areas will increase because we have not yet calculated the area of occupation of site DL-2, probably the largest site occupied during this period. Moreover, because of our ongoing study, I have demurred from providing total population estimates for this and subsequent periods, but plan to do so in our forthcoming final report.26 As in our previous report,27 we will attempt to work out settlement dynamics by implementing the probabilistic model of site abandonment and creation devised by Dewar for this and the Sasanian period.28

During the Parthian period, there is a striking paucity of sites in the piedmont zone and central portions of the alluvial plain zone, suggesting a heavy dependence on canal irrigation from the rivers and a lack of focus on dry-farming. However, the few Parthian sites recorded on the piedmont are perhaps misleading owing to the absence of diagnostic pottery and the ephemeral remains of tent camps characterising pastoralist occupation. The likelihood of transhumant pastoralists occupying the piedmont zone during this and subsequent periods is discussed below. Measurable site areas for occupations of only one period were usually determinable. However, the site area for each period of occupation on multi-component sites was difficult and resulted in only rough estimates. Therefore the following data for all periods should be seen as approximate. During the Parthian period the site areas ranged from 0.02 hectares to about 50 hectares, with a mean of c. 7.0 hectares. Dwelling areas

Neely and Wright in preparation. Neely and Wright 1994: 200–10. 28  See Dewar 1991. 26  27 

63

Sasanian Archaeology

Figure 16. Site DL-194: a dry-farming homestead in the piedmont zone constructed during the Sasanian period. The structures shown at the northern extreme of this map form part of a hill-top site

Four observations were made relative to the Parthian period settlements. However, due to as yet uncertain identification of unglazed pottery, the third and fourth observations may more appropriately apply to the Sasanian period, and therefore are discussed in detail in the following section. First, the number of sites increases from the previous periods.29 Second, most sites decreased in size from previous periods. Third, this is the first period in which we find evidence of small alluvial plain irrigation farming homesteads (described below) with a well-defined structure and compound arrangement (see Figure 11). Finally, this 29 

is the first period in which evidence of piedmont dryfarming homestead complexes (described below) with a well-defined structure and compound arrangement (Figure 16) are found. Water management and irrigation systems The dating of Parthian and Sasanian canals has only been tentatively accomplished through association with adjacent sites. Most of the canals assigned to these periods were either in continuous use or were refurbishments of canals evidently dating to earlier times. Thus, the dimensions (i.e., length, width, depth) of the canals at any specific point in time were

Wright and Neely (eds) 2010.

64

J.A. Neely: Parthian and Sasanian Site and Settlement Patterns The sites were smaller, more numerous, and more dispersed across the landscape than in previous periods. Sites also varied greatly in size from small one-family farmhouses (e.g., DL-10: Figures 9–10), some of which were less than 100 sq m in floor area, to an urban centre (DL-2) of about 112 hectares. For the reasons outlined above, the following data should be seen only as approximations. For this period, site areas range from 0.005 hectares to about 112 hectares, with a mean of 8.8 hectares. Dwelling areas range from 0.003 hectares to 5.0 hectares, with a mean of 0.69 hectares. It is expected that the dwelling area maximum range and mean size will increase because the area of occupation of site DL-2, the largest site occupied during this period, has not yet been determined.

extremely difficult, if not impossible, to determine. As a result, the dimensions we present here are wellconsidered estimates which must await excavation for verification. Maintaining an already long-standing practice, the placement of sites was in close proximity to canals, presumably to obtain water for domestic needs. Although somewhat later and located about 120 km to the south-east of Deh Luran, the canal and inverted siphon system at the site of Jundi Shapur is perhaps the most technologically sophisticated example of the use of canal systems to meet the need for domestic water as well as the demands for irrigation.30 In association with the late period sites, the 1969 survey found each of the three canal systems (the Mehmeh, the Ab-i Garm, and the Dawairij) to be over 25 km in length. The primary canals were large, ranging from approximately 5 to 20 m in width, with an average width of about 10 m.

The mean site area of about 8.8 hectares is 1.8 hectares larger than the mean area of the Parthian period sites. This, in addition to the increase in the total number of apparently contemporaneously occupied sites (from 31 to 108), suggests an increase in population. Because of our ongoing study, I have not provided total population estimates in this paper, but plan to do so in our final report.31

The Sasanian and 7th century Islamic periods (AD 225–c. 700) Because of the difficulty in distinguishing Sasanian pottery from that of the 7th century Islamic period, I have, for the time being, combined these two periods for this paper. This segment of time is represented by the largest number of sites (108) recorded on the 1969 survey. Table 1 documents that two sites (2%) from this period are represented by only glazed pottery, 81 (75%) were found with only unglazed wares, and 25 (23%) were recorded as having both glazed and unglazed pottery. 17 (16%) of the 108 sites documented for this period were re-occupations of earlier sites which had not been occupied during the Parthian period. The west and east clusters each had five re-occupied sites, while the central cluster had seven.

When compared with other studies of sites dating to the Parthian and Sasanian periods,32 the majority of the Deh Luran Parthian and Sasanian sites were small. Wenke reports that there are few settlements in any period on the Susiana plain which are less than one-half hectare in area,33 but the Deh Luran plain has many with areas of one-half hectare and less. It has not been determined if the presence of these small sites is unusual, or if their discovery was due to the near absence of alluviation on the plain for the last several hundred years and the relatively undisturbed condition of the plain at the time of the 1969 survey.

Site and settlement patterns On the alluvium, the Sasanian and 7th century Islamic site and settlement patterns, in general, were not radically different from those of the preceding period (compare Figures 15, 17). However, the number of sites had increased, with the north-west portion having, by far, the densest grouping. However, the settlement pattern and site density on the piedmont changed dramatically. A comparison of Figures 15 and 17 indicates an increase from two to nine dated sites, but an examination of Figure 18 and Appendix map 1B in Neely and Wright (1994), gives a better idea of the true site density as most of the piedmont sites lack diagnostic pottery and remain attributable only to Parthian/Sasanian/ Islamic times. The majority of the piedmont sites were the dry-farming homesteads described below.

Most of the sites on the Deh Luran alluvial plain conform to the description provided by Adams for sites in the Diyala region ‘low and sprawling, with irregular shapes and indefinite contours … occupational remains extend in thin bands for considerable distances along old canal levees or crop up sparsely at intervals separated by apparently uninhabited areas’.34 Because of indefinite site boundaries and the close proximity of sites one to another, it is entirely possible that some features recorded in 1969 as multiple sites were originally part of single larger sites. Large structures consisted of platforms and low mounds of earth constructed of unfired and fired brick. Small structures are usually represented by foundations of cobbles, boulders, and occasionally stone slabs, which today lie partially Neely and Wright in preparation. e.g., Adams 1965; Wenke 1975/76; 1987. 33  Wenke 1987: 261. 34  Adams 1965: 73. 31  32 

30 

Adams and Hansen 1968: 59–63.

65

Sasanian Archaeology

Figure 17. The Deh Luran plain, with the distribution of dated sites for the Sasanian and the 7th century Islamic periods (AD 225–c. 700): owing to the small scale of this map, the site numbers have not been indicated.

buried in the alluvium of the plain and the clay eroded from the upper walls. Based on a few preserved upper wall fragments, and ethnographic building techniques seen in Deh Luran, the upper walls of most structures and compounds were probably constructed of bricks of straw-tempered unfired clay. The upper walls of some less well made structures and compounds may also have been constructed with straw-tempered unfired ‘turtlebacks’ (i.e., irregular large globs of clay with nearly parallel longitudinal sides shaped by scraping, a convex upper surface, and a concave lower surface).

evidently parallels both Adams’ Diyala findings and Wenke’s Susiana observations,35 this situation contrasts with other areas of Sasanian occupation: for example, on the Mughan steppe in north-west Iran and southern Azerbaijan36 and the Gorgan plain,37 many of the sites, both large and small, have been found with fortification walls. An important contribution of the Deh Luran survey was the recording of the small Partho-Sasanian compound homesteads (Figures 11, 16) which especially characterised the western and central portions of the piedmont and the western alluvial plain area. The majority of these compound units were located in previously occupied portions of the Deh Luran plain, but many were constructed on previously unoccupied land. These compounds may represent the single or

With the exception of the compound sites (see below), the walled communities recorded by our Deh Luran survey appear to have been constructed before or after the Partho-Sasanian occupations. DL-2, evidently the largest site (c. 112 hectares) on the Deh Luran plain during the Sasanian period, has no evidence of being a walled community, and was sited in a broad canyon not suitable for defence. While this Deh Luran pattern

Adams 1965: 73; Wenke 1981: 313; 1987: 255. Alizadeh and Ur 2007; Ur and Alizadeh, this volume. 37  Kiani 1982b. 35  36 

66

J.A. Neely: Parthian and Sasanian Site and Settlement Patterns

Figure 18. Detail map of the north-central portion of the Deh Luran plain: note the canal system (DL-5) with its drop-tower gristmills.

extended family versions of the larger earlier and later walled sites that have been found on the plain. A question arising is whether these compounds were unique to Deh Luran, or if they were present elsewhere and, for various reasons, have not been recognised or recorded? That question has recently been answered by St John Simpson, who referred me to the work of Elisabetta Valtz at Tell Mahmud, a site located a little over 1 km east of Tell Yelkhi in the Hamrin Dam Project

area of east-central Iraq and about 275 km north-west of Deh Luran.38 Valtz briefly reports the excavation of a complete structure at Tell Mahmud which is quite similar to some of the Deh Luran structures with compound walls (e.g., Figures 11, 16).39 Although the Tell Mahmud example is about twice the size of its 38  39 

67

Valtz 1985; Invernizzi 1980: fig. B; see also Fedele, this volume. Valtz 1985.

Sasanian Archaeology Deh Luran counterparts, the date of the structure, its location and nature of the exposed complex of rooms within the compound, and the orientation of the structure and compound are quite similar. It would be interesting to learn if the Tell Mahmud structure was built on the mound/tell of an earlier occupation as was found with site DL-241 (see Figures 13–14). The economic and socio-political relationships between the occupants of these widely separated areas with similar structures deserve further study.

in the western part of the plain, and the introduction of dry-farming water and soil management technology to the piedmont zone. These modifications to the water management systems allowed the inhabitants of the Deh Luran plain to occupy and use additional surface areas of the plain. An estimated 400 sq km, or approximately 40% of the total area of the plain, was under cultivation during the Sasanian period. Of the approximately 40,000 hectares of land being utilised, about 30,000 hectares were under irrigation, while about 10,000 were piedmont dryfarming areas utilising the water harvesting techniques noted below. Thus, the Sasanian and 7th century medieval periods appear to have been the apex in terms of the total amount and intensity of land use as well as maximum population density on the plain. In spite of the different nature of the survey data available from the Diyala, Upper Khuzistan (Susiana), Mughan steppe, and Deh Luran regions, the patterns of intensity of land use, technological maximisation, and high population densities appear to be similar.

It has been noted that there is a distinct possibility that the transhumant round practised today by the Lurish pastoralists between the Khorrambad valley and Deh Luran plain extends back in time perhaps as early as 4800 BC.40 It seems quite possible that pre-Partho-Sasanian pastoral encampments on the Deh Luran piedmont and alluvium may have been tent sites as they are today, such as Hole found at Tepe Tula’i.41 Considering this long history of transhumance in this region, we should take under consideration that a transhumant pastoral group or groups, living on the Deh Luran plain primarily during the winter, occupied at least some of the piedmont and alluvium homesteads. Hole has noted that the Mehmeh Phase (4800–4600 BC) alluvial plain houses mapped at Tepe Ashrafabad have a general plan layout which ‘resembles that of a tent rather than a mud-walled house’,42 and perhaps indicates, as is found today, that some of the transhumant pastoralists spent seasonal residence in settled villages. Mortensen illustrates structures and compounds forming the village of Kahreh that he attributes to the Lurs in the Hulailan valley, about 100 km north of Deh Luran, that are quite similar in plan view to those illustrated in Figures 11 and 16.43 Also, the structures with associated compounds which largely characterised the village of Deh Luran during our fieldwork bore a close resemblance to those illustrated in Figures 11 and 16, but were larger in size and area. The distinct possibility that pastoralists inhabited piedmont and alluvial sites immediately raises the related, and very important, question of how much the subsistence and economy of the plain’s inhabitants depended on domestic animals?

Canals and qanats appear to have increased in number during this period. New systems of canals and qanats, tapping springs and the subterranean aquifers, appear on the piedmont (e.g., Figure 18).44 The water harvesting of piedmont runoff may also have been introduced during this period.45 In addition, if not introduced before,46 it seems probable that it was during this period that some of the late Deh Luran qanat systems were constructed to incorporate an ingenious modification of the qanat technology. This modification was to obtain seepage waters from the Mehmeh and Dawairij rivers, rather than to tap underground aquifers as traditional qanat technology does. The modified qanats did not take water directly from the rivers, but ran parallel to them at distances from 10 to 50 m for stretches of 200 m to almost two km, and then turned in toward the centre of the alluvial plain. Water was obtained from the rivers as it percolated through the soil and rock of the channel banks into the qanat. This variant technology may well have been contrived to serve four purposes. First, it greatly reduced the amount of silt carried into the systems, a problem which must have been enormous in canal systems taking water directly from the Mehmeh and Dawairij rivers by means of diversion dams or weirs. Second, the percolation of the water through the riverbanks filtered out vegetal matter and minerals in suspension. Third, it is conceivable that this filtering process resulted in the deposition of dissolved minerals in solution through evaporation as hardened calcium carbonate (caliche) layers at the top of the aqueous zone. Fourth, although the inhabitants of Deh Luran

Water management and irrigation systems Modifications in the water management technology were evidently correlated with variations in the topography, settlement pattern, and the apparent population increase. These modifications consisted of the expansion of old canal and qanat systems, the addition of new canals/qanats in nearly all portions of the plain, the construction of terraces and check dams Hole 1978; 1979; 1987; Hole, Flannery and Neely 1969: 349–50. Hole 1974; 1987: 83. 42  Hole 1987: 36; Neely and Wright 1994: 88–94. 43  Mortensen 1993: figs 6.50–6.51. 40 

See Neely and Wright 1994: appendix map 1B. e.g., canal DL-330 in Neely and Wright 1994: appendix map 1B. 46  Neely and Wright 1994: 200; Wright and Neely (eds) 2010.

41 

44  45 

68

J.A. Neely: Parthian and Sasanian Site and Settlement Patterns

Figure 19. Low dry-farming terraces located at the north-west corner of site DL-194 and bordering the drainage forming the site’s western boundary (see Figure 16), looking north-east

Rainfall could easily have been augmented with water diverted from nearby canals, and it is conceivable that the labour-intensive practice of hand-watering may have been used on the terraces upslope from the canals.48

were undoubtedly unaware of it, the filtration of water through riverbanks has recently been documented as serving to remove microbes and other pollutants (e.g., heavy metals) from water supplies.47 Three of these processes reduced the frequency of the need to clean the systems, and slowed the soil salinisation which probably gradually affected crop production. The fourth benefited the health of the inhabitants of the now densely populated plain.

Dry-farming water management systems The piedmont micro-environment was converted from an evidently sparsely inhabited zone to one of rather dense occupation with the introduction of a new array of water management technologies based on water harvesting during the Sasanian period, although there is some evidence that this process may have been initiated in the Parthian period. As seen in Figures 17–18,49 this occupation was focused in the western and central parts of the piedmont.

In the western portion of the plain, a second, less impressive but nevertheless important, water management technology was also apparently introduced at this time. This involved the construction of numerous low terraces and check-dams. The low drylaid stone terraces were associated with the alluvial plain sites, near many of the intermittent drainages dissecting the plain, and frequently constructed on the banks to parallel the drainage. Small dry-laid stone check-dams were constructed within many of the drainages. These terraces and check-dams appear to have functioned in a manner similar to the dry-farming features to be described below for the piedmont zone. Both the terraces and the check-dams would have served well for the retention and annual rejuvenation of soils as well as maintaining a higher level of moisture in the soils. However, considering their locations vis-àvis canals, it is perhaps more likely that they functioned in combination with more usual irrigation techniques. 47 

Terraces and check-dams were introduced to conserve and renew soils, as well as more efficiently distribute and retain rainfall and runoff waters on sections of the piedmont by ingenious systems of water harvesting. Low terrace walls of unmodified dry-laid cobbles and boulders were built at intervals varying with the slope of the terrain and following the contours of the land (Figure 19). An integral part of this system were 48  See Caran and Neely 2006; Castetter and Bell 1942; Doolittle 2000; Kirkby 1973; Neely 2005a; 2005b; Neely and Caran 2011; Neely, Caran and Winsborough 1990. 49  See also Neely and Wright 1994: appendix map 1B for undated sites not cited in Table 1, nor illustrated in Figure 17.

Hubbs (ed.) 2004; Ray, Melin and Linsky (eds) 2002.

69

Sasanian Archaeology

Figure 20. Check-dams, or cross-channel terraces, located in the drainage forming the western boundary of DL-194 (see Figure 16), looking north-west

small check-dams, or cross-channel terraces, of drylaid, unmodified cobbles and boulders, constructed at right angles to the flow across intermittent drainages dissecting the piedmont (Figure 20). Water and soil washing from the piedmont slopes would flow into the drainages. The check-dams functioned to retain soils and retard runoff so that water would thoroughly soak these soils. Because the small plots behind the checkdams had relatively deep soil deposits, and the fact that the plots received greater amounts of rejuvenating soils and water than the slope terraces, these small areas probably had greater crop productivity and reliability. Such features were likely also constructed to retard headword erosion and the deterioration of the piedmont slope fields. The only negative aspect of these dry-farming systems was that the terraces and the dams were exposed to destruction by heavy runoff. However, in Deh Luran most were found to be in a remarkably good state of preservation.

unit of houses, corrals, farming terraces, and checkdams is usually clearly separated by some 25 to 100 m from the next. In many cases this separation is further defined by an intermittent drainage, or by a low wall of unmodified, dry-laid cobbles and boulders. Sites apparently associated with the piedmont homesteads were situated on nearby hilltops (Figure 16). These hilltop sites were comprised of several structures, and were found distributed one for every four to six piedmont homesteads. Their locations would have made them good defensible refuges, although there is no other evidence to foster that interpretation. From these dry-farming homestead complexes we may derive information beyond the apparent technological and architectural data. For example, the ratio of habitation area to the area of cultivation may be determined, and thereby we may obtain an idea of the area needed to sustain a more accurately specified population through the use of dry-farming techniques. By comparing these ratios with crop productivity estimates, we may be able to determine if the piedmont homesteads utilising seasonal dry-farming techniques were able to derive a surplus, and perhaps address such questions as: were there enough crops to use for trade or barter? In addition, this ratio may be compared with the irrigation agriculture homesteads on the alluvial plain to determine relative crop productivity. It is, of course, recognised that such calculations are very tentative due to the problems in estimating crop productivity, living area, and population size, as well as our poor

The piedmont agricultural terraces and check-dam systems clustered in distinct units associated with one or more habitation structures (usually having a walled courtyard/compound), one or more small ‘storage’ structures, and frequently one or more rectangular structures reminiscent of cattle pens or ‘corrals’. Figure 16 is a plane-table map of site DL-194 and illustrates a classic example of one of these units. These are thought to be distinct socioeconomic units that may be referred to as: ‘dry-farming homesteads’. Justification for this proposition is the fact that each 70

J.A. Neely: Parthian and Sasanian Site and Settlement Patterns salts and plant root rot. The northernmost secondary canal supplied site DL-274. Excess water drained into the other secondary canal to be utilised by site DL-275, or passed to the next homestead or hamlet unit to the south-east. This system may have drained into the salty marshes lying to the south-east, or into another canal system running south-east from the Mehmeh river to supply water to the southern segment of the plain. One very important function of the latter, southern canal system mentioned may have been to drain the salty marsh to retard salt buildup and possibly to reclaim some of the marsh for agriculture.

knowledge of Partho-Sasanian farming techniques and land tenure. There is also the question of the relative importance of dry-farming vis-à-vis herding. However, they do provide interesting first approximations that may be refined as data increase and improve. Using site DL-194 (Figure 16) as an example of a piedmont homestead, we find the total living area (including the walled courtyards/compounds) of the two structures to be 1,088 sq m, or 289 sq m if we only consider the area of the definable rooms. The ratio of total living area (including compounds) to the area under dry-farming cultivation (totaling 16.88 hectares) at DL-194 is about 1:155 sq m, or about 1:584 sq m if only the area of the definable rooms is used.

Returning to the comparison of piedmont site DL-194 with alluvium homesteads DL-274 and DL-275, it must be noted that the comparison is not between equivalent sets of data. Regrettably, it was not possible to find homesteads on the alluvium which displayed both field areas and clearly exposed measurable structures. The total living areas of DL-194 and DL-195 (the house mounds and surrounding areas having building debris and a relatively dense sherd scatter) were approximately 7,174 sq m and 9,190 sq m, respectively. The areas under irrigated cultivation are estimated at 124.1 ha for DL-274 and 202.2 hectares for DL-275. The ratio of total living area to the area under irrigated cultivation is 1:173 sq m for DL-274 and 1:220 sq m for DL-275. Sites DL274 and 275 would have been large enough to support a maximum of six and eight compounds respectively, the size of those found at DL-194. If excavations reveal the nature (i.e., number of structures and rooms) of the structures comprising DL-274 and 275, or if similar sites can be located with preserved structures visible on their surfaces, population estimates can be calculated. Unfortunately, since such information is not available, a population estimate for these two sites has not been attempted.

Sites and structures of this type and size are not usual in most of the population size calculations conducted for Middle Eastern sites.50 Until we complete our study of the late periods of Deh Luran plain occupation,51 we are unable to more accurately work out the settlement dynamics for this period. However, until we can achieve that goal, I have chosen a reasonable method from a study from the American South-west52 for estimating populations based on the number of rooms present. From this method, a population of between 22 and 25 individuals for DL-194 may be estimated. These two figures represent the extremes, calculated assuming that both structures were occupied contemporaneously and that half of the rooms were used as living quarters. Despite the use of a method from outside the Middle East, I believe these two population figures are more accurate than using other methods such as Naroll’s cross-cultural method, which projects a population of 29 persons,53 or the Naroll ‘corrected’ method which projects a population of between 38 and 61 persons.54 A comparison may be made between the preceding settlement unit and contemporary homesteads or hamlet units based on irrigation agriculture. These latter sites were found on the alluvium near the Mehmeh river in the south-western quadrant of the Deh Luran plain. Figure 21 illustrates these two units, DL-274 and DL-275; composed of small, low house mounds and associated fields irrigated by canals. Secondary canals which derived water from primary canal and qanat systems divided the units from one another. Smaller feeder/field canals branched off the secondary canals to form a dendritic pattern in the cultivated fields of each homestead or hamlet unit. The secondary canals also apparently supplied water to several similar sites, and probably also drained excess water from the fields to retard the buildup of

Spring-fed canals with drop-tower gristmills As with the Mehmeh and Dawairij rivers, the waters of the Ab-i Garm springs (DL-170, see Figure 18) were managed from an early period.55 These springs are located high in the piedmont near its juncture with the Kuh-i Siah range of the Zagros mountains. While there are a number of other springs similarly located along this break in the topography, evidently only the Ab-i Garm sources, north-north-east of the present town of Deh Luran, were managed in the manner to be described. Perhaps beginning during the Chogha Mami Transitional phase (c. 5400–5200 BC), but quite likely by the Khazineh phase (c. 5000–4800 BC), the management of Ab-i Garm waters appears to have been focused on

e.g., Adams 1965: 23–25; Pasciuti and Chase-Dunn 2002. Neely and Wright in preparation. 52  Hill 1970. 53  Naroll 1962. 54  Brown 1987. 50  51 

55 

71

Neely and Wright 1994: 187–88.

Sasanian Archaeology

Figure 21. Sites DL-274, DL-275: irrigation agriculture homesteads on the alluvial plain constructed during the Sasanian period

slight modifications of its natural drainage channels.56 The first recognisable canal excavation (DL-5: see Figure 18) conducted to carry Ab-i Garm waters appears to have taken place during either the Sargarab phase (c. 4000–3750 BC) or more likely the Early Uruk period (c. 3750–3500 BC) to convey waters to the site of Sargarab.57

Garm channel were particularly broad or deep, and would thereby require the excavation and maintenance of many metres of additional canal to follow the contours properly, an aqueduct of mortared masonry was constructed to span the secondary drainage. The small canal at first paralleled the Ab-i Garm natural drainage and then continued to course southward while the natural drainage of the Ab-i Garm took a more south-westerly course (Figure 18).

This small canal system was apparently modified and expanded during the Sasanian period.58 North of DL169 where secondary natural drainages joining the Ab-i

The spring waters were diverted into the canal system either at the spring-heads or just downstream. There was no evidence as to exactly how this diversion was accomplished, but a diversion dam or weir seems most

See DL-276A and 276B in Neely and Wright 1994: appendix map 1B. DL-169: Neely and Wright 1994: 130–38; Wright et al. 1975. 58  Neely 2011. 56  57 

72

J.A. Neely: Parthian and Sasanian Site and Settlement Patterns

Figure 22. A perspective drawing of a segment of canal system DL-5 near its point of origin: lush grasses within a very shallow linear depression indicate the presence of the canal, the bed of the Ab-i Garm drainage lies in front and to the right of the drop-tower gristmill, looking south-east

southward as the natural Ab-i Garm drainage changed course toward the south-west (Figure 18), involved a more laborious method of construction. Here, trenches were excavated into the piedmont slope so as to duplicate the gently graded canals present in the upper segment of the system. As this canal grade was gentler than that of the piedmont, the trenches excavated were necessarily deeper at their upper ends and became proportionately shallower as they coursed southward. The towers were set into the trenches at their deep northern extremes. Similar ‘buried’ gristmills were found in the Sohar region of Oman.59 This portion of the canal system probably was built in response to the need for additional gristmills, the domestic water needs of the adjacent Sasanian and later communities (DL3, DL-4, DL-168, DL-312/313, 135, and 123 (see Figure 18),60 as well as the irrigation needs of fields located on the alluvium immediately south of the central and eastern piedmont zone. Prior to the completion of this canal system, the part of the alluvial plain for which it eventually supplied water would have been useful only

likely. These canals were relatively small, averaging about 1.5 m wide by 0.6 m deep, and had been excavated into both faces of the banks of the natural drainage that carried Ab-i Garm waters down slope to the alluvial plain. The canals carefully followed the contours of the banks to retain a very gentle yet efficient grade. On the east bank in the northern portion of the system, test trenches were excavated through the canal at right angles to the flow to permit an accurate measurement of the gradient (Figure 22). The canal grade was found to average about 0.1 %. In the northernmost three km of the system, well-made tower-like structures of mortared masonry were built against and partially into the steep face of the drainage banks (Figures 18, 22). These tower-like structures were attached to water-driven gristmills constructed at irregular intervals ranging from approximately 50 to 600 m apart, with an average distance of about 275 m separating these features. The towers received the canalised water and dropped it some 6.5 m to provide the force to drive the gristmills, and then empty into a continuation of the canal system. The remaining six km of the system, in which the canal was carried

59  60 

73

Costa and Wilkinson 1987: 56–76. See also Neely and Wright 1994: appendix map 1B.

Sasanian Archaeology for seasonal dry-farming and grazing. Rainfall runoff from the piedmont zone may well have augmented the Ab-i Garm waters flowing through canal DL-330, a probable further extension of the DL-5 canal which coursed along the toe of the piedmont towards the south-east.61 Several other canals (e.g., DL-6) and qanats (e.g., DL-161) also apparently fed into that piedmonttoe canal. That canal, in turn, appears to have drained into and augmented canal DL-121 about 2.5 km northeast of Tepe Guran (DL-34), to supply additional water for domestic uses and the irrigation of fields paralleling the canal further to the south-west.62

an absolute date of AD 680+150.66 The extended use of this canal and gristmill system is suggested by an underglaze painted sherd (sample OxL-1348), recovered from the spoil bank of the DL-5 canal in Site DL-3 (Fig. 18). It was dated in a similar manner to AD 1750+100.67 A previous brief account of this drop-tower gristmill has been followed by a more detailed report.68

Large sections of the system were in poor condition due to natural erosion, localised alluviation, stone robbery, and filling-in with earth and debris resulting from modern agriculture and herding. The system was traceable for a straight-line distance of about nine km, within which the remnants of 22 drop-towers were found. However, as noted above, it appears that the canal may have extended south-eastward along the toe of the piedmont for about another 15 km.

Observations and conclusion

Only one other example of a drop-tower gristmill was recorded during the 1969 survey. That was situated in a deep trench cut into the alluvial plain, located between qanat/canal systems DL-121 and 122, and driven by the waters of one or both of those two systems. This droptower gristmill is situated about two km due east of the large site of DL-34.63 Similar ‘buried’ gristmills have been noted in Iran and elsewhere.64

Site and settlement patterns

This interesting gristmill technology has been found elsewhere in the Middle East,69 as well as in Europe,70 and in the New World71 and examples are still being reported.72

The observations and conclusions presented in a previous publication concerning the later occupations of the Deh Luran plain have been modified and augmented through subsequent analyses,73 and with the help of colleagues who have provided excellent comparative data to place Deh Luran into perspective within the Partho-Sasanian spheres.

During Partho-Sasanian times, the Deh Luran plain displays site and settlement pattern changes which generally parallel events documented for the Diyala region of east-central Iraq,74 the Upper Khuzistan (Susiana) plain of west-central Iran,75 the Mughan steppe of north-western Iran and southern Azerbaijan,76 and elsewhere.

The excavation of one of the towers (Figure 23) in the northern portion of the system revealed that the canals supplied water to the towers which concentrated its flow to provide waterpower to drive millwheels for grinding grain. This drop-tower gristmill construction is technically termed the ‘Arubah penstock’.65 Water management features and systems are notoriously difficult to date. While far from a satisfactory means, the general policy has been to date them through associations with nearby dateable habitation sites. Fortunately, a more direct method of dating was possible with the DL-5 gristmill system. Fragments of datable diagnostic pottery were found imbedded in the mortared masonry of the excavated gristmill of the DL-5 system. These fragments indicated a date range from the Sasanian to the 7th century Islamic period. One of these fragments (a blue glazed sherd: sample OxL-1349) was assayed at the Research Laboratory for Archaeology and the History of Art at Oxford University by means of Optically Stimulated Luminescence, and has provided

However, there are differences. Site locations along canals on the Diyala plain,77 Susiana plain,78 and Mughan steppe79 seem to be more strategically situated, for instance at the branchings of major canals, than those of the Deh Luran plain.80 Simpson noted cyclical settlement patterns in the area of the Eski Mosul Dam on the upper Tigris: ‘The results suggest clear peaks and troughs in settlement densities at different periods: whereas Seleucid, Late Sasanian and MiddleLate Islamic settlements were relatively frequent, those of Parthian/Roman–Early Sasanian and Early Islamic Hill 2006. Hill 2006. 68  Neely 1974; 2011. 69  e.g., Avitsur 1960; Beazley 1967; 1977; Costa and Wilkinson 1987; Gardiner and McQuitty 1987; Harverson 1993; Roaf 1999/2000; Wulff 1966: 280–82. 70  e.g., Beazley 1963; Goudie 1886. 71  e.g., Gritzner 1974; Neely 1999. 72  e.g., anonymous 2006. 73  Neely 1974. 74  Adams 1965: 69–83. 75  Adams 1962a: 116–17; Wenke 1975; 1987. 76  Alizadeh and Ur 2007; Ur and Alizadeh, this volume. 77  Adams 1965: fig. 5. 78  Wenke 1975; 1975/76. 79  Alizadeh and Ur 2007; Ur and Alizadeh, this volume. 80  See Neely and Wright 1994: appendix maps 1B and 1D. 66  67 

See Neely and Wright 1994: appendix map 1B. See Neely and Wright 1994: appendix map 1B. Neely and Wright 1994: appendix map 1B. 64  Bonine 1982: 148; Costa and Wilkinson 1987; Harverson 1993: 153– 54; Wulff 1966: 282. 65  Avitsur 1960. 61  62  63 

74

J.A. Neely: Parthian and Sasanian Site and Settlement Patterns

Figure 23. A perspective cross-section drawing of an excavated drop-tower gristmill complex: this is one of several such gristmills forming an integral part of canal system DL-5 near its point of origin at the Ab-i Garm springs, looking north-east

date were rare’.81 As of this stage of our study, we have not been able to discern similar cyclical perturbations in settlement densities on the Deh Luran plain, but this may be the result of our less refined ceramic chronology.

this may not have been the situation on the Deh Luran plain. For the reasons noted above, an assessment of the effects of this relatively rapid population and infrastructure growth on the Deh Luran plain will be presented in our forthcoming final report.85

Both Adams and Wenke observe that while population densities apparently reached their peaks during the Parthian period in the areas they studied, it was during the Sasanian period that the landscape, as well as the settlement and subsistence systems, were most affected and modified.82 The findings of the Deh Luran survey concur with their assessments of the Sasanian period yet the data suggest that it was also during the Sasanian period when the population densities of the Deh Luran plain peaked. Wenke supports this finding in his discussion of the Deh Luran plain.83 Furthermore, Wenke also observes that, despite apparently rapid population growth, fertile irrigable areas were not exploited and there is little to suggest that population pressure was a problem on the Susiana plain.84 However,

Agriculture and pastoralism In order to better understand the Partho-Sasanian imperial dynamics we need to know the processes by, and extent to which, farmers and herdsmen from the highlands were assimilated into the lowland economies and socio-political systems.86 The presence of two distinctive types of sites and agriculture/water management technologies, namely the irrigation-based agricultural communities on the alluvium and the dry-farming/pastoral communities on the piedmont, lead one to envision the existence of two distinct, but very probably interrelated and interdependent, economic and settlement systems on the Deh Luran plain. The former of these was most

Simpson 1996a: 99. Adams 1965: 73; Wenke 1975; 1987. 83  Wenke 1987: 256. 84  Wenke 1981: 310. 81  82 

85  86 

75

Neely and Wright in preparation. Hole 1980; Wenke 1987: 253.

Sasanian Archaeology likely a sedentary occupation while the latter may have been of a more transhumant-based nature with sedentary nodes as documented in historic times for the Lurs of the Deh Luran plain. This proposal is supported by Adams’ observation that: ‘Not intensive, but extensive, periodically shifting agriculture with a large component of husbandry is and has always been the system best adapted to Mesopotamian conditions’.87

Water management and irrigation systems On a pan-Middle Eastern scale, it is interesting to note that areas located between rivers (e.g., Mesopotamia, Susiana, Mughan steppe, and Deh Luran plain) were frequently chosen by groups for agricultural expansion and intensification. However, this choice was most logical considering the availability of water and rich alluvial soils.

Perhaps following a transhumant pattern between Deh Luran and the Khorramabad valley (Luristan [Lorestan] Province) as indicated by Hole’s research, the pastoralists would have occupied the plain during winter and sustained themselves through pastoralism and dry-farming.88 Sites such as DL-194 suggest that some of these pastoralist/agriculturalists may have led more sedentary lives. It is also conceivable that some of the pastoralists may have occupied settled communities on the alluvium and participated in irrigated agriculture as they do today. This may have been a scenario that the Partho-Sasanians developed to incorporate transhumant groups as contributing members into the empire. It would have been an approach to maximise the use and productivity of the Deh Luran plain as an integral part of the ParthoSasanaian empire with minimal disruption of the long established pattern of transhumance. Such a scenario presents a contrasting variation to the more traditional cyclical either/or picture of Middle Eastern subsistence systems.89 However, this tentative explanation begs questions regarding the processes involved in such a co-residence such as the amelioration of traditional antagonistic relationships90 and the construction of permanent houses and dry-farming infrastructure (i.e., terraces and check-dams) for which we have no earlier evidence in Deh Luran. These are questions which are difficult to answer through archaeology alone.

The majority of the Partho-Sasanian water management and irrigation technology systems found on the Deh Luran plain are not unique. Most have earlier origins on the Deh Luran plain and elsewhere in the Middle East. Those which have been found and recorded in other regions of the Parthian and Sasanian empires often appear at a much greater scale of size and complexity. The only aspect of the water management technology which appears unusual for the Deh Luran plain is the possible modification of some of the qanat systems to offtake waters through a seepage process from flowing rivers. However, the observations by Kirkby that the down-cutting process of the Deh Luran riverbeds began by the 2nd millennium BC and that this process undoubtedly affected the locations of canal offtakes,92 suggests an earlier invention and perhaps negates the possibility of this modification as having been originated by Partho-Sasanian engineers. One may obtain a more complete idea of the nature and density of the sites vis-à-vis canal and qanat systems of the Deh Luran plain by referring to Neely and Wright.93 However, when compared to the larger sites and canal networks of Mesopotamia94 and the Susiana plain95 one sees the relatively limited scale of the Deh Luran systems. Figure 21 provides a detail of the relationships of two small irrigation agriculture sites, as well as the secondary and tertiary/field canals in association. Satellite imagery has provided additional details on the larger sites of the Deh Luran plain,96 as it has elsewhere in the Middle East.97 It is hoped that details of the smaller sites and water management systems on the plain will be forthcoming as the technology improves and becomes more available and cost effective.

The well-documented and apparently contemporaneous use of irrigation agriculture, dry-farming and pastoralism as integral parts of the subsistence system and economy mark the Deh Luran plain and analogous areas as locations where questions regarding the processes of change, development, and maintenance of such a mixed economy and its socio-political milieu can be effectively studied. The need for study and increased knowledge of the sedentary vis à vis transhumant/ nomad interactions has been well stated by Adams.91

There is a distinct probability that the Deh Luran shallow salty marsh micro-environmental zone is a largely manmade phenomenon. Data from excavations indicate the presence of a marsh on the plain which provided comestible flora and fauna.98 The species of flora and Kirkby 1977; Kirkby and Kirkby 1969. Neely and Wright 1994: appendix map 1B and 1D; also see Wright and Neely (eds) 2010. 94  e.g., Adams 1965: figs 4–5; Simpson 2000: fig. 12. 95  Wenke 1975: maps 24–28. 96  Wright and Neely 2010: pls 2–6. 97  e.g., Alizadeh and Ur 2007; Challis et al. 2002–04; Menze, Mühl and Sherratt 2007; Ur and Alizadeh, this volume. 98  Hole, Flannery and Neely 1969. 92  93 

Adams 1975: 5. Hole 1974; 1978; 1979. e.g., Abdi 2003; Alizadeh and Ur 2007; Salzman 2004; Wilkinson 2003. 90  Alizadeh and Ur 2007; Buccellati 1966. 91  Adams 1978. 87  88  89 

76

J.A. Neely: Parthian and Sasanian Site and Settlement Patterns periods. The exact role played by Deh Luran in the larger picture remains largely unknown, yet it has been documented here, as well as by Wright and Neely, that there was a continuity in settlement patterns for the plain as they were generally similar to the more central and dominant economic and socio-political regions of the Middle East from the early empires through the Sasanian period.105 Based on the glazed pottery found during the 1969 survey, David Hill sees the Deh Luran plain as being well integrated into the later Mesopotamian economic system.106 He observes that the presence of glazed pottery produced in Mesopotamia during the Parthian, Sasanian, and early Islamic periods indicates a wide range of trade contacts and communication between the peoples of Deh Luran and the larger communities in the Tigris–Euphrates basin. On the contrary, the settlement pattern changes recorded for Deh Luran suggest that in some respects its marginal role continued and that it followed the re-organisation taking place somewhat earlier in the other regions closer to the centres where policies relating to the Parthian and Sasanian empires were developed. The results of these apparently conflicting, but not unreasonable, findings indicate that neither pottery (most especially glazed pottery) nor settlement patterns may be used alone to evaluate the status of regional integration. From the data at hand, it appears that the Susiana plain was the earliest of the large regions to see this late re-organisation,107 with central Mesopotamia following.108 Deh Luran evidently more closely follows the dating of central Mesopotamia, but it is not clear if the re-organisation and development of these two regions was contemporaneous. Adams observes that lags in re-organisation and development are ‘in line with the Sasanian pattern of shifting economic investments from one zone to another’.109

fauna found indicate that ‘sweet’ waters characterised the marsh, and settlements in this area point to the marsh area being viable for occupation. However, the settlement patterns of the 8th through 10th centuries indicate a slow abandonment, and thereafter the plain appears to have been almost unoccupied. This trend could very well have resulted from the increasingly heavy use of mineral-laden water of the Mehmeh and Dawairij rivers for irrigation culminating in ParthoSasanian times, which slowly salinated the soils and made cultivation of the lands less and less productive. Thus, the shallow, salt marsh as seen during this survey is, in fact, very probably a relatively recent phenomenon most likely exacerbated by the large influx of population in Partho-Sasanian times and perhaps not fully formed until the 10th century Islamic occupation. Generally similar human-induced areas of salinisation resulting from irrigation outfalls and affecting abandonments have been documented for the Amuq plain of southeast Turkey99 and around the Shatt al-Nil between Baghdad and Babylon.100 With the preceding in mind, it seems likely that, in addition to possible socio-political reasons, there was a geophysical reason why the Deh Luran plain was slowly depopulated. The many canals of the Deh Luran plain very likely generated a riparian micro-environment along their courses. Such riparian zones would have provided a ready supply of edible and useful plants, and would have been a haven for animals which would have found the zone a source of both water and food prior to kill or capture for human use. This riparian microenvironment phenomenon along canals has been observed in the American south-west,101 along the courses of the canals in the Tehuacan Valley of southern Mexico,102 and elsewhere. Reorganisation and change on the Deh Luran plain

While the Parthians were evidently successful in conducting their political systems, Greek, Latin, and Hebrew documents suggest the Parthian ‘empire’ was mostly an unstable coalition of vassal states brought periodically under Parthian control.110 It was not until the second half of the 4th century that there was a transformation of political control and administration throughout much of the Middle East which strengthened the economic and socio-political systems.111 Be that as it may, the historical record indicates that the Parthians, and, to a much greater extent, the Sasanians, were quite successful at integrating member ethnic groups, inaugurating taxation and conscription, expanding and intensifying agriculture and water management,

During its long history of human occupation, the Deh Luran plain has been a place of continuity and change as well as contrasts and contradictions. The marginal position occupied by Deh Luran relative to the economic and socio-political systems in operation across the greater Middle East during the early empires (c. 2600–2100 BC) has been documented elsewhere,103 yet the plain has maintained a semblance of importance due to the presence of large, but historically poorly known, sites and its probable role as a way/posting station on the Assyrian/Achaemenid royal road.104 This contrast of marginality versus importance appears to have persisted into the Parthian and Sasanian

Wright and Neely (eds) 2010. Hill 2006; this volume. 107  Wenke 1987. 108  Adams 1981: 183. 109  Adams 1981: 183. 110  Wenke 1981: 306. 111  Simpson 1996a: 88. 105 

Wilkinson 2000: 176–77. 100  Gibson 1974: 15. 101  Neely 2001a; 2005a; Neely and Murphy 2008. 102  Neely 2001b; 2005b. 103  Wright and Neely (eds) 2010. 104  Lendering 2011; Wilkinson et al. 2005. 99 

106 

77

Sasanian Archaeology and competing in trade. Archaeological evidence and written sources indicate that the Sasanians, perhaps emulating and expanding upon a policy initiated by the kings of the Neo-Assyrian empire,112 moved tens of thousands of people, sent engineering missions into the most under-developed parts of their empire, and literally re-shaped the land surface of large areas of the Middle East.

works including irrigation systems and other local and empire-supporting infrastructure.118 This programme could have been carried out on the Deh Luran plain for political and/or economic reasons. Politically, it could have been conducted to provide a location for the settlement of a newly conquered population, or to relocate part of the rapidly expanding population from densely occupied areas within the empire such as Upper Khuzistan (Susiana) and the Diyala.119 Simpson also observes that: ‘population transfers appear to have been effected in order to bolster community loyalty in newly acquired frontier zones’.120 I might add to this list of possible reasons that the transfer of certain groups might also have served to remove rebellious factions from potentially hostile areas and resettle them in locations where they lacked the numbers and the local prestige to do harm. Confirming this population relocation programme, Wenke states that ‘it is not at all unlikely that Deh Luran was chosen for imperially directed development schemes’.121 The construction of numerous unwalled sites may well be a reflection of organisation and protection by a dominant central government, and a changing economy that required a more dispersed residence.122 Contrasting with this proposition of rather strict control by the Sasanian empire, I propose that the occupants of Deh Luran had a degree of local autonomy probably most like that of the ‘smallholder’ or ‘householder’ as described by Netting.123

As noted, there are easily recognised changes in the settlement patterns of the Deh Luran plain beginning in the Parthian period and evidently accelerating in the Sasanian period. These changes (i.e., an increasing number of sites, smaller sites, a wider distribution of site components, and the appearance of new architectural forms (e.g., unwalled sites and the construction of compounds) in previously little used areas of the plain (e.g., the piedmont zone) all suggest rather drastic modifications of the economy and/or the socio-political systems. These modifications are similar in many respects to those reported for the same periods for regions of Mesopotamia,113 the Susiana plain,114 and the Mughan steppe.115 The question of why the above-noted changes took place must be addressed. The possibility of internal population growth and/or random influxes of populations from outside the plain seem unlikely due to apparent rapid changes, the magnitude of change, and observations of roughly contemporaneous population stability in adjacent regions. For example, Wenke notes widespread evidence that ‘indicates relatively stable and low rates of population growth’ for south-west Asia between c. 700 BC and AD 640.116

A correlated economic purpose for this proposed Partho-Sasanian programme might have been to use Deh Luran as a ‘breadbasket’ area to provide additional land for the production of badly needed foodstuffs for the rapidly expanding local population as well as that of the empire,124 or to produce foodstuffs in large quantities for trade and taxes. Wenke notes that: subsistence agricultural systems in many areas gave way, under imperial direction and funding, to the cultivation of foodstuffs the major value of which could be realised only with a centralised system of management, transport, distribution, and, especially, taxation.125 It was the great agricultural productivity (and the tax potential of this activity) of the regions chosen by the Sasanians for development that was a major force in their political history.126

Considering the foregoing, it is proposed that population growth, the expansion of existing and introduction of new methods of water management and irrigation, and the socio-political developments reflected in the changing settlement and site (community) patterns from c. 210 BC to AD 640 on the Deh Luran plain are a direct result of processes of planned expansion promoted initially by Parthian leaders, but drastically expanded by the Sasanian government. The documentation available indicates that the Sasanians undertook an intensive well-planned expansion programme to build the economy and population, and thereby the power and importance of their empire. As expressions of this directed change,117 the Sasanian kings fostered the rebuilding and resettlement of entire communities, the incorporation of large numbers of war prisoners within the empire, the encouragement of population growth by providing incentives for marriage and childbearing, and the planning and construction of large-scale public

In either case, or perhaps a combination of both, the expansion on the Deh Luran plain was made possible, 118  Adams 1965: 69–71; Howard-Johnston 2008; Nöldeke 1973; Rawlinson 1885: 484–85, 488. 119  Adams 1965: 69–71; Neely 1974. 120  Simpson 1996a: 88. 121  Wenke 1987: 257. 122  Wenke 1981: 313. 123  Netting 1993. 124  Rawlinson 1885: 488. 125  Wenke 1987: 259. 126  Wenke 1987: 253.

Ur 2005; Wilkinson et al. 2005. Adams 1965; 1981: 179–81. Wenke 1975; 1987: 258. 115  Alizadeh and Ur 2007; Ur and Alizadeh, this volume. 116  Wenke 1987: 257. 117  See Spicer 1962; Spicer (ed.) 1961. 112  113  114 

78

J.A. Neely: Parthian and Sasanian Site and Settlement Patterns unlike Adams’ findings on the Diyala plain, these data suggest that there evidently was also a high positive correlation and continuity between the Sasanian and early Islamic economic and socio-political systems on the Deh Luran plain.

likely with the assistance of Sasanian engineers, by adapting existing and introducing new water management and irrigation techniques. Van Roggen and Adams have noted the probable use of captive Roman legionnaires as labourers in the construction of irrigation and water management systems during the Sasanian period.127 Schwarz goes further by stating that Roman engineers captured by Shapur I designed the irrigation systems necessary to support large populations.128 In contrast, considering the water management infrastructure designed and constructed prior to and by the Sasanians in Deh Luran and elsewhere, it seems unlikely that Roman engineers played a significant role in designing the systems.

Several scholars, including Adams, have stated that a continuous process of sociopolitical integration or ‘centralisation’ of power resulted in, or played a major role in, changes toward urbanisation and increased population density.130 Attempts to define the nature of this centralisation have placed the emphasis of research on the ancient centres of economic and socio-political power in the Middle East. I submit that it would be enlightening to approach this problem from the opposite direction and by studying the more marginal regions we may obtain new, and quite likely different, perspectives as to the role and status of the various parts comprising the larger sphere of the centralisation process. For example, to what degree did the centralised authority control relatively marginal regions such as the Deh Luran plain and the Mughan steppe? I suspect there was a degree of economic and socio-political autonomy enjoyed by the inhabitants of these regions which has not yet been fully recognised or appreciated.131 The questions that immediately arise are: what was the degree of autonomy enjoyed and how was it expressed?

A simple, but hopefully evocative, comparison of continually occupied and/or re-occupied sites provides some further reinforcement of the proposition made concerning the directed changes taking place on the Deh Luran plain in Partho-Sasanian times. It also suggests a continuity of Partho-Sasanian economic and socio-political policy into the 8th and 9th century Islamic period. As noted above, eight (25.8%) of the 31 sites identified as Parthian occupations on the Deh Luran plain were continually occupied or re-occupied sites founded in earlier periods. However, 30 (96.7%) of these sites showed continuity of occupation or were re-occupied during the subsequent Sasanian period. Continuing this line of inquiry, we found that 59 (88%) of the 67 sites occupied during the 8th and 9th centuries had been occupied during the preceding Sasanian period. Using these data, which will probably be modified somewhat with our ongoing attempts to refine the temporal placement of the Deh Luran sites, we may tentatively compare the relative impact of economic and socio-political change from pre-Parthian to Parthian, Parthian to Sasanian, and Sasanian to early Islamic times. This comparison is based on Adams’ proposition that the pattern of founding early Islamic sites near abandoned Sasanian sites, but on previously unoccupied land: ‘suggests that the Sasanian abandonment was associated with a social upheaval sufficient to break off the tradition of residence at most of the Sasanian sites’.129 The above comparative data on continuity of occupation and/or re-occupation suggest that on the Deh Luran plain there was a significant difference between pre-Parthian and Parthian economic and socio-political mores, and that there was a high positive correlation and continuity between Parthian and Sasanian practices. However,

As Deh Luran is now, and for most of its long history of occupation probably was, an out of-the-way, marginal and provincial region, characterised by a somewhat less-than-ideal environmental setting, it also bodes well for providing well-preserved interesting and valuable information on the processes of acculturation involved in directed technological, economic and socio-political change that led to the economic and social expansion of a large and developed political unit – the ParthoSasanian empire. Further detailed investigations in this and analogous areas should also provide data on aspects of the operational systems of the Parthian and Sasanian periods not readily accessible through the study of large nuclear centres. In spite of its continued potentials, it is unlikely that I will return to the Deh Luran plain considering the current political conditions. It is my hope that the work I have accomplished will be of use to those who work there in the future.

Van Roggen 1905: 168; Adams 1962a: 116; 1965: 69, 82–83. Schwarz 1896: 346–47. 129  Adams 1965: 81. 127  128 

130  131 

79

Adams 1965: 73, 82–83. Neely 2011.

The Sasanian Colonisation of the Mughan Steppe, Ardebil Province, North-West Iran Jason Ur1 and Karim Alizadeh2 1

Department of Anthropology, Harvard University Peabody Museum, 11 Divinity Avenue, Cambridge, MA 02138. 2 Department of Anthropology, Grand Valley State University, Allendale, Michigan 49401.

Introduction In terms of archaeological excavation, the Sasanian period has been rather neglected by archaeologists, who have focused disproportionately on the remains of the earliest states and empires. On the other hand, the landscapes of the Sasanian empire have been studied in detail and are remarkably well interpreted. This situation has resulted from a fortuitous combination of three factors. Most obviously, Sasanian landscapes were monumental, both the symbolic landscapes of carved rock reliefs and the economic landscapes of state-sponsored irrigation infrastructure. Within the long sweep of history in the Near East, the Sasanian period was relatively recent, and therefore the various cultural and natural transformative processes have had less time to damage, obscure, or remove elements of its landscapes. Finally, we are fortunate that these landscapes coincide with some of the earliest archaeological surveys where pioneering field methods were developed, particularly in the Diyala region, the central Mesopotamian plains, and Khuzestan. From these surveys, the outlines of a state-directed agricultural intensification scheme of previously unseen proportions have emerged. The most dramatic example is the massive Nahrawan canal, which Adams estimates re-directed the entire output of the Diyala river onto the plains north of the capital at Ctesiphon.1 On the alluvial plains to its south and west, Sasanian engineers created a vast interlocking network of canals which intermingled the waters of the Tigris and Euphrates and maximised the agricultural output of the plain to much greater degree than the similarly minded (but much earlier) Ur III state at the end of the 3rd millennium BC.2 A similar expansion of population and irrigation infrastructure occurred in nearby Khuzestan which permanently reconfigured the hydrology of the plain even after the system fell out of use.3 The productivity of these large plains had been known and exploited for millennia prior to their agricultural colonisation under Sasanian impetus, but Adams 1965. 2  Adams 1978; 1981; 2006; Morony 1994; Simpson 2000. 3  Wenke 1975/76; 1989; Alizadeh et al. 2004. 1 

Sasanian Archaeology (Archaeopress 2022): 80–89

records of tax revenues preserved by Islamic writers allow a quantification that is difficult or impossible from archaeological data alone.4 Along with the intensified use of traditionally productive lands, settlement and agriculture were extended into marginal or previously under-exploited areas. For example, archaeological survey in the Deh Luran plain showed an incredible expansion of settlement numbers from the Parthian period to the Sasanian period.5 Settlement was not limited to the alluvial zone in the centre of the basin, but also moved up into the agriculturally marginal piedmont to the north-east, where it was associated with small scale but elaborate irrigation channels and water mills. Field research on Sasanian settlement and land use has concentrated on the areas around imperial political centres in south-western Iran and southern Iraq, but recent research is recovering similar patterns of intensification and extensification throughout the Sasanian realm, and particularly along its borders. At its western frontier with Rome, the dry-farmed plains of northern Mesopotamia witnessed a similar expansion of settlement numbers.6 Recent research on the northeastern frontier has demonstrated that the 200 km wall (the misnamed and often mis-dated ‘Alexander’s Wall’) was only one component of an elaborate built landscape of defensive and irrigation features.7 In this paper we present another Sasanian landscape, in this case on the north-western frontier. Through field research and analysis of satellite imagery and aerial photography, we have documented a system of fortified settlements and canals on the Mughan steppe of Ardebil province, north-western Iran (Figure 1). We will describe this archaeological landscape and the special challenges it presents, and attempt to place it within the broader social and political context of the southern

Christensen 1993. Neely 1974; this volume. 6  Wilkinson and Tucker 1995; Ur 2002; Simpson 1996a. 7  Nokandeh et al. 2006; this volume. 4  5 

Jason Ur and Karim Alizadeh: The Sasanian Colonisation of the Mughan Steppe

Figure 1. The southern Caucasus and north-western Iran in the Sasanian period, based on Kettenhofen (1993) and Hewsen (1992)

Caucasus in the 1st millennium and in the wider world of the Sasanian empire.8

two thirds are within the (now independent) Republic of Azerbaijan.11

The Mughan steppe

The climate and geomorphology of the plain is within the bounds of generally reliable rainfed cereal cultivation, but Mughan has traditionally been given over to pasture. Mughan and adjacent areas were the favourite winter grazing lands of the Mongols,12 and later Timur also wintered in the area. Since the early 18th century it has hosted the winter camps of the Shahsevan tribal confederacy, who migrated annually between pastures in Mughan and on the slopes of Mount Sabalan.13 However, in the mid-20th century, this long– term economic mode began to shift. An initial irrigation system of almost 20,000 hectares was installed in 1951. It was expanded to 90,000 with the joint Soviet-Iranian creation of a dam across the Aras near Aslanduz in 1971. The best of the former Shahsevan grazing grounds are now under cultivation.14

The Mughan steppe is a geographically diverse region in the north-western corner of the Islamic Republic of Iran, defined at the north by the Aras (Araxes) river and including the broad plain to its south as well as the foothills east of the Qara Su river leading up to Mount Sabalan in the Ardebil region.9 The plain itself was formed by fluvio-lacustrine sediments during the last glacial maximum, when the Caspian sea extended to inland to the south-west. Today the Aras flows in a braided channel in a broad but incised floodplain. Rainfall averages almost 300 mm per year, with mild winters and dry summers.10 Since the closure of the Russian-Iranian frontier in 1884, the south-western third of the plain falls within Iran and the north-eastern This paper was first published in Tehran (Ur and Alizadeh 2013) but has been lightly edited and included here because of its signifance and relevance to many of the other papers in this section. 9  Schweizer 1970; Minorsky 1993. 10  Schweizer 1970: 86–92. 8 

Tapper 1983. Smith 1999. 13  Tapper 1979; 1997. 14  Schweizer 1974; Tapper 1997: 309. 11  12 

81

Sasanian Archaeology Although travellers as early as the 10th century had reported on the remains of the plain, archaeological research in Iranian Mughan only began in 2004 with soundings at Ultan Qalasi, a 33 hectare fortified city of the late Sasanian and Islamic periods.15 Subsequently, Nader Tepesi, a high mound on edge of the river terrace, has been subjected to intensive surface collection16 and excavation. In January 2005, we began a new programme of landscape investigations to place these targeted excavations into a broader geographical framework. Although our research is complicated by the impact of the steppe’s modern irrigation system, our preliminary results have shown that the recent replacement of a pastoral landscape with a stateimposed irrigation agricultural landscape represents the reappearance of an economic mode already present in the Sasanian period. Elsewhere we have presented general overviews of the project and a discussion of the landscape dynamics of pastoralism and irrigation agriculture.17 Here we present a more detailed examination of Sasanian period settlement and land use.

invasion in AD 395 which overran Roman Mesopotamia and extended almost to Ctesiphon,22 both the Sasanians and the Romans were keenly aware of this threat and made great efforts to protect the easiest passes through the mountains at Daryal and along the Caspian coast at Derbend (see Figure 1). The latter is a system of five fortifications from just north of the Apsheron peninsula near Baku to the great system of walls at Derbend.23 Given the speed at which these mounted invaders could move, the 300 km distance between Derbend and Mughan provided little security. Epigraphic documentation for settlement in Mughan in the Sasanian era is rather sparse, although several routes of communication through it are documented. The major route ran between Ardebil north to Bajarvan, across the Aras to Baylaqan, and on to Bardha’a, which was the political centre of Albania. Baylaqan is identified with Oren Kala, a square fortified settlement complex 33 km north of the Aras, in the Republic of Azerbaijan.24 Ultan Qalasi represents the ruins of Vardanakert, a place known through Armenian geography.25 This place continued to be settled into at least the 10th century, when it was well known to Arab geographers as Warthan, a thriving market town with a wall and a Friday mosque.26

Mughan in the Sasanian world In the 1st millennium AD, Mughan was known as Balasagan (P’aytakaran in Armenian).18 During the Parthian period, Balasagan was part of Armenia. After a brief period of independence in the early 4th century, Balasagan became part of the Sasanian empire with the partition of Armenia in AD 387, administratively attached to Adhurbadagan (Iranian Azerbaijan), where it remained until at least the 6th century.19 However, in the Sasanian division of the world into ‘Iran’ and ‘non–Iran’, Balasagan and the lands of the southern Caucasus were considered to be non-Iranian (anērān), although Azerbaijan, its neighbour to the south, was considered to be Iranian.20 Balasagan was a generally stable territory within the empire, but was occasionally caught up in intrigues between Albania, Armenia and the Sasanian state.21

The archaeological landscape of Sasanian Mughan The plain is dominated by the remains of Ultan Qalasi, a large settlement complex perched on the edge of the Aras floodplain (Figure 2). Early travellers and later archaeologists have been drawn to the 33 ha rectangular fortification.27 Its south-eastern side extends 720 m, its north-eastern side is about 504 m, and its southwestern side is about 320 m. Ultan’s builders used the curving terrace edge to define its north-western side, which extends 745 m. There is extramural settlement to the south-west, south and east of this citadel. Excavations in Ultan Qalasi took place in February and November 2004 with the initial aim of establishing the periods of occupations of the site.28 To this end, one small (2 x 6 m) test trench (Trench 1) was excavated at the site. Trench 1 was located on the section created by bulldozer on the south-west side of the citadel. The trench was excavated to a depth of more than 5 m to virgin soil and contained materials from the earliest phases of the site.

In addition to its position on a cultural frontier between the Iranian world and the lands of culturally and linguistically non-Iranian peoples, Mughan remained a dangerous political frontier zone throughout the Sasanian period, despite the fact that the Sasanian sphere extended over Albania and Armenia to its north. Mughan was astride the easiest invasion route of nomadic groups from the northern side of the Caucasus (variously Alans, Huns, and Khazars). Since a disastrous Alizadeh 2004. Mohammadi 2004. 17  Alizadeh and Ur 2006; 2007. 18  Chaumont 1988; Brunner 1983; Hewsen 1992. 19  Hewsen 1992: 255. 20  Marquart 1931; Gignoux 1987; Frye 1963: 206. 21  Chaumont 1985; 1988.

See Greatrex and Lieu (eds) 2002: 17–19. Frye 1977; Minorsky 1958: 85–89. 24  Iessen 1959. 25  Hewsen 1992; Kettenhofen 1993. 26  Le Strange 1905: 176; Minorsky 1958. 27  e.g., Monteith 1833: 29–30. 28  Alizadeh 2006.

15 

22 

16 

23 

82

Jason Ur and Karim Alizadeh: The Sasanian Colonisation of the Mughan Steppe

Figure 2. Topography of Ultan Qalasi (contour interval 25 cm)

The phases of Trench 1 are numbered from 1 to 4. Phase 1 comprised the walls of the earliest fortification and related layers, while Phase 2 comprised the upper/ new fortifications. Phase 3 consisted of structures built inside the wall, and surface deposits made up Phase 4. In phase 1, the earliest fortifications were constructed of mudbrick and were preserved to nearly 5 m. During its life, they encountered two major floods of the Aras. In Phase 2, the upper fortifications were constructed atop the deposits of the second flood event, again by mudbrick. In Phase 3, an enigmatic small construction was built above one of upper courses of the lower (Phase 1) fortifications on its interior face. Small pottery assemblages from Trench 1 attest to occupation of the site from the late Sasanian period in the first phase to the late Sasanian/early Islamic period in the second phase. The third phase possibly belongs to the Seljuk period.

was also a square fortified complex on a long canal from the Aras.29 The Mughan Steppe Archaeological Project (MSAP) started a field survey component in 2005, and it quickly became clear that Ultan was only the largest of a series of rectangular fortified complexes on the steppe. We initially identified places of interest from declassified intelligence satellite imagery of the U.S. CORONA programme, which has emerged as a powerful survey tool in areas where archaeologists do not have access to low level aerial photographs.30 In Mesopotamian archaeology, CORONA has been used not only to identify sites but to reconstruct entire landscapes of roads and tracks31 and canal systems.32 The landscape transformations wrought by the installation of the irrigation system in 1971 made these historic photographs invaluable: the scenes used by the MSAP were from missions in May 1968 (Mission 1103) and May 1970 (Mission 1110), thus predating most of the levelling and canalisation of the plain.

Beyond the wall to the south-west is an extensive area of settlement cut by ancient irrigation channels (Figure 3). The extramural areas to the south have been badly damaged by the 1951 irrigation scheme, and more recently, gravel extraction has disturbed the western suburban area visible in CORONA photographs. Despite these disturbances, the traces of several canals are visible through ground survey and satellite images. One of the main canals surrounds the citadel like a moat and flows to the Araxes river at its north-eastern end. In size and settlement morphology, Ultan is very similar to its contemporary at Oren Kala (ancient Baylaqan), which

Sites and landscape features identified from CORONA photographs were then visited on the ground. In our initial field season, our primary concern was to develop our abilities to interpret the site signatures on the Iessen 1959. Reviewed in Fowler 2004. 31  Ur 2003; Altaweel 2003. 32  Pournelle 2003; Hritz 2004; Ur 2005. 29  30 

83

Sasanian Archaeology

Figure 3. Ultan Qalasi and associated canal features. CORONA satellite photograph (1103-1057DF073, 5 May 1968) courtesy of the US Geological Survey (modified by the authors) Table 1. Fortified Sasanian structures on the Mughan steppe. Measurements apply to fortified components only; site complexes also include much larger outer unwalled settlement areas (not yet measured), and Site 15 included two fortified structures. This table includes only those sites visited by the MSAP in 2005 Site No.

Name

Shape

Dimensions

Area (hectares)

Condition

-

Ultan Qalasi

Rectilinear

720 x 504 m

33.0

Intact

14

Qishlaq Qaravol

Square

100 x 90 m

0.90

Levelled

15

(unknown)

Both Square

30 x 30 m; 25 x 30 m

0.09; 0.08

Levelled Levelled

16

Nadir Tepe

Circular

90 x 120 m

1.08

Cultivated

18

Kara Tavara

Rectilinear

80 x 90 m

0.72

Cultivated

19

Aq Mazar

Rectilinear

100 x 160 m

1.60

Intact

20

(unknown)

Square

50 x 50

0.25

Levelled

21

Qala Yeri 1

Square

c. 50 x 50 m

c. 0.25

Intact

24

(unknown)

Square

50 x 60 m

0.30

Levelled

Aras floodplain or positioned at the interface of the flat steppe and the foothills to the south. The survey was able to visit nine of these sites (Table 1) but our imagery analysis shows that many others exist within Iranian Mughan and to the east in the Republic of Azerbaijan. Aside from Ultan Qalasi, Oren Kala, and a large site just beyond the Azerbaijani frontier (a good candidate for the location of the city of P’aytakaran), the other sites have modest fortified components, rarely as much as 100 x 100 m and often closer to a quarter of that size.

imagery, so our coverage was non-systematic, with the exception of a few sampling zones.33 In CORONA photographs, the large Sasanian settlement complexes were easily recognised (Figure 4). At the time the photographs were taken, the flat or concave fortified components were generally uncultivated, and often had signatures of recent Shahsevan campsites on them. They were arranged either along the terrace edge of the 33 

Alizadeh and Ur 2007.

84

Jason Ur and Karim Alizadeh: The Sasanian Colonisation of the Mughan Steppe

Figure 4. Sasanian fortified settlement complexes at Site 19 (Aq Mazar, on the Aras river terrace edge) and Site 16 (Nadir Tepe). CORONA satellite photographs (1103-1057DF073 and F074, 5 May 1968) courtesy of the US Geological Survey (modified by the authors)

Figure 5. Fortified structures with extramural settlement remains at Sites 15 and 18 (Kara Tavara). CORONA satellite photograph (1103-1057DF074, 5 May 1968) courtesy of the US Geological Survey (modified by the authors)

When terrace or foothill topography did not dictate otherwise, the ideal plan appears to have been square. Wall height rarely exceeded 2 m, especially on the ploughed sites; the exception is Nadir Tepe (Site 16), whose great height appears to be due to its position on a foothill spur extending out into the plain. In several cases (Sites 14, 15, 20, 24) the earth grading that accompanied the 1971 irrigation system had removed

all topography, leaving the CORONA scenes as our only record. On levelled and preserved sites alike, sherd scatters on the fortified sites were remarkably light, compared to surface assemblages elsewhere in the Near East. We attribute this condition to either low ceramic consumption or short term occupation.

85

Sasanian Archaeology

Figure 6. The major ancient feeder canal on the Mughan steppe, with associated sites and offtakes

Sherd scatters were not limited to the areas of the fortified structures. In most cases, these were accompanied by light but site-density surface assemblages in adjacent extramural areas. The unwalled components have all been ploughed and often also levelled, so it is difficult or impossible to determine their internal structure from either sherd distribution or topography. In many cases, however, CORONA scenes can be employed to approximate the extent of such settlement, which has a mottled or heterogenous appearance compared to adjacent unsettled fields (Figure 5). As with the surfaces of the fortified site components, these areas had very light sherd scatters.

the stratified sequences from the excavations at Nadir Tepe.

The pottery collected from the surfaces of these sites correlate well with the forms found in the first phase of the soundings at Ultan Qalasi, which date to the later Sasanian period. Conspicuously lacking from our collections are the distinctive glazed sherds known from Ultan’s early Islamic levels. There are hints of earlier occupation at some sites (e.g., Achaemenid and Parthian sherds at Site 14), so we hesitate to rule out pre-Sasanian occupation at any of these sites until we have a reliable typology of diagnostic types from earlier periods, especially since a similar pattern of fortified settlements has been well documented in the Armenian Late Bronze Age.34 Local ceramic sequences for this region of north-western Iran are not readily available, but future ceramic analysis will be based on

The major sites along the Aras river terrace (Ultan Qalasi, Aq Mazar) have been in the midst of modern settlement and irrigation since at least the 1950s, and their associated canals are preserved only in fragments as a result. The sites along the foothill edge, however, are strung along a major feeder canal which can be traced from the Aras near the modern town of Aslandouz to beyond the frontier with the Republic of Azerbaijan (Figure 6). Using CORONA photographs, the canal can be traced for 80 km within the boundaries of Iran. The situation is complex near the canal head, where the remains of multiple systems are visible, but in the eastern half of Iranian Mughan, only a single feeder canal was in use (or has survived). The fortified complexes themselves appear to command weirs which

34 

In all cases, these settlement complexes were in close association with the remains of irrigation canals. Unfortunately, the 1971 irrigation system removed most traces of the canals, except in some places where the planners consciously incorporated them as drains. Therefore we have relied heavily on geo-referenced CORONA satellite photographs which predate the completion of the modern irrigation system. CORONA has been very useful for reconstructing ancient irrigation systems elsewhere.35

Smith et al. 2004.

35 

86

Ur 2005.

Jason Ur and Karim Alizadeh: The Sasanian Colonisation of the Mughan Steppe

Figure 7. Dendritic patterns of canals and offtakes in the central Mughan steppe. CORONA satellite photograph (11031057DF074, 5 May 1968) courtesy of the US Geological Survey

took water from the main feeder and moved it north and north-east out onto the plain. The pattern of channels at these offtakes is often difficult to understand from the images; their apparent redundancy may result from consecutive rebuilds.

We presently know a second major system exclusively from CORONA photographs. It lies almost entirely within the Republic of Azerbaijan, although its canal head is to be located six km north-east of Parsabad (see Figure 6). We suspect that it is also to be dated to the Sasanian period because the signature of this feeder canal on our imagery is very similar to the signature of the main canal within Iranian Mughan. It is associated with several fortified settlement complexes, including an impressively walled 14 hectare settlement that is a good candidate for the city of P’aytakaran,36 should our chronological hypothesis prove correct.

The main offtakes extend north onto the plain and in several places are preserved on CORONA in remarkable detail. Dendritic patterns of branching canals are particularly clear on the central plain (Figure 7). North of Site 20, fields as narrow as 20–25 m can be discerned (Figure 8). At present we have been unable to identify any of these smaller canals or field systems on the ground, nor have we been able to recover any settlement traces among them; in both cases we attribute this situation to the heavy transformation of the plain by modern irrigation.

The clear association between the fortified settlement complexes and the major elements of the irrigation system suggest that most of the traces visible on CORONA imagery are to be dated to the Sasanian period. The total of all visible canals must, however, represent a palimpsest of multiple non-contemporary systems, although the degree to which this is true cannot yet be determined. There is a historical tradition, going back to Sharaf al–Din’s Zafar-nama, that in AD 1401 Timur dug two canals in the area, one from the Aras leading to the city of Baylaqan on the northern side of the river, and a second from the Qara Su, a right bank tributary

The position of the main canal head remains a difficult question. It is possible to follow the main channel back the point where it meets the Aras floodplain, a point 6 km downstream from Aslandouz. If the canal head was near this location, it has been removed by the Aras, which flows directly against the terrace edge at this point, now some 15 m below the ground level of the plain.

36 

87

Hewsen 1992: 255–58.

Sasanian Archaeology

Figure 8. Traces of ancient irrigated fields in the region north of Site 20 and an uncollected settlement complex with two fortified structures. The white lines are modern tracks and the polygonal structure at left is a fishery under construction. CORONA satellite photograph (1103-1057DF074, 5 May 1968) courtesy of the US Geological Survey

of the Aras, leading across the Mughan steppe.37 Tax revenue documents suggest to Tapper that irrigation on Mughan must have been productive in the 1720s.38 A fuller understanding of the dynamics of the postSasanian landscape will have to await systematic archaeological and geomorphological survey in future seasons.

I in the 6th century.39 It seems logical that these walls would be protecting more than client or buffer states, but rather would provide security for the state’s investment in revenue-producing agricultural systems, and now fieldwork in Mughan and Gorgan is recovering just such agricultural systems.40

The chronology of the Sasanian agricultural colonisation of the Mughan steppe is presently dependent on surface pottery, but it is possible to suggest a more precise placement, and to attempt to place the settlement system within a socio-political context. The reinforcement of the northern frontiers at Derbend and Gorgan is best attributed to Khusrau

If a late Sasanian date is accepted, the repeated patterns we find in the settlement might reflect known elements of Sasanian society at the time. Complexes with one or more fortified structures and extramural settlement in association with irrigation system would approximate the archaeological signature of the dastkart. In its original meaning, it signified a household, but in later Sasanian times its use expanded such that ‘le dastkart

Tapper 1997: 72; Le Strange 1905: 178. 38  Tapper 1997: 126.

39  Minorsky 1958; Frye 1977: 15; Nokandeh et al. 2006; see Omrani Rekavandi et al., this volume. 40  Alizadeh and Ur 2007; Nokandeh et al. 2006.

37 

88

Jason Ur and Karim Alizadeh: The Sasanian Colonisation of the Mughan Steppe goodness of the soil’.46 Abu Dulaf ’s number is certainly an exaggeration, but his general impression of a plain with many abandoned settlements, the walls of which have not had time to collapse into a tepe, would agree with the preliminary results of our 2005 reconnaissance.

correspondait à un domaine rural comprenant maison d’habitation, bâtiments divers, canaux, puits. Mais il peut égaliment signifier une parcelle de terre cultivée par un seul esclave; plus tard enfin, il servit à désigner les grands domaines fortifiés où s’élevait un château et qu’habitaient les nobles ou le shah’.41 The expansion of these estate complexes on the Mughan steppe might be a result of the noted reforms of Khusrau I, who undermined the power of the most powerful families by promoting the dehqans, a class of landholding minor nobles.42

Certainly by the time of the Mongol invasions but probably one or two centuries earlier, Ultan Qalasi was abandoned. The plain became a favoured pastureland, first by various Kurdish and Ghuzz Turk groups, then later as winter pasture for the armies of the Mongols.47 By the early 18th century the descendants of these groups had coalesced into the Shahsevan confederacy,48 whose winter campsites were abundantly scattered over the uplands and the plain in the late 1960s.49

The collapse of the Mughan irrigation system and the rise of pastoralism The Islamic conquest of the south-eastern Caucasus region appears not to have impacted irrigation along the terrace edge. The excavations at Ultan Qalasi show that its fortifications continued in use and were repaired.43 Our soundings to date are too small to present a broad picture, but Islamic historians of the 10th century describe the town (now known as Warthan) as a bustling economic centre with a large mosque outside the city walls.44 On the northern side of the Aras, the city of Baylaqan remained an important centre.45

Conclusion The Sasanian state-directed expansion of settlement and irrigation agriculture is well documented from landscape studies in the Diyala, the southern Mesopotamian plain, and Khuzestan. However, it was not limited to these longstanding areas of agricultural productivity and political power. Areas peripheral to the Iranian cultural sphere but under the political control of the Sasanian state, such as the Mughan steppe, were also part of the state’s expansion strategy. While it is clear that we are not dealing with a system on the order of the Nahrawan canal, at 80 km the Mughan system approximates the length of the mashruqan canal (now the Gargar river) below Shushtar in Khuzestan.50 Presently, historians and archaeologists view the northern frontier of the Iranian linguistic and cultural sphere as a political periphery, but the Sasanian state was willing to invest considerable resources in its agricultural infrastructure. These new projects in previously understudied regions promise to give us a more representative picture of Sasanian empire.

The 80 km long irrigation system which ran along the edge of the uplands appears not to have survived far beyond the late Sasanian period, however. Our survey recovered none of the Islamic sherds which characterise the upper layers of the soundings at Ultan Qalasi on the fortified settlement complexes of the plain. Again we are aided by Islamic written sources, in this case the traveller Abu Dulaf, who crossed the steppe in around AD 950. In his account, he states that in this plain there are 5,000 villages or more, all in ruins, although their walls and buildings stand upright without decay in view of the

Pigulevskaja 1963: 152–53; see also Gignoux 1994. Tafazzoli 1994; 2000; Rubin 1995. 43  Alizadeh 2004. 44  Le Strange 1905: 176. 45  Bosworth 1989.

Minorsky 1955: 36. Smith 1999; Tapper 1997: 39. 48  Tapper 1979; 1997. 49  Alizadeh and Ur 2007. 50  Alizadeh et al. 2004.

41 

46 

42 

47 

89

The Army and Urbanism at the Sasanian Empire’s Northern Frontiers: Fieldwork on the Linear Barriers, Fortresses and Cities at the Margins of the Gorgan Plain (Results of a Joint Project Between the Iranian Cultural Heritage, Handicraft and Tourism Organisation and the Universities of Edinburgh and Durham) Hamid Omrani Rekavandi (HO), Eberhard W. Sauer (EWS), Tony J. Wilkinson (TW),† Ghorban Ali Abbasi (GA), Roger Ainslie (RA), Francesco Caputo (FC), Mohammad Ershadi (ME), Morteza Fattahi (MF), Nikolaus Galiatsatos (NG), Kristen Hopper (KH), Julian Jansen Van Rensburg (JJ), Eve Macdonald (EM), Majid Mahmoudi (MM), Amin Nazifi (AN), Jebrael Nokandeh (JN), Chris Oatley (CO), Seth Priestman (SP), James Ratcliffe (JR), Kourosh Roustaei (KR), Esmail Safari Tamak (EST), Bardia Shabani (BS) and Lucian Steven Usher-WilsoN (SU) Our paper is based on a joint project between the Iranian Cultural Heritage, Handicraft and Tourism Organisation (ICHHTO), the Iranian Center of Archaeological Research (ICAR) and the Universities of Edinburgh and Durham. It is based on a lecture delivered at the 10th Gathering of Iranian Archaeology in Bandar Abbas in December 2008. We are indebted to Dr Mohammad Mortezai, the director of the Iranian Center of Archaeological Research (ICAR), for his kind permission to reprint the papers delivered at this international conference, with some amendments, in this volume. We are also grateful for his advice and help since the first year of the project, in 2005. We would like to thank Dr St John Simpson for his invitation to include this contribution to his Sasanian Archaeology, as well as for his advice and support throughout the project. We are very grateful to Dr Seyed Taha Hashemi, the vice-director of the Iranian Cultural Heritage and Tourism Organisation (ICHHTO) and the head of the Research Department of the ICHHTO, and to Dr Hassan Fazeli, then the director of the Iranian Center of Archaeological Research (ICAR), for their kind permission to continue with our joint project from 2006 to 2009 and for their kind and essential help. We are very grateful to Mr Sayed Mohammed Beheshti, then the director of the Centre of Research of the ICHHTO, and the late Dr Massoud Azarnoush, then the director of the ICAR, for their permission to commence our joint project in 2005 and their interest and support. To Dr Fazeli we are also very grateful for his kind invitation to present our research at the conferences in Tehran and Bandar Abbas. We are indebted to Dr Seyed Mehdi Mousavi, then the vice-director of the Research Department of the ICHHTO, for granting permission to process samples in the UK. We would like to thank Mr Fereidoon Faali, the director of the Golestan ICHHTO, for his help, advice and personal interest in our work. The help offered by Mr Karim Alizadeh, Mrs Leyla Kashiha, Mrs Leyla Safa’ie, Mrs Mojgan Seyedin and Mrs Azam Tohidlou, members of the international section of the ICAR, has been invaluable. We are also indebted to Mr Fereydoun Unagh, the director of ICHHTO at Gonbad-e Kavus, and members of the local ICHHTO office for facilitating our research project in many ways. Without the exceptional efforts of the team none of this could have been achieved. In addition to the authors of this and previous reports, we are most grateful for the excellent drawings compiled by Maryam Hussein-Zadeh, Mohaddeseh Mansouri Razi and Mrs Esmaili-Sade. We are also grateful to Mr Abdolhussein Badpa, the driver of the Gorgan Wall base, and would like to acknowledge the essential contribution by many academic supporters (notably Dr Vesta Sarkhosh Curtis and Dr John Curtis), our workmen and drivers, space does not allow to list. We are most grateful for the generous support by the ICHHTO, the AHRC, the British Institute of Persian Studies, the Carnegie Trust for the Universities of Scotland, the Iran Heritage Foundation, the Stein Arnold Exploration Fund of the British Academy and the support received from Edinburgh University’s School of History Classics and Archaeology.

Introduction Fieldwork in the Gorgan Plain, as part of a joint project between the ICHHTO and the University of Edinburgh, has taken place from 2005 to 2009 (Sauer et al. 2013) and from 2014 onwards. This paper focuses on the results of the 2008 season which explored the Great Wall of Gorgan and a military campaign base (Qal’eh Kharabeh) in its hinterland (Figure 1). Urban development on both sides of the barrier was explored via excavations at the Sasanian Archaeology (Archaeopress 2022): 90–102

vast city of Dasht Qal’eh, contemporary to the Wall, and Qelich Qoineq, a town long abandoned when the wall was built. Tent enclosures, canals and industrial activity: fieldwork at Qal’eh Kharabeh (EWS, EST, RA, ME) Excavations and geophysical survey at Qal’eh Kharabeh (Figure 2) established that tent enclosures seem to densely occupy much of the eastern half of the

Hamid Omrani Rekavandi et al.: The Army and Urbanism

Figure 1. The Great Wall of Gorgan, Sasanian military installations and other sites mentioned in the text

Figure 2. Magnetometer survey in Qal’eh Kharabeh with the location of our trenches L, M, N and R (by Abingdon Archaeological Geophysics and the ICHHTO, notably Roger Ainslie, Chris Oatley, Majid Mahmoudi and Mohammad Ershadi)

91

Sasanian Archaeology compound and that they extend right to the central north-south road. This unique discovery suggests strongly that the vast 42 hectare fortress was occupied by a temporary military garrison, probably the Persian field army. These rows of enclosures are absent from the western half, where we encountered irregular scatters of high magnetic anomalies (perhaps areas of industrial activity). We can only hypothesise on the reasons for the curious bipartite division of the compound. It is possible that parts of the interior space were reserved for production, processing or repair of essential commodities or equipment, other parts accommodated perhaps the garrison’s horses and pack animals (or, less probably, served as a defensible retreat for the local population with their livestock and movable belongings in times of crisis). One might wonder whether defensive logic would not have been an argument against confining all combat troops to one half of a large compound, thus increasing the response time in the case of any surprise attack against the unoccupied half. Yet, the open landscape would probably have allowed spotting any approaching hostile force in time to man the ramparts, so that there may have been no need to encamp soldiers close to all sections of the defences. It is, of course, nevertheless possible that seemingly empty parts of the interior space were occupied by troops, perhaps of a different origin to those occupying the tent(?) enclosures, living in untraceable structures, e.g. unenclosed yurts or tents.

Figure 3. The impressive square fortification of Qal’eh Kharabeh and the Gorgan Wall with Forts 28 and 29. Corona satellite image by courtesy of the US Geological Survey

In contrast to such temporary accommodation, the substantial walls of the Qal’eh, with clear traces of projecting towers and gates, were almost certainly built of mud brick (a building material demonstrably used for the wall surrounding the citadel, explored in Trench L in 2007). Qal’eh Kharabeh’s wall circuit of c. 2.5 km would have required significant quantities of mud brick and substantial amounts of water to make mud brick out of the locally available soil and any added plant material.

east gate, and close to a possible butt end revealed by magnetometer survey (Figure 2: Trench R). We encountered mid greenish grey clayey silt deposits, whose greenish hue, extensive iron-staining and, more or less, level bottom suggested that they had been deposited whilst forming the bottom of a water-filled feature, most probably indeed a canal. An earth bank, c. 0.75 m high and c. 1.25 m wide at the bottom and c. 1 m at the top, which must have formed one northern side of the raised canal, was visible in the east-facing profile, but scarcely recognisable in the west-facing profile. The greenish grey clayey silt deposits extended from 0–13.20 m north in the east-facing profile, but only to c. 10.80 m in the west-facing, perhaps suggesting that the canal was indeed starting to curve towards the butt end, identified on the basis of the geophysical survey, and that the water thus was not channelled into the Qal’eh through the east gate, unless by means of a siphon. Similar to our section through the canal parallel to the Gorgan Wall in 2005,1 non-waterlogged ditch fills and the natural (and the earth bank in Trench R) differed little in colour and composition, probably as a result of the scarcity of decayed organic matter in this semiarid environment. While the similarity of such deposits made it difficult to recognise the earth bank, there is no

The satellite image of Qal’eh Kharabeh (Figure 3) shows two raised linear features, running between the north and south gates and the west and east gates respectively, intersecting in the centre of the compound and dividing it into four almost symmetrical quadrants. Parts of these linear banks appear as parallel double lines on this photo from space (notably the one running from the north gate to the centre of the complex), and they still form distinct causeways today. We had thus considered the possibility before that they served a dual function, as roads and raised canals, perhaps tapping the old canal cut by the Gorgan Wall and by Fort 29, and that they provided the water needed for mud brick production. We excavated a partial section through the west-eastrunning of these raised features, near the Qal’eh’s

1 

92

Nokandeh et al. 2006: 141–46.

Hamid Omrani Rekavandi et al.: The Army and Urbanism doubt that the causeway is indeed formed by a broad raised canal, extending in the south beyond the limits of our trench.

by the Wall-builders and in any case at most a century before or after. The presence of regularly spaced hillocks along the defences of the Qal’eh, the remains of projecting mudbrick towers, and the presence of internal canals are further architectural parallels between the Wall forts and the Qal’eh and further arguments for tentatively assigning both to the same building programme, or at least the same architectural tradition. The evidence suggests that Qal’eh Kharabeh’s occupation was much shorter than that of the Wall forts. Qal’eh Kharabeh was thus not part of a defencein-depth system, but probably largely or completely abandoned for most of the Wall’s history as an active defensive barrier.

The geophysical survey revealed a further possible linear feature halfway between the central west-east causeway and the southern defences of the Qal’eh. We do not know its precise function, but it may have been a drainage ditch under a minor road or a ditch marking an internal division of the compound (Figure 2). The excavation of a ditch along the axis of symmetry of the compound’s southern half suggests in any case that much effort had been invested into the neat organisation of space. Date and function of the hinterland campaign bases (EWS, EST, RA, ME)

The theory that it may have been a failed urban foundation seems unlikely. The landscape survey by Hamid Omrani Rekavandi and Tony Wilkinson and the pottery studies by Seth Priestman suggest that the sites north of the Wall had been abandoned long before its construction. The hypothesis that we are dealing with a major programme of resettlement of people, who had lived north of the Wall before, is thus almost certainly false. While it is in theory not beyond possibility that other territories could have provided a postulated town with inhabitants, an explanation of Qal’eh Kharabeh as a military foundation, occupied for a limited time only, seems much more plausible.

We obtained one date for a radiocarbon sample from the fill of the canal in Trench R and three from Trench M, in the centre of the Qal’eh excavated in the previous season. These dates have now confirmed what we had postulated on the basis of architectural similarities between the Gorgan Wall forts and the hinterland compound of Qal’eh Kharabeh and on the basis of Seth Priestman’s pottery typology. Seth Priestman already had suggested, on the basis of his analysis of the pottery from Qal’eh Kharabeh and before scientific dates had been available, that the site was occupied for a short period, perhaps in the earliest phase of the Wall’s history. This hypothesis has gained considerable strength in the light of the scientific dating: the occupation of Qal’eh Kharabeh, the construction of the Gorgan Wall and the earliest occupation of its associated forts are so close in time that radiocarbon dating places them within the same time group: our most recent radiocarbon dates (all four dating to the period between AD 406 and 540 at a 95.4% confidence level) have now established beyond doubt that Qal’eh Kharabeh was occupied in the 5th and/or early 6th century. As the earliest and latest likely dates for the construction of the Gorgan Wall fall within the same plateau of the calibration curve and are also more than a century apart, we cannot be sure that the Wall and the Qal’eh are indeed contemporary. It is also possible that Qal’eh Kharabeh could be a few years, decades or even generations earlier than the Wall (e.g., that it was constructed during a war which preceded the decision to protect the Gorgan Plain by means of a defensive barrier) or vice versa. We cannot be sure whether it played some role in events leading up to the decision to build the Gorgan Wall, during its construction or in the early years of its existence. The square layout of the compound, the temporary nature of the structures in the interior, the comparatively small quantity of finds and (in contrast to Fort 4) the lesser depth of the stratigraphy suggest also that we are dealing with a short-lived site – perhaps constructed

It appears that Qal’eh Kharabeh (much larger than any contemporary installations in the Roman empire) was probably a campaign base for the Persian field army, rather than housing frontier troops, let alone a civilian population. Indications to suggest that its function was military include its regular geometric design, other architectural similarities to the Wall forts (e.g., projecting towers and raised canals), the presence of numerous possible tent enclosures in neat rows, the site’s location just 2 km south of the Wall with the main gate facing the Wall and its apparent short occupation. A site visit to Gabri Qal’eh helped us to put the results achieved at Qal’eh Kharabeh into a broader context.2 This c. 585 x 585 m large fortress, about 7 km southeast of the central section of the Wall, displays various features, which strikingly resemble those observed at Qal’eh Kharabeh: it has a square ground plan, a prominent corner citadel and a raised linear feature (probably a raised canal as well) leading through a gate into the site. It thus is tempting to assume that its original foundation date and function match those of Qal’eh Kharabeh. It will, however, be much more difficult to prove or disprove this hypothesis, or explore the site in any detail, as it appears to have experienced extensive re-occupation in the Islamic period. The earliest layers are thus likely to be buried deeply and 2 

93

Kiani 1982b: 56–57, pls 26.2–27.2, 28, figs 30–31.

Sasanian Archaeology beyond the reach of geophysical survey methods; it may only be possible to reach them, like in the city of Dasht Qal’eh (to be discussed below), in deep sondages. A third site, Qal’eh Gug A,3 shows the same characteristics and is likely to belong to the same category of campaign bases(?), as does Qal’eh Daland.4 The results achieved at Qal’eh Kharabeh are thus not just of relevance to this one individual site, but likely to apply to at least three other sites. If we are right in thinking that they served as military compounds, then their sheer scale sheds interesting light on the extent of deployment of troops and labour in the Gorgan Plain around the time when the Wall was built – or on events not too long before or after. Whilst it cannot be excluded that they might have housed local units temporarily, it now seems much more plausible, given their standardised plan, to regard these sites as campaign bases for large units of the Sasanian field army. The size of the compounds and the density of the enclosures suggests that each would have held a garrison of perhaps 10,000 men, even if there can, of course, never be any certainty regarding the accuracy of such estimates.

Wall once was or what its upper sections looked like, as the ruined wall provided a convenient quarry until no remains were left above ground. The contemporary or nearly contemporary Tammisheh Wall survives in its best-preserved section to a height of c. 2.80 m,6 probably the minimum height, excluding any breastwork. Bivar and Fehérvári report that in a trench excavated in the 1960s the Wall survived to a height of 4 m,7 though the published photos make one wonder whether this refers to the preserved Wall or the elevation of the bank over the silted-up ditch. As only four courses of an excavated tower survived, the stated height must at least include the clay foundations of the Wall. Parts of the Sasanian Derbent Wall, however, reached more than 6 m in height.8 Its different construction method (built of stone, with a rubble core and limestone facing), the physical distance to the Gorgan Wall of over 700 km and its potentially slightly later date9 make it hazardous to assume that it was necessarily built to similar specifications as the Gorgan and Tammisheh Walls. It seems unlikely that the Gorgan Wall, mortarless and without foundation, would have been much higher than the Derbent Wall, and it is equally improbable that it would not, at the very least, have reached the height of the best-preserved section of the Tammisheh Wall. Whether the Wall was just 3 m high or, possibly, over 6 m, plus perhaps another 1.5 m of parapet would have made a major difference to its physical appearance, its defensive capabilities and, presumably, its deterrent effect on any potential intruders.

The height and design of the Gorgan and Tammisheh Walls (EWS, HO, JJ, EM, FC, BS, AN) Earth-moving operations in preparation for the widening of the road from Aq Qal’eh to Incheh Borun had revealed a section through the Gorgan Wall close to the intersection between the road and the Wall (at N 37°07.130’/ E 54°28.856’). We took advantage of the opportunity to clean and record this partially exposed stretch of Wall (Trench Q), which seemed all the more important as its destruction, in the course of the ongoing construction works, was thought to be imminent. As in the previously recorded sections,5 the Wall was five bricks, i.e., c. 2 m wide. Most of the bottom and parts of a second course survived. It was overlain by a soft mid-reddish brown silt deposit with c. 25% brick fragments, undoubtedly the fill of a robber trench. A scatter of, perhaps discarded, fragments of brick south of the Wall, at the same level as its lowest course, suggests that it had been built on the ancient land surface and that it lacked any foundations. It appears that no foundation trench was considered necessary on flat terrain, neither for reasons of stability, nor defensibility. It seems unlikely that potential enemies had experience in undermining walls; the canal on the Wall’s north side (especially as long as it carried water) would have made any such operation difficult in any case, if not impossible.

While we do not know yet, what height the Gorgan and Tammisheh Walls once reached, the 2008 season has demonstrated that the clue to solving this question may be found at the bottom of the Caspian Sea. Julian Jansen Van Rensburg, Francesco Caputo and Hamid Omrani Rekavandi carried out an underwater survey of the remains of the Tammisheh Wall, once built on land and now flooded, as a result of a rise in the sea-level of the Caspian Sea. The survey established that the remains of the Tammisheh Wall run into the Sea for several hundred metres further than so far known. It appears that this section of the Wall collapsed under the waves of the rising Caspian Sea before it had fallen prey to demolition works. Within a test transect of just 10 x 2 m (less than half of the c. 25 m wide bank of collapsed Wall rubble) brick fragment were systematically collected and weighed. On land, the Gorgan Wall was preserved to a height of 1.47 m (13 courses) in Trench E and, as in some other sections known to us, was built solidly of bricks and about five bricks wide,10 even if in places the Wall

The Wall’s foundations are, of course, easier to explore than its third dimension. We do not know how high the

Nokandeh et al. 2006: 152–53. Bivar and Fehérvári 1966: 42, pl. IIa–b. 8  Powell 2008. 9  See Frye 1977: 12; Harmatta 1996: 82. 10  Nokandeh et al. 2006: 137, fig. 14. 6  7 

Kiani 1982b: 42, pl. 14.1, figs 30–31. Kiani 1982b: 58–59, figs 30–31, pl. 30.1. 5  Nokandeh et al. 2006: 135–38. 3  4 

94

Hamid Omrani Rekavandi et al.: The Army and Urbanism There was nothing to suggest that the third category of site (unlike the latter two, outside the protective bounds of the Walls and probably not built from scratch, but evolving over a period of time) was chronologically or functionally related to these defensive barriers. Yet, fieldwork north of the Gorgan Wall promised to reveal settlement expansion and retraction prior to its construction. Numerous factors can lead to the abandonment of settlement in economically marginal land; these include a population decrease or displacement as a result of political or economic developments, epidemics, changes in land management or climate or devastation and insecurity caused by full-scale war or small-scale raiding. With no sufficiently precise pottery chronology yet being available, it seemed important to obtain independent dating evidence for these historical turning points in the settlement history of the Gorgan Plain and to verify whether or not the abandonment of sites might even be close in time to the construction of our defensive barriers. If so, and this is no more than one of many possible models at present, settlements north of the later Wall may hold the clue to the historical circumstances behind the Wall’s construction.

appears to have been built on an earthwork foundation. Brick dimensions and weights vary slightly in different sections of the Gorgan and Tammisheh Walls and, for those from the underwater survey, the weight of shell incrustations and sea water in their cavities has to be taken into account. Using the rough average of 20 kg per brick, then each (c. 0.11 m high) course of the Gorgan Wall would have required c. 250 kg of bricks per metre. Bivar and Fehérvári, however, observed that the Tammisheh Wall was built on a foundation ‘exclusively of clay’.11 The section still well preserved today has an earth core and an only partially preserved brick facing on either side. Whether the section excavated by Bivar and Fehérvári had a foundation ‘exclusively of clay’ or whether any brick facing had been robbed out, is hard to tell, but it seems that parts or all of the Tammisheh Wall consisted of an earth core with a brick facing on both sides. Such a building technique would reduce the bricks required per metre to c. 100 kg for each course. As the team of divers recovered 1.315 tons of bricks from a 10 x 2 m transect (some 40% of the width of the brick scatter), the overall quantity of surface bricks from a 2 m wide section through the Wall rubble probably would have been in the order of c. 3.3 tons (or half as much for a 1 m wide transect). This quantity would account for little over 16 courses. Yet, it ought to be borne in mind that these are just the surface bricks and that more were embedded in deeper layers.

Urban development in the hinterland of the Gorgan Wall (EWS, GA, SU, KR, HO, SP) According to medieval sources, Yazdgird II (438– 457), Peroz (457–484), Kavad I (488–96, 499–531) and Khosrow I (531–579) carried out major construction work in the Gorgan plain, including the establishment of new towns. Yet, the location of these new urban foundations is unknown.12 To gain some knowledge of the development of existing urban centres in the hinterland of the Gorgan Wall, or the foundation of new ones from scratch, seemed important for a number of reasons. The size, appearance and location of new cities promised to shed light on the economic prosperity and population size of the area and the extent of state interest and expenditure. We also hoped that it might contribute to a better understanding of the overall strategy to defend, develop and aggrandize a rich and fertile, yet vulnerable, frontier territory.

Urbanisation in the frontier zone (HO, EWS) In 2008, whilst continuing to work on Qal’eh Kharabeh, the Gorgan and the Tammisheh Walls, our main focus shifted to the relationship between the former linear barrier and one representative each of three different categories of sites: • Military compounds south of the Wall of a geometric layout • Urban settlements south of the Wall of a geometric layout • Urban settlements north of the Wall of a nongeometric layout. Only the first category of site (i.e., the campaign bases) had been explored already before. It shared regularity of design with the second category, and similarities between these two types of sites and the Gorgan Wall forts, made it tempting to assume that they originated at a similar date and that they were the brainchildren of the same architects. To test this assumption and to explore the original purpose and later development of these major compounds promised to yield significant new insights into military strategies to defend this frontier region, its economic prosperity and the scale of (royal?) building programmes in the Gorgan Plain. 11 

We identified Dasht Qal’eh, a c. 3 km2 large city (Figure 4) as a possible key site, in the light of its extraordinary dimensions, its position just 10 km behind (south/ south-east) of the Wall and, most importantly, possible similarities with known Sasanian sites. These included a prominent corner citadel, in a position which resembled that of Qal’eh Kharabeh’s equivalent, and its remarkably regular overall plan. While, unlike Qal’eh Kharabeh and other compounds thought to represent Sasanian campaign bases, its plan is not square, its wall circuit is certainly sufficiently regular to leave

Bivar and Fehérvári 1966: 42, pl. IIa–b.

12 

95

Keall and Roaf 2000.

Sasanian Archaeology

Figure 4. A Sasanian urban foundation, perhaps ancient Gorgan: view from Dasht Qal’eh’s citadel to the north-west, over the walls and massive moat surrounding the three square kilometres large city

little doubt that it represents a carefully planned foundation (category 2 of the types of sites referred to above), whether from scratch or over an older existing settlement. Our aim was thus to establish whether or not there is a chronological link between the establishment of forts and fortresses with a geometric layout along the Gorgan Wall and that of an even much more extensive civilian city, equally with geometric layout and, like Qal’eh Kharabeh, with a corner citadel.

urban foundation, or an even earlier town which was substantially rebuilt then, seemed possible and worth exploring. We were also intrigued by monumental brick structures in the east of the town, consisting of parallel rows of high magnetic anomalies, probably brick pillars, laid out on an orthogonal grid. These major structures had been discovered by Babak Amin Pour and Hamid Omrani Rekavandi via magnetometer survey, yet no part had ever been excavated. The scale and regularity of these rows of pillars, reminiscent of colonnaded streets, made one wonder what their age and precise function was. It thus seemed essential to excavate a test trench to obtain independent dating evidence; only scientific dating had the potential to decisively date these structures and the earliest occupation of the city, thus establishing what, if any, role the Sasanians played in its foundation or later development.

It was Kiani who first recognised the importance of this major urban centre.13 We also owe the first detailed plan and systematic excavations to him. He initially identified the site tentatively with Sirynx, a major city in the late 3rd century BC according to Polybius’ description14 – mainly, it seems, on the basis of the vast size and evident importance of Dasht Qal’eh15 – and postulated that it was the regional capital as early as the 6th century BC. More recently, however, he argued that Sirynx was much further west,16 a conclusion also reached by other scholars, even if there is no consensus where Sirynx was,17 nor are there much grounds for confidence that the evidence currently available will allow a certain identification. The foundation date and ancient name of Dasht Qal’eh thus remain unknown and the theory that it might represent a Sasanian

We resurveyed parts of the brick rows detected by Amin Pour and Omrani Rekavandi to enable us to locate a promising spot for excavation.18 Our trench (Trench S) was centred on a well-defined high magnetic anomaly, correctly identified, it turned out, as a brick pillar. Scientific dating suggests that it is late Sasanian, dating to the 6th or early 7th century. It is possible that it formed part of a covered bazaar street, almost certainly with mudbrick buildings filling the rectangular spaces between these urban traffic arteries. Built of square bricks of uniformly 31–33 cm diameter and on a solid

Kiani 1982b: 48–52, pls 17.2–22.1, figs 30–31. Polybius The Histories, 10,31,6–13. 15  Kiani 1982b: 52. 16  Kiani 2003: 148–49. 17  Bivar 2003: 151–52; Herrmann 1932; Walbank 1967: 237, 241. 13  14 

18 

96

Amin Pour and Omrani Rekavandi 2007.

Hamid Omrani Rekavandi et al.: The Army and Urbanism surroundings, and the extent to which the yet undated moat outside the city walls was filled with water. Finds from underneath the pillar’s foundations were of preSasanian age, on top of a Bronze Age to Early Iron Age settlement mound, based on Seth Priestman’s analysis of the pottery. We stopped excavation at 5.06 m below the present surface (5.15 m below the site benchmark), as a result of hitting the water table (at 5.07 m below the site benchmark), time constraints and safety concerns, without having reached natural soil. It is important to stress, of course, that the earliest deposit in our trench need not necessarily date the foundation of the city for a number of reasons: • Since we stopped shortly after having reached the water table (which may be, as a result of irrigation, higher today than it had been when the site was first occupied – and probably lower than at the time when the moat around the city was filled with water), we do not know how much deeper occupation deposits in our trench reach and how much, or how little, the age of the earliest deposit differs from the lowest, encountered in our trench • The earliest deposit encountered within a small sondage (the deepest part just measuring c. 1.1 x 1.3 m) need not date to the time of the city’s foundations, as parts of its interior could initially have been unoccupied • The geometric town wall circuit could have been erected over an earlier town (as has been suggested for Darabgird),20 in which case the earliest occupation layer in Trench S might be earlier than the town’s regular defences.

Figure 5. Trench S in Dasht Qal’eh: the deep sondage north of an excavated brick pillar. The segments of the scale each measure 0.50 m

mortared foundation, there seems no doubt that the pillar formed part of a carefully planned monument, built from new bricks (unlike a possible gate on top of the citadel, employing bricks of various sizes, probably robbed from existing structures). Already Kiani found pottery of the 13th and 14th centuries on the citadel, as well as, at deeper levels, wares he tentatively attributed to the early 1st millennium BC, the Parthian and the Sasanian periods.19

While it was impossible to obtain radiocarbon dates for the earliest occupation, five samples obtained from higher levels suggest strongly that the geometric city dates to the 5th or early 6th century.21 Perhaps it was the early 5th century ancient Gorgan. The, seemingly single-phase, wall of the main town, built of 38 x 38 x 8 cm large bricks,22 a size closer, but not identical to, those used in the nearby sections of the Gorgan Wall, has not been dated – but is likely to be contemporary to the earliest Sasanian horizons in Trench S.

Next to the pillar, we dug a sondage, Trench S (Figure 5), which revealed that not just the citadel, but also the town itself (or at least the area in the south-east we explored in our trench) was occupied, intermittently, over a very long period of time. The foundations of the pillar partially overlay a deep pit, almost certainly a well. West and north of the probable well we encountered a sequence of undisturbed deposits. At 3.30–3.45 m depth (below the site benchmark) we encountered a sterile alluvial deposit, which may attest to a period of flooding. The water table is likely to have fluctuated substantially over time, depending on irrigation in the 19 

It appears that Dasht Qal’eh’s defences, and the fertility of the surrounding land, made it an attractive place to re-occupy, thus explaining the extensive built-up of Sasanian occupation layers over a prehistoric mound. The pillared structure in the south-east of the city dates to the late Sasanian phase of its history. The latest occupation of the citadel is Islamic, and the differences Huff 2008: 40–41. Sauer et al. 2013: 401. 22  Kiani 1982b: 48. 20  21 

Kiani 1982b: 52.

97

Sasanian Archaeology in the sizes of the bricks employed suggest that the monument on the citadel is not contemporary with the pillared structure at the foot of this artificial mound.

those beneath. It is also noticeable that lime flecks are far fewer in the lower deposits of the site, i.e. those predating the postulated rammed-earth house, than layers associated with its occupation and more recent phases of re-occupation of the site. These early occupation layers, separated by sterile natural deposits of windblown sediments, seem to derive from repeated episodes of ephemeral occupation, before substantial housing was erected. The geophysical survey and the remains of the walls visible in the sections of our trench suggest that there was just a single phase of substantial architecture. Yet, it is clear that the site had been occupied intermittently for some time before. The nature of the site’s occupation after the collapse of the rammed-earth house is unclear. Higher occupation layers, associated with a high concentration of pottery and lime flecks, suggest further phases of permanent occupation, which, however, are not apparent from the geophysics plot or any traces of surviving architecture within our trench. The extensive study of the pottery by Seth Priestman, both from Trench P and a surface survey across the site, suggests that there was little change and that all can be attributed to a single ceramic phase.

An abandoned city north of the Gorgan Wall (EWS, KH, EM, MM, RA, ME, HO, SP, KR, TW, JR) While defended sites in the most fertile sections of the Gorgan plain, established under Sasanian rule or before, often remained occupied, intermittently or continuously, for centuries into the Middle Ages, many sites north of the Gorgan Wall seem to have been abandoned permanently some time before its construction. Just a few, such as the prominent site of Tokhmaq Tappeh, show signs of extensive later occupation. On the basis of the landscape survey by Hamid Omrani Rekavandi, Tony Wilkinson and Kourosh Roustaei, an urban settlement, Qelich Qoineq, has been selected for magnetometer survey, topographical survey and excavation. This nucleated site of irregular shape (c. 900 x 1,200 m diameter) consists of substantial mounds, with a prominent Qal’eh, surrounded by a moat, in the centre. It is about 2 km north of the Wall, not far from Fort 27 (Figure 6). Our magnetometer survey has revealed the street and housing pattern of this large town in fine detail. Superimposing the magnetometer plot (Figure 7) over our topographical survey (Figure 8), it appears that the mounds represent collapsed architecture, the depressions, roads, the moat and other unoccupied spaces between blocks of housing. We dug a sondage (Trench P) in the corner of a large house in the area of what appeared to be one in a row of rooms facing the street. Our excavation revealed the outer – possibly rammed earth – walls of this building, as well as earlier and later occupation horizons. While these walls were not particularly distinct, they were evidently constructed on a level surface and they are abutted by horizontal layers, notably an associated occupation deposit in the interior. The absence of any traces of mud bricks within these features, as well as the manner in which they appear to have disintegrated at ground level, suggests that we are dealing with rammedearth walls, a widespread construction technique in the ancient world, and still employed in some parts of the world today.23 Later, an internal mudbrick wall was built on top of occupation debris, which had built up in the interior of the house. The house was thus occupied for some time and architecturally modified at a late stage of its occupation. After the collapse of the house’s outer walls, the area of our trench appears to have seen further phases of human habitation, as suggested by occupation layers, rich in finds, at a higher level. The occupation deposit associated with the rammed earth house was of greater thickness than 23 

The radiocarbon dates provide no firm answer to this question. They indicate that the site’s occupation (in the area of Trench P at least) lasted for a maximum of c. 385 years, but could equally be confined to a much shorter period of time. The stratigraphically youngest sample dates to 761–416 BC at 95.4% confidence, the oldest, to 791–543 BC, the one, in between to 756–415 BC. We thus have a precise time frame for the expansion of settlement, based on complex irrigation systems, into the steppe north of the wall and its eventual abandonment. The abandonment of this large town (probably dependent on the irrigation system north of the Wall) dates to a period long before the Wall was built and thus is historically unrelated. It nonetheless sheds fascinating light on a period of history roughly a millennium earlier. The absolute dating of the pottery at Qelich Qoineq to the 8th–5th centuries BC is a significant step and promises to yield significant insights into the period of occupation of sites yielding a similar finds spectrum. It is also of relevance to explain the course of the Gorgan Wall, which enclosed all the major Sasanian sites in the Gorgan Plain. The major settlements to the north had been in ruins when it was built. The historical context of the Gorgan Wall (EWS, HO) Radiocarbon samples from the 2007 and 2008 seasons have now allowed us to further narrow down the date of the Gorgan Wall’s construction and that of the occupation of Fort 4. Particularly interesting for the construction of the Gorgan Wall is a sample from the

Herrmann 1999: 48–49, 51.

98

Hamid Omrani Rekavandi et al.: The Army and Urbanism

Figure 6. Satellite image of Qelich Qoineq, a major town north of the later Gorgan Wall, with the location of our magnetometer and topographical surveys. Corona satellite image by courtesy of the US Geological Survey (modeified by the authors; see Figures 7–8 for the results of the magnetometer and topographical surveys)

before AD 429 in the west and/or lasted beyond AD 534 in the east. Two additional samples are being dated and may allow a further refinement of the chronology. Our results are strikingly similar to the date for the construction of the Tammisheh Wall of AD 402–537,25 as well as to the above-cited dates for the occupation of Qal’eh Kharabeh.

bottom fill of the brick kiln west of Fort 30, dating to AD 342–534. In conjunction with the date obtained from the bottom fill of a brick kiln near the east end of the Wall of AD 429–574,24 this results allows us to narrow down the Wall’s likely construction to about a century, i.e. to the time between AD 429 and 534. As one sample is from the western section of the Wall, the other from near the opposite end, we cannot as yet, however, exclude the possibility that building the Wall started 24 

Nokandeh et al. 2006: 161.

25 

99

Nokandeh et al. 2006: 161.

Sasanian Archaeology

Figure 7. Qelich Qoineq: Results of the magnetometer survey (by Abingdon Archaeological Geophysics and the ICHHTO), revealing the layout of the town, its housing blocks and streets, and the location of Trench P (see Figure 6 for the location of the survey)

Figure 8. Qelich Qoineq: Topographical survey (by James Ratcliffe, Kristen Hopper and Amin Nazifi) (see Figure 6 for the location of the survey)

100

Hamid Omrani Rekavandi et al.: The Army and Urbanism Our radiocarbon dates suggest strongly that the construction of the Gorgan Wall and the occupation of Qal’eh Kharabeh fall some time between the early 5th and the first third of 5th century; OSL dates would add support for a 5th century date (prior to AD 486?) for the construction of the Gorgan Wall.26 While an earlier date indeed seems more likely, we should not as yet exclude the possibility that the large-scale construction of military infrastructure in the Gorgan plain dates to the late 5th or early 6th century.

AD dates Event 420–438

440s 453/454 457–484

In 2008 we used archaeomagnetic dating in addition to radiocarbon and OSL dating. The first of these methods was pioneered during this season by David Greenwood and Dr Cathy Batt of the University of Bradford and promises to be a significant step in establishing an archaeomagnetic dating curve for Iran, which may help to date other sites. At the same time Seth Priestman, Mohaddeseh Mansouri Razi and others systematically processed and analysed pottery found during survey and excavation at a range of sites. Once the scientific dates will be available, it should be possible to calibrate the pottery sequence. All of these dating techniques taken together may allow us to narrow down the date of the Gorgan Wall even further.

457–484

484

488–496 499–531

531–579

Important historical events in the Gorgan plain in the 5th and 6th centuries (EWS, HO) Table 1. Important historical events in the Gorgan Plain in the 5th and 6th centuries: this short summary is largely based on translations and secondary literature, in addition to those specifically cited, notably Bosworth (1999), Browne (1905) and Thomson (1991). AD dates Event

420

26  27 

Yazdgird II resided for seven years at Gorgan and established a new town in the area (Schindel 2004: 386) Yazdgird II passed through Gorgan on his way to Nishapur (Schindel 2004: 386)

the 9th/10th century author al-Tabari (874, 895) attributes foundation of a town and fortification works near Gorgan to Peroz; the 13th century author Ibn Isfandiyár (1,2) refers to the ‘Moat of Peroz’ (probably the Gorgan Wall); the 14th century author al-Mustawfi (trans. Le Strange 1919: 156; see Le Strange 1915: 159/ ١٥٩) equally attributes the construction of the Gorgan Wall to king Peroz (Kiani 1982b: 11) coin evidence suggests start or marked increase in trade between the Sasanian Empire and the Perm region, probably via the Caspian Sea and the Volga and Kama rivers during the reign of Peroz (AD 457–484) or shortly after, while his coins still dominated circulation (Goldina and Nikitin 1997)

460s–474 Peroz’s first war against the Hephthalites in the area of Gorgan (Procopius, Wars I.3.1–3); the king received here an ambassador from the Eastern Roman Empire, Constantinus (Priscus, Fr. 41,3)

The creation of the Gorgan Wall (in the 5th or, possibly, early 6th century) was preceded by generations of intermittent warfare at the Persian Empire’s northern and eastern boundaries, even if the textual and numismatic evidence is often too sketchy and imprecise to allow any detailed reconstruction of the precise territories affected.27 Historical records do not allow us to pinpoint exactly the period of the Wall’s construction, yet they certainly leave no doubt that the Gorgan plain was of major strategic importance for most of the 5th and early 6th centuries, as the following brief overview may demonstrate:

410s

Varahran V passed through Gorgan on a major campaign against enemies in the north or east of the Persian Empire (Schindel 2004: 365–66; Farrokh 2007: 211–12)

Yazdgird I (AD 399–420) opened, late in his reign, a new mint at ancient Gorgan (Schindel 2004: 335), perhaps Dasht Qal’eh

Peroz’s second war against the Hephthalites: he stayed and gathered troops in the Gorgan area, followed by the king’s defeat and death (Procopius, Wars I.4.10; Łazar, 85/ 155)

Kavad I established close diplomatic ties with the Hephthalites and was responsible for an extensive building programme in the area (alTabari, 895); he founded, according to 13th/14th century author al-Dimishqui, the sea-port of Abaskun at the south-east of the Caspian Sea (Mehren ed. 1874: 314) Khusrau I is credited with the construction of ‘towns, castles, fortified mounds, and many other buildings, which would serve as a protection for the people of his lands, where they might seek refuge from the enemy in the event of a sudden attack’ (Bosworth 1999: 152) in the vicinity of Gorgan by al-Tabari (895); the 16th century author Siyâqi-Nizâm attributes the construction of the Gorgan Wall to Khusrau I (Kiani 1982b: 11). Perhaps the brick pillar colonnades at Dasht Qal’eh may belong to this phase of history.

It is perfectly possible that Ibn Isfandiyár is right in attributing the Gorgan Wall to Peroz. This is perhaps indeed the most likely option, considering the length and gradual escalation of the confrontation with the Hephthalites during his reign. Yet we cannot exclude as yet a construction before or (less probably) after Peroz’s time. Whatever the date of initial construction, the grand system of Walls, forts and fortresses need not have

Yazdgird I died at Gorgan (Tabari, 850): royal presence suggests increased military/strategic importance

Nokandeh et al. 2006: 159–62. Schindel 2004: 245–47, 335, 344; Bivar 1983: 211–12.

101

Sasanian Archaeology been completed during a single king’s rule. It is perhaps more probable that it was developed over generations and this also seems to be reflected in the historical accounts, notably those of Tabari. Much still needs to be explored, e.g. whether sites, like Qal’eh Kharabeh, were campaign bases preceding the construction of the Gorgan Wall, or hinterland compounds occupied during its initial stages of construction or occupation, though not for long. Similar questions arise concerning cities in the hinterland of the Wall.

The length of the fort’s occupation thus could have amounted to a minimum of about 70 years or up to a maximum of almost 250 years. Despite this range of possibilities, it is clear now that the fort was occupied for at least three generations, but more probably for a century or two. While it may seem dangerous to extrapolate the period of occupation of the wall forts from two trenches in one of them, it would have made little strategic sense just to maintain a garrison in few forts, nor were they attractive for civilian reoccupation, in the light of their marginal geographic location. It is thus not certain, but likely that the results obtained from Fort 4 are representative for the history of the Gorgan Wall as a whole. If so, this major defensive barrier appears to have maintained a garrison for a considerable period of time, starting during the 5th or early 6th century and extending into the 7th century AD, the very century when the Sasanian empire itself ceased to exist.

The length of occupation of Fort 4 (EWS, GA, EM, SP) Two new samples from Trench J are of considerable interest to the history of Fort 4 and its water supply. One sample, from the middle fill of a gully, has been dated to AD 422–542. This suggests that Fort 4 could have been constructed at the same time as the Wall (though it is also possible that it was a few years, decades or even a century later). It is in any case no later than the AD 540s at the very latest. The gully might have been amongst its earliest installations. From the upper fill of the gully we obtained a sample dating to AD 543–638. While the relative date of the samples is consistent with their position in the stratigraphy, the broad margin of error seems to leave all options open, from the two samples being contemporary to being over 200 years apart. One would be inclined to think though that the gradual filling-up of the gully extended over a period of time. Taking into account all the dates from Fort 4, including those from 2006,28 the chronology of deposition of organic material at the site can be summarised as follows: Table 2. Chronology of samples from Fort 4 Date AD Comment 422

earliest likely date of earliest dated sample

542

latest likely date of start of occupation

604

earliest likely date of latest sample and, as deep in the stratigraphy of Trench H, suggesting that occupation continued well beyond

655

28 

latest likely date of latest sample (NB: the AD 666 terminus ante quem from a Trench H sample is probably cancelled out by two samples above with termini ante quos of AD 644 – unless we are dealing with re-deposited material)

Omrani Rekavandi et al. 2007: 125.

102

A Temple Among the Fields: Preliminary Reconstruction of an Agricultural Landscape in the Serakhs Oasis, Southern Turkmenistan Barbara Kaim Institute of Archaeology, Warsaw University, Poland.

Excavations by the Polish-Turkmen expedition were conducted at the site of Mele Hairam between 1997 and 2001. This site consists of one larger mound (c. 40 x 30 m) rising nearly 7 m above the surrounding plain (at its highest point it reaches 269.2 m above sea level), and four smaller mounds ranging between 2 and 3.5 m in height scattered in the western part of the site (Figure 1). Excavations of the main mound have revealed the remains of a fire temple (Figure 2). Thus far, six temple rooms have been uncovered of which some were closely connected to the cult, while others are believed to have acted as subsidiary chambers. The coins recovered from the excavations range from late Parthian issues struck in Margiana in the 2nd century to Sasanian coins minted by Varahran II (AD 276–293) and Shapur II (AD 309–379). Reconstruction of the architectural history of the temple will be possible only upon the completion of the excavations. The architectural features of the temple discovered at Mele Hairam are not, however, the subject of this paper, as this and the most important finds have been outlined in other preliminary reports.1 A very important yet very difficult problem is the reconstruction of landscape around the temple and livelihood of the temple service. Currently the reconstruction is at its initial stage, so the results presented herein are of preliminary nature only but builds on a previous paper published on the Parthian settlement patterns in the Serakhs oasis.2 Key datasets used for this outline study are the archaeological ground survey information gathered by archaeological teams, topographical maps of the region as well as satellite images. The analysis and assessment of these already collected data will facilitate defining both the quality and quantity of data to be gathered at the next stage of investigations.

1  Kaim 2000; 2001; 2002; 2004; 2010; 2012; 2016; 2018. This project was supported by Grant N 10900731/0426 awarded by the Ministry of Science and Higher Education. 2  Kaim 2008.

Sasanian Archaeology (Archaeopress 2022): 103–108

The site of Mele Hairam is located some 15 km east of the modern town Serakhs, the main administrative centre of the oasis, in the sub-delta of the Tedjen river (Figure 3). In the past, the region of Serakhs oasis was densely populated as evidenced by numerous large and small mounds breaking the flat landscape. However, although the territory of the Serakhs oasis is very rich in archaeological monuments, it has not previously aroused any great interest among scholars. The history of archaeological investigations in the region is therefore very brief. Prior to the first regular excavations, several surveys and expeditions were launched by the Institute of Turkmen History. These focused on Islamic monuments scattered in the immediate vicinity of Serakhs, in particular two mausolea: one of Yarty Gumbez and the other of Abu Fazl.3 A.A. Marushchenko, while excavating at Old Serakhs in 1953, collected evidence that the site had been inhabited continuously at least from the end of the 2nd millennium BC up to 1832.4 During the same season, A.A. Marushchenko carried out excavations of the eneolithic settlement in the old bed of the Tedjen river, 102 km north of Serakhs. In 1953–1956, a branch of Southern Turkmenistan Archeological Complex Expedition led by K.A. Abdykov studied the sites situated along the road that links Serakhs and Merv.5 In 1964 a team from the Turkmenian Academy of Science led by O. Orazov began a seven successive year investigation in the Serakhs oasis. During these campaigns, test trenches were opened on several mounds and an archaeological survey was carried out with the aim of establishing the regional land use and settlement patterns in the Serakhs oasis. A very schematic archaeological map of the region was then produced (Figure 4). Due to technological limitations, the map, or rather a sketch, is not sufficiently accurate, to put it mildly, and what is worse, its description Masson 1953: 9, 65–66. Marushchenko 1956. 5  Abdykov 1959. 3  4 

Sasanian Archaeology

Figure 1. Mele Hairam: contour map of the site (drawn by M. Wagner)

104

Barbara Kaim: A Temple Among the Fields

Figure 2. Plan of central part of fire temple at Mele Hairam (drawn by M. Wagner)

Figure 3. Map showing location of Mele Hairam (drawn by M. Wagner)

105

Sasanian Archaeology

Figure 4. Archaeological map of the Serakhs oasis (after Orazov 1973: 23)

contains many departures from the geographical reality.

what he regarded as the Sasanian stratum. We were not able, however, to confirm this claim during our works. Nevertheless, as a result of this very short archaeological activity in the Serakhs oasis, some 60 archaeological sites dating from the 1st millennium BC to the 19th century have been identified. It should be noted, that the areas along the main reaches of the

In 1996 the Polish-Turkmen team commenced the investigation of the fortifications of Old Serakhs on the southern side of the citadel, and precisely where O. Orazov had stopped his excavations upon reaching 106

Barbara Kaim: A Temple Among the Fields Tedjen river, which forms the present Turkmen-Iranian border, and along the stream called Shorkel or Khorkhor, which delineates the eastern border of the oasis, still remain totally unknown archaeologically.

the Karman-yap, and a few km further down, Salyryap bifurcates into two big systems: Kichi-aga-yap and Ata-yap, with many channels, both narrow and wide, of which some are still functioning.

There is no doubt that the economy in the oasis was traditionally based upon herding and agriculture. The agricultural production in Serakhs oasis was and still is almost totally dependent on irrigation. The canal systems criss-cross the vast expanse of the alluvium and represent structures of different periods of the past. The access to water was essential, and the control and management of the region’s water resources was always an important factor in the history of the oasis. The main water source for the Serakhs oasis is the Tedjen river which originates in the Kuh-e Baba range of central Afghanistan. This river is called the Hari Rud in Afghanistan and is the main water supply for the irrigation system in the Serakhs oasis. The river crosses the Serakhs oasis and flows north to the Tedjen oasis, before disappearing into the desert sands. Tedjen’s regime is characterised by a main spring flood, with several peaks, caused by heavy rainfalls. More than occasionally rains can lead to short-term floods during winter, while in July–November the river dries up and its water does not even reach Serakhs.

An issue of possible climatic change in central Asia during the last 10,000 years remains undecided. Lisitsina and Gerasimov provide evidence for the climate stability in the said period.11 However, the late P.M. Dolukhanov argued in favor of desiccation of the environment over the last four millennia.12 In any case it seems certain that for thousands of years the Serakhs oasis region as well as any other central Asian areas did not receive more rainfall than it does today (annual precipitation at about 172 mm). Moreover, the rainfed rivers as Tedjen, are totally dependent on yearly climatic changes and show very unstable behaviour. So al-Yaqubi’s and Istakhri’s reports, if precise, may be concerned with extremely dry years. Istakhi wrote that there had been few cultivated fields, and that pasturage concentrated in the environs of the town,13 what could be explained as an effect of drought. Climatic conditions similar to the present day only enabled farming when supported with artificial irrigation. On the other hand, a tendency is observed of re-activating and renovating of old temporarily abandoned canals rather than constructing new ones. It is thus conceivable that a canal along which archaeological site in located, follows the course of an ancient canal. Therefore, given the basic layout of well dated settlements, it is possible to establish their relationship to the irrigation canal, and thus correlate the history of both.14

The irrigation system employed in the Serakhs oasis is still poorly studied. Scholars know nothing but its early history and some development of its channels and levees. Any available written sources recording the relevant information mostly date to later periods. Al-Yaqubi noted that there had been not any canal or stream in Serakhs and city dwellers had to draw water from wells.6 This account seems to be confirmed by Istakhri, the Persian geographer, in his Kitāb al Masālik wa ‘l-mamālik where he state that the fields were watered with rain waters.7 However, in the late 10th century, al-Muqaddasi wrote about the Sanfaghar canal carrying water from the Tedjen river;8 we are also told by the same author that the head of the canal was about 2 farsakhs south of the city. Identical information is provided also by ibn Rustah.9 In the 15th century, according to Hafizi Abru, there were five channels distributing water in the Serakhs oasis.10

The earliest occupation in the oasis area dates to the eneolithic period. The site excavated by A. Marushchenko is located next to an ancient bed of the Tedjen river. It seems that early farmers used uncomplicated techniques such as canal levee breaks or sluices to control water flow into fields or settlements.15 Taking into account that a site dated to the 3rd and 2nd millennia BC has been discovered on the other side of the river, near the village of Cheshmeh Shur,16 it is highly possible that the apparent lack of settlements of this period in the Serakhs oasis is due to the inadequacy of archaeological research. A discovery of a ‘miniature column’ at Mele Hairam,17 this being a distinctive artifact known from several Bronze Age sites of the Iranian plateau, Central and Southern Asia, seems to confirm this supposition.

As in the past, the present network of irrigation canals distributes water supplied by a channel called Salyryap. The origins of this canal are not yet clear (several old beds are visible on the satellite images) and some field investigations are required to establish its history. About 10 km downstream the channel feeds two secondary branches, these being the Khana-yap and

Lisitsina 1978; Gerasimov 1978. Dolukhanov 1988. 13  Istakhri 1927: 273. 14  Adams 1962a: 112–13. 15  Orazov 1973: 52–53. 16  Malek Shahmirzadi 1987: 73. 17  Wagner 2005. 11  12 

al-Yaqubi 1892: 279. Istakhri 1927: 272–73. 8  al-Muqaddasi 2001: 268. 9  Ibn Rusta 1892: 173. 10  Hafizi Abru 1317/1938: 162. 6  7 

107

Sasanian Archaeology The scarcity of information is also evident for the settlement history and patterns existing in the following periods up to the first half of the 1st millennium BC. At Old Serakhs A. Marushchenko identified some pottery fragments dated to this period.18 Three other contemporary sites have been identified by O. Orazov, these being Movlek depe on the north border of the oasis and Akcha depe and Besh depe in the northcentral part.19 In the vicinity of Movlek depe situated some 36 km north of Serakhs, O. Orazov identified a dry canal bed of 6–7 m wide. If this identification is accepted, this must have been a very long canal indeed crossing the Serakhs oasis in the first half of the 1st millennium BC. It is not unlikely because in the much better studied area, i.e., the Tedjen delta, the irrigation network of the 4th millennium has been revealed,20 and it is generally accepted that within the first half of the 1st millennium BC all of the deltas of the main rivers of central Asia were already provided with large-scale irrigation systems.21

The site of Mele Hairam is situated between two irrigation canals. Their relatively large and high (3–4 m) levees as well as the lack of any traces of old canals traces in the vicinity suggest that these two canals were active at the time of the temple’s activity. The course of these two canals could delimit territorial boundaries of the temple property: in the literary evidence there are clear indications that a founder of a fire usually endowed it usually with lands.22 Two fragmentary ostraca found at Mele Hairam deal with the lease of a vineyard and delivery of wine to the treasury. This is of course an indication of how one of possible ways in which fields watered by canals were exploited. Zooarchaeological evidence provides some additional information about the subsistence activities of people inhabiting this part of the Serakhs oasis. The majority of the excavated animal-bone from Mele Hairam has been identified as belonging to mammals and a smaller number of birds. The mammalian osteological material consists mainly of domestic animals, i.e., sheep and goats, while cattle bones occupy the second place. The presence of cattle bones proves that the region once must have been abundant in water resources as cattle require large quantities of this. Wild animals were represented by a few goitred gazelle bones. Until recently, this species of gazelle, known locally as the djeiran, was frequent in all the deserts and semi-deserts of Asia, where it was a common game animal. The gazelle bones excavated at a number of sites of different periods across central Asia confirm that people hunted gazelles for centuries.23 The presence of gazelle bones at Mele Hairam furnishes evidence that hunting was also practised by the inhabitants of the region in this late period, albeit not as frequently. There are also indications for birds as a supplement to the diet of the temple service: although some bones are not diagnostic, others belong to chickens.

The small number of settlements benefiting from this channel is surprising however: namely Movlek depe, Akcha depe, located about 9 km north of Serakhs, and Old Serakhs, whereas Besh depe is in the vicinity of old Tedjen’s bed. It seems more probably that that the channel reaching Movlek depe was a natural distributary of the main river but this problem requires further research. Based on the currently available evidence, the late Parthian period seems to shows some settlement and land use intensification when compared to the preceding period. Three sites dated to the late Parthian period have been identified by O. Orazov. At present this cluster may be extended by the site of Mele Hairam and Old Serakhs. All these sites are connected with the Kichi-aga-yap system carrying water from the Salyryap channel, and all these sites are located in the north central part of the oasis. Old Serakhs is near a point where the main Kichi-aga-yap channel divides into three main branches. This location indicates that Old Serakhs was a main centre of the oasis at least in this period, if not earlier. Any possible relationship between Serakhs and Mele Hairam is to be established.

The Zoroastrian fire temple discovered at Mele Hairam was closely tied to the agricultural territory surrounding this imposing building. The reason for the abandonment of the temple is perhaps to be sought in a dramatic change in the natural environment in this part of the Serakhs oasis.

Marushchenko 1956: 174–75. Orazov 1973: 41–45, 95–98. 20  Lisitsina 1969; Kohl 1984: 76–91. 21  Gubaev, Koshelenko and Tosi (eds) 1998; Besenval and Isakov 1989; Askarov 1992. 18  19 

22  23 

108

See Nöldeke 1973: 111–12; Boyce 1968. Formozov et al. 1961: 81; Harris and Masson et al. 1993.

Excavations at Kirpichli depe in Dehistan A.E. Atagarryev Ashgabat, Turkmenistan.

Introduction1 Despite its historical importance, particularly during the Parthian, Sasanian and Islamic periods, the province of Dehistan (in south-west Turkmenistan) remains little explored in terms of its archaeology. Pioneering surveys were carried out in the 1960s and these showed that this region had a complex pattern of settlement based on irrigation agriculture which began during the Iron Age and was concentrated along canal off-takes from the massive Shahdiz canal which in turn took its water from the Atrek river.2 This early period of occupation has been the main focus of interest and is referred to as the Archaic Dehistan period.3 After a break, settlement resumed during the Sasanian period but now was concentrated in the northern part of Dehistan: the number of sites is greater than before and they tend to be fortified. This has been interpreted as evidence for sedentarised nomad occupation by Hephthalites and then Tchols prior to the Arab conquest in AD 716/717.4 Orta Depeslik, Khanly depe and Geoktchik depe are among the few late period sites to be excavated here.5 During medieval times, this northern region grew even more prosperous and a major city developed at the site of Misrian where there are important standing medieval remains,6 and extensive archaeological investigations were carried out here by the author’s father.7 In 1991/92, the author conducted the first excavations at the extensive site of Kirpichli depe and the results provide important new information which allows reevaluation of earlier excavations in this region. The extensive archaeological site of Kirpichli depe is located 18 km from the medieval city of Misrian. Kirpichli depe measures 2 km in length by about 150 m across. It is 1  Due to unforeseen circumstances, these excavations were never resumed and, apart from a short conference abstract (Atagarryev 1992), this report is likely to constitute the only publication of these brief but important excavations. The text was kindly translated by Dr G. Tsetskhladze. Identifications and comments on the coins, bullae and seal were provided by the late S.D. Loginov and the late A.B. Nikitin who also kindly made and supplied the photographs, and the drawings, catalogue and comments on the small finds, glassware and pottery added by the editor at the author’s kind request in order to place this important material within a wider context. 2  Atagarryev and Lisitsina 1970; Lecomte 1999. 3  e.g., Muradova 1991. 4  Masson 1961. 5  Sarianidi 1952; Lecomte 1999: 147–66. 6  Pugachenkova 1953: 192–227. 7  Atagarryev 1986.

Sasanian Archaeology (Archaeopress 2022): 109–126

orientated south-east/north-west, consists of as many as 42 different mounds of various sizes and is situated on natural gravels (Figure 1). Four trenches were excavated on four separate mounds during 1991/92 and the remains of mudbrick architecture were found in each (Figure 2). The results are summarised below. The excavations Trench 1: Mound 33 Eleven rooms belonging to a large mudbrick structure were excavated, either in full or at least partially and the building probably extended further to the west. The rooms were rectangular and of different sizes. The mudbricks measured 36 x 36 x 8 cm. The finds consisted of spindle whorls (Figure 11, table 5), astragali, a glass backgammon piece (Figure 12), a glass bead, partly reconstructable glassware (Figure 15.8), pottery (Figure 18.5) and carbonised plant remains. • Room 1: 3 x 2.5 m; in the eastern corner some carbonised gunshut [a local plant] was found at a depth of 60 cm. • Room 2: 4.5 x 2.5 m; a small pottery jar (khum), missing its upper portion, was found close to and midway along the eastern wall and a cluster of approximately 20 astragali were found in an ashy area in the southern corner. • Room 3: 6 x 2.5 m; many mudbricks were found in the centre of this otherwise poorly preserved room. • Rooms 4–5: the walls of these longer southern rooms were poorly preserved but room 4 measured 6.5 m long and 3 m across. An earring and a glass bead were found next to one wall of this room which was also the findspot of two glass vessels (Figure 15.8). • Room 6: 1.5 x 1 m; no artefacts were found in this small narrow room at the corner of the building. • Room 7: 4.5 x 3.3 m; the foundations of this room were of mudbrick, each measuring 35 x 35 x 7 cm, and there was a square brick platform on the western side. • Room 8: about 11 x 6 m; carbonised grain and gunshut were found here, probably once contained in a cloth bag.

Sasanian Archaeology

Figure 1. General plan of the site of Kirpichli depe showing the extent and alignment of the different mounds

• Room 9: 2.8 x 1 m; some carbonised gunshut was found in the western corner. • Room 10: 15 m long; the north wall was not found and this might have been a courtyard area. • Room 11: 3.2 x 1 m.

• Room 1: 4 x 3 m; these walls survived only to an approximate height of 10 cm but remains of carbonised cotton bolls were found close to the north-west wall. • Room 2: 6 x 3 m; the wall was built of a double row of mudbricks but additional (subsequently destroyed) brick construction was found in the western corner. • Room 3: 3.5 x 3 m; the northern wall was either not preserved or it opened onto a courtyard on this side; fragments of metal objects were found within this room. • Room 4: 6.5 x 3 m; several metal objects were found here, plus several joining pieces of a glass

Trench 2: Mound 35 The corner of a large building was located here and a long row of six rooms, all constructed of mudbricks measuring 36 x 36 x 8 cm, was excavated to floor-level. Glassware (Figure 15.11), pottery (Figures 17, 18.2–3, 19), fragmentary metalwork, spindle whorls and carbonised plant remains were found. 110

A.E. Atagarryev: Excavations at Kirpichli depe in Dehistan

Figure 2. Plans of the architectural remains excavated on Mound 33 (Trench 1), Mound 35 (Trench 2), Mound 36 (Trench 3) and Mound 37 (Trench 4)

vessel (Figure 15.11), three large crushed pottery vessels (Figure 17) and some carbonised grain. A fireplace was situated in the north-west corner. • Room 5: 6 x 3 m; carbonised grain, a small pottery jug (Figure 19) and numerous spindlewhorls were found here. • Room 6: 4 x 3 m; the north-west wall was not preserved and no finds were recovered here apart from pottery (Figure 18.3).

following: 75 = AD 726, 80 = AD 731. There is a noteworthy absence of any marginal inscriptions characteristic of the 8th century hemidrachms of Tabaristan on the obverse of these coins. Separate coin finds Almost all of these coins are rather badly preserved and the loss of metal therefore must be taken into account when considering their weights (Table 2).

Trench 3: Mound 36

Stamp seal

Two rooms were excavated, the eastern and southern walls of which had been destroyed. Remains of carbonised grain were found.

A veined carnelian ellipsoid stamp seal was also found at the site (Table 4; Figures 5, 9.1). The face was engraved with the representation of a plump rounded bird walking left, with each leg indicated by a pair of engraved lines, one indicating its beak, three horizontal parallel lines indicating the tail and five oblique lines indicating its breast. The scratch-style correlates to one of five different styles, each reflecting a different technique of engraving.8 It may be typologically dated to the late Sasanian period but an exact dating is not possible. It is possible that this form of stamp-seal continued to be made and used for up to a century or perhaps more after the Arab Conquest as they are reported from the single-period 8th century site of Jazirat al-Hulayla9 and three surface finds from the early medieval port of Siraf,10 both in the Persian Gulf.

Trench 4: Mound 37 A single room constructed of mudbrick and measuring 3.55 x 5.2 m. on the interior was excavated, with a doorway on the north side and possibly a second on the opposite wall. A fragmentary pottery jar (khum) was found next to the west wall; astragali were also recovered and remains of reeds observed. Coins, bullae and seal The numismatic finds comprise a hoard of eight Sasanian and Tabaristan drachms (Table 1: Table 2: nos 9–12; Figures 3–4). The hoard

Brunner 1978: 134. Sasaki 1996: 179, 186, fig. 13, pl. 37. Whitehouse 1972: pl. XIId; there is no stratigraphic evidence to support a Sasanian origin for this site, nor has any Sasanian pottery or glassware been identified in a thorough re-assessment of the finds in the British Museum. 8 

The name of the ruler represented on the last four coins – plhw’ – is most probably to be transcribed as Farrox. If we take it for granted that the post-Yazdgird era of Tabaristan was intended, the dates correspond to the

9 

10 

111

Sasanian Archaeology Table 1. The hoard from Kirpichli depe Mint

D mm

Th mm

Wgt gr.

Die point

Fig.

sy (?) = Shirajan

23.5

1.0

2.03

4 o’clock

3

bys = Bishapur

24

1.0

2.38

4 o’clock

3

24.5

1.0

1.90

9 o’clock

3

4

Spahbed Farrox (c. AD 720s–730s); obverse: king’s bust to the right, behind head: GDH ‘pzwt; in front plhw’ = Farrox; two dotted rims; on the margins: stars within crescents at 3, right: tpwlst’n = 6 and 9 o’clock; no marginal inscriptions. Tabaristan Reverse: fire altar with two attendants standing frontally by its sides; left: pnchpt’ (t) - year 75; three dotted rims; four groups of three dots and four stars within crescents on the margins

24

0.8

1.96

3 o’clock

3

5

The same, but the date: hst’t = year 80

tpwlst’n = Tabaristan

24

0.8

1.96

3 o’clock

3

6

The same; date: hst’t = year 80; broken

tpwlst’n = Tabaristan

23

0.7

1.85

2 o’clock

3

7

The same; date on the reverse indistinguishable

tpwlst’n = Tabaristan

23.5

1.0

2.64

?

4

tpwlst’n = Tabaristan

24

1.0

1.83

?

4

Inv. Nr. Description 1 2 3

8

Hormizd IV (AD 579–590), drachm; date: dw’cdh = year 12

Khusrau II (AD 590–628), drachm; date: TLT’ = year 3 Khusrau II (AD 590–628), drachm; date: TLT’ = year 3

The same; date on the reverse indistinguishable

my (?) = Meshan

Table 2. Single coin finds from Kirpichli depe Inv. nr. Description 9

10

11 12 1  2 

Sasanian: Ardashir III (AD 628–630), drachm. Obverse: king’s bust to the right, in front: ‘rths[ ; rev.: T]LYN = do – year 2. Fragment and about one half of the coin is missing

Mint

D mm

Th mm

Wgt gr.

Die point

Fig.

[‘yl]’n = Susa

23

0.7

0.97

3 o’clock

4

1.1

0.68

?

4

1.5

1.04

12 o’clock

4

1.0

1.69

12 o’clock

4

Sasanian: Khusrau II (AD 590–628), drachm. Only the name hwslwb is legible on the obverse. Fragment: 15 x 13 mm

Chorasmian bronze, first half of the 8th century.1 Obverse: portrait of a king to the right; reverse: tamga, traces of Chorasmian inscription. Fragment: 18.5 x 11 mm Chorasmian bronze, first half of the 8th century2

19.5

Vajnberg 1977: 62–63, pl. XIX, type G/11. Vajnberg 1977: 61, pl. XX, type G.V/10 = King Savshafarn, second quarter to mid-8th century.

112

A.E. Atagarryev: Excavations at Kirpichli depe in Dehistan

Figure 3. Coins from Kirpichli depe

113

Sasanian Archaeology

Figure 4. Coins from Kirpichli depe

Table 4. Stamp seal from Kirpichli depe Description Stamp seal bearing a scratch-style representation of a plump rounded bird walking left; ellipsoid with an oval sealing surface (10 x 7.5 mm) and a waisted bore perforated from two directions (perforation D 3.5 mm); light orange carnelian with white veining

L W Ht Figs mm mm mm

15

12.5

13

5, 9.1

Figure 5. Face of the stamp seal from Kirpichli depe

114

A.E. Atagarryev: Excavations at Kirpichli depe in Dehistan Table 3. Impressed clay bullae from Kirpichli depe Description Clay bulla with an oval sealing (10 x 7.5 mm) representing an ibex standing to the right before a tree; string hole perforated through the back

Fragment of a clay bulla with three different sealings, of which only one is well-preserved: a. Official seal, circular impression (diameter 20 mm). A Middle Persian inscription - centre: gwlg’n (in two lines), the name of the city and province of Gorgan; along the border: y’tkgwby W d’twb’l ZY dlgws’n, defender of the poor and judge b. A partly preserved oval sealing c. 9 mm). A winged horse standing to the right c. A partly preserved sealing (c. 15– 17 mm). Only the edge, with traces of an inscription

L W Th Figs mm mm mm 22

40

21

37

12

17

6, 9.2

Figure 6. Clay bulla from Kirpichli depe with the impression of a personal seal representing a standing ibex

7–8, 9.3

Clay bullae Two impressed clay bullae were also found at the site (Table 3; Figures 6–9). The official seal on the second of these resembles late or post-Sasanian bullae which bear the sealings of judges of Gorgan, and found in that region at Tureng tepe and Tepe Kabudan.11 Other finds A small but interesting collection of other finds was made at the site, some from excavation and the remainder from surface collections. They are summarised below, beginning with the small finds. Small finds Despite the small sample size, this group is a rather interesting collection (Figure 9). They include two very distinctive press-moulded fired clay figurines. The first was made of light brownish clay and had numerous finger-tip impressions on the reverse from where it had been pressed into a mould before drying and firing. It was complete apart from its head and the lowermost portion and represented a standing figure with arms folded across its chest and wearing a long dress with numerous parallel folds or creases below the waist with the upper portion highlighted with an oval panel with a serrated edge which was probably intended to

Figures 7–8. Clay bulla from Kirpichli depe with the impression of an official seal in the centre

represent a separately stitched-on embroidered panel (Figures 10.1, 11). If this interpretation is correct it may represent the oldest depiction of this style of women’s dress in this region. It also constitutes the latest dated example in southern Turkmenistan of a very long and widespread central Asian tradition of making small press-moulded kiln-fired representations of standing female (and less often male) figures, who are shown in full dress and often holding an item such

Frye 1970a: 84, no. 776; Gignoux 1978: vol. II, no. 6.1; Gignoux and Gyselen 1987: 139-40, IBT 3–4 (Iran Bastan Museum, Tehran), 72, MOT 7–8 (Mochiri collection), 105, KP 3–4 Kevorkian); Gyselen 1987.

11 

115

Sasanian Archaeology Table 5. Spindle whorl dimensions from Kirpichli depe Description

D

perf D

Findspot

Figs

Spindle whorl; chipped sherd disc; smoothed edges

3

0.6

-

11: top row

Spindle whorl; chipped sherd disc; smoothed edges

2.8

0.5

-

11: top row

Spindle whorl; chipped sherd disc; smoothed edges

2.6

0.5

-

11: top row

Spindle whorl; chipped sherd disc; smoothed edges

2.3

0.5

-

11: top row

Spindle whorl; chipped sherd disc; incompletely perforated; chipped edges

2.3

0.5

-

11: top row

Spindle whorl; chipped sherd disc; smoothed edges

2.5

0.5

-

11: top row

Spindle whorl; chipped sherd disc; smoothed edges

2.5–2.8

0.5

-

11: second row

Spindle whorl; chipped sherd disc; smoothed edges

1.6–1.8

0.3

-

11: second row

Spindle whorl; chipped sherd disc; smoothed edges; broken

2.3

0.5

-

11: second row

Spindle whorl; chipped sherd disc; smoothed edges

2.1

0.4

-

11: second row

Spindle whorl; chipped sherd disc; smoothed edges; broken

3.2

0.3

-

11: second row

Spindle whorl; chipped sherd disc; smoothed edges; broken

5

1

-

11: second row

Spindle whorl; fired clay; flat top and bottom

2.6

0.5

-

11: third row

Spindle whorl; fired clay; flat top and bottom

2.5

0.4

-

11: third row

Spindle whorl; fired clay; flat top and bottom

2.5

0.4

-

11: third row

Spindle whorl; fired clay; flat top and bottom; chipped on one edge

2.3

0.4

-

11: third row

Spindle whorl; fired clay; flat top and bottom

2.3

0.3

-

11: third row

Spindle whorl; fired clay; flat top and bottom

3.2

0.5

Trench 1

11: fourth row

Spindle whorl; fired clay; flat top and bottom

3.2

0.5

-

11: fourth row

Spindle whorl; fired clay; flat top and bottom

2.6

0.5

-

11: fourth row

Spindle whorl; fired clay; flat top and bottom

3.2

0.5

-

11: fourth row

Spindle whorl; fired clay; flat top and bottom; cracked and partly broken

3.7

0.8

-

11: fourth row

Spindle whorl; stone disc; flat top and bottom

2.2

0.3

Trench 1

11: bottom row

Spindle whorl; stone disc; flat top and bottom; broken

3

0.7

-

Spindle whorl; stone disc; flat top and bottom

3

0.7

-

Spindle whorl; stone disc; flat top and bottom

4

0.9

-

11: bottom row

1.9

0.4

-

-

Spindle whorl; glass disc; flat top and bottom; covered with opaque weathering layer; complete

116

11: bottom row

11: bottom row

A.E. Atagarryev: Excavations at Kirpichli depe in Dehistan than their decorated counterparts but have not been so sought after by collectors.

as a mirror or flower. The styles of these are sufficiently varied as to indicate strong local craft traditions and multiple centres of production aimed at mass domestic consumption judging by the very different types made and used locally in the Merv oasis,12 Sogd13 and Bactria.14 The second fragment had a compact yellowish fabric and smooth surface, and possibly represented the lower portion of a figure wearing a swirling dress or robe with two vertical folds down the centre (Figure 10.2).

A small quantity of metalwork was found, including a small concentration in room 4 in Trench 2. The metal finds included large heavy flat and trilobate tanged iron arrowheads, cylindrical bars and other fragments (Figure 17), simple copper alloy bangles and two copper alloy finger rings, one with a flat bezel (Figure 14). Other finds included unmodified marine shells which were presumably used as miniature containers (Figure 14: top left), a glass ‘eye’ bead (Figure 9.5) and a fragment of an unidentified incised and bevelled stone object with a flaked reverse (Figure 9.7) complete the inventory of recorded small finds. This was carved from a smooth greyish purple stone: the precise material is uncertain but it resembles an inscribed piece of slate illustrated from the Gorgan plain site of Shah Tepe.21

Many small plain spindle-whorls were found at the site and these were made of re-used potsherds, fired clay, stone and one was made of glass (Table 5; Figures 9.6, 11). All were plain and the commonest were those made from re-used potsherds. They suggest that spinning was a common household activity and the existence of local textile manufacture is proven by the identification of numerous carbonised cotton bolls from the excavations at this site. The discovery that cotton was cultivated here implies that the settlement lay within a landscape of carefully managed irrigation as cotton is a waterhungry summer crop. Moreover, the discovery supports the independent archaeobotanical and literary evidence from Sasanian and early medieval contexts at ancient Merv,15 as well as other carbonised finds from Kafyrkala,16 and Kara tepe in Uzbekistan,17 that these three regions of central Asia sustained important cotton textile industries during Late Antiquity and the early medieval period.18

Glassware A number of plain glassware diagnostics (i.e., rims and bases) were recovered, although mainly from the surface of the site rather than from excavation (Figure 15, Table 6). All of the rim sherds are illustrated here. A small number of fragments belonged to bowls, one of which had a lightly spiraled design made by dipping the gather into a mould with vertical indentations and twisting the gob while re-blowing (Figure 15.1); two others were plain (Figure 15.2–3). The remainder mostly belong to small bottles with flared out or thickened folded-in rims (Figures 15.4–13); the form of one of these was partially reconstructed from sherds found in Trench 2 (Room 4). All were free-blown. These vessels are partly paralleled at the Gorgan plain sites of Shah tepe22 and Tureng tepe.23 Similar bottles are also illustrated from Munchak tepe II and Penjikent in Tajikistan, both dated between the 7th and 8th centuries,24 and Qasr-i Abu Nasr in Fars province.25 Most of the bases consisted of low push-up bottoms in a range of semi-transparent light bluish fabrics. There was little evidence for decoration but two body sherds were decorated with applied trails (Figures 15.14–15): somewhat similar crimped trailing to one of these is found elsewhere on small bottles, for instance at Qasr-i Abu Nasr where a late Sasanian or early Islamic date was suggested for them.26

A large concentration of astragali was found in one room in Trench 1 and others were found in a room excavated in Trench 4. These may have been used in divination but are more likely to represent the remains of throwing games and these continue to be popular among children in Turkmenistan. Modified sheep astragali were common finds in Sasanian and later contexts at ancient Merv and in one case as many as six were found in a single context in Erk-Kala.19 A slightly chipped plain dark blue domed glass gamingpiece for backgammon was also recovered from one room in Trench 1 (Figures 9.7, 12). This type of gamingpiece was sometimes decorated in a mosaic (so-called millefiori) glass technique and a number of these are represented in different collections although most lack archaeological provenance and are therefore poorly dated;20 the plain versions were probably more common

The fabrics were mostly semi-transparent and consisted of light blue (Figures 15.4, 15.7, 15.13), blue (Figure 15.12), light blue-green (Figure 15.15), dark /

Pugachenkova 1959; 1962; Simpson et al. forthcoming a. Meshkeris 1977. 14  Abdullaev 1996; Zavyalov 1981. 15  Nesbitt 1994; Boardman 1999. 16  Yu. Buryacov, pers. comm. 1991. 17  Brite and Marston 2013. 18  Bulliet 2009; see also Section 3, introduction. 19  Simpson et al. forthcoming a. 20  e.g., Fukai 1977: 60, pl. 51; Hasson 1979: 20–21, fig. 37; Jenkins 1986: 52, 56, no. 62; Christie’s 1991: no. 127; Ancient Orient Museum 1998: 98, cat. II-84; also Nerazik and Rapoport 1981: 111, no. 28. 12  13 

Arne 1945: 366, pl. LXXXVIII, fig. 724.M. Arne 1945: 365, pl. LXXXII, fig. 684. 23  Period VIIC: Boucharlat and Lecomte 1987: 181, pl. 103.14–15. 24  Mandelshtam and Pevzner 1958: 313, fig. 21; Belenitsky 1958: 134, fig. 31. 25  Whitcomb 1985: 154, 156–57, fig. 58. 26  Whitcomb 1985: 154, 156–57, 206, fig. 58.j, l, s, pl. 45. 21  22 

117

Sasanian Archaeology

Figure 9. Small finds from Kirpichli depe

118

A.E. Atagarryev: Excavations at Kirpichli depe in Dehistan

Figure 10. Press-moulded fired clay figurine

Figure 11. Small finds from Kirpichli depe: spindle whorls made of reused potsherds (top two rows), fired clay (third and fourth rows) and stone (bottom row); 5 cm scale

Figure 12. Small finds from Kirpichli depe: dark blue glass backgammon piece

119

Sasanian Archaeology

Figure 13. Small finds from Kirpichli depe: ironwork; 5 cm scale

deep blue (Figures 15.1-2, 6), dark green (Figure 15.8) and olive green glass (Figure 15.3); others were covered with a heavily pitted opaque creamy (Figure 15.10) or grey-brown weathering (Figures 15.6, 15.14) which prevented the underlying fabric being visible. None were scientifically analysed but the type of weathering suggests at least some were probably plant ash glasses. Deep blue glass has also been found in 7th century late/post-Sasanian contexts at Merv, Ctesiphon and Kush; olive or dark green was also popular for a class of bottles found in 8th century contexts from Jebel Sais to Kafyrkala. Finally, grey-brown weathering is typical of much glass found in 7th or 8th century contexts in Iran and the Persian Gulf, including Kharg island, Jazirat

al-Hulayla, Fulayj and Sir Bani Yas, and compositional analyses show this to be plant ash glass.27 The place (or places) of manufacture of the Kirpichli depe glassware is uncertain. There are some parallels with pieces published from north-west and southern Iran but whether there was a glass industry in either region by the 7th or 8th centuries is uncertain. Nevertheless, this publication adds to the small corpus of glassware reported from late Sasanian and postSasanian archaeological contexts in this region. A small blown glass flask with a short cylindrical neck 27 

120

Phelps, Simpson and Freestone 2018.

A.E. Atagarryev: Excavations at Kirpichli depe in Dehistan

Figure 14. Small finds from Kirpichli depe: marine shells (top left), decorated glassware, glass spindle whorl and backgammon piece (right), copper alloy bangles and finger rings (bottom); 5 cm scale

and sagging shoulders, and decorated around the body with pulled warts was excavated at Orta Depeslik and dated between the 6th and 8th centuries:28 it has close but undated parallels from the early Sasanian cemetery at Abu Skhair in south central Iraq29 and excavations at Shemakha in Azerbaijan.30 Moreover, an unspecified number of fragments belonging to ‘bottles and cups, usually in green or grey colours’ were reported from M.Y. Kiani’s excavations of Fort 12 on the Gorgan Wall. These were said to have been found in Parthian levels but this dating must be seriously questioned as the single illustrated piece belongs to a tall cylindrical tube

decorated with cut facets and the contexted parallels from level VI at Tell Baruda in Iraq and the ‘burnt room’ on the citadel of Qasr-i Abu Nasr indicate that this very distinctive form dates to the early 7th century.31 A single piece of what was described as ‘yellow cut glass’ was found in a layer of fallen roof tiles in a trench excavated in the Bānsarān Fort on the Tammisheh Wall and which was attributed a possible pre-Islamic date by the excavators although this remains to be verified in the light of the later dating of the Gorgan wall.32 31  Kiani 1982b: 36–37; see Negro Ponzi 1987: 272, fig. C, 348, type Q; Whitcomb 1985: 155, 158–59, fig. 59e, pl. 43: top left; see also Ghasemi, Noruzi and Rezaei, this volume. A late Sasanian date for the Gorgan Wall itself is now beyond doubt: see Omrani Rekavandi et al., this volume. 32  Bivar and Fehérvári 1966: 43.

Masson 1961: 39–41, fig. 4. Negro Ponzi 1972: 231–32, fig. 21, no. 40. 30  Simpson 2015a: 92, n. 78, fig. 16.3. 28  29 

121

Sasanian Archaeology Table 6. Catalogue of the glassware found at Kirpichli depe Description

RD

BD

profile of bowl with lightly spiraled mould-blown reblown ribbing, deep blue with 8 fine pitting; two joining sherds

Findspot

Figs

-

15.1

rim of bowl, deep blue

c. 10

-

-

15.2

rim of bowl, olive green

c. 15– 18

-

-

15.3

rim of bottle, light blue

3

-

-

15.4

rim of bottle, semi-transparent with flaking opaque grey-brown weathering layer 5

-

-

15.5

rim of bottle, dark blue

c. 5

-

-

15.6

rim of bottle, light blue

3

-

-

15.7

rim of bottle, dark green

3.2

-

trench 1, room 4

15.8

-

-

15.9

5

rim of thick-walled bottle rim of small jar, heavily pitted with opaque creamy weathering layer

4

-

-

15.10

rim of small jar; four joining sherds

6.8

-

15.11 (part)

lower portion of small jar, cut down pontil measuring 2.7 x 2.9 cm; two joining sherds

trench 2, room 4

-

5

trench 2, room 4

15.11 (part)

rim of small jar, blue; three joining sherds

7

-

-

15.12

rim of small jar, light blue

c. 11–13 -

-

15.13

body with pinched trail, semi-transparent with opaque grey-brown weathering layer

-

-

-

15.14

body with applied trail, semi-transparent light blue-green with fine pitting and partly flaked off opaque white weathering layer

-

-

-

15.15

base, transparent green with partly cut down pontil mark; push-up H 0.8 cm

-

-

surface

-

base, transparent green; push-up H ca 0.6 cm

-

-

surface

-

base, semi-transparent green with partly cut down pontil mark; push-up H 0.4 cm -

-

surface

-

base, semi-transparent light olive green; push-up H 0.4 cm

-

-

surface

-

base, opaque grey-brown weathering layer; push-up H 0.15 cm

-

-

surface

-

base, semi-transparent light green; push-up H 0.8 cm

-

-

surface

-

Finally, six sherds belonging to small bottles or other containers and three different bowls decorated with cut facets were found in 6th century Sasanian contexts in the Phase VI fort on the summit of Tureng tepe.33 The same quantity was found in the subsequent Phase 33 

VII A–B, dating to the 7th–8th centuries, whereas the quantity increases and the types change in Phase VII C, dating to the 9th century, when they represent 5.6% as opposed to 1–3% of the combined number of pottery/ glass diagnostics.34

Boucharlat and Lecomte 1987: 172–73, pls 99: 6–9, 156a–c.

34 

122

Boucharlat and Lecomte 1987: 176, 178–82, pls 100, 102–103.

A.E. Atagarryev: Excavations at Kirpichli depe in Dehistan

Figure 15. Glassware from Kirpichli depe

• Hemispherical lids with arched handles (Figure 17) • Hemispherical lids with pull-knobs (Figures 18.2–3) • Flat-based open bowls (Figure 18.1) • Small jars of various shapes (Figures 19–20) • Large jars (Figures 18.4–7) • Tall-necked burnished jars (Figure 18.8)

Pottery An interesting range of handmade burnished and wheel-thrown plainware vessels were found, some represented by whole or restorable forms (Figures 16–21). No attempt was made to quantify the pottery but the following notes illustrate some of the key types represented at Kirpichli depe: • Flat-based lids with perforated bases (Figure 16)

The low hemispherical lids with simple pull-knobs are paralleled by examples from Shah tepe level I 123

Sasanian Archaeology

Figure 16. Pottery from Kirpichli depe: lid with flat string-cut base indicating it had been thrown from the hump; partially blackened, H 3.5, RD 11.5, BD 3 cm

Figure 17. Pottery from Kirpichli depe: complete hemispherical lid with arched handle; wheel-thrown; reddish brown, partially blackened interior; H 8, RD 16.8 cm; Trench 2, room 4

Figure 18. Pottery from Kirpichli depe: 1, flat-based bowl, wheel-thrown pale yellowish, RD 15 cm; 2, lid with pull-knob handle, wheel-thrown buff with slightly pinkish surfaces; 3, lid with pull-knob handle, handmade friable reddish-brown, Trench 2, room 6; 4, jar with finely incised decoration on the neck, possibly handmade dark reddish brown with chocolate brown slip, vertical burnish on the exterior, handle broken off at the rim, RD c. 9 cm; 5, large jar with incised decoration on the shoulder and a pair of opposing handles, wheel-thrown buff ware, RD 11 cm, Trench 1; 6, casserole, wheel-thrown light brownish with paler surfaces and wavy incised decoration, RD uncertain; 7, large jar, handmade coarse dark grey core with cream/pink exterior and oxidised red interior, RD c. 22 cm; 8, jar with wavy incised line round the shoulder; possibly handmade; dark grey fabric with chocolate brown burnished exterior and dark grey reduced lower interior, RD 9 cm

124

A.E. Atagarryev: Excavations at Kirpichli depe in Dehistan

Figure 19. Pottery from Kirpichli depe: small jars, restored from sherds; left: possibly wheel-thrown; reddish brown with lattice burnished body and two blackened patches on the lower body below each handle; H 12.2, RD 9, BD 7.5 cm; right: possibly wheel-thrown; light reddish brown with partially blackened surfaces; H 15, D 7.5, BD 7.8 cm; both from Trench 2, room 5

and Tureng tepe (especially Period VII A/B).35 More elaborate pull-knobs were found in Period VII C at the latter site. However, hemispherical lids with arched handles differ slightly from examples published from Shah tepe and Tureng tepe.36 Flat-based flaring lids with a string-cut perforated base were also recovered, and a series of flat-based open bowls complete the range of open forms. Larger vessel types included casseroles and tall-necked burnished jars: the latter are known from Shah tepe and Tureng tepe (especially Periods VI A–VII A/B).37 Other closed forms, sometimes with a pair of handles, are illustrated here and are partially paralleled at Tureng tepe Period VII A/B.38 Decoration was limited: simple incision was occasionally found on the neck/shoulder junctions of jars but burnishing was common. This usually consisted of vertical strokes on the necks of the vessels with horizontal or lattice strokes on the bodies. This style of surface treatment was both functional and decorative as it reduced porosity, facilitated cleaning and emulated the dull reflectiveness of metalwares.39 It was popular across northern Iran during the late 2nd and early 1st millennia BC (early Iron Age) and continued throughout the subsequent Achaemenid and Parthian periods with regional variations of form and occasionally strong

Figure 20. Pottery from Kirpichli depe: semi-complete medium-sized jar with a pair of incised lines around the shoulder, one surviving handle of originally an opposing pair and a slightly sagged base; possibly wheel-thrown then paddle and anvil applied to compact the lower body; slightly friable dark reddish brown with smoothed exterior; H 23.5, RD 10.5, BD 9.5–10 cm; surface find

Arne 1945: 366, pl. LXXXVI, figs 711–12; Lecomte 1987: pl. 71.6. Arne 1945: 366, pl. LXXXVI, fig. 709; Lecomte 1987: pls 71.9-10, 141.4. 37  Arne 1945: 366, pl. LXXXVII, fig. 714.a–b ; Lecomte 1987: pls 58.2, 69, 70.1–2, 140. 38  Lecomte 1987: pls 65.2, 66. 39  Simpson 1997b. 35  36 

125

Sasanian Archaeology hints of the influence of local schools of toreutic on the contemporary pottery skeuomorphs.40

underlined by the fact that the same administrator is named on bullae found at the two sites.43 The close cultural connections between these two parts of Hyrcania have been emphasised by other scholars such as O. Lecomte,44 and provide additional context for the building of the Gorgan wall which controlled access by effectively bisecting the plain into two with a deep noman’s-land acting as a buffer zone or vorlimes between it and the settlements in Dehistan.45 The excavations at Kirpichli depe therefore provide important new data for understanding the relations between the two areas and the archaeological transition from the Sasanian to early Islamic periods in this cultural border region of south-west Turkmenistan and north-east Iran.

Conclusion Kirpichli depe is an important and extensive postSasanian site dated by the numismatic evidence to the early/mid-8th century. This attribution is independently supported by the associated pottery and glassware. Related material has been excavated at two sites in the Gorgan plain of north-east Iran: Shah tepe and Tureng tepe.41 The material from Kirpichli depe particularly closely parallels that from Period VII A/B at Tureng tepe for which a 7th–8th century date was suggested by the excavators,42 and this is further

See Haerinck 1983: 191–96. Arne 1945; Boucharlat and Lecomte 1987. 42  Boucharlat and Lecomte et al. 1987.

Gyselen 1987: 190, pl. 163c. Lecomte 2007: 304. 45  See Omrani Rekavandi et al., this volume; Sauer et al. 2013.

40 

43 

41 

44 

126

The First Season of Excavations at the Sasanian Site of Tole Qaleh Seyfabad (TQS), South-west Iran: Preliminary Report Parsa Ghasemi,1 Reza Noruzi2 and Azizallah Rezaei3 2

1 PhD Candidate in Archaeology, University of Paris 1-Panthéon Sorbonne, UMR 7041 ArScAn-APOHR; MA in Archaeology from Tarbiat Modares University, Research fellow of the Iranian Centre for Archaeological Research (ICAR), 3

ICHHTO of Fars.

MA in archaeology from Tarbiat Modares University, Researcher of Fars ICHHTO

Abstract: Tole Qaleh Seyfabad is one of the most important Sasanian sites in south-west Iran, located south of the major Sasanian city of so-called Bishapur. The authors conducted an initial season of excavation at this site during the autumn of 2014. The main aims of this fieldwork were to prevent further damage to the site, get a better understanding of the site formation processes, and obtain a more detailed picture of a major site in the Sasanian period in the heartland of its empire. Four trenches were opened in the south-west portion of the site and uncovered part of a large regular building which had several functions, including official, storage and workshops. In addition, a large number of Sasanian clay bullae related to the advanced administrative system were found, plus various kinds of pottery, fired clay, stone, metal and glass objects which attest commercial and economic activities; proof of nearby arable agriculture was also found. This preliminary report discusses the importance of the site during the Sasanian period and presents some of the most significant finds. The results indicate that this site was a major active administrative, economic and commercial centre of the Sasanian province of Bishapur, in the heartland of this empire and the abandonment of it must be dated to the late Sasanian period prior to AD 651. Key words: Tole Qaleh Seyfabad (TQS), Sasanian period, Bishapur Province, archaeological excavation, Kazerun, Iran.

Introduction1 The Sasanian empire was one of the powerful empires of Late Antiquity in the east. This empire was founded by Ardashir I, who was born in city of Istakhr in Fars state, southern Iran, and was crowned in AD 226. The territory of the Sasanian empire during four centuries (c. AD 224–651) enclosed much of the Middle East and central Asia. After the rise of the Sasanian empire, most of south-west and central Asia were politically and economically dominated by them. In terms of complicated cultural criteria, the Sasanian political, economical and social entities were more sophisticated and powerful than the earlier governments. Existence of a powerful administrative and managerial hierarchy that administered the central state’s policies gave rise to development and florescence of a great empire. Based on the existing data, the Sasanian made fundamental changes in the administrative and political 1  We should thank Mosayeb Amiri, Alireza Askari Chaverdi, Hamide Chubak, Marjan Mashkour, Kouroosh Roustaei, Fereidun Biglari, Rika Gyselen, Remy Boucharlat and Ebrahim Gezelbash and all the members of the excavation team: Amir Bahrami, Abbas Razmpush, Hosein and Nima Ghasemi, Hasel Gharagezlu, Moslen Rezaei, Zohreh Zehbari, Sahar Jalalian, Leila Ghasemi, Fatemeh Dehghan and colleagues in the office of Cultural Heritage in Kazerun County and Bishâpûr Research Foundation, especially Behnam Askari, Noruzali Kashkoli and Manochehr Moeini. We are also grateful to officials and residents of village of Seyfabad who all kindly helped throughout the project. Finally we wish to thank Dr St John Simpson and Donald Whitcomb who gave generously of their time to read the text of this paper and made helpful suggestions.

Sasanian Archaeology (Archaeopress 2022): 127–140

entities of their territories. Through development of trade and industries, and cloth, glass and silk production they established a number of significant commercial centers that were directly managed by themselves. Besides, they developed agriculture and taxed farmers that all resulted to a complex administrative system.2 Such an efficient administrative and economical system paved the way for establishment of largescale urban settlements. The southern parts of Fars neighboured the Persian Gulf and because of their economic, commercial and political significance, the Sasanians established a large number of settlements there. The region between Firuzabad, Kazerun and Borazjan in the south-west portion of the Sasanian heartland is densely dotted with settlement sites. Despite this high archaeological potential, little research has been done here. Prior to the Iranian Revolution of 1979, investigations included work at the major sites of Bishapur and Firuzabad;3 these investigations were continued after the Revolution by a number of Iranian archaeologists.4 Tole Qaleh Seyfabad (henceforth TQS) is one of the key Sasanian sites in this region and is located at the very heart of the Sasanian empire in south-west Pigulevskaja 1963. Salles and Ghirshman 1936; Ghirshman et al. 1956, 1971; Vanden Berghe 1961; Sarfaraz 1969; Huff 1972; 2011b; 2014. 4  Mehryar 2000; Amiri 2009; Noruzi 2005; 2010; Ghasemi 2008; 2009; 2012a; 2012b. 2  3 

Sasanian Archaeology

Figure 1. Satellite photograph of south-west Iran giving the location of Tole Qaleh Seyfabad (TQS) and other important sites of the Sasanian period

Ramijan, Shahjan, Mooz, Dadin, Upper Khomarjan, Lower Khomarjan, Tirmardan and so on’.6

Iran, in the centre of Kazerun County. It is situated in the rural district of Balian, 330 m north of the new Seyfabad village on the natural mounds known as Tol-e Pahn (Figures 1–2). The site is 40 m higher than its surrounding plain and located 868 m above sealevel, and covers an area of about 150 hectares. TQS is situated 29 km south-east of Bishapur city, 7 km southeast of Kazerun city, 109 km north-west of Ardashir Khurreh/Firuzabad, 35 km north-west of Jereh, 88 km west of Qasr-i Abu Nasr and 214 km north of Siraf. The site was partly damaged by a gas pipeline which was constructed through it in the 1980s. In 2005, as part of surveys conducted in the Kazerun plain, this site was registered by Parsa Ghasemi with the site code KZ 92 (Figure 3a).5

Istakhri describes the territory of Shapur Province as: ‘Sabur, Kazerun, Janjan, Basal Chafteh, Dezboz, Khest and Kamarj, Hendijan, Dazin, Sur, Mahruyan and Arjan, Bazarang, Belad-e Shapur, Reyshahr, Jenabeh, Seeniz’.7 Ibn Balkhi names Bishapur, Jereh, Ghandjan, Khesht and Komarej, Anburan, BashtGhuta, Masaram, Bazrang, Kholar, Kazerun, Nobanjan, Shoab-e Bavan, and cities between Fars and Khuzestan as territory of this province.8 Based on the Sasanian historical and archaeological evidence such as seals and administrative bullae, the name of this province is Bishapur. In all administrative bullae recovered from TQS, Bishapur was written as a province but we do not actually know exactly where the capital of this province was.9

According to the Islamic sources, during the Sasanian period Kazerun was part of Bishapur province which was then the fourth and smallest province in ancient Fars. The territory of Bishapur province is mentioned in Ibn Khordadbeh’s Masalek al-Mamalek which was written in the 9th century:

The city of Shapur I in Tange Chowgan (now called Bishapur) was founded in 254 by the second king of the Sasanian dynasty, Shapur I. The Sasanian basreliefs on cliffs in Chowgān valley and the Middle Persian inscription on the pillars in the centre of this city illustrated that it was active during the reigns of Shapur I (AD 242–272) to Shapur II (AD 309–379). Until

‘Bishapur Shapur and its capital Nobandejan with its villages including Khesht and Kamarj, Kazerun, Jereh, Dasht-e Barin, Hendijan, Darkhuvaydud and Tanbuk, Khubazan, Meydan Mahan, Gonbad,

Ibn Khordadbeh 1992: 31. Istakhri 1992: 88, 97. 8  Ibn Balkhi 1994: 341. 9  Ghasemi et al. forthcoming. 6  7 

5 

Ghasemi 2008.

128

Parsa Ghasemi et al.: The First Season of Excavations at the Sasanian Site

Figure 2. Topographic map of TQS and neighbouring villages (1:25,000 scale)

129

Sasanian Archaeology

Figure 3. Aerial photograph of the excavated area in the southern part of the TQS site, showing the location of the four trenches excavated in the 2014 season to the south and north of the gas pipeline route

now, this was the only site in Kazerun County to have been excavated.10

lake and its sources are rainwater and eleven springs which mainly flow from the foot of the mountains to the north of the lake.

Landscape of the study area

The local population is mostly engaged in animal husbandry and farming, and normally rely on wells for their agricultural water supply. The region today has over 2,000 hectares of farmland over which dry farming used to apply to a small portion. However, droughts in recent years have changed the proportion and, in areas around TQS, 70% of the land is exploited via dry farming rather than irrigation which is usually conducted with modern deep wells. Wheat, barley and summer crops are the main crops.14 The annual rainfall in Kazerun is 291 mm which is indicative of the region’s semi-dry climate. Most of this precipitation falls in the cold season of the year with 55% in winter, 36% in autumn, 9% in spring and as little as 0.01% in summer, thus the region has a rainy season during the cold period of the year and a warm and dry season in summer. Considered as one of the steppic regions of the Zagros, its vegetation is composed of small plants and bushes. Trees are scarce and limited to a small number

The site of TQS is located in the centre of the Seyfabad plain surrounded by mountains of the southern Zagros. These mountains, like those in other parts of Fars province, are aligned north-west to south-east. The main mountains are Parishan, Kohmare Sorkhi, Qebleh and Dadin.11 The geological formations of the Zagros which surrounds the plain belong to the late Cretaceous and Quaternary periods. Geologically the area is part of the folding zone of the Zagros mountain range.12 The region enjoys deposits of various minerals, including alum, gypsum, limestone and sulphur.13 Parishan Lake with an area of 4,400 hectares lies 3 km east of TQS and is the habitat for many different species of animals and migrant birds which travel to Iran from Siberia, Canada, Denmark and North Africa. Parishan is a fresh-water 10  Salles and Ghirshman 1936; Ghirshman et al. 1956; 1971; Sarfarāz 1970; Mehryar 2000; Noruzi 2005; Amiri 2009. 11  Mozaffarian 1994: 26–30. 12  Shahrabi 1994. 13  Mozaffarian 1994: 31.

14 

130

Abbasi 1995; Saghafi, Alinia and Alizadeh 2010: 2.

Parsa Ghasemi et al.: The First Season of Excavations at the Sasanian Site

Figure 4. Views of the site of Tole Qaleh Seyfabad: a, view of the north-east side of the excavated mound and the gas pipeline which runs through it; b, position of Trenches I and II prior to excavation

of types including wild figs, wild almond and cedar. However, the highlands near Parishan Lake are densely covered with oak.15

problems made it very difficult to select a proper spot for excavation but finally the north-west side of the hill was chosen as it was less damaged (Figure 4b).

Excavation

The general aim of the excavation at TQS was to conduct a salvage operation in order to prevent further looting and damage to the site, and to facilitate identification of layers of different periods in order to determine a reliable chronology for the site. Firstly, a topographic plan was made of the site and the neighbouring structures were surveyed in order to set the site into a clearer cultural and historical context. Preliminary surveys proved that the site covers about 160 hectares with significant traces on its surface including the excavated spot mentioned in this paper.

The site has suffered extensive damage over the past four decades through a combination of deliberate destruction and the construction of the Kazerun–Shiraz gas pipeline which caused disturbance to the upper levels and the complete flattening of the central part of the hill. In the upper layers the traces of bulldozer blades and illegal looting pits are still visible. These 15 

Abbasi 1995; Saghafi, Alinia and Alizadeh 2010: 11.

131

Sasanian Archaeology

Figure 5. Trench I: a row of three rooms with a corridor in front

Prior to excavation, the surface of the site was mapped and several temporary benchmarks (TBMs) were established. TBM 1 was located near trench 1 in the north-east part of the site and is 863 m above sea-level. Here we applied a context-based method of recording layers (e.g., context 1001) and loci for structural features (e.g., locus 1001). As trench I was the first to be excavated, the contexts began with 1000 (the surface layer) and ended at 1011 (virgin soil). In recording the finds from each context, we have used the abbreviation TQS, which stands for Tole Qaleh Seyfabad, the name of the context and a four digit number (e.g., TQS 1001– 0032). The various excavated layers generally belong to the Sasanian period but as the finds have not yet been studied in detail, we cannot be more precise at present. Two trenches were excavated north of the Kazerun– Shiraz gas pipeline route at a spot where a number of clay bullae had been reported previously, and these trenches were aligned in a north-east/south-west direction (TQS. I–II). Two additional trenches were also excavated on the southern and eastern sides of the gas pipeline route in order to investigate a partially exposed outer wall revealed by previous illegal excavations (TQS. III–IV).

illegal excavations have been made since the 1980s. The main reason for choosing this spot for excavation was the occasional discovery of 27 clay bullae in the area.16 Trench I was the most important trench to be excavated during this season. It initially measured 6 x 10 m but was extended to a 10 x 10 m square to reveal more of the architecture which consisted of brick, stone and plaster construction. Virgin soil was reached at a depth of 2 m and a total of 12 contexts were recorded (Figures 4b–5). A large number and variety of finds were made here. As many as 304 clay bullae were found in three contexts (1000, 1001, 1004). A whetstone, mortar and a good number of ground stone items, including a rotary quern and a stone sieve (Figures 6a, c–d), and a few metal objects were also found. The latter included an iron axe, a complete and a fragmentary copper alloy ladle (Figure 6e), a badly corroded and therefore unidentified copper coin and numerous corroded objects. In addition, there were several plaster lids, two circular sheets made of unfired clay, an unfired clay cake, two conical unfired clay objects and many decorated beads made of stone, plaster and glass, pieces of two free-blown glass tubes and a glass tube decorated with circular facets, several decorated shells, four pottery vessels, the head of a fired clay animal figurine, pottery pendants, a cubical die and a backgammon counter (Figure 6b).

Trench TQS. I This was situated in the north-west part of the site and north of an unmetalled road in an area where numerous

16 

132

Askari Chāverdi and Cereti 2013: 182; Barfi et al. 2013.

Parsa Ghasemi et al.: The First Season of Excavations at the Sasanian Site

Figure 6. Selected finds from Trench I: a, whetstone; b, cubical die and backgammon counter; c, limestone mortar; d, ground stone; e, copper alloy ladle

(Figure 9b), three whetstones, numerous stone and glass beads (Figure 10a), a glass necklace or bangle, three pottery spindle-whorls, 18 complete pottery vessels (Figure 10b), several glazed pendants and objects similar to pottery spindle-whorls, and numerous fragments of corroded metal and metal objects. The latter consisted of an iron pick and an iron axe, a large tripod-footed spouted copper alloy pot with a narrow opening and capacity of 50 litres (Figure 11a), two broken iron daggers, four rings, two sickle-like metal objects (probably parts of a wooden door), two Sasanian monograms as two back to back crescents, a two litre copper alloy handled beaker (Figure 11b), metal nails, a traditional iron plough, a copper alloy bell, a pair of

Trench TQS. II In order to investigate architectural spaces in trench I, a new trench measuring 10 x 11m was opened on the northern and north-east side of trench I (Figure 3a). 30 contexts were recorded (Figures 6–7). An additional 67 clay bullae were discovered (contexts 2002, 2005, 2006, 2007, 2010, 2014, 2018). Other finds included the bottom of a stone vessel, an object resembling an engraver’s chisel, four stone mortars, 14 grinding stones, a footed stone plate or lid of a vat, a stone base, four rotary querns (Figure 9a), a personal Sasanian stamp-seal bearing a monogram on the face 133

Sasanian Archaeology

Figure 7. Aerial photograph of Trench II showing the excavated architectural remains

Figure 8. Aerial photograph of architectural remains discovered in Trenches I and II in the north-west portion of the site

134

Parsa Ghasemi et al.: The First Season of Excavations at the Sasanian Site

Figure 9. Selected finds from Trench II; a, rotary quern; b, banded agate stamp-seal (H 1.4 x D 1.5 mm, wgt 6 g) with a monogram engraved in double-lines (inscription in Middle-Persian in semilapidary script from 3 to 7 o’clock: abestān ō yazdān, ‘Confidence in the gods’)

Figure 10. Selected finds from Trench II: a, decorated beads; b, large storage vat found in the eastern part of this trench

Trench TQS. IV On the eastern side of the site, previous illegal excavations had exposed the upper portions of a thick wall constructed of stone and plaster. This trench, measuring 9 x 5 m across, was excavated in order to understand this feature; eight contexts were recorded, and date-stones, potsherds, the handle of a copper alloy vessel and a copper coin were found. Two plastered stone walls were uncovered, one of them with a southwest/north-east direction and extending for 2.42 m, surviving 2 m high and 1.42 m wide; the second was aligned east-west, measured 2 m long, 2.3 m high and 1.15 m wide. Together, they formed part of the architectural spaces inside and outside the main wall. Excavation along the centre of the east-west wall led to the discovery of an entrance with an arched lintel constructed of stone and gypsum, a set of steps in front and a plastered floor (Figure 15).

tweezers, the copper alloy base of a pricket stand or lamp (Figure 11c), several dove-tail joints, a circular iron sheet, a metal pin, and a small cosmetic container. Other finds included two glass tubes – one decorated with oval facets in low relief (Figure 12) – circular plaster plates which were probably used as lids to some vessels along with other conical plaster objects, several decorative shells with holes in them and a handmade clay plate. Trench TQS. III In the south-east part of the hill, a 7 x 5 m trench was excavated at an almost untouched spot in order to determine the conditions of the inner court of the site (Figures 13–14). Due to the lack of considerable deposit, the dimensions of the trench were reduced to 1.5 x 1.5 m and virgin soil was reached. Three contexts were recorded; potsherds and a stone mortar were among the finds.

Discussion The archaeological finds from Bishapur to the northwest of TQS shows that it was a royal city with some 135

Sasanian Archaeology

Figure 12. Trench II: complete green glass tube with a surface weathering layer, made by free-blowing and decorated with relief facets

I, proving this to be the most important archive room: of these, 75 are administrative and the rest are official or personal. Indeed most are personal but these have not yet been studied. Preliminary studies on this important archive suggest that TQS was probably the most important administrative centre in Bishapur province during the Sasanian period. With the studies of the Middle Persian inscription on the administrative bullae, several districts of Bishapur province and a district of Ardashir Khurreh province were identified. Moreover, with these discoveries, the subdivisions of administrative system of the Bishapur province are revealed. This excavation shows that the name of TQS probably was Ābād-Šābuhr because as many as 58 of the administrative bullae contain the name of the ĀbādŠābuhr district on their seal impressions.18 A large number of metal, stone, plaster and glass objects were found in these excavations. Many may have been made locally but production workshops or industrial areas have not been identified at the site, at least as yet. However, it is clear that this site had many functions: in room 2, the only finds were bullae which indicates that this room was the main archive whereas evidence for juice production and probably grinding grain from room 3 suggest it was used for food processing.

Figure 11 Selected finds from Trench II: a, large tripod-footed spouted copper alloy pot with a pair of handles; b, copper alloy handled beaker with two-litre capacity; c, the base of a pricket stand or lamp similar to one found at Qasr-i Abu Nasr

About 50 kg of carbonised dates were discovered in various contexts in trenches I, II and IV, with a great number inside storage jars in rooms 4 and 5 (Figure 16). Evidence for producing juice and a number of bullae were also found in room 4. Stone mortars and grinding stones are the most important workroom objects were found in this site. They were produced locally but there is no evidence to indicate exactly what they were used for. A total of 512 pieces of bone weighing 4 kg were also found during the excavations: these items are being studied by Dr Margareta Tengberg and Dr Marjan Mashkour respectively. Finally, in addition to the metal objects mentioned above, 807 metal fragments with a total weight of 18.6 kg were found in this season. Other objects such as glass tubes undoubtedly arrived to TQS by trade activities; a few numbers of tubular glass vessels with facet-cut decoration and a single glass

monumental buildings but probably was not a centre of administration and commerce. In contrast, it appears that TQS was a commercial centre and was well placed at the junction of ancient roads leading to the Sasanian cities of Istakhr and Qasr-i Abu Nasr to the north-east, Ardeshir Khurreh, Jereh, Ghandjan/Dasht Barin and Siraf to the east and south-east, the Persian Gulf and cities of Rishahr, Deylam, Jenaveh (Gonaveh), Tavaj (Tuz/Borazjan) to the south-west, and the Sasanian cities of Jundi Shapur, Veh Ardashir and Ctesiphon to the west. The evidence of these roads in Kazerun County was the subject of earlier research by the author.17 A total of 371 clay bullae were found from this season of excavation at TQS. Over 300 were in room 2 in trench 17 

Ghasemi 2008; 2009.

18 

136

Ghasemi et al. forthcoming.

Parsa Ghasemi et al.: The First Season of Excavations at the Sasanian Site

Figure 13. Trench IV and the excavated test trench

Figure 14. Aerial photograph of Trench IV and excavated architectural remains

137

Sasanian Archaeology

Figure 15. Trench IV: entrance

tube decorated with relief facets probably indicate the significance of this place in trade issues. As many scholars believe that the main centre of the Sasanian glass workshops were in southern Mesopotamia,19 probably such glass objects reached other parts of the empire through trade.

adding to previous finds made at Bandian Dare Gaz22 and Qasr-i Abu Nasr.23 The architectural remains found in trenches I and II belonged to relatively small rooms with the principal walls made of plastered stone whereas the partition walls were constructed of fired brick and mudbrick. The entrances to the rooms were narrow and their floors were plastered. The foundations of the inner partition walls were made of small stones over which square bricks had been laid. There were also rectangular blocks made of small stones and plaster used in corners of walls. A large number of metal items were found, including nails, dovetail joints and other pieces which probably belonged to wooden doors. The nature of destruction of the walls and the arrangement of the debris on floors indicate that this complex was probably destroyed by natural causes such as an earthquake; following this destruction, and as recently as 40 years ago, local inhabitants used these materials to build their own houses. The excavated discoveries prove that this site was a well connected centre and some of the finds resemble those unearthed in earlier excavations at Qasr-i Abu Nasr, near Shiraz.24 Around the outer

Over 22,000 potsherds and 22 complete pottery vessels were also found; many belonged to large storage vessels. These finds along with the discovery of 371 clay bullae, indicates that the site functioned as a place for collecting, processing and storage of products. A piece of backgammon was also found; backgammon is a board game which is believed to have been invented in the Sasanian period. The Middle-Persian text Kārnāmag ī Ardaxšēr ī Pāpakān associates the invention of the game with Ardashir I (AD 224–241), the founder of the Sasanian dynasty, whereas in the Middle Persian narrative Wičārišn ī čatrang ud nihišn ī nēwardaxšēr [Explanation of Chess and the Invention of Backgammon] it is Bozorgmehr, the vizier of Khosrow I Anūshīravān (AD 531–579), who is credited with the invention of the game.20 This piece was made from the bone of a large animal.21 A stone die was also found, Simpson 2014a. Schädler et al. 2009. 21  Marjan Mashkour, pers. comm.

Rahbar 2007a: 133. Whitcomb 1985: fig. 74d. 24  Whitcomb 1985.

19 

22 

20 

23 

138

Parsa Ghasemi et al.: The First Season of Excavations at the Sasanian Site

Figure 16. Trench II: the remains of carbonised dates were found at the bottom of this jar in the south-west part of room 4

part of the site there are thick walls with semi-circular towers made of stone and plaster which proves the site had been carefully fortified; parts of this fortification are visible on aerial photographs.

found in the site and these are generally believed to have appeared in the last two centuries of Sasanian period and continued to be used into early decades of the Islamic era.

Conclusion

This important site provides significant new information on administration with evidence for trade and agriculture. The bullae from this first season of excavation also provide evidence for the wider administration and show how it was sub-divided into smaller districts under the supervision of Zoroastrian clerics belong to the Bishapur province (Figure 16). Moreover, these bullae also add the names of some previously unattested administrators and districts of Bishapur province. Intensive agriculture was of great importance in this Sasanian province and the excavated evidence from this site for the collection, storage and processing of agricultural products, especially dates, illustrate the commercial and economic significance of Bishapur province. Dates were grown on a large scale in this province as a result of irrigation. They must have been exported but there is also evidence from this site for the production of date juice or syrup. All kinds of dates were found here in archaeological contexts and it is suggested that they reached here from many different places across the province, hence the names of different districts on the administrative bullae. In addition, a large collection of Sasanian pottery was found, much of

The Sasanian empire was one of the powerful empires in the eastern part of the ancient world and differed from its contemporary western rivals in their political, religious, administrative and economic approaches. Excavations in archaeological sites belonging to this period, either within or outside Iran, illustrate the grand policies of this empire. The site of TQS is one of the few excavated sites to be explored so far near the birthplace of this empire in south-west Fars. The architectural remains indicate the use of small rooms for storage, archives and processing workshops; they were constructed from locally available materials of gypsum, stone, mudbrick and fired brick. There are many similarities between the material culture from this site and other Sasanian sites. Pending radiocarbon dates on the excavated plant remains and seeds, an exact date has not yet been established. However, we can be sure of the significance of this site during the Sasanian period as it had expanded to an area of 160 hectares. Moreover, a large number of bullae were 139

Sasanian Archaeology which consisted of storage jars. The presence of rotary querns and other ground stone proves additional food processing, most likely the grinding of barley or wheat. The Sasanian name of TQS was probably Ābād-Šābuhr, but this remains a hypothesis until further research on the bullae has been conducted (Figure 17). However, it is almost certain that the site was the main economic and administrative centre in Bishapûr province. The main reason for the expansion of this site was its location on the junction of trade and royal roads connecting the centre of Fars with the Persian Gulf and the northern and western parts of the Sasanian empire.

Figure 17. An administrative bulla from TQS with the seal impression of Ābād-Šābuhr administrative district

140

Sasanian and Early Islamic Settlement Patterns North of the Persian Gulf Ahmadali Asadi,1 Seyed Mehdi Mousavi Kouhpar,2 Javad Neyestani,2 Alireza Hojabri Nobari3 Ph.D student in archaeology, Tarbiat Modares University. Assistant Professor, Department of Archaeology, Tarbiat Modares University. 3 Associate Professor, Department of Archaeology, Tarbiat Modares University. 1

2

Abstract: Recent archaeological surveys of the Iranian coastal hinterlands of the Persian Gulf have identified a remarkable growth in the number and size of sites during the Sasanian period. The results of the surveys show that this expansion was mostly due to large-scale investment in cultivation and irrigation systems. Due to the unfavourable environmental circumstances of most of these areas, governmental support for the creation and maintenance of irrigation systems was necessary. A network of fortified sites, found almost anywhere where Sasanian sites are located, demonstrate the marked presence of central government in the hinterlands. After the Sasanian period, archaeological evidence shows that there was a sharp decline in the number of settlements. Many of the sites, particularly those dependent on irrigation systems, were abandoned soon after the Arab invasion. On the coastline too, most of the traditionally important Sasanian ports, such as Rev Ardashir on the Bushehr peninsula, declined and remote military ports, such as Siraf, began an artificial economic growth.

Introduction There has been little research into the effect that the collapse of the Sasanian empire had on settlement patterns on the Iranian plateau. The general hypothesis is that the economic system remained unchanged after the Islamic invasion and some researchers even contend that the economic system improved.1 Recent archaeological surveys in the hinterlands north of the Persian Gulf provide some important new information on this matter. The results from most of these surveys in the hinterland share a common feature in that they illustrate an increase in the number of sites during the Sasanian period and the reduction, and in some cases collapse, of sites following the arrival of Islam. One might conclude from these recent surveys that new settlement patterns were introduced during the Islamic period due to the disappearance of settlement patterns particular to the Sasanian dynasty. In this paper we will examine in detail Sasanian policy in relation to economic investment in the hinterlands north of the Persian Gulf. Archaeological evidence reveals traces and remains of huge Sasanian investment in agriculture and extension of irrigation in the gorges in the mountainous areas south of Fars province. However, the development of agriculture in the aforementioned area differs from what the Sasanians carried out in the Susiana plain and Mesopotamia where huge dams and massive irrigation channels, some of which still survive, allowed the Sasanians to develop agriculture on a large scale. 2 However, in the mountainous region 1  2 

Whitcomb 1979: 5; Sumner and Whitcomb 1999: 314. Wenke 1975/76; 1981; 1987; Adams 1965; 1981.

Sasanian Archaeology (Archaeopress 2022): 141–152

the situation was different: here there were neither large plains nor big rivers like the Euphrates, Karun or Tigris. It appears that Sasanian agricultural policy in mountainous regions was small investment on a large scale and to reach this aim, governmental control and presence was a necessity. After the collapse of the Sasanian dynasty this agricultural system was no longer sustainable. Using archaeological evidence from recent surveys in various areas north of the Persian Gulf we will now look at settlement patterns, the agents affecting and the changes after the Islamic period. Geographical setting of the hinterlands and its implications As we will see, owing to differences in the geographical and climatic settings between the Persian Gulf hinterlands and the northern and central region of Fars province, varied settlement patterns are witnessed through time. The climatic situation in the hinterlands is variable and the environment was a key factor in how and where human settlements were established. A survey in the hinterlands show that natural factors such as poor rainfall, salty plugs and rivers, mountains, variable climate and access to suitable water play an important role in how and where human settlements were established. The Zagros mountains, as we travel south, change direction from north-west/south-east to east/west as we reach Larestan and Bandar Abbas. It is significant that here, where the region is known as Tangestan, the Zagros highlands rise immediately from the Persian

Sasanian Archaeology

Figure 1. Distribution of salty plugs along the northern side of the Persian Gulf (after Fisher 1968: 31, fig. 10)

considerable disadvantage – Rūdshūr (or Rūd-i-shūr) means ‘Salt river’ … In some places an original salt plug may have been eroded away so that the rocks have fallen back into the solution hallow, producing a chaotic local landscape of jumbled rock strata and irregular topographic outlines’.3

Gulf and there is therefore no intervening coastal plain. A good description of this is provided by Fisher: ‘Ridges front much of the coast in parallel ranks, producing a closed coastline devoid of indentations that could serve as harbours; and the small size of rivers and their deeply indented valleys greatly reduce the possibilities of irrigated agriculture and the occurrence of natural pasture. Poor in resources, without major routes that could assist development, Tangistān has only scanty and small settlements that produce dates, a little millet and barley, and vegetables ... Eastwards again, fronting the re-entrant of the Gulf of Hormuz, is a physiographically distinctive region whose relief is much influenced by the presence of numerous salt domes or plugs that have erupted from considerable depth through overlying strata ... As a semi-plastic substance, the salt has been forced through overlying rock formations and some plugs now stand at heights of 2,000 to 4,000 ft above the surface [Figure 1], appearing as pinkish-white or yellowish masses ... The salt itself has little if any commercial value, and the extra salinity imparted by drainage from the areas where plugs occur is a

Moving to the east, the sudden change in direction of the present-day coastline, which now runs northeastward from Bandar-e Lingeh, means that, in sharp contrast to the area west and north of Lingeh (including Tangistan), ‘the component ridges of the Zagros are aligned at right angles to the coast, which is consequently much more accessible and “open” with better harbours. A number of coastal settlements have thus grown up, with distinctly larger numbers of people than in the area farther west’.4 The most striking factor separating this southern Zagros area from other regions is that climate rainfall is very much lower compared with the northern and middle regions of Zagros and consequently there are 3  4 

142

Fisher 1968: 30–31. Fisher 1968: 32.

Ahmadali Asadi et al.: Sasanian and Early Islamic Settlement Patterns Whitcomb in 1973.13 In 1973 a survey was undertaken by a joint archaeological mission by the Siraf excavation mission and the Iranian Archaeological Centre in the Siraf hinterlands.14 30 years later new surveys were initiated by the Iranian Centre for Archaeology with the aim of completing the archaeological map of Iran. The most recent surveys were carried out in 2004 in Lar and Lamerd districts by Askari,15 Bushehr hinterlands by Carter and others in 2006,16 Bastak district by Asadi in 2007,17 and Ghavbandi district by Askari in 2007.18 The new surveys seem to provide a more precise understanding of settlement patterns during the Sasanian period in the southern Fars province. In the rest of this paper we will study the results of these surveys in finer detail although the interpretations put forth here are preliminary and will need to be reviewed in the near future.

major changes in the nature of topography and human responses. In addition to a reduction in the overall amount, rainfall tends to be distinctly more unreliable and sporadic with a tendency for heavy but short falls which are irregular both within a single season and from year to year.5 In order to compensate for this, irrigation systems dominate the layout of cultivation, with heavy use of spring-fed ditches and qanats and, in areas of lower altitude where temperatures can be very high, covered tanks and pools are increasingly used.6 The unfavourable natural settings in coastal and hinterland sites reduced the possibility of establishing human settlements in these areas. Before the formation of empires, settlement patterns in the hinterlands were dependant mainly on access to water sources from springs and wet gorges. In historical periods too much of the archaeological evidence, such as ancient water storage structures and irrigation systems in the hinterland areas, shows that issues of water supply have been the main challenge for human life throughout history in the areas under discussion here. Due to the salinity of the rivers, particularly in the downstream areas today, there is no human settlement close to the rivers and, according to the distribution maps of ancient sites, a similar situation prevailed in antiquity.7 The only land which provides favourable conditions for the establishment of settlement are the foothills between the rivers and the mountains8 and, as we will see in this paper, most of the ancient settlements in the hinterlands are found in these areas.

Bushehr district The Bushehr plain is the most westerly point of this research. Compared to eastern areas such as Lamerd, Mohr and Bastak, the Bushehr plain is geographically and environmentally distinct and one of only a handful of fertile plains on the Persian Gulf coastline. Sweet water rivers flow through the plain today and, unlike most of the other hinterland areas where salty rivers prevent human settlements being established nearby, here the major settlement was concentrated along the rivers.19 Although today some of the rivers have become salty, they are still used to cultivate date palms and cereals.

Previous research

The Bushehr peninsula and its hinterlands were surveyed by a joint Iranian/British archaeological team in 2004, with the aim of identifying the nature of the Persian Gulf hinterland settlements and establishing a chronological and cultural framework for the region.20 The survey included the areas between Bushehr, Gonaveh, Ahrom and Daleki (Figure 2), and covered an area extending 90 km north-west to south-east and 60 km north-east to the south-west.21 During this survey 56 sites were identified from the Chalcolithic (5,000 BC) to the late Islamic periods (15th–18th centuries), of which the majority belonged to the Achaemenid to Sasanian periods.22 The results of the survey indicate that the Elamite period, with 10 identified sites, marked the first major expansion of settlement in the region,23 perhaps

For researchers of the Sasanian period the Persian Gulf has always been an important topic, especially in relation to Sasanian maritime trade. Scholars such as Williamson, Whitehouse and Daryaee have written on this subject.9 However, their analyses of the archaeological evidence do not consider the hinterlands in detail. This is due to the lack of adequate archaeological surveys of this region until recently. Pioneering antiquarian research arising from finds at various places across the Bushehr peninsula was carried out by some of the 19th century employees of the East India Company and Indo-Persian Telegraph Company but their results were not fully appreciated until recently.10 The first surveys of the hinterland were conducted by Stein11 and, following a long hiatus, research was resumed in the Bushehr plain first in 1970 by Prickett and Williamson,12 and then by

Whitcomb 1987. Whitcomb 2009: 95. 15  Askari Chāverdi and Azarnoush 2004. 16  Carter et al. 2006. 17  Asadi 2010. 18  Askari Chāverdi 2008. 19  Carter et al. 2006: 75. 20  Carter et al. 2006: 63. 21  Carter et al. 2006: 68. 22  Carter et al. 2006: 64. 23  Carter et al. 2006: 71. 13  14 

Fisher 1968: 28. Fisher 1968: 28. 7  Asadi 2010: 27. 8  Asadi 2010: 27. 9  Williamson 1972; Whitehouse 2009; Whitehouse and Williamson 1973; Daryaee 2003. 10  Simpson 2007a. 11  Stein 1937. 12  Prickett and Williamson 1970. 5  6 

143

Sasanian Archaeology

Figure 2. Map of the Bushehr plain with identified ancient sites (after Carter et al. 2006: 64)

recorded in Carter’s survey as site number BH29–44. Carter interpreted these as the remains of a Sasanian city. There are also more Sasanian sites encircling BH29–44.29 Without doubt, the Sasanian port of Rev Ardashir, located midway along the Bushehr peninsula coastline, was the most important factor in the growth of settlements in the Bushehr hinterlands. Following his surveys conducted with Martha Prickett between 1969 and 1971, Andrew Williamson concluded that the remains of the site of Rishahr, located 6 km south of modern Bushehr, to be the Sasanian port of Rev Ardashir.30 According to historical sources, this port was one of the foremost ports of the Sasanian period and was established by Ardashīr i Pāpakān as part of his policy for control of maritime trade in the Persian Gulf.31 The size of pre-Islamic settlement in the Bushehr peninsula was remarkable, with some researchers estimating the total to be over 450 hectares,32 and the majority of this belongs to the Sasanian era. Using Carter’s Sasanian distribution map, we can assume that the Sasanian city situated north of Borazjan (BH29–44) played a central role, situated as it is in the centre of so many other

on account of the middle Elamite centre at Liyan.24 After the Elamites the next large period of growth in settlement was during the Achaemenid/Parthian period, with 32 sites recorded.25 The importance of the Bushehr plain during the Achaemenid period is illustrated by the number of Achaemenid palaces found in the area. The Sasanian period is represented by 36 sites which mark the peak of settlement in the region.26 With the arrival of Islam we see a very marked decline in settlements in the region with only nine sites found in the plain.27 Sasanian settlements are mostly concentrated southeast of the Dalki river (Figure 3). Most of the sites with Achaemenid/Parthian pottery also bore Sasanian material and Carter’s team states that, while the classes associated with the Achaemenid/Parthian pottery may have also been used during the Sasanian period, many of the sites currently assigned to both the Achaemenid/ Parthian and Sasanian horizons may date solely to the latter period.28 The most important Sasanian sites are a complex of mounds located north of Borazjan and Carter et al. 2006: 71. Carter et al. 2006: 94. 26  Carter et al. 2006: 94. 27  Carter et al. 2006: 71. 28  Carter et al. 2006: 71, 96. 24 

Carter et al. 2006: 96. Whitehouse and Williamson 1973: 39–41. 31  Piacentini 1985: 60; Miri 2009: 28–34. 32  Whitehouse and Williamson 1973: 37.

25 

29  30 

144

Ahmadali Asadi et al.: Sasanian and Early Islamic Settlement Patterns

Figure 3. Distribution of sites of the Achaemenid/Parthian (A) and Sasanian (B) periods in Bushehr plain (after Carter et al. 2006: 65–66)

to find the name of the city that supported the port during the Achaemenid and Sasanian periods, using Islamic, Roman and Pahlavi written sources. Nearkhos, admiral to Alexander, mentions a city named Taoxn in his travelogue ‘from Sind to Tigris’. Taoxn was located close to the river Granis, which also had a Persian royal palace situated nearby.37 Later still, the 8th or early 9th century Pahlavi text Shahrestānīhā-ye Ērānshahr [Provincial Capitals of Iran], claims that ‘The city of Tūz [tūzag] was built by Humāy, the daughter of Čihr Āzād’ and finally, in the early Islamic sources a town called Tawwaj or Tawwaz appears.38 We can assume that these names – Achaemenid Taoxn, Sasanian Tuzag, Islamic Tawwaj/Tawwaz – were one and the same with a slight alteration in the location of the city.

Sasanian settlements, and was of significance for the Rev Ardashir port. Carter writes as follows: ‘Bushehr may well have acted primarily as a port and entrepôt, while the inland town lay within the heart of an agricultural landscape surrounded by other Achaemenid to Sasanian sites, some of which also attained considerable proportions. The town may have been highly significant in co-ordinating the provision of subsistence or surplus commodities to Bushehr and the redistribution of goods from the port to the major centres in inland Fars’.33 The destruction of Rev Ardashir during attacks by invading Arabs in the early 7th century must have been catastrophic for the city, and archaeological evidence suggests both the port and city were abandoned simultaneously.34

Using aerial photographs, Donald Whitcomb identified an old canal which he attributed to the Achaemenid period but a more recent ground survey in 2004 failed to locate this feature.39 As we shall see below, the Sasanians established a large network of irrigation canals and structures in the hinterland gorges. Due to the fertility of the soil in the Bushehr plain there must have been more structures related to farming. Moreover, when considering the agricultural systems put in place elsewhere across the Sasanian empire, there is a need for more intensive surveys to locate equivalent evidence for irrigation systems on the Bushehr plain. In their 2004 survey, Carter and his team also attempted to reconstruct some of the roads between the Bushehr plain and the interior of Fars province.40 Due to the importance of Rev Ardashir during the Sasanian period,

This symbiotic relationship between city and port predates the Sasanian period. As previously noted, using Carter’s site maps (Figure 3), it becomes clear that the distribution of Sasanian sites in the Bushehr plain is similar to that of the Achaemenid/Parthian periods.35 It appears that prior to the foundation of Rev Ardashir in the 3rd century there was a pre-existing port serving the Bushehr peninsula. Whitehouse and Williamson speak about a possible Achaemenid port on the southern-most edge of the peninsula.36 If we accept the existence of an Achaemenid port on the peninsula, one can postulate the importance of this for the expansion of Achaemenid/Parthian settlements. It is possible Carter et al. 2006: 98. Carter et al. 2006: 97. 35  Carter et al. 2006: 66. 36  Whitehouse and Williamson 1973: 35.

Arrian, Indica 39.3. Daryaee 2002: 20, 51; Kianrad 2005: 108; Miri 2009: 42. 39  Carter et al. 2006: 67. 40  Carter et al. 2006: 96.

33 

37 

34 

38 

145

Sasanian Archaeology

Figure 4. Satellite map of the Siraf hinterlands with the map of identified sites (after Whitcomb 2009: 78)

there must have been a busy communication route to the north which may have passed many Sasanian sites. It was perhaps the only important road to inland Fars throughout the entire Iranian coastline from Bushehr to the straits of Hormoz.

– two in Jam and one in Dejghah – were attributed to the Sasanian and early Islamic periods.45 Of all of these Sasanian sites, five consisted of mounds with a height of 50 m or more,46 which leads one to believe that more intensive surveys in this region might discover more small Sasanian sites. Unlike the situation in the Islamic period, there are no concentrations of Sasanian sites along the trade roads from the port of Siraf to the north. Whitehouse interpreted this in the light of Sasanian Siraf having a primarily military, rather than commercial, function whereas trading ports – as Siraf was to become in the Islamic period – required more supporting settlements and structures.47 It appears that the economy of the valleys behind Siraf, like Lamer and Bastak, as we will see below, during the Sasanian period were possibly related more to the control of agriculture in the interior settlements.

Siraf port The hinterlands of the port of Siraf were studied in 1973 by an Iranian/British team led by Whitcomb (Figure 4). They surveyed the eastern part of the Jam valley, the western end of Galehdar valley, and the road between the Jam and Dejgah valleys.41 This research centred on the nature and extent of settlements during the early Islamic period (9th–11th centuries) when Siraf was at its height and was famous for its trade.42 However, some Sasanian sites were noted during the course of the work. The results of this survey indicate that the road between Siraf and the north – especially from Jam to the Dezhgah – was full of caravanserais and water storage facilities.43 In contrast they noted that there were no structures related to this road in Galehdar where instead they noted fortified sites which differed from the Sirafi structures.44 In terms of dating, they defined one site in Galehdar, one in Jam and three sites in the valley as Sasanian, and an additional three sites

Lamerd, Mohr and Darolmizan valley The Lamerd and Mohr districts are the natural continuation of the Galehdar valley to the east, and to the north, Alamard dasht district runs parallel to Lamerd and Mohr (Figure 5). Three perennial rivers, the Mehran, Alamard dasht and Darolmizan, flow through these plains. The first research in these plains was undertaken by Stein (1937) and then Heinz

Whitcomb 2009: 78. Whitcomb 2009: 78. 43  Whitcomb 2009: 93–94. 44  Whitcomb 2009: 93–94. 41 

Whitcomb 2009: 94. Whitcomb 2009: 83–92. 47  Whitehouse 2009: 12.

42 

45  46 

146

Ahmadali Asadi et al.: Sasanian and Early Islamic Settlement Patterns

Figure 5. Satellite map of the Lar and Lamerd, Bastak, Siraf and Gavbandi districts

the additional 15 non-occupational structures.55 This increase in sites during the Sasanian period is striking and an examination of their distribution shows that, whereas prior to the Sasanian period most settlements were in the foothills where habitation conditions were favourable, during the Sasanian period settlements are found evenly distributed across the valley and in areas which are less favourable for habitation (Figure 6).56 The archaeological evidence therefore illustrates the advanced nature of Sasanian irrigation techniques and land management which rendered it possible for them to increase settlements in such large numbers. The remains of irrigation systems are reported everywhere close to Sasanian sites and some of these would not have been habitable had water not been transferred from remote gorges. One might postulate that there was a degree of forward planning with initial construction of irrigation systems which made possible the development of settlement across the valley. It seems that better use of land and water was a policy of central government. In his report, Askari refers to about 20 Sasanian fortified sites which are located near the settlements in the foothills of the plains.57 Doubtless these are a sign of governmental presence in the region and their role might have included tax collection, provision of security and management advice, particularly over the development of irrigation systems. Askari Chāverdi argues that the increase in settlements in the valleys is also associated with trade from the port of Siraf.58 However, during this period Siraf

Gaube (1980) who discovered more than 30 sites in these plains.48 In his 1973 survey of the hinterlands of Siraf, Whitcomb also included the Galehdar valley in his research, locating 15 additional sites.49 However, the most extensive survey was carried out by Askari Chāverdi between 1998 and 2001, and this located a total of 43 sites. Combining his results with those of Stein and Gaube, Askari argues that the total number of sites in these valleys is 76.50 However, some of the sites discovered by Askari Chāverdi in Galehdar may overlap with those discovered by Whitcomb as his research had not been published when Askari Chāverdi undertook his survey. Of a total of 76 sites and mounds, four are Achaemenid, 12 are Parthian and 30 are Sasanian, plus another 15 Sasanian structures of various functions including architectural, water storage and one building for burial.51 A total of 50 Islamic sites were located (some of which were re-occupation of earlier sites), of which 19 were dated more precisely to the early Islamic period.52 According to the survey report by Askari, under Sasanian rule there is an increase in the population in areas which were already populated, and settlements were also established in the foothills to the south and south-east of the surveyed area where previously there had been no habitation.53 The results show a three-fold increase in sites in the Sasanian period compared to the Parthian period and a ten-fold increase compared to the Achaemenid period.54 Moreover, this figure excludes Askari Chāverdi and Azarnoush 2004: 6. Whitcomb 2009: 86–90. 50  Askari Chāverdi and Azarnoush 2004: 6. 51  Askari Chāverdi and Azarnoush 2004: 10–12. 52  Askari Chāverdi and Azarnoush 2004: 12. 53  Askari Chāverdi and Azarnoush 2004: 10–12. 54  Askari Chāverdi and Azarnoush 2004: 10–12. 48  49 

Askari Chāverdi and Azarnoush 2004: 12. Askari Chāverdi and Azarnoush 2004: 15. 57  Askari Chāverdi and Azarnoush 2004: 6. 58  Askari Chāverdi and Azarnoush 2004: 9. 55  56 

147

Sasanian Archaeology

Figure 6. Distribution of sites of the Sasanian period in Bushehr (after Carter et al. 2006: 66), Lar, Lamerd (after Askari Chāverdi and Azarnoush 2004: 12) and Bastak districts (after Asadi 2010: 17)

was primarily a military base rather than a commercial port and the authors do not believe the increase in sites to have been affected by the presence of this military base. Furthermore, the areas surveyed by Askari were not on the trade routes to the north. In conclusion the increase in the number of sites in the Lar, Lamer and Darolmizan valleys (as surveyed by Askari Chāverdi) is due to an increase in agriculture supported by central government.

caravanserais, guard towers and mosques.60 The 70 sites include four Achaemenid, 17 Parthian, 47 Sasanian and 38 Islamic sites (Figure 6).61 The results of the Lamerd and Bastak surveys are very similar. Both regions see an increase in settlements during the Parthian period followed by a noticeable peak during the Sasanian period and then near collapse during the Islamic period. During the Sasanian period in the Bastak region there is more than a two-fold increase in settlements and fortified sites compared to the preceding Parthian period. During the Sasanian era there are 29 settlements in the plains and foothills, and 18 fortified sites in the mountains of Bastak.62 32 of the Sasanian sites were newly established and some sites measured 10 hectares or more.63 One of the settlements in the vicinity of the Mehran river, known locally as Shahr-e Sibeh, covered 30 hectares and is the largest Sasanian settlement in the Bastak district.64 As in the Lar and Lamerd districts, remains of irrigation systems and structures related to water management were also discovered.65 In some

Bastak district The Bastak district, located east of Lamerd, is the final district to be discussed here. Geographically Lamerd and Bastak are very similar. The Bastak district is composed of two salty river basins, namely Mehran in the south-west and Rod-e Shoor (‘salty river’) to the north-east. Prior to the archaeological survey undertaken by one of the authors in 2007,59 this area had not been archaeologically researched. A total of 200 sites and mounds were noted in Bastak, of which 70 were mounds, fortified sites, settlements and cemeteries, with the remainder composed of structures such as bridges, water storage installations, qanats, canals, 59 

Asadi 2010: 8. Asadi 2010: 9. Asadi 2010: 17–18. 63  Asadi 2010: 18–20. 64  Asadi 2010: 18. 65  Asadi 2010: 18–24. 60  61  62 

Asadi 2010.

148

Ahmadali Asadi et al.: Sasanian and Early Islamic Settlement Patterns Table 1. Sasanian and early Islamic sites discovered on surveys along the northern Persian Gulf District

No. of No. of early Sasanian Publication Islamic sites sites

Bushehr

34

Mohr, Lamerd 30 and Dezhgah

9 19

Bastak

47

10

Siraf Hinterlands

8

14

Comments

Carter et al. 2006

Askari Chāverdi and Azarnoush 2004 Asadi 2010

Whitcomb 2009

Whitcomb reports three sites (two in Jam valley, one in Dezhgah) which were dated to both early Islamic and Sasanian periods, thus are counted twice here

Figure 7. Distribution of sites of the early Islamic period in Bushehr (after Carter et al. 2006: 66), Lar, Lamerd (after Askari Chāverdi and Azarnoush 2004: 13) and Bastak districts (after Asadi 2010: 21)

instances, canals extending up to 10 km in length can be seen, and everywhere we see old qanats, known locally as gabri (i.e., ‘Zoroastrian’), which hints at their age. In many cases the location and general shape of fortified sites in the Bastak district are similar.66 For 66 

example, most of the fortified sites are oval, generally measure between 0.5 and 1 hectare in area, and are situated in the mountains nearest to the settlements in the foothills.67 With regard to their function, again we must focus on security, collection of taxes, irrigation

Asadi 2010: 18.

67 

149

Asadi 2010: 18.

Sasanian Archaeology and management of settlements, especially due to the lack of trade roads linking these settlements and fortified sites. In other words in the Bastak district the increase in settlements is due to the improvement of agricultural and irrigation systems which would not have been possible without governmental support. Sasanian to early Islamic transformation in the discussed areas In most of the surveyed areas, the settlement distribution maps indicate a sharp reduction in the number of sites from the Sasanian to the early Islamic periods. The quantity of sites for both periods, in the area discussed, is shown here (Table 1). In the early Islamic era the Bastak district lost 80% of its Sasanian sites which is the highest reduction in all of the surveyed areas (Figure 7). Based on Carter’s site tabulation, the reduction in the Bushehr hinterland is approximately 70%,68 and in Lar, Lamerd and Darolmizan gorges Sasanian sites diminished by 35%. The only district which apparently shows an increase in the number of sites from the Sasanian to early Islamic period is the Siraf highlands and, as has already been explained, this was due to the importance of the port at Siraf and its trade roadway to the north.

Figure 8. The occupation diagram of Bastak (Asadi 2010: 29) and Lamerd – Mohr districts (after Askari Chāverdi and Azarnoush 2004: 14)

In the Bastak district many Sasanian sites in the foothills, as well as virtually all of the fortified sites, were suddenly abandoned and there are no signs of any diagnostic early Islamic sherds.69 Based on a re-survey of these sites, it is clear that in many cases there are no water resources in their vicinity and the survival of these settlements without a network of channels or water storage structures must have been difficult, if not impossible.70

of the Congregational Mosque above. General evidence shows that the Siraf uprising in the early Islamic centuries was due to the annihilation of other Persian Gulf ports, in particular Rev Ardashir in the Bushehr peninsula. Compared to the exceptional situation of Rev Ardashir, Siraf never had the capability to convert its Sasanian military role to be a trade centre for a long time. As Whitcomb writes:

Few of the Sasanian sites which show continuity from the Sasanian to the Islamic period are located in favourable environmental areas. In other words, the settlement pattern in the Bastak district in the early Islamic era appears to have returned to that of the preParthian period. This situation is also reported in the Lamerd and Mohr plains although without the degree of collapse seen in the Bastak district, probably owing to the slightly better environmental situation of the Lamerd, Mohr and Darolmizan plains (Figure 8).

‘Once a port [Siraf] ceased to be simply a small fishing village inhabited by the coastal population (historically Arab), its social and economic structure took a quantum leap to an order of complexity which resulted from (and was only supportable by) involvement in long-range trade and by a parasitic relationship with at least one urban centre in the hinterland. This was due to the fact that, in economic terms, there was no regional infrastructure of settlements which could give rise naturally to the port as a regional centre; instead, the port as an urban entity was an artificial implantation, becoming a nexus round which a regional settlement system subsequently developed’.71

The excavations at Siraf show that the Sasanian buildings, particularly its castle, were destroyed, presumably in Arab attacks. However, it seems that soon afterwards the site was re-used by the invaders. Whitehouse writes that there is a short gap between the destruction of the Sasanian castle and the building

The decline of Siraf in the later Islamic centuries was an unavoidable process that was merely accelerated by the catastrophic earthquake in AD 977. Moreover,

Carter et al. 2006: 71. Asadi 2010: 20. 70  Asadi 2010: 28. 68  69 

71 

150

Whitcomb 2009: 96.

Ahmadali Asadi et al.: Sasanian and Early Islamic Settlement Patterns time of the Sasanian era was amplified through the involvement of central government. Physical evidence of this central government in the Bastak, Lamerd and Siraf hinterlands is evident in the presence of multiple fortified sites near settlements. The existence of these fortified sites, in addition to providing security for the settlements, was also important for planning and managing the irrigation systems and mobilising the population in the realisation of governmental plans. It is difficult to estimate the level of farming surpluses in the hinterland areas: due to the poor soil fertility of most of the discussed areas (apart from the Bushehr plain), one can postulate that the products were used merely in the interior of Fars province.

in the Bushehr hinterlands Carter noted a drop in site numbers during the early Islamic period and added that: ‘The study recently undertaken of the Williamson Collection proves this point conclusively (Priestman 2005a), indicating a drop of over half the number of sites between the 6th–9th and 9th–11th centuries A.D. Williamson himself recognised this trend, having recorded just seven sites of the 9th–14th centuries A.D. period across the Peninsula, with a combined area of 15 ha., compared with 450 ha. of the Sasanian period’.72 This conclusion was reinforced during the 2004 survey following a brief inspection of Hazar Mardom on the Bushehr peninsula and a casual examination of pottery from Rishahr. Both sites appear to be devoid of common Umayyad and Abbasid period indicators, such as appliqué decorated Alkaline Glazed Ware or ‘Samarra horizon’ wares.73 As noted above, at the same time as the Sasanian port of Rev Ardashir fell into decline, most of the Sasanian settlements on the Bushehr plain were abandoned.

There is no detailed research on trade in the Persian Gulf hinterlands during the Sasanian period. However, analysis of the surveys undertaken in this region do not show any concentration of sites related to the trade roads west of Bushehr plain to the Hormuz strait. The lack of trade routes may in part be due to the east/west trend of the mountains and the difficulty in establishing roads in such terrain which inhibit the movement of goods from ports such as Siraf. In other parts of Fars province the evidence for trade relations between sites is also not particularly strong. Whitcomb’s research on Qasr-i Abu Nasr shows a limited relationship between this and nearby sites:

Conclusions As Wenke has noted, ‘historical and numismatic evidence tell us much about the Parthians’ and Sasanians’ roles in changing the political and social complexion of South-west Asia, but it is imperative that archaeological research also be applied to these problems, particularly in analysing regional settlement patterns, demographics, agricultural systems, trade, craft production, and other factors which, while neglected textually, seem to be crucial elements in these cultural changes’.74 Due to the lack of historical records relating to the Persian Gulf hinterlands in the Sasanian period, archaeological evidence is the only way to study the changing settlement patterns in this region and time. The combined results of the surveys in these areas show that, as in Susiana and Mesopotamia, there was a major increase in settlements in the Persian Gulf hinterland during the Sasanian period. Water management in the hinterland, even today, is a key factor in establishing and sustaining settlements. The increase in settlement during the Sasanian period is most likely to have been powered by governmental assistance in land management and irrigation, without which only a small percentage of the hinterlands is suitable for agriculture or settlement. This may well explain the lack of sites prior to the Parthian period, during which time a form of local management and irrigation must have been initiated, which by the

‘The sealings thus suggest that, at least on an official level, Qasr-i Abu Nasr interacted mostly with Shiraz and secondarily with the district capital of Jur. Connections with other district capitals were very limited, and Qasr-i Abu Nasr did not play the centralized role that medieval Shiraz did. Rather the sealings suggest that Qasr-i Abu Nasr was a regional town, perhaps typical of many, its influence limited to the plain of Shiraz. The identity of this fortress and its relationship with Shiraz (both the name and the city) must be postponed until the rest of the archaeological evidence from this site has been presented’.75 If the same situation prevailed in other parts of the Fars province it might be possible to state that the economy of this province during the Sasanian period was based on agriculture with very limited trade between the sites in the interior parts of the province. However, much more research is required to test this hypothesis. Owing to the necessity of governmental management and support for most of the hinterland settlements during the Sasanian period, their collapse in the early Islamic period is not surprising. We do not know much in detail about the situation created by the arrival of the Arab invaders, but the abandonment of most of

Carter et al. 2006: 97. Carter et al. 2006: 97. 74  Wenke 1981: 304. 72  73 

75 

151

Whitcomb 1985: 15–16.

Sasanian Archaeology the Sasanian fortified sites shows that the Sasanian administrative system which controlled the hinterland districts did not survive. It seems that there was further devastation of sites in important populated centres such as the Bushehr peninsula. Owing to the decline of Rev Ardashir, the port of Siraf temporarily replaced it and became famous in the early medieval centuries. As Carter has noted, the reduction of sites in the Bushehr

plain (as well as in other hinterland districts), contrasts with Whitcomb and Sumner’s hypothesis about early Islamic settlements increasing in the interior parts of Fars province.76 Although the rate of abandonment may have been lower in the interior and northern districts of Fars compared to the hinterland areas considered here, their assumption that the number of sites in Fars increased after the Sasanian period must be reviewed.

76 

152

Carter et al. 2006: 97; see Whitcomb and Sumner 1999: 314.

Bushehr, Dashtestan and Siraf: the Transformation of the Maritime Trade Network in the Middle Persian Gulf Seth, M. N. Priestman Department of Archaeology, Durham University, South Road, Durham DHI 3LE.

Introduction In one of the key works on maritime trade in the Indian Ocean, K.N. Chaudhuri notes that ‘the rise of Siraf … is well attested, but curiously enough never properly explained’.1 Here I wish to revisit this old question through the presentation of selected results from new investigations including a study of finds from Andrew Williamson’s survey of southern Iran,2 a pilot survey of the Bushehr hinterland,3 and recent study of excavated finds from Siraf in the British Museum.4 The aim of this paper is to reconsider the timing, significance and potential reasons for the development of Siraf as one of the main emporia within the Persian Gulf, drawing on my research and on recent advances in the state of knowledge regarding the late Sasanian/early Islamic transition within the region. In particular, I wish to advance the argument that the origins of Siraf can be best explained in relation to the decline of Bushehr, itself linked to a broader realignment in the structure and organisation of maritime trade within the central Persian Gulf and wider Indian Ocean at this period.5 Chaudhuri 1985: 48. Priestman and Kennet 2002; Priestman 2003; 2005a; 2006; forthcoming a. 3  Carter et al. 2006. 4  Priestman 2007; 2021; forthcoming b. 5  This paper represents a substantially reworked version of one published in the proceedings of the International Congress of Siraf Port, 14th–16th November 2005, Bushehr (Priestman 2005b). It has developed out of a number of separate research projects and detailed discussion with different individuals over a long period. Although it will not be possible to acknowledge the many sources of support and inspiration, certain individuals and organisations deserve special mention. In particular I am indebted to St John Simpson and Derek Kennet for innumerable discussions, reading and providing invaluable comments on various drafts of this paper, and access to papers that remain in press. Robert Carter kindly invited me to participate in the Bushehr Hinterland Survey, allowed me to make use of the results of this work here and has shared many thought on the subjects covered in this paper. Similarly, Roberta Tomber has been a source of much information. I would like to express my gratitude to Vesta Sarkhosh Curtis for informative discussions concerning the Sasanian coinage from Siraf. The British Institute of Persian Studies, The British Museum and the University of Durham have supported different stages of my research which are all directly incorporated here. Finally, I would like to express my sincere gratitude to the organisers of the ‘Congress of Siraf Port’ mentioned above for inviting me to present at an extremely memorable and personally 1  2 

Sasanian Archaeology (Archaeopress 2022): 153–177

In fact, a partial answer to Chaudhuri’s question had already been provided in David Whitehouse and Andrew Williamson’s earlier paper on Sasanian maritime trade which juxtaposes the evidence obtained from investigations at Bushehr with those from Siraf.6 Elsewhere, Williamson made a more explicit case for the connection between the rise and fall of the two ports: ‘a similar decline in the area of sites in the Bushehr Peninsula occurred [as that in the sites around Siraf in the 12th–13th centuries] before the 9th to 10th century A.D., and suggests that the immediate predecessor of Siraf is one of several large sites near Bushehr’.7 In addition: ‘The archaeological material gathered so far indicates a striking concentration of coastal settlement by period. For example, preliminary survey indicates that within 50 kms. of Siraf there is a string of large and contemporaneous sites occupying in total more than three times the area of the major site. However, occupation has been very much reduced in this region at all later periods. The Bushehr Peninsula too, has many Sasanian and very early Islamic sites, three of which measure over 1 km. square. Yet the area seems to have rapidly sunk from this great prosperity, since there is very little settlement between the 9th and 15th centuries A.D. This would seem to suggest that the records of families migrating from one city to another as the trade centres shifted might represent a major movement of population’.8

formative event. This paper is born directly out of the paper that I presented there and many of the discussions I had during the event, particularly with the late Sarah Jennings, Willem Floor, Mark Horton and Lawrence Potter. 6  Whitehouse and Williamson 1973. 7  Williamson 1970a: 6. 8  Williamson 1970b: 4.

Sasanian Archaeology Behind the question of a proposed structural realignment in trade and switch of port location from Bushehr to Siraf lies the far bigger one concerning the significance of maritime trade activity during the Sasanian period, and the extent to which such activity may have changed in structure and volume during the early Islamic period. It is unnecessary to enter into a detailed assessment of these issues here and, in some ways, the interacting ‘life-cycles’ of Bushehr and Siraf can be explored in isolation. On the other hand, in order to provide a context for this discussion, it is necessary to summarise some of the salient points of the debate. The traditional perspective has been to regard the 9th–10th centuries as a period of significant development in the scale and intensity of maritime trade contacts between the Persian Gulf and other areas of the Indian Ocean.9 Certain evidence, such as the widespread circulation of pottery, including those exported from east Asia or from the Persian Gulf, seems to support the conclusion. At the same time, from a historical perspective, it must be acknowledged that a real disparity exists between the Sasanian and early Islamic periods in terms of the quantity of evidence available, and this may in part contribute towards the impression of dramatic change in the scale of activity around the 9th century and corresponding to the increase in written sources. Partly in response, there has been a trend over the last four decades to assert the importance of a Sasanian maritime presence within the Persian Gulf, and to emphasis aspects of continuity across the Sasanian/ Islamic transition.10

Piacentini’s argument and sees this as part of an even broader strategic process which enabled the Sasanian empire to monopolise the Indian Ocean economy and exclude Roman competition between the Red Sea and India.13 Another perspective is to see Sasanian monopolisation of the Persian Gulf less in terms of trade and motivated more by the desire for political expansion and the need to secure valuable resources within the Arabian peninsula.14 Based upon historical evidence alone, it is difficult to support or refute the validity of these respective arguments without going into considerable detail and reference to the original sources. Approaching the topic from a different perspective, one thing that is certain is that any significant state-sanctioned engagement in maritime trade during the Sasanian period would have left tangible archaeological evidence which should be detectable through systematic investigation. Since Whitehouse and Williamson presented their evidence, further information for this has been relatively slow to emerge and the results ambiguous. Arguably, the most important recent contribution to our understanding of Sasanian commercial activity in the Indian Ocean comes from a study of pottery found in India.15 Tomber revisited 31 sites and their archived collections where Roman amphorae had been reported. These were concentrated mostly along the west coast, predominantly in the northern half of the subcontinent. Surprisingly, at as many as 18 of these sites, some or all of the ‘Roman amphorae’ transpired to be Torpedo Jars originating via the Persian Gulf.16 Torpedo Jars have a tall elongated form, a closed hole-mouth with a thickened rim, pointed base and asphalt lining. Although not yet confirmed, the typical reddish or yellow sandy fabric is believed to originate somewhere within the Tigris/Euphrates/Karun alluvial basin (i.e., southern Iraq and south-west Iran). The class is found widely distributed throughout the Persian Gulf with a predominantly coastal distribution.17 Coastal findspots starting in the north and moving clockwise include: asSabiyah,18 al-Qusur on Failaka island,19 Kharg island,20 various survey sites along the north shore including a major concentration around Bushehr,21 Jazirat alGhanam,22 Khatt,23 Kush,24 Jazirat al-Hulayla,25 Mleiha,26

The most direct and original treatment of the issue of Sasanian maritime trade is the seminal paper written by Whitehouse and Williamson.11 They argued that the origins of the wide-ranging contacts witnessed between the Persian Gulf and regions such as east Africa, south, south-east and east Asia in the early Islamic period can be traced back to the Sasanian period. In support of this hypothesis, they assembled a wide range of Arab, Aramaic and Chinese written sources, together with the archaeological evidence of early occupation levels at Siraf and Williamson’s impressive discovery of a large-scale Sasanian coastal settlement at Bushehr. The Whitehouse/Williamson thesis has attracted widespread support in many quarters. Placing heavy dependence on their ideas, Piacentini argues that the initial Sasanian conquest of southern Iran by Ardashir I was strategically motivated by the desire of securing access to the Persian Gulf in order to develop its economic potential.12 Daryaee repeats much of

Daryaee 2003: 9. Morony 2001/02: 37. Tomber 2007. 16  Tomber 2007: 977, fig. 6, table 1. 17  Priestman 2005a: 112, fig. 24. 18  Carter et al. 1999: fig. 3: SB4/1, SB2/1, SB5/2. 19  Bernard and Salles 1991: fig. 7: 26; Kennet 1991: fig. 6: 1021. 20  Stève and Julien 2003: pls 13: 2–3, 15: 4. 21  Priestman 2005a: 111–13, 208–209, 339–41, fig. 24. 22  de Cardi 1972: fig. 2: 30. 23  Kennet 1998: fig. 5: 14. 24  Kennet 2004: 63, ‘Type 74’, fig. 36. 25  Sasaki and Sasaki 1996: fig. 49c–e; 2000: figs 14d, 16a, pl. 9: JHU98-4. 26  Benoist, Mouton and Schiettecatte 2003: 69, fig. 8: 10. 13  14  15 

Hourani 1947; 1951: 61–64; Chaudhuri 1985: 36. Whitehouse 1971a; Williamson 1972; Whitehouse and Williamson 1973; Zang Jun-yan 1983; Piacentini 1985; 1992; Gropp 1991; Morony 2001/02; Daryaee 2003; Stève 2003; Kervran 2004; Kervran, Hiebert and Rougeulle 2005. 11  Whitehouse and Williamson 1973. 12  Piacentini 1985: 57–58. 9 

10 

154

Seth, M. N. Priestman: Bushehr, Dashtestan and Siraf and Sir Bani Yas.27 Some finds have also been reported from beyond the Persian Gulf at Sohar in Oman,28 Masirah island,29 Qana in Yemen,30 Ras Hafun,31 Manda32 and Shanga in east Africa,33 and Anuradhapura in Sri Lanka.34 Tomber’s research provides the first indication of the widespread occurrence of these transport vessels within south Asia.

excavations at Khatt40 and Kush,41 were specifically designed in part to resolve the long-term regional ceramic chronology, and has resulted in a far better appreciation of some of the main varieties of common coarsewares and glazed pottery forms associated with different stages of the 5th–8th centuries.42 Building on this framework, it has been possible to define a range of commonly encountered material associated with the first centuries of the Islamic period at different sites, including Kush,43 Hulayla (Area D),44 Sir Bani Yas,45 and al-Qusur.46 While the picture is one which is still rapidly evolving, the results enable a re-dating of a significant number of sites in the Persian Gulf where secure evidence of Sasanian occupation had previously been identified. This includes most of the known examples of so-called ‘Nestorian’ churches in southern Iraq, eastern Arabia and the island of the Persian Gulf,47 together with the significant Sasanian occupation sequences from the regional centres of Qala’at al-Bahrain, Sohar,48 and Siraf, discussed further below. Considered more broadly, this exercise necessitates a reconsideration of the nature of settlement and economic activity during the Sasanian period. Following a systematic review comparing archaeological indices of Hellenistic and Sasanian activity in eastern Arabia, including attested site occupation, coin finds and burials, Kennet concluded that the Sasanian period is actually marked by a prolonged episode of economic decline.49 This conclusion provocatively swings the argument back the other way in terms of the debate over the strategic involvement of the Sasanian empire within the Persian Gulf region as a whole. Kennet’s argument is based essentially on the absence of evidence, but whether this will withstand long-term scrutiny remains unclear. What is certain is that it re-opens the debate and forces one to critically re-examine existing evidence of Sasanian activity within the Persian Gulf wherever it has been previously reported. However, it is based upon evidence from the Arabian coast and islands in the Persian Gulf, raising another question as to whether the same situation applies to the opposite coastline, and it is this point which forms the basis of this paper.

Torpedo jars have long been regarded as a typical Sasanian type-fossil.35 On the face of it, therefore, Tomber’s research seems to provide evidence for an active phase of commercial exchange, most likely in luxury foodstuffs, between the Persian Gulf and south Asia during the Sasanian period.36 However, there are problems with this interpretation. The production of Torpedo Jars seems to have continued uninterrupted from the Parthian period up to around the 10th century, as Adams himself acknowledged when he began using this type as a diagnostic indicator of Sasanian sites found on his surveys in central and southern Iraq.37 While some of the south Asian finds do come from contexts which can be securely attributed to the 4th–6th centuries, the wide dating of this type of pottery leaves open the possibility that some of the sites in India belong to a stage contemporary with the early Islamic period. Closer study of Torpedo Jar form profiles, combined with petrographic analysis of the fabrics and chemical analysis of the asphalt lining, may enable this long-lived tradition to be further subdivided and its constituent components distinguished and more closely dated.38 At present the discussion still rests on the finds with which Torpedo Jars are associated in south Asian contexts, and unfortunately many of these are themselves imprecisely dated. One of the major factors hampering more precise discussion of the Sasanian/Islamic transition within the Persian Gulf has been the inability to closely define pottery belonging to this period. A long-term research project concentrated in Ras al-Khaimah (United Arab Emirates), which involved regional survey39 and Carter 2008: figs 13: 1–6, 8–10, 19: 10–11. Kervran 2004: figs 10: 15, 11: 6, 22: 3, 24: 2. 29  Weisgerber and al-Shanfari et al. n.d.: 86, 98, fig 174, 206: two complete examples were found lying side by side in the coastal sand near a large rocky coastal outcrop and each containing an estimated 25 litres but the asphalt lining excludes the proposal that they were ‘probably used to store water on ships’. 30  Sedov 2007: fig. 4.22: 5–7. 31  Smith and Wright 1988: 121, figs 4: e, 5: a–c, 7: a–b, e, 9: a–g. These authors compare the examples from Ras Hafun with the ‘classic’ Torpedo Jar tradition of the Persian Gulf but the drawing suggest that these are actually an earlier variant of the tradition well represented by examples from Mleiha and ed-Dur. 32  Chittick 1984: 88, figs 43: a–d, 45. 33  Horton 1996: 296. 34  Stern et al. 2008: 3, fig. 2. 35  e.g., Adams 1965: 132, Period 12, Types K–L. 36  Tomber 2007: 982. 37  Adams 1965: 132. 38  Carter et al. 2011. 39  Kennet 1994; 2003. 27 

The Williamson Survey

28 

Introduction The starting point for this particular enquiry is the pottery recovered from Andrew Williamson’s survey of Kennet 1998. Kennet 1997; 2004. Kennet 2002; 2004. 43  Kennet 2004. 44  Sasaki and Sasaki 1996; 2000. 45  Carter 2008. 46  Patitucci and Uggeri 1984; Kennet 1991. 47  Simpson 2018a; Kennet 2007: 89–94; Carter 2008: 71–72, 97–103. 48  Kennet 2007: 96–100. 49  Kennet 2005; 2007; 2008. 40  41  42 

155

Sasanian Archaeology

Figure 1. Map showing the location of those sites located from Williamson’s survey of southern Iran together with major rivers and ground above 500 m (shaded in grey)

southern Iran and the Persian Gulf between 1968 and 1971. Since October 2001, research has been carried out by the author, under the supervision of Dr Derek Kennet with generous support of the British Institute of Persian Studies, on the substantial portion of some 17,000 sherds which Williamson exported to the UK.50 Williamson was prevented from personally bringing this work to completion following a fatal accident during fieldwork in Oman in April 1975. Had his death been avoided, our understanding of Persian Gulf archaeology in later historic periods would likely be substantially different to the point at which it stands today. What he did accomplish during his short career was, in its own right, a significant and unique achievement: over the space of three years and involving some 18 months of fieldwork, Williamson undertook a very extensive survey of most of the Persian Gulf coast of Iran, from Bushehr to Jask and inland through Fars in the area between Bushehr, Shiraz and Sirjan, and in Kerman from the Minab plain through the valley systems leading up towards Jiroft (Figure 1). Throughout his survey, he focused specifically on the Sasanian and Islamic periods, although for most of the project he was accompanied by Martha Prickett who recorded evidence for prehistoric occupation wherever it was

encountered.51 The type of survey which Williamson undertook was by no means detailed, and the nature of the information that he collected was basic: just small samples of finds, mostly pottery, collected from the surface of sites.52 However, what the collection does provide is the opportunity to look at the distribution of sites and particular groups of pottery over a very wide geographical area, and this information can be used to compile generalised models of commercial interaction and long-term economic change operating on a large scale.53 Emporia ‘migration’ Despite the problems encountered in extracting information from what is an incomplete and rather poorly organised paper archive,54 the coherence of some of the results emerging from the detailed classification and analysis of the pottery inspires confidence in the overall integrity of the data. One of the most striking developments that it has been possible to chart what appears to be a long and continuous sequence involving successive major emporia within the Persian Prickett c. 1971; 1986: 508–16. Priestman 2004: 30. 53  Williamson 1987; Priestman 2005a: chapters 3–4. 54  These issues are discussed in detail in Priestman (2005a) and will be made more widely available with the forthcoming publication. 51  52 

50 

Priestman and Kennet 2002; Priestman 2003; 2004; 2005a; 2006.

156

Seth, M. N. Priestman: Bushehr, Dashtestan and Siraf situated on the northern shore of the Persian Gulf, and that the primary function of this was as a conduit for goods between the major centres of manufacture and consumption in Mesopotamia and southern Iran and the wider Indian Ocean trade network. Over time, as one centre fell into decline, another rose to power and, with each change, there appears to have been a regular trend whereby the trade emporia shifted progressively southwards. It is possible that this pattern was simply governed by chance,63 but if it was guided by a coherent structuring principle, then this may be related to wider developments in the organisation of Indian Ocean trade, a point explored in greater depth below.

Gulf with a periodic ‘migration’ from one emporium to the next.55 This migration appears to start within the upper Persian Gulf, and over a period of about 1800 years or more, moves towards the lower end. Siraf was one of these major ports which, at one time, played a dominant role in regulating the flow of trade through the Persian Gulf. The events which mark the end of Siraf ’s period of dominance are reasonably well known. An earthquake in 977 may well have caused substantial devastation from which it was difficult to recover, and this in turn made Siraf more susceptible to attack. At the same time, it is clear that the site continued to be occupied after these events,56 although by the early 11th century it came under increasing competition from the emerging centre of Kish, and with a shift in emphasis towards a direct north/south trade route between Kerman and the Persian Gulf coast after the fall of the Buyids in the mid-11th century, the rulers of Kish were able to succeed militarily and most of the trade shifted further south.57 Kish itself became a major emporium which was able to dominate trade through the Persian Gulf between the 11th–13th centuries,58 although it came under increasing pressure from the emerging economic and political influence of Hormuz, situated within the mouth of the Persian Gulf on the edge of the Minab delta.59 During the early 14th century, the city of Hormuz was relocated to Jarun island, probably to make it less vulnerable to attack.60 Certainly there must have been good reason for this move as that island has very limited water resources, and the added expenditure needed to maintain a settlement there must have been significant. At about the same time that the city was refounded, Hormuzi forces won a significant battle over the navy of Kish and following these events, the port at Kish is thought to have fallen into rapid decline with most of the trade shifting to the mouth of the Persian Gulf,61 although the pottery recovered by Williamson indicates that the decline of Kish may have been more protracted than previously believed.62

Returning to the origins of Siraf, it is clear that the port was part of a wider trade network from an early stage. Even in the earliest levels excavated, before any major architectural developments in Site A and before the construction of the Great Mosque at Site B, substantial volumes of east Asian pottery were arriving at the site.64 These earliest imports mostly consist of green glazed Dusun jars (DUSUN: Figure 2),65 painted stoneware Changsha bowls (CHANG: Figure 3), and three more unusual categories: one of large handmade jars with a black fly-ash glaze (STO.BG1), another of light-green glazed bowls with large square patches removed from the interior (STO.PLG),66 and small jars with a black speckled grey-green glaze with patches of brighter duck-egg blue on the surface (STO.BLU). The last category is of particularly crucial significance. Only three sherds are represented amongst the body of east Asian imports from excavations at Siraf in the British Museum.67 The same class is also represented at Sohar in Sector II from the surface.68 Both the distinctive glaze and an elongated vertically pierced handle are characteristic of the Sui or early Tang dynasties and date to within the 7th century.69 These finds represent the earliest east Asian imports recovered from anywhere in the Persian Gulf.70

The long-term sequence involving the succession of Bushehr, Siraf, Kish and Hormuz is undoubtedly oversimplified, and there are likely to have been many factors involved in the evolution of the trade network within the Persian Gulf over this long period. In essence, however, there appears to be good evidence to suggest, both historically and archaeologically, that at any one time there was only one major trade emporium

There is a one in six chance of outcome which actually occurred. Whitehouse 1968: 18; 1969: 59; 1979: 56. 65  Throughout the text veramic classes are referred to by their class title and class code in parentheses, although once a class has been introduced within a section, it is referred to just by the class code. Class codes represent an acronym of the full class title which is used as shorthand for that class category. Similarly, types are referred to by the type codes they have been assigned in the Kush study (Kennet 2004) or during study of finds from Siraf in the British Museum (Priestman forthcoming b). A list of class and type codes referred to in the text appears at the end of the report (Fig. 12), along with illustrations of some of the main varieties of ceramic material discussed (Fig. 13). 66  Whitehouse 1973: 244–49, pls 18.1, 18.2, 18.3. 67  BM 2007,6001.2739, 6814, 9024. 68  Kervran 2004: fig. 37: 20. 69  Priestman and Krahl forthcoming. 70  Rougeulle’s proposal of a 5th century date for sherds of ST.BG1 is incorrect (Rougeulle 1996: 160), and based on the erroneous dating of the French excavations at Sohar (Kervran 2004). 63  64 

Priestman 2005a: 142–45. Whitehouse 1975: 269 57  Potter 2005: 37–38. 58  Whitehouse 1975: 268. 59  Williamson 1973; Morgan 1991. 60  Bakhtiari 1979. 61  Williamson 1973: 57. 62  Priestman 2005a: 149, fig. 38. 55  56 

157

Sasanian Archaeology

Figure 2. Dusun jars from Siraf dating to the late 8th–10th centuries, similar to those from the Belitung shipwreck.

Figure 3. Changsha bowls from Siraf dating to the late 8th–9th centuries, similar to those from the Belitung shipwreck

158

Seth, M. N. Priestman: Bushehr, Dashtestan and Siraf conclusions relating to the development of Sasanian maritime trade.

Further evidence for early trade contact between east Asia and the Persian Gulf comes from the fortuitous discovery of a shipwreck off the island of Belitung in the western Java Sea dated to after 838,71 most likely close to the mid-9th century.72 This ship was built of wood from India, constructed following an Arab/Iranian or Indian design, and loaded with a cargo of 56,500 Changsha bowls stacked inside Dusun jars.73 Although it is difficult to prove, there is every reason to believe that this ship was bound for the Persian Gulf. If so, it would confirm what has already been described historically, and what is indicated from the archaeological evidence from Siraf, namely that voyages were being made to China by merchants from the Persian Gulf at this early date.

Bushehr and its hinterland Bushehr A short pilot season of multi-period field survey was carried out in the hinterland of the Bushehr peninsula between 23rd November and 18th December 2004, concentrating on the district of Dashtestan between Ahram in the south and Gonaveh in the north (Figure 4).80 The Bushehr peninsula is the site of a series of archaeological discoveries stretching back to the beginning of the 19th century. The British Residency in the Persian Gulf – located at Bushehr with a summer residence further south at Sabzabad – employed or hosted a succession of political and military officials, some of whom took an antiquarian interest in the local archaeology.81 One of the recurrent finds were cemeteries containing linear alignments of Torpedo Jars or carved limestone ossuaries, both containing human remains. Nine different burial sites of this nature have now been reported.82 In 1913, the French Archaeological Mission to Persia sent a delegation to Bushehr to conduct an investigation of the area. This included an excavation at Tul-e Peytul (ancient Liyan), a site they were drawn to by the report of cuneiform inscribed bricks,83 and excavated previously by Andreas in c. 1876.84 Pézard described the painted pottery of the mound as ‘Chalcolithic’, although in fact it belongs to the Kaftari horizon of the late 3rd/early 2nd millennium BC.85 Sir Aurel Stein spent some time at Bushehr in 1933 at the end of his second Persian expedition, awaiting permission to continue his expedition into Fars.86 As this could not be easily obtained, the exploration eventually had to be postponed to the following season. Perhaps due to the unusual situation of being thwarted, Stein’s assessment of the area is uncharacteristically dismissive:

The available archaeological evidence from a large number of port sites distributed around the rim of the western Indian Ocean provide evidence for ceramic exports from east Asia or the Persian Gulf from the 7th century at the earliest, and mostly from the late 8th century onwards. This indicates a significant realignment of trade during the early Islamic period, beginning seriously within the Abbasid period. At the same time, there are hints that the Persian Gulf was linked to the Indian Ocean maritime system at an earlier date too. We have already referred to the evidence of Torpedo Jars. Small quantities of pre-Islamic Alkaline Glazed Wares emanating from southern Iraq (ALK.T/YG) have been reported from a number of sites, including Ras Hafun,74 Berenike75 and Aksum,76 as well as Pattanam on the western coast of India.77 These modest sources of evidence are likely to grow with further investigation of the appropriate sites and regions, and as the relevant categories of pottery become better known and more widely recognised. It therefore appears that the emporia system was not itself entirely new. As has already been discussed, Whitehouse and Williamson were able to present their own arguments to suggest that long-distance maritime trade was already flourishing during the Sasanian period.78 Williamson’s detailed survey of the Bushehr peninsula was significant in making this point. In 2004 the present author was able to pursue these questions further through fieldwork conducted within the Bushehr area as part of the joint University of Durham/ Iranian Centre for Archaeological Research pilot survey of the Bushehr hinterland.79 This work produced important results which supplement Williamson’s

‘The effects of this … delay might have been less felt by me if Bushire itself had offered more The work was carried out by a team co-directed by Dr Robert Carter (Oxford and Durham University) and Mr Hossein Tofighian (Iranian Centre for Archaeological Research). Other members of the team in the field included Mr Hameed Zareh and Mr Biladi (Bushehr Cultural Heritage Handicrafts and Tourism Organisation), Dr Ben Horton and Andy Kemp (University of Pennsylvania) and the present author (Durham University). Acknowledgement of the sources of support and assistance are given in the fieldwork report of this season: Carter et al. 2006. 81  Simpson 2007a; 2019c. 82  Simpson 2007a: 153. 83  Pézard 1914. 84  Simpson 2007a: 155. 85  Potts 2003: 159. 86  Stein 1937: 234–41. 80 

Guy 2001/02. Krahl, pers. comm. 2009. Flecker 2000; 2001: 345–48. 74  Smith and Wright 1988: 121, figs 5: d–e; 7: c–d, f–k. 75  Tomber 2008: 51–52, fig. 10. 76  Wilding 1989: 315; Phillips 2000: 326–27, fig. 284; see also Simpson, this volume: 256. 77  Shajan et al. 2008; Cherian et al. 2009: 237. 78  Whitehouse and Williamson 1973. 79  Carter et al. 2006. 71  72  73 

159

Sasanian Archaeology

Figure 4. Map showing the distribution of sites (marked in black) surveyed and recorded during the Bushehr Hinterland Survey

opportunities for useful fresh observations … But apart from the scanty remains of Rishahr … the only site so far traced is the ruined mound … east of … Sabzabad’.87

and the historically attested port of Rev Ardashir had already been made,88 Williamson was able to effectively end a long-running dispute over this city’s location by setting out the archaeological evidence for its identification. Williamson spent at least a month carrying out a survey of the peninsula over three separate visits from 1969–1971,89 during which he recorded at least 89 separate sites. Apart from an 80 x 120 m site (H200) situated 6 km north of Halileh on the road to Bushehr, with ‘painted pottery characteristic of Early and Late Ubaid together with a well fired straw tempered coarse ware’,90 Williamson’s most significant discoveries were extensive ruins towards the middle

Despite this criticism, Stein produced a detailed account of various aspects of archaeology on the peninsula which was only superseded in the late 1960s when Andrew Williamson conducted the first systematic archaeological investigation of the area. Williamson seems to have been the first to recognise the extent of archaeological remains on the Bushehr peninsula and particularly to appreciate their significance within a Sasanian context. Although the link between Bushehr

Lockhart 1960: 1341. Williamson 1971a; 1971b. 90  Williamson 1970c: 7; Prickett and Williamson 1970: 1. 88 

87 

89 

Stein 1937: 235.

160

Seth, M. N. Priestman: Bushehr, Dashtestan and Siraf

Figure 5. Distinctive olive-green Alkaline-Glazed Ware (ALK.YG) of the 5th to 7th centuries used as the main type-fossil by Williamson for dating the Sasanian settlement at Bushehr

one essentially restricted in character to central Iraq, with very few obvious points of association with the repertoire circulating within the Persian Gulf.95 Perhaps fortuitously, therefore, Williamson’s general dating seems to be backed up by more recent dating evidence. At Kush a few examples of the same type (Type 64) occur in Periods I–II, suggesting a date range between the 5th–7th/8th centuries (Figure 5).96 Potentially also significant is the absence of Type 64 from the assemblage at Sir Bani Yas, which has been re-dated on basis of the pottery and glass to between the mid-7th and mid-8th centuries.97 This might allow the potential range of Type 64 to be narrowed to the period before the mid-7th century, although a similar pattern of absence from other sites with larger pottery assemblages of the Sir Bani Yas horizon would be needed to strengthen the argument.

of the peninsula at Rishahr and at the southern tip of the peninsula between Halileh and Hazar Mardom. Between these two major settlements, he also traced an almost continuous scatter of sites strung out along the higher ground facing the shore.91 Williamson estimated that across the Bushehr peninsula there were 450 hectares of archaeological mounding simultaneously occupied during the mid to late Sasanian period.92 Cumulatively, this is larger than any other settlement he encountered on the Persian Gulf coast. The main evidence cited in support of this assertion is the occurrence of a single late Sasanian ceramic type-fossil assigned to the 5th–7th centuries and found at all sites. This type, an Alkaline Glazed Ware basin with lug handles, light olive-green glaze and distinctive bifurcating rim,93 was attributed by Williamson to the period between the 5th–7th centuries as it was absent from either 3rd–5th century levels at Veh Ardashir in Iraq, or the earliest levels at Siraf.94 This dating by absence is not particularly satisfactory, although of course Williamson was simply – and quite reasonably – using what limited information he had. In the case of Veh Ardashir, there is a real problem in the fact that the Sasanian assemblage appears to be

The scale of the archaeological discoveries which Williamson made at Bushehr, together with the absence of any sites of comparable size or date along the 95  Information based on Laura Cordera’s unpublished paper ‘Pottery from Veh-Ardashir: continuity and change in production during the Sasanian period in Mesopotamia’ in the Mega-Sites session of the 7th International Congress on the Ancient Near East held in London, 12th–16th April 2010. I am also grateful to Dr Cordera for informative discussions concerning the regional distribution of Sasanian pottery traditions. 96  Kennet 2004: 29–31, fig. 5, table 16. 97  Carter 2008: 71; Phelps, Simpson and Freestone 2018.

Whitehouse and Williamson 1973: 37, fig. 4. Williamson 1971/72: 35. 93  Priestman 2005a: 234–38, ‘ALK: 29–30’, pl. 91. 94  Williamson 1971/72: 10–11. 91  92 

161

Sasanian Archaeology remainder of the Persian Gulf coast, make a compelling argument for the identification of Rishahr with Rev Ardashir.98 This port city is described historically as being one of the principal ports of the Sasanian empire, founded in the 3rd century by Ardashir I, apparently as part of a strategic policy of commercial expansion within the Persian Gulf,99 and regular trade contacts are said to have been maintained with India from this period on.100 It was in this city that the seat of the metropolitan of Persis within the Church of the East was established between the 6th–8th centuries, from where bishops were ordained throughout the main cities of southern Iran, as well al-Bahrain, Socotra, southern India and Sri Lanka.101 By the later Sasanian period, settlement on the Bushehr peninsula appears to have reached its apogee. This was followed by a phase of substantial settlement collapse at some point between the 6th and 9th centuries.102 Williamson recognised this trend as he recorded just seven sites of the 9th–14th centuries, with a combined area of 15 hectares, a marked contrast to the previous total of 450 hectares.103

Dashtestan plain down to Shif. From Shif he proposed that it would have been carried out on an aqueduct straight to the Bushehr peninsula, via Shif Island, and then channelled through pottery pipes to the settlements of Rishahr and Halileh.106 The presence of this system suggests a significant eastwards extension of a system of major organised state intervention in agricultural production demonstrated elsewhere by the extensive Sasanian irrigation networks of central and southern Mesopotamia107 and south-west Iran.108 The existence of the Angali canal seemed to resolve the issue of lack of water on the peninsula, a question which Williamson believed to have addressed through the discovery of a number of well heads, some of which he estimated would have had to be sunk up to 100 m to reach the watertable.109 Results of the Bushehr Hinterland Survey The first season of the Bushehr Hinterland Survey set out to address a number of questions. Before reaching the field, the survey area was divided into 14 zones broadly corresponding to different geomorphological areas based on colour contrasts visible on regional composite ETM satellite images.110 The area of the investigation was deliberately situated on the mainland away from the Bushehr peninsula as the potential to work there was restricted. In addition, Williamson’s investigation of Bushehr was already relatively complete compared to the limited information available for the district of Dashtestan. In the field an attempt was made to visit as many of the 14 zones as possible in order to assess their variable archaeological potential and characterise the range of archaeological sites within the overall district. At the same time, more carefully focused survey was carried out to answer specific questions related to two particular periods: namely the Ubaid and Sasanian periods. In relation to the latter, some attempt was made to explore the relationship between the urban settlements on the Bushehr peninsula and the occupation in its immediate rural hinterland, partly involving an attempt to substantiate the existence of the Angali canal on the ground.

Given the scale of the archaeological remains identified by Williamson on the Bushehr peninsula, and the important role of the port city suggested by the written sources, the port must have been supported by a substantial regional infrastructure. Part of this would have been the routeways designed to carry goods between the core centres of the Sasanian administration in inland Fars, including Bishapur, Firuzabad, Istakhr and Darabgird, and the coastal port and centre of Sasanian maritime economy at Rev Ardashir itself.104 In turn Rev Ardashir must have depended on local agricultural production capable of supporting the immediate population, providing supplies for shipping and perhaps even a surplus for an export market. On the peninsula, the potential for agricultural production is severely limited by the lack of arable land and running water but its immediate hinterland in Dashtestan represents one of the few areas adjacent to the northern shore of the Persian Gulf which has a substantial cultivable plain and ready source of irrigation. Following a short field survey in 1973, Donald Whitcomb identified what he believed to be significant evidence directly linking Bushehr and its rural hinterland.105 Inspection of aerial photographs of the area made after the survey revealed a substantial linear feature which he interpreted to be a canal taken off the Dalaki-Rud, north of Borazjan, and running for over 40 km across the

One of the first significant discoveries arising from this survey was the realisation that the Angali ‘canal’ may have been mis-identified.111 Despite the presence of a distinct linear feature visible on even relatively lowresolution satellite images, inspection at several points along its length revealed none of the typical features associated with a canal such an upcast levy. At a bend in the Dalaki-Rud, where the canal was supposed to have

Williamson 1971/72: 34–35. Piacentini 1985: 60. 100  Gropp 1991: 86. 101  Gropp 1991: 85. 102  Williamson 1970b: 4; see below for further discussion of the dating of this event. 103  Prickett and Williamson 1970: 1. 104  Whitcomb 1984: 333; Daryaee 2003: 6. 105  Whitcomb 1987. 98  99 

Whitcomb 1987: 311. Adams 1965; 1981. Neely 1974; Wenke 1975/76. 109  Williamson 1971/72: 35. 110  Carter et al. 2006: fig. 3. 111  Carter et al. 2006: 67. 106  107  108 

162

Seth, M. N. Priestman: Bushehr, Dashtestan and Siraf been taken off, there is actually a modern pumping station providing water for a small irrigation channel. At the same point, there is also a sizeable, mainly late Sasanian, site (BH45). However, before pump technology a dam would have been required to raise the level of the water up by some 12 m to reach the height of the surrounding plain and no evidence of such a structure could be found. Moreover, even if a dam had been constructed, the flow of the river appeared highly seasonal and for the majority of the year, too weak to support a 40 km canal. A visit to the opposite end of the canal proved more conclusive. At the canal’s terminus with the sea at Shif a concrete jetty has been constructed providing access to the fishing community on Shif island. During the 17th–19th centuries, when Bushehr served as one of the main Persian Gulf ports for the Dutch and English East India Companies,112 Shif provided an acess point for boats carrying goods on and off the peninsula, in preference to the slower and more treacherous route across the sabkha flats separating the peninsula from the mainland.113

Figure 6. View looking out across the open stretch of water to Shif Island: half of the distance that would need to be crossed by an aqueduct (photograph: author)

during the First World War and erected between Shif to Borazjan by the British Indian army to support the operations of the Persian Field Force in 1918.116 After the war the railway continued to be used for freight and civilian transport until 1921 when it was dismantled and transported back to India. However, the railway hypothesis still does not resolve the lack of visible evidence for the feature on the ground, and it is by no means certain what vestige might remain a century later of a light railway that was in place for as little as three years. Alternatively, the route may also have been one that was used in antiquity with its course being marked by a hollow way which would be more difficult to discern on the ground.117 Ultimately further fieldwork is needed to resolve the question.

Set back from the modern causeway and jetty, a series of low archaeological mounds were examined, including a group with abundant pottery of the 12th– 14th centuries (BH5B-E) and a more substantial mound belonging to the mid to late Sasanian period (BH5A). The scale of the archaeological remains suggests that Shif was an important part of the infrastructure of the Bushehr peninsula in earlier periods, as it went on to do between the 17th–19th centuries. Looking out from Shif towards Bushehr there is a wide, open and relatively deep stretch of water which would pose a significant challenge even to modern bridge-builders (Figure 6). It seems inconceivable that an aqueduct ever crossed this stretch of water. Instead, it seems far more likely that Shif acted as a ferry point throughout the major periods of occupation on the Bushehr peninsula. This point was reinforced by an examination of Shif Island, which produced only a minor background of Sasanian pottery (BH53-55) and no indication of substantial stone or brickwork associated with an aqueduct. Finally, it has been observed that what Whitcomb identified as water pipes114 were actually alignments of torpedo jar burials, correctly identified as such by those who recorded them in the 19th century.115

Turning to the issue of rural settlement on the Dashtestan plain, the Bushehr Hinterland Survey produced a number of important results. During attempts to identify the ‘Angali canal’, several points across the Dashtestan plain were examined. While some locations produced no finds, in the centre of the plain, c. 500 m to the south of the modern village of Mokhi, two areas of c. 60 m north/south x 100 m east/west (BH2) and c. 300 m north/south (BH3) were recorded which were covered with a thin scatter of small heavily fragmented and badly abraded Achaemenid and/ or late Sasanian coarseware sherds118 which bore the typical signs of a manuring scatter associated with intensive agriculture.119 Apart from this field scatter, little further attempt was made to explore areas on the extensive alluvial plain immediately inland from Shif. This zone is likely to provide a number of difficulties for archaeological investigation for much is under modern cultivation. At the same time the landscape

Taken together, the evidence related to the linear feature identified by Whitcomb and seen on recent satellite images indicates that it cannot have been a canal. One possible explanation for the feature is that it marks the course of the 53 km long light railway commandeered from the Indian Powyan Steam Tramway Company

Floor 2004: 180, 194; Hughes 1994. Carter et al. 2006: 67; Wilkinson 2003: 111–17, figs 6.10–14. 118  See below on the problems of distinguishing the common coarse wares of the Sasanian and Achaemenid periods. 119  Wilkinson 2003: 55–57. 116 

Lockhart 1960. 113  de Planhol 1990: 570. 114  Whitcomb 1987: 315. 115  Simpson 2007a: 156. 112 

117 

163

Sasanian Archaeology Table 1. Class and Type codes referred to throughout the text displayed in alphabetical order and showing: their full title; where the code is first published and described; the class/types associated dating and believed place of origin and the equivalent class code that appears in Kennet (2004) Class Code

ALK.T

Name

CHANG

CLINKY

CONG.C

Type 64

Clinky Fired Earthenware

 

   

 

CONG.RG

TURQ.3-4

Iraq

TURQ.5

Iraq

TURQ.3-4

Class (Priestman forthcoming b); 5th–mid 8thC Type (Kennet 2004)

Iraq

TURQ.1-2

Class (Kennet 2004)

4th–6thC

Iran

Class (Carter et al. 2006)

c. 3rdC BC–4thC Iran

Class (Priestman forthcoming b); mid 7th–mid Type (Kennet 2004) 8thC

Yellow-Green Alkaline Glazed Ware

CONG.G

Iraq

Type 72 Turquoise-Blue Alkaline-Glazed JR5 Ware

Changsha Ware

Kennet Class

Class /Type Code Origin and Description

BR26 ALK.YG

Origin

Type

 

Class and Type (Priestman forthcoming b)

Class and Type (Priestman forthcoming b)

Class (Carter et al. 2006)

Class (Carter et al. 2006)

Green Glazed ‘Dusun’ Ware

 

Class (Kennet 2004)

EGG.PI

Plain or Incised Eggshell Ware

 

Class (Priestman forthcoming b)

FINLIM

Fine Sandy Lime-Spalled Red Ware

LISV

Class and Type (Priestman forthcoming b)

HARGE HARLIM HARMIC HONEY IRPW

REBROS SLIP.HB

Fine Orange Painted Ware

Hard Lime Spalled Red/Black Ware

STO.BG1 STO.BLU STO.PLG

  LISV

Coarse Red Black Painted Ware   Honeycomb Ware

 

Gritted Red/Brown Slipped Ware

LISV

Indian Red Polished Ware

SLIP.HR

SMAG.B

 

   

 

Black Iron-Glazed Stoneware Jars, Group 1 Green Glazed Stoneware Jars with Blue Highlights Patchy Light-Green Glazed Stoneware

 

mid 8th–late 10thC

mid 8th–late 10thC

Class (Priestman forthcoming b) late 8th–9thC

DUSUN

FOPW

Date

c. 3rdC BC–4thC Iran

c. 3rdC BC–4thC Iran mid 8th–late 10thC

 

 

 

China Unchanged Iraq

EGG

5th–mid 8thC

Iran

 

Class (Kennet 2004)

3rd–6thC

Iran

Unchanged

Class and Type (Priestman forthcoming b)

7th–early 9thC Iran

Class (Carter et al. 2006)

c. 3rdC BC–4thC Iran

 

SMAG

Class (Priestman forthcoming b) ?

India

 

Class (Kennet 2004)

India

Unchanged

Class (Kennet 2004)

8th–9thC

Class and Type (Priestman forthcoming b)

mid 8th–late 10thC

Class (Carter et al. 2006)

Class (Carter et al. 2006)

Class (Carter et al. 2006)

1st–8thC

Iraq  

Unchanged  

c. 3rdC BC–4thC Iran

 

c. 3rdC BC–4thC Iran

 

c. 3rdC BC–4thC Iran

 

 

Class (Priestman forthcoming b) late 8th–9thC

China  

 

Class (Priestman forthcoming b) 7thC

China  

 

Class (Priestman forthcoming b) late 8th–9thC

China  

mid 8th–late 10thC

Class (Priestman forthcoming b)

TORP.S

Class (Priestman forthcoming b) 3rd–mid 8thC

 

Iraq

 

Iraq

TORP

This area requires further intensive survey in order to establish whether areas like the field scatter at Mokhi are widespread or whether significant areas of the plain were brought under cultivation during the Sasanian period. Towards the back of the plain, before reaching the precipitous chain of hills that flank its northern edge, the ground begins to break up, undulating in a

is largely featureless and the whole area subject to heavy processes of alluviation, which may well have masked even relatively recent archaeological sites. It is striking that the survey did not reveal any sites earlier than the Sasanian period below the 20 m contour.120 120 

Unchanged

mid 8th–late 10thC

TORP.RG Red Grit Tempered Torpedo Jars   Sandy Torpedo Jars

China Unchanged

Carter et al. 2006: 75.

164

Seth, M. N. Priestman: Bushehr, Dashtestan and Siraf

Figure 7. A selection of some of the main vessel types and classes referred to within the text in relation to the late Sasanian/early Islamic transition: a, ALK.T/BR25; b, ALK.T/ Type 72; c, ALK.YG/Type 64; d–e, ALK.T/JR5; f–g, CLINKY; h–i, HARLIM; j–k, IRPW; l, HONEY

low series of ridges and gravel fans. It was in this area that most of the archaeological discoveries were made.

from around the 5th century onwards, a wide range of classes could be confidently identified on the basis of comparison with the classification developed on the basis of the Williamson Collection,121 itself linked to the chronology provided by excavations at Kush.122 The most abundant and diagnostic classes for this period are Hard Lime Spalled Red/Black Ware (HARLIM), Clinky Ware (CLINKY), Large Incised Storage Vessels (LISV), Torpedo Jars (TORP.S), Honeycomb Ware (HONEY) and particular vessel types associated with Alkaline Glazed Ware (ALK.T/YG) (see Table 1; Figure 7). Within the later Sasanian and early Islamic range (c. 5th–8th centuries), predominantly earlier or later occupation

A total of 56 sites were identified and recorded during the Bushehr Hinterland Survey, rising to 73 if sub-sites are included. Although these ranged in date from the 5th millennium BC to the 20th century, there was a clear concentration of sites dating within the range of the Achaemenid to late Sasanian periods. Dating was undertaken largely on the basis of the surface pottery and during three weeks of survey, 3,192 diagnostic sherds were collected as ‘grab samples’ from most of the sites visited. At the end of the project, all of the pottery was classified, described and illustrated. The treatment and chronology of the post-Elamite material needs some further explanation. For the material dated

121  122 

165

Priestman 2005a. Kennet 2004.

Sasanian Archaeology

Figure 8. Site count showing the number of new sites, abandoned sites and sites which had continuity of occupation between periods, based on the results of the first season of the Bushehr Hinterland Survey

could sometimes be differentiated by the presence of CLINKY and ALK.T Type 64 for the earlier dated sites, and ALK.T types JR5 and BR25 and HONEY on the early Islamic sites. At the time of the survey, none of the specialists involved were as alert to the definition of the 7th–8th century assemblage as we are today.123

that he must have come to the same conclusion regarding the dating of this material and excluded it deliberately on chronological grounds. This conclusion seems to be borne out by the presence of a selection of sherds of SLIP.HR and SLIP.HB separated out from the other finds from Bushehr and left among a collection of 12,648 sherds of purposefully excluded material from Williamson’s surveys now housed at the National Museum in Tehran, and examined by the author in April 2005.124

In addition to the range of reliably identified later Sasanian and early Islamic classes, a lot of the coarsewares were of varieties poorly represented in the Williamson Collection and otherwise little known from elsewhere. From the existing body of knowledge, it is clear that these pre-date, rather than post-date, the c. 5th–8th century assemblage which has been described, although they were often found together on the same sites. The fact that a few sites did provide an uncontaminated cross-section of these earlier classes supports the anticipated conclusion: we are dealing with a chronologically distinct body of the material with a strong regional pattern of continuity in site occupation. The most important classes defined during the study which are representative of this earlier horizon include: HARGE, SLIP.HR, SLIP.HB, SMAG.B, CONG.G, CONG.RG and CONG.C. These classes are represented on 46 of the 73 sites recorded on the Dashtestan plain and were also observed widely distributed on the surface of the Sasanian settlement of Hazar Mardom on the Bushehr peninsula.

Regarding the dating of the earlier assemblage, some close parallels exist for the fairly generic folded rimmed bowls, particularly associated with CONG.G125 with material from Tal-i Malyan.126 Similarly, a jar with a sharply carinated neck, which sometimes develops into a projecting collar and associated with SMAG.B, finds a good comparison with one illustrated from the same assemblage.127 The pottery from Tal-i Malyan comes from a series of pits containing coarseware wasters associated with one of the few Sasanian pottery production sites so far excavated, and the kiln itself dated to around AD 300 on the basis of radiocarbon and thermoluminescence samples.128 Other good parallels exist for the carinated necked jar type with a number published from the excavations at Hājīābād.129 That site is dated by the excavator to between the 5th and 6th centuries, although this is based largely on pottery comparanda and is open to adjustment. Parallels with

Williamson would have encountered these same classes during his extensive survey and mapping of settlement on the Bushehr peninsula. Their near absence from the large body of survey finds exported to the UK suggests 123 

Priestman forthcoming a. Carter et al. 2006: fig. 13: 16–17. Alden 1978: figs 5: 13–15, 6: 22–24. 127  Carter et al. 2006: fig. 13: 1–2, 11; see Alden 1978: fig. 6: 6. 128  Alden 1978: 79. 129  Azarnoush 1994: fig. 164: e; 172: c–d, 175: r. 124  125  126 

See for example Carter 2008 for a recent treatment of this period.

166

Seth, M. N. Priestman: Bushehr, Dashtestan and Siraf Tal-i Malyan, in particular, seem perfectly consistent with the notion of continuity of occupation within the Bushehr area, with the pottery described representing earlier Sasanian occupation. More disconcertingly, good parallels for individual types within these same classes exist with securely attributed pottery of Achaemenid date from Qala’at al-Bahrain and Pasargadae.130 This suggests either that the pottery classification developed during the project is inherently flawed, or that styles only changed to a limited degree across the entire span of the Achaemenid, Hellenistic and early Sasanian periods. Amongst the earlier assemblage, it is important to note the absence of any clear parallels with pottery characteristic of the Hellenistic period. This suggests that either there was a major episode of abandonment across the region during this period, with most of the same sites being re-occupied during the Sasanian period or, as is perhaps more likely, there was a conservative local tradition which changed little between the Achaemenid and early Sasanian periods. The results of the pilot survey indicate that the Bushehr and Dashtestan areas offer significant potential for charting the long-term development of common coarsewares from the Achaemenid to early Islamic periods. For now, all that can be ascertained with confidence is a division in site dating between the later Sasanian period and a rather ill-defined episode which may span all or part of the Achaemenid to early Sasanian periods. Even with this crude handling, it is informative to consider the evidence for site occupation densities by period, based even on the results of the pilot survey (Figure 8).

Figure 9. Sasanian architectural fragment found amongst the palm gardens next to Tul-e Shahid

and late Sasanian pottery, although two appeared to be dominated by material of later date. On the ground between the mounds the pottery scatter continued mostly uninterrupted, but became patchier towards the south and east where the site continued for an undetermined distance as a series of low hummocks. No architectural remains were noted, and an exposed bulldozer section through BH8A confirmed that the mounds were natural and indicated that the cultural layer was thin. At Isanandeh a single archaeological mound was noted, measuring 40 x 30 x 3 m, with at least nine courses of mudbrick and a plaster floor exposed in an eroding section (BH18A). The structure was of monumental proportions and, from what could be discerned from the surface, appeared to belong to a solid foundation platform rather than standing walls. Pottery picked up from the surface was mostly later Sasanian with a smaller background of earlier pieces. To the north of the structure were a further series of mounds and a flat area scattered with predominantly late Islamic pottery, but also a significant background of Achaemenid to early Sasanian and later Sasanian material (BH18B–C).

Although no attempt was made to carry out systematic area coverage, and the material collected represents ‘grab samples’ from only those sites most easily identified, the scale of the Achaemenid to Sasanian settlement across all areas of the survey area is suggested both by the raw site count data and by the constant background of Achaemenid to Sasanian coarse and slipped wares encountered on most of the sites visited, even where an earlier or later period represented the main horizon of occupation or use. Amongst those sites recorded, there are a few larger ones of predominantly Achaemenid to Sasanian date which deserve particular mention: Shif (BH5A), described above, Khosh Ab (BH8A-C), Isanandeh (BH18A-C), Tul-e Shahid (BH22), and an urban settlement of approximately 300 hectares north-east of Deh Qa’ed (BH28-45).

Tul-e Shahid (BH22) is situated in the middle of thick palm gardens and consists of a prominent mound of 80 x 35 x 3 m. Most of the centre of the mound had already been removed during excavation by the local archaeological service, but the results remain unpublished and access to the archived report could not be arranged at the time. Excavation of the central portion of the mound had exposed the remains of substantial mortared fired brick and stone architecture surviving for several courses in the section. Very little pottery was left on the surface but that which was recovered was mostly of Achaemenid to Sasanian date. This dating may be supported by a large fluted column

The site of Kosh Ab, translated as ‘sweet water’ or ‘good water’, consists of three low mounds measuring c. 100 x 5, 150 x 3 and 50 x 1.5 m respectively, each covered by a moderate to dense pottery scatter (BH8A–C). All three mounds produced Achaemenid to early Sasanian 130 

See Carter et al. 2006: 94–96 for discussion.

167

Sasanian Archaeology

Figure 10. Map showing the location of areas of significant mounding that delimit the area of the Achaemenid/Sasanian town at Deh Qa’ed (after Carter et al. 2006: fig. 14). Note the fortified enclosure of Tul-e Khandagh at BH29 and the bend in the DalakRud at BH45

Figure 11. View looking north-west along the top of the massive mudbrick walls of Tul-e Khandagh (photograph: author)

168

Seth, M. N. Priestman: Bushehr, Dashtestan and Siraf hinterland, the economic configuration of the region needs to be reconsidered as instead it appears that the Dashtestan plain was not a hinterland in the sense that say, the Diyala region has been interpreted in relation to Baghdad.134 Instead this was the land between two population centres. Bushehr may have acted primarily as a port and entrepôt, while the inland town at Deh Qa’ed lay within the heart of an agricultural landscape surrounded by other late pre-Islamic sites, some of which also attained considerable proportions. Although it is probably not wise to push historical parallels too far, it may be that the basic topography of the region has to a certain extent dictated its economic configuration. If so, then the town at Deh Qa’ed may have played a similar role in relation to Rev Ardashir and Istakhr, as Borazjan did for Bushehr and Shiraz during the 17th– 19th centuries, when Bushehr once more became a significant centre of maritime trade on the northern shore of the Persian Gulf. Interestingly the role that Borazjan played in later periods included a range of functions from a market centre, where locally produced commodities were pooled for export through Bushehr and where imports were redistributed; a staging post on the caravan route between Bushehr and Shiraz; and a centre of local political authority which was crucial to the maintenance of security in the Dashtestan district and to the mediation of power between the competing interests of the governing elites in Shiraz and Bushehr.135

base or capital noted amongst the palm gardens to the edge of the mound (Figure 9). Similar examples are known from other sites within the region and are regarded as being of Sasanian style,131 but the author is unaware of any published parallels to support this identification. Of all the Achaemenid to Sasanian sites noted during the 2004 survey season, the most significant discovery was made at the end of the linear feature previously identified as the Angali canal.132 Close to the bend in the Dalaki-Rud, where Whitcomb believed the canal was taken off, and just north-east of the village of Deh Qa’ed, a substantial dispersed city-site was identified covering a total area of over 300 hectares (Figure 10). This site has been under investigation by Professor Ali Akbar Sarfaraz and Mr Ismael Yaghma’i over the course of three decades, although it does not appear to be widely known.133 During the 2004 Bushehr Hinterland Survey, a rapid inspection was made of the surface remains across the site in order to attempt to delimit its extent and to establish its main period of occupation. The collections consist entirely of material dated to within the range of the Achaemenid and Sasanian periods, although the largest mounds roughly occupying the centre of the site (BH32A/B) produced a greater concentration of finds related to the later Sasanian period, and roughly dating to the 5th–7th centuries. For the majority of the site, including the massive mudbrick fortification of Tul-e Khandagh, with sides measuring c. 200 x 140 m and 15–20 m high (Figure 11), finds with Achaemenid to early Sasanian parallels appeared to predominate. For the latest occupation, a range of coarsewares where recovered including SMAG and LISV, which could potentially date as late as the early 9th century. However, the complete absence of ‘Samarra horizon’ glazed wares or appliqué decorated Alkaline Glazed Ware suggest that the city was abandoned by the mid8th century at the latest, and possibly during the 7th century.

Sasanian economic decline As has been discussed before, the prosperity of Bushehr is thought to have ended by the mid-8th century and possibly slightly earlier. The study recently undertaken of the Williamson Collection confirms this and the number of settlements drop by over half from the 6th–9th to the 9th–11th centuries (Figure 8). If only those sites are counted where more than 10% of the assemblage relates to the periods in question, thereby excluding sites with only minor background components, then the drop in site numbers is even more dramatic.

Configuration of port/hinterland relations The presence of a substantial late pre-Islamic town at the back of the Dashtestan plain, close to Borazjan and about 40 km inland from Bushehr with the possibility of an ancient route linking the two, provides a very different picture of the Bushehr hinterland than assumed previously. Instead of Bushehr being seen solely as a node of maritime contact which served the economic demands of the heartland cities of Sasanian Fars, and supported in turn by a dependent rural

These results raise the question of the likely cause of this dramatic collapse, especially after what appears to have been a long and sustained period of growth starting within the Achaemenid period or earlier, and continuing all the way through to the later Sasanian period. Interestingly, the pottery suggests that at the same time that the major settlement centre of Bushehr fell into decline, the inland city north-east of Deh Qa’ed was also abandoned, as were all other important later pre-Islamic sites within the area (see Figure 8). If Bushehr was intimately linked to its hinterland, as proposed previously, then it appears that it was not just

Zareh, pers. comm. 2004. Whitcomb 1987. The author is grateful to Dr Sarfaraz and Mr Yaghma’i for the information that they kindly shared about the site, and for allowing the project to conduct a rapid supplementary study of the site prior to the publication of their own detailed investigations. 131  132  133 

134  135 

169

Adams 1965. Floor 2004: 180, 190–91.

Sasanian Archaeology the settlement on the peninsula that collapsed but also the entire regional settlement infrastructure.

Bushehr and its hinterland (Figure 7). They include the following:

Dating the event

Type 64: Light olive-green alkaline-glazed ware basins with a distinctive bifurcating rim dated to between the 5th–8th centuries as discussed above,141 or perhaps on the negative evidence from Sir Bani Yas to the 5th– early 7th centuries.142

In order to find an adequate explanation for these events, it is necessary to examine the dating evidence more closely. In the analysis of the Williamson Collection and Williamson’s own observations of the dramatic decline in aggregate settlement area on the Bushehr peninsula from 450 to 15 hectares, this decline is placed in about the 9th century.136 In both cases this has been determined by comparing the number of sites with typical Sasanian type-fossils and the wellknown repertoire of ‘Samarra horizon’ glazed wares. If one wishes to narrow down the timing further, it is necessary to look again at the transition between the late Sasanian period and the well-defined technical and stylistic break marked by the introduction of the ‘Samarra horizon’ in the early part of the 9th century.

• Type 72: A small alkaline-glazed bowl with a high gentle carination and simple rounded rim.143 At Kush this type only occurs in Period III dated to the late 8th–early 9th centuries.144 The type is also well represented at Sir Bani Yas in an assemblage lacking appliqué decorated Alkaline Glazed Ware (JR5) suggesting that it should be placed earlier than the late 8th century.145 A factor which appears to confuse the use of this type is that one finds a very similar form occurring in Hellenistic/Parthian contexts, although minor differences between the two are discernable.146 • JR5: Appliqué decorated Alkaline-Glazed Ware jars with a short collar neck and several ‘butterfly’ handles attached around the shoulder.147 At Kush the type first appears in the sequence in Phase E-03 at the end of Period II, dated to between the 7th–8th centuries.148 This suggests that while the type has been widely referred to as ‘SasanianIslamic’,149 they are actually an exclusively postSasanian phenomenon.150 It is also important to note at Kush the introduction of the type two phases earlier than the earliest examples of the ‘Samarra horizon’. While JR5 and ‘Samarra horizon’ glazed wares are often found together, the Kush evidence contributes towards the now quite widely recognised understanding that this type first begins to circulate before the beginning of the 9th century. The type then continues in circulation until the end of the Samarra-horizon period. At Shanga in east Africa, JR5 tails off dramatically after Phase 8 in the Trench 6-10, suggesting a decline in its circulation during the 10th century.151 This conclusion appears to be supported at Sharma in Yemen, where only a few sherds occur on the surface or in the

So far post-Sasanian/pre-‘Samarra horizon’ pottery has not been widely recognised, despite the availability of several useful archaeological assemblages covering this period, such as those from Tulul al-Ukhaidhir, Jazirat al-Hulayla (Area D) and Sir Bani Yas.137 The continuing confusion over certain aspects of the dating at Siraf, such as the reported existence of Sasanian levels below the Great Mosque foundation platform or the dating of the platform itself, have contributed significantly towards the uncertainty surrounding the period.138 Recent excavations at Kush, combined with regional survey and a test trench at Khatt, have begun to redress the issue.139 It seems that the main defining attributes of much of the pottery of the 7th–8th centuries was not a radical break from late Sasanian traditions of glazed pottery manufacture, but rather a gradual shift of stylistic attributes such as surface treatment and particularly vessel form. Such changes often only become discernable through quantified analysis, for it is this which allows one to see the gradual tailing off of typical late Sasanian types and their replacement over time by recognisably early Islamic types.140 For the purposes of surface survey, where seriation is not possible, the main tool available is to consider the presence or absence of specific combinations of classes or types. Within the present discussion a small number of reasonably well-established chronological markers, mostly mentioned above, can be used to illustrate the key relationships relating to the abandonment of

Kennet 2004: 29–31, fig. 5, table 16, ‘Type 64’; Priestman 2005a: 234–38, pl. 91, ‘ALK: 29–30’; Williamson 1971/72: 10–11. 142  Carter 2008. 143  Kennet 2004: 30, fig. 5, ‘Type 72’; Priestman 2005a: 234–38, ‘ALK: 12’. 144  Kennet 2004: table 16. 145  Carter 2008: 89, fig. 10: 1–6. 146  Kennet 2007: 97–98. 147  Priestman 2005a: 234–38, ‘ALK: 32’. 148  Kennet 2004: 29–30, ‘TURQ.5’, table 3. 149  e.g., Horton 1996: 274–77. 150  Kennet 2004: 29–30, table 3, ‘TURQ.5’. 151  Horton 1996: 277, fig. 197. 141 

Priestman 2005a: 92, fig. 15; Williamson 1971/72: 37. Finster and Schmidt 1976; Sasaki and Sasaki 1996: 64–81, figs 43– 54; Carter 2008. 138  See Whitehouse 1968: 15; 1979: 52; Tampoe 1989; Kennet 2004: 31–32, tables 18–19. 139  Kennet 1994; 1997; 1998; 2001; 2003; 2004. 140  See for example the transition between CLINKY to SMAG in Periods I–II: Kennet 2004: table 42. 136  137 

170

Seth, M. N. Priestman: Bushehr, Dashtestan and Siraf lowest levels which can be dated to the late 10th century.152 • SMAG: Hard, lime-spalled, well-sintered, ‘clinky’, red or grey bowls, jars and jugs of various sizes. This class represents the most common coarse ware of the 5th–early 9th centuries at Kush.153 The class peaks in circulation in Phase W-04 dated by a single radiocarbon sample to after the mid-7th to early 8th century. It then appears to go out of circulation in the latter half of Period III, before the introduction of the ‘Samarra horizon’ in the early 9th century.154 Evidence from the Williamson, Bushehr Hinterland and Hormuzgan surveys suggest that this class is extensively distributed on sites along the northern shore of the Persian Gulf.155

Tawwaj (BH12), several km to the north of Deh Qa’ed. Inspection of surface remains at Tawwaj, which cover a mounded area of at least 900 x 600 m, indicate that when activity was resumed, it was across the whole site simultaneously. From the close succession in the pottery sequence between the abandonment of the city at Deh Qa’ed and the beginning of the sequence at Tawwaj, and the proximity of the two sites, it seems possible that their respective fall and rise was connected and that a substantial component of the population transferred from one to the other. The Islamic conquest The traditional explanation for the decline of Sasanian Rishahr, following al-Baladhuri’s account, is that the city was sacked during the Arab conquest.158 Both the dating and the dramatic nature of the event seem to fit well with the accumulating body of archaeological evidence from the Bushehr peninsula and its hinterland.159

The recorded observation of these four common period markers in the Bushehr area enable us to more closely date the period of Bushehr’s economic decline. From three separate visits to the extensive remains at Halileh and Hazar Mardom in 2004 and 2005, and from detailed examination of the eroding shoreline section through the mass of remains at Rishahr in 2005,156 what is striking is the profusion of SMAG, Type 64 and other late Sasanian markers such as CLINKY, TORP.S and LISV. As already discussed, some of these classes continue in circulation into the 9th century, but what is notable is the complete absence of either ‘Samarra horizon’ glazed wares or of JR5 and Type 72, and exactly the same pattern has also been observed at Shif (the proposed Sasanian ferry point to Bushehr), Deh Qa’ed and all other substantial late pre-Islamic sites in the Bushehr hinterland. The main exception to this pattern is Tawwaj. Here there appears to be an episode of earlier occupation represented by a number of Elamite and Achaemenid markers. This is followed by a break in occupation during the Sasanian period, and then a mass of JR5 together with late SMAG similar to that from Siraf,157 and a very small quantity of ‘Samarra horizon’ glazed wares. What we therefore appear to have is the complete abandonment of Rev Ardashir, together with the associated centre of Deh Qa’ed and other elements of the local infrastructure, before the mid-8th century and possibly somewhat earlier, i.e. during the 7th or early 8th century. This was then closely followed by a major reoccupation of the site of

As a background to the historically attested events surrounding the first cross-Gulf conquest, it is pertinent to reconsider the results of recent analysis of longterm settlement development during the Sasanian and Islamic periods based on the evidence from the Williamson Collection.160 One of the most striking features is evidence for a significant contrast between the upper and lower Persian Gulf, with relatively low numbers of sites recorded for the Sasanian period on the Minab plain, followed by a progressive increase up to a peak during the much later period of Hormuzi dominance in the 14th–17th centuries.161 As we have already seen, Bushehr and neighbouring Dashtestan appear to follow a reverse pattern with site numbers starting at a high point during the Sasanian period and then falling into decline thereafter (Figure 11). Within this context it seems noteworthy that the first cross-Gulf campaign occurred when a force, acting independently from the caliphate under the command of the governor of al-Bahrayn, led a raid against Qeshm Island and from there against the mainland.162 The inference which might be drawn from this is that the initial foray across the Persian Gulf was one which was essentially sent through the ‘backdoor’ into an area of relatively low population density. Again, two years later, in a more concerted effort, this time with official sanction from Medina, an Arab force crossed the Persian Gulf from Julfar and seized Qeshm Island.163

Rougeulle 2005: 226–27. Kennet 2004: 62–63, table 3, ‘SMAG’. 154  Kennet 2004: table 3. 155  Priestman 2005a; Carter et al. 2006; Kennet et al. 2006. 156  I wish to express my sincere gratitude to the organisers of the Internal Congress of Siraf Port providing me the opportunity to examine the ruins at Rishahr Halileh and Hazar Mardom before the conference began. I would also like to thank Mark Horton and Fred Aldsworth for the lively and informative thoughts shared while examining the impressive shoreline sections. 157  Personal observation based on the author’s study of the Siraf Collection in the British Museum. 152  153 

Before continuing with the account of the events which followed, a further detail meriting consideration is the Whitehouse and Williamson 1973: 42; Whitcomb 1987: 332–33. Williamson 1971/72: 37; Priestman 2005a; Carter et al. 2006. Priestman 2005a. 161  Priestman 2005a: 91–92, fig. 16. 162  Hinds 1984: 41. 163  Hinds 1984: 41–42. 158  159  160 

171

Sasanian Archaeology timing of these attacks. The deeper context to Arab expansion is provided by the combination of a Sasanian administration weakened by factionalism, institutional fatigue and the rising influence of Byzantium, set against a degree of opportunistic success164 and by Arab forces unified by the message of Islam and motivated by their leaders’ call for jihad with all the spiritual and economic rewards which this could offer.165 More specifically, the campaign of the governor of al-Bahrayn was motivated, it is reported, by intelligence that the Sasanian military was in the process of amassing a naval force in the Persian Gulf to counter any further advance.166 This detail carries significant implications. Of course the inroads of Arab armies into southern Mesopotamia at this time must have placed a strain on Sasanian military resources. However, the Persian Gulf coast was the empire’s only major maritime frontier, and the fact that it was not already heavily militarised either implies that Sasanian maritime activity in the Persian Gulf hardly existed previously, or that Sasanian trade required little protection, perhaps due to the total vacuum created by overarching Sasanian dominance with the region. The two interpretations are in complete contradiction, and we are not in a clear position to determine which is the most likely. If we do assume that Sasanian trade was developed, even partially, and that it was not dependent on a significant naval presence, the situation may have been one involving the regular movement in bulk staple commodities with only relatively low-return potential. Such a picture would fit with the image of state-regulated trade167 and with the archaeological discoveries at sites such as Rishahr, Halileh and Hazar Mardom which have produced few high-status exotic imports, but did yield large quantities of standardised transport amphorae and a limited repertoire of glazed vessels presumably originating in southern Mesopotamia. The nature of this assemblage suggests regular formalised exchange on a large scale between one of the major centres of food production within the Sasanian empire and the main node of redistribution within the Persian Gulf at Bushehr. From Bushehr, goods were then presumably moved in a number of directions, including – as Tomber’s evidence suggests – India.168 As will become clear below, this pattern appears to be one which was radically transformed during the early Islamic period.

in the region of 30,000–40,000, somewhere between Tawwaj and Rishahr.169 Whatever the actual scale of the force, the result of the battle was a decisive victory for the invading army, one that al-Baladhuri compared in scale and significance to Qadissiya which was one of the battles that marked a turning point in the Arab conquest of southern Iraq.170 Again there are details in the account that suggest a great deal about the pre-existing system of Sasanian state influence within the region. Recalling the organisational structure previously described of Bushehr, the town close to De Qa’ed and various centres in inland Fars, it is surely significat that the Arab army chose not to attack or blockade the port but instead sever its line of provision and communication. Without Tawwaj the whole regional infrastructure was crippled and the Sasanian army no longer had any option but to turn out from Rishahr and engage the enemy in what may already have been apparent, a futile cause. Given the complete absence of Sasanian evidence from the site identified as Tawwaj today, it seems doubtful that this is the same site as that initially captured by Arab forces. An alternative scenario could be proposed that fits well with the archaeological evidence, and which goes some way towards explaining apparent ambiguities in the historical records concerning whether Tawwaj was captured and developed as a misr (a garrison city) before or after the main battle against a Sasanian army close to Rishahr.171 Although the argument may be difficult to prove, it seems likely that name of Tawwaj was transferred from the Sasanian city close to Deh Qa’ed to the site of the early Islamic city. If so, the initial Arab conquest would have been of the first site and, following the defeat of a Sasanian army at Rishahr, the city re-founded about 10 km to the north to a position next to a second tributary of the Rud-e Hilleh, the Rudkhaneh-ye Shapur, where a city was founded with planned streets, several mosques, a large military garrison and a population which is said to have included a significant number of settled Arabs.172 For a short while, Tawwaj was used as a base from which to launch further raids on the remaining Sasanian strongholds in Fars. Ultimately, however, having broken the economic artery leading between Bushehr and inland Fars, the fortunes of Dashtestan must have waned and, as we have seen, rural settlement collapsed and Tawwaj itself was largely abandoned by the early 9th century. At this point the economic configuration of the Persian Gulf region shifted decisively away from Dashtestan towards Siraf.

Returning to the events associated with the Arab conquest, during the second cross-Gulf campaign, after successfully defeating a Sasanian force sent from Hormuz, the Arab army moved by land to capture Tawwaj. From Tawwaj the Arab army went on to engage a large Sasanian army, with troop numbers reportedly

Hinds 1984: 41–42. Troop numbers are likely to be highly exaggerated and seem particularly suspicious for a region such as Bushehr which does not appear to have been very significant militarily. 170  Zarrīnkūb 1975: 21. 171  Hinds 1984: 43–44. 172  Hinds 1984: 44. 169 

Abu Ezzah 1979: 58. Zarrīnkūb 1975: 17. 166  Hinds 1984: 42. 167  Piacentini 1985: 60; 1992: 137. 168  Tomber 2007. 164  165 

172

Seth, M. N. Priestman: Bushehr, Dashtestan and Siraf least 70 m across and 32 m wide, contained within a 1m thick curtain wall. The compound contains an elaborate network of over 50 rooms, traversed by narrow earthen streets.179 The overall plan is highly irregular. The landward side of the compound is completed by the exterior wall of a substantial fort entered through a monumental gate protected by rounded towers. This leads into a gatehouse and from there into a cobbled courtyard with rooms ranged around the external walls.180 The inner enclosure clearly belongs to a highstatus fortification or fortified palace.181

The rise of Siraf In comparison to Bushehr, the historical and topographical configuration of Siraf could not be more different. Bushehr is situated in an open and relatively indefensible island landscape which benefited from an extensive cultivable hinterland on the mainland. The port appears to have developed gradually over several hundred years and to have been closely integrated into a state-organised administrative structure which encompassed the surrounding territory. Rev Ardashir may have been developed as a commercial arm of the state,173 and trade activity may have been stateregulated. In contrast, Siraf is situated in a confined landscape on a narrow sloping strip no more than 0.5–1 km wide with the first of several precipitous chains of hills rising immediately behind.174 Access to the site was only possible through mountainous terrain along narrow tracks plied by pack animals. Irrigation and arable land were limited and while every available plot of land was brought under cultivation during the period of the city’s florescence,175 most of the basic provisions needed to sustain the city’s inhabitants had to be brought by sea or transported along the narrow pack route for at least 19 km from the inland valleys of Jam and Galehdar.176

Regarding the date of these structure, a number of architectural features have been pointed to as indicators. The worn condition of the walls of both fort and outer enclosure led the excavators to believe that they were already in ruins when the mosque was built.182 Such an assessment is a difficult one to make, and even if correct, we need not assume that it would take long for a building to fall into such a condition of disrepair. Other highlighted features include four separate phases of construction within the residential compound, and the fact that the buildings of phase 4 come almost up to the walls of the fort suggests that it no longer performed an effective defensive function and the latter therefore must have been built at an earlier date.183

The succession of Siraf over Bushehr raises a number of issues and the linkage of events depends to a degree on the interpretation of the dating. The argument has already been presented for the collapse of Rev Ardashir and the Sasanian infrastructure associated with the port around the time of the Arab conquest of Fars during the earlier 7th century. At Siraf, the earliest dated east Asian imports have been attributed to the 7th century, as discussed above, although the first significant influx appears to be from the later 8th century. This provides a strong indication of the site’s emerging importance as a centre of maritime trade. However, concealed directly below the foundation platform of the Great Mosque at Site B, and projecting out over a larger area, are a series of structures of potentially even earlier date.177 These were first recognised during deep excavation in the third season, and were specifically targeted after clearing out the mass of fill from the Great Mosque foundation platform.178 The structures uncovered consist of an outer residential compound measuring at

These elements are relative and for an abolute dating we must turn to the finds as there are no radiocarbon dates. Phase 4 within the residential compound appears to belong to the later 8th century as lead coins sealed below the floors of a number of buildings resemble those of the 8th and early 9th centuries from the Great Mosque foundation platform.184 Similarly, there are strong connections in terms of the pottery assemblage as the pottery recovered from phase 4 of the residential compound is indistinguishable from that derived from the foundation platform of the Great Mosque. The most distinctive elements of this assemblage include: appliqué decorated Alkaline Glazed Ware (JR5), moulded eggshell ware (EGG.PI), and some imports from east Asia, including Dusun jars (DUSUN) and painted Changsha bowls (CHANG). Most of these classes survive into the 9th century, but the complete absence of ‘Samarra horizon’ pottery, places this assemblage in the immediately preceding period of the later 8th or very early 9th centuries. Moving down through the sequence, phase 3 within the residential compound produced very limited dating evidence, apart from a single gold Byzantine coin minted between 651 and 659:185 an

Piacentini 1985. Whitehouse 1968: 3, fig. 2. 175  Wilkinson 1974: 129. 176  Whitcomb 2005: 104–105. 177  Whitehouse 1971a: 262. I first examined the issue of the Sasanian dating of Siraf in a paper entitled ‘The Sasanian origin of Siraf?’ presented at ‘The Persian Gulf in Prehistory and History’ conference held at the University of Durham on 1st–2nd July 2008. I wish to thank the organisers of the conference: the Iranian Cultural Heritage, Handicrafts and Tourism Organisation and the British Institute of Persian Studies for inviting me to participate. 178  Whitehouse 1970: 8. 173  174 

Whitehouse 1971b: 4–5, fig. 1, pl. I: b; 1972: 70, fig. 3, pl. III: a–c. Whitehouse 1971b: 5, pl. I: a; 1972: 70, fig. 3, pl. II: c–d. 181  Whitehouse 1971a: 262. 182  Whitehouse 1971a: 262 183  Whitehouse 1972: 71. 184  Whitehouse 1972: 70. 185  Whitehouse 1972: 70. 179  180 

173

Sasanian Archaeology intrinsically valuable coin would not have depreciated significantly in value and may have remained in circulation for some time after it was minted. In any case, the single coin suggests a date within the Islamic period. Phases 1 and 2 produced coins which have been attributed to the Sasanian period, although none can be assigned to a specific ruler;186 the same phases contained a number of sherds of Indian Red Polished Ware (IRPW), and a single sherd of Fine Orange Painted Ware (FOPW). Together the excavators regarded these as evidence of a very early Sasanian foundation for the residential compound ‘within the first few centuries AD’.187 Initially it was only the residential compound that produced any dating evidence but excavations during the final season revealed that there were two phases of construction associated with the fort.188 Sealed beneath the later of these were two sherds of a distinctive class of black glazed stoneware (STO.BG1). At present there is limited independent dating evidence for this class: the same kind of ceramic has been recovered from levels at Sohar associated with alkaline-glazed pottery and other classes which may be attributed to the 7th–8th centuries.189 Given the fact that a range of other east Asian classes first appear in the Persian Gulf no earlier than the 7th century, and mostly towards the late 8th century, it seems inconceivable that this particular class stands out alone. The implication in terms of the dating of the fort is that the building was still in use and undergoing reconstruction up to the period shortly before the construction of the Great Mosque. Far from the fort being redundant, it was being maintained at the same time as the redevelopment of the residential compound in phase 4.

sometime after 840/41.192 Although similar coin hoards are not particularly common within the Persian Gulf, greater numbers, particularly from western Russia contain late Sasanian issues as a significant proportion up to around 812, after which they rapidly decline.193 The second feature to note regarding the early coinage from Siraf, is that just four of those attributed to the Sasanian period can be assigned to a specific ruler and those which have been identified tend to fall within the mid-Sasanian period, around the 4th–5th centuries. The majority of the coins, including those which have been roughly dated, have almost no distinguishing marks or are illegible. The identification of these coins to the Sasanian period is based largely on their size, thickness and fabric. Nine of the 16 Sasanian coins from Site B fall within Lowick’s entirely illegible category: the most characteristic feature of all is that they are thick and very small in size, therefore different from the typical thin flans of late Sasanian and early Islamic coinage.194 Lowick’s identification seems convincing as this same distinctive coin form is shared by a few legible examples attributed to the reign of Yazdgird I and Yazdgird II.195 Saying this, it should be admitted that we have little information on low denomination coinage of this period and the possibility that small, thick, locally minted coins circulated in other periods probably cannot be completely excluded. Where the argument for a Sasanian foundation of Siraf is weakest is in the identification of the pottery and this is the field where significant progress has been made. If Siraf had been occupied during the Sasanian period, one would expect there to be a significant body of Sasanian pottery recovered from the excavations but this does not appear to be the case. A recent study of finds from Siraf in the British Museum contributes towards our understanding of this issue as this collection contains just over 10,000 sherds.196 Although this is only a small fraction of the total number of pieces originally excavated it is large enough to gain a reasonable overview of the main varieties represented within the assemblage.

The strongest element of Whitehouse’s argument for a Sasanian foundation date for phases 1 and 2 comes from Lowick’s identification of a group of Sasanian coins.190 Altogether there are 36 coins from Siraf which have been attributed to the Sasanian or ArabSasanian periods. Four of these were found on the surface. Around half of the stratified coins come from scattered locations distributed across the various excavations; the remainder come from Site B. The first point to note about the Sasanian coins is that just two are silver, and one of these had been perforated for use as a pendant and was found in association with a burial.191 For the silver coinage we know that late Sasanian issues continued in circulation up until at least the mid-9th century, as is indicated for example by the Sināw hoard from central Oman which contains a small proportion of Sasanian coins and was deposited

Among the collection, there are about a hundred sherds belonging to the early levels at Site B. These are clearly distinct in character and chronology to the rest of the pottery represented in the collection. Diagnostic components amongst this group are Indian Red Polished Ware (IRPW), Alkaline Glazed Ware (ALK.T) Type 72 (and other contemporary types not defined during the Kush study), sandy Torpedo Jars (TORP.S) similar to those from Bushehr and Sir Bani Yas, and Small Grey Vessels (SMAG). The 7th–8th century dating of Type 72

Whitehouse 1972: 70. Whitehouse 2009: 8. 188  Whitehouse 1974: 7. 189  Kervran 2004: figs 12: 24, 17: 7, pl. 22; see Kennet 2007: 97 for the dating. 190  Lowick 1985: 2, cats 5–40. 191  Lowick 1985: 2, cat. 5. 186  187 

Lowick 1983: 199. Lowick 1983: 202. 194  V. Sarkhosh Curtis, pers. comm. 2008. 195  Lowick 1985: cats 6–8. 196  Priestman forthcoming b. 192  193 

174

Seth, M. N. Priestman: Bushehr, Dashtestan and Siraf variety represented amongst the finds from the Siraf excavations in the British Museum by the class FINLIM. This class has important characteristics in common with the later classes of LISV (HARLIM and REBROS), which dominate the Siraf assemblage and which are common in east Africa. This includes a fabric containing abundant angular red or black mudstone platelets and fragments of lime which spall on the surface. The main difference with FINLIM is that it has a finer grained structure and vessel forms dominated by wide mouthed vats with a different range of more rounded rim profiles. FINLIM represents the dominant class of LISV on sites in the Bushehr area and supporting Whitehouse’s identification of the class as being predominantly earlier. At the same time, the results from Kush suggest that LISV became increasingly common as one moves into the 7th–9th centuries,210 and similar vessel forms to those represented at Siraf are also represented at Sir Bani Yas and Jazirat al-Hulayla.211 In other words, the ‘Massive Ware’ from the citadel need not necessarily be Sasanian and could indeed form part of the same preSamarran early Islamic horizon identified elsewhere.

and the extension of SMAG into this period has already been discussed. Another important category in this context is IRPW. Traditionally this has been regarded as a reliable type-fossil for the Sasanian period or earlier,197 but the excavations at Kush alter this picture. Although IRPW seems to have come into circulation in India from the 1st century, the main concentration of the class within the stratified sequence at Kush is between phases E-03 and E-04, dated between the 7th–8th centuries.198 A similar dating may apply to IRPW found elsewhere within the Persian Gulf as this class has been found in similar dated contexts at Sohar199 and Jazirat al-Hulayla.200 The occurrence of TORP.S is of less clear chronological significance. What can be stated with confidence is that the relevant sherds from the early levels of Site B are distinct in a number of respects to the later variety of Torpedo Jar (TORP.RG) which came to dominate the assemblage from Siraf and which also appears to be characteristic of Torpedo Jar assemblages from other sites dated from the latter half of the 8th century onwards, including Manda201 and Sohar Level V.202 The main characteristics of this later group are thicker walls, a less sandy pale fabric, and a pronounced lip on the interior of the rim. The earlier orange sandy TORP.S therefore provides a further indication of a pre late-8th century date for these early levels, but could itself be dated anywhere back to around the 2nd/3rd century, when one again seems to encounter a change in Torpedo Jar fabric and vessel morphology.203

Across the site, sporadic finds of Sasanian date have been recovered in six other trenches. The coins have already been mentioned, and these are probably the mostly securely attributable artefact of the Sasanian period, particularly in the form of low denomination issues. Added to this are a few ‘Sasanian style’ stamp seals,212 a further cache of Sasanian seals reportedly from ‘near Siraf ’ discovered by local villagers,213 and a pair of gold earrings with pearl settings dated stylistically to the Sasanian period.214 Within the cemetery there are large numbers of standard rock-cut graves, as well as larger tombs cut into cliff faces. One of these was inspected by Dr Lumsdaine and Mr L. Osborne who visited Siraf in 1835 on board the H.C. Corvette Falkland.215 They noted the presence of a number of different individual skeletons adding support to the conclusion that the larger tombs represent the site of collective burials and a rite usually associated with Zoroastrian burial customs. This in itself is of interest for understanding

The architectural sequence revealed below the Great Mosque at Site B represents the most important component of the evidence previously adduced for a substantial pre-Islamic occupation at Siraf. After Site B, the next most important area is the ‘citadel’.204 This is situated on a high fairly inaccessible promontory on a ridge overlooking the site. One of the most frequent and diagnostic surface finds from this area are fragments of ‘Massive Ware’ jars,205 which are distinct from those manufactured at Siraf but found in large numbers at Manda206 and a number of other sites in east Africa, including Shanga207 and Unguja Ukuu.208 Although the pottery described from the citadel at Siraf has not yet been illustrated, it is clear from the description that the reference is to LISV,209 and to a distinct

firing, surface treatment, etc. 210  Kennet 2004: 58, table 3. 211  Carter 2008: 86–87, figs 14: 2–5, 15; Sasaki and Sasaki 1996: fig. 50: JHU95-1; 2000: figs. 15: JHU98-5, 18: JHU98-58. 212  Whitehouse 1972: 70, pls XIId–e. 213  Whitehouse 1971a: 263–64. 214  Whitehouse 1972: 70, pl. XIIc. Recent examination of the excavated stamp seals reconfirms that they are of Sasanian type. However there is less known about pictographic seals and they fall within a package of Zoroastrian iconography which could conceivably last into the Islamic period (Simpson, pers. comm. 2009). One such ‘Sasanian’ seal has been recovered from the 7th/8th century site of Area D at Jazirat al-Hulayla (Sasaki 1996: 179, 184, fig. 13, pl. 37). The reported discovery of Sasanian stamp seals from near Siraf is not convincing particularly in relation to the lack of a securely attributed provenance. Similar examples of the gold earrings are depicted on Sasanian rock reliefs but belong to a type which may have survived into the Islamic period (Simpson, pers. comm. 2009). 215  Kempthorne 1856: 139.

e.g., Williamson 1972: 100, ‘Type 2’; Whitehouse 2009: 8. Kennet 2004: 65–66, table 3. 199  Kervran 1996: 38, figs 3: 2–6, 4: 5–6, 11–13. 200  Sasaki and Sasaki 1996: fig. 48: JHU95-144. 201  Chittick 1984: fig. 43: a–c 202  Kervran 2004: figs 22bis: 3, 24: 2. 203  Benoist, Mouton and Schiettecatte 2003: 69. 204  Whitehouse 2009: 8–9. 205  Whitehouse 2009: 9. 206  Chittick 1984: 83–84, figs 40a–e, 41a, e–g, 42a, pl. 38a–c 207  Horton 1996: 296–97, figs 220.a, d. 208  Juma 2004: 6.17.11. 209  LISV is categorised in Kennet’s classification as a class (Kennet 2004: 58). Actually it is a broad group of vessel types which are spread over a number of individual classes defined on the basis of fabric, 197  198 

175

Sasanian Archaeology the ethnic and religious composition of the community living at Siraf, although as Whitehouse observed, there was a lengthy period where Zoroastrians and Islamic converts co-existed,216 and the tombs do not therefore constitute reliable evidence of Sasanian occupation. In terms of the pottery, there are a few important markers for an earlier phase of activity. Most notable of these are a rim sherd of FOPW from Site B,217 and a second body sherd noted among the Siraf Collection in the British Museum,218 also from Site B. Based on the evidence from Kush, it is clear that FOPW continued in circulation later than the date proposed on the basis of the finds from ed-Dur,219 but still does not appear to extend beyond the 5th/6th centuries.220 Other early pieces of pottery represented in the British Museum include two sherds of ALK.YG Type 64, both from Site K.221 As discussed previously, the dating of this type is likely to extend into the Sasanian period, but could equally continue as late as the mid-7th century.

be explained either as a ‘residual baggage’ of a newly founded occupation established very close to the beginning of the Islamic era, or as traces of a smallscale earlier settlement. What seems far less likely is that these traces total up to an ‘extensive’ settlement, or one that was in any way functionally equivalent to what the city developed as from around the late 8th century. The functional dichotomy which Piacentini outlines between an earlier military site and a late great port could still stand.223 What seems less likely is that Siraf ’s ‘military phase’ was one occurring within the Sasanian era. The changing maritime order of the middle Persian Gulf Although we have considered the question of a possible Sasanian foundation for Siraf in some detail, it is not fundamental to the argument presented here. It is clear now that the site underwent a radical transformation of function after the decline of Bushehr, and in this sense it may be seen to have assumed the role and status of the latter port. However, it is also clear that if a later dating for the origins of Siraf is accepted, this has a number of important implications. Firstly, the Siraf assemblage as a whole becomes a much more useful one as we can exclude significant contamination through residuality of earlier sherds. Secondly, if the first significant activity at Siraf occurred in the form of a palace or fort during the 7th/8th centuries, this fits well with the timing of the changes wrought by the Arab conquest and the dismantling of Bushehr and its broader infrastructure as discussed above. What remains to be considered is whether the role adopted by Siraf was a new one, or whether it simply inherited the main elements of a preexisting system of commercial exchange. Important in this context is the precise timing of the transition between the two sites which we have attempted to narrow as far as the available pottery dating evidence will allow. Even with these attempts, different scenarios remain open. Was there, for example, an immediate switch in trade from Bushehr to the newly founded port of Siraf? Or was there a period of a century or more which saw a significant lull in maritime trade within the Persian Gulf with a gradual tailing off of activity around Bushehr and Dashtestan drawn out by the resilience of particular elements of the local resource base? During that phase, there may have been only limited trading activity at Siraf. Such a scenario might fit with the evidence for a rapid dismantlement of major state infrastructure built up around the port of Bushehr and the fairly lengthy sequence represented at Siraf by a relatively modest scale military outpost with a gradually expanding civilian population.

Whitehouse states that: ‘Sasanian Siraf emerged from the excavations as a substantial but elusive site: substantial because, apart from the citadel and the building beneath the congregational mosque, we found Sasanian material in six widely dispersed locations (Sites D, F, K, L, M and O); elusive because we learned very little about its topography and development’.222 Once one removes all evidence which is either ambiguous or can now be securely attributed to the immediate post-Sasanian period on the basis of recent advances in pottery dating, the existence of a Sasanian site becomes even more elusive. What we have from the main excavated area covering the early part of the sequence in Site B, and to some extent from the citadel too, is an assemblage containing no east Asian imports, and most typically Type 72, IRPW, TORP.S, SMAG, FINLIM and south Asian pottery with strongly everted rims and simple black painted decoration (HARMIC). This assemblage is now becoming increasingly widely recognised from sites such as Kush, Sir Bani Yas, Hulayla, Sohar and now Siraf. Although we cannot exclude the possibility that some of these features originate in the Sasanian period, the assemblage instead appears to fall within the 7th–mid-8th centuries. In terms of the origin of Siraf, it seems entirely conceivable that this material fits largely within the early Islamic period. The few incontrovertible Sasanian finds could then Whitehouse 2009: 9. Whitehouse 1974: 7. 218  BM 2007,6001.3045. 219  Potts 1998: 209. 220  Kennet 2004: 61–62, table 3. 221  One is from the surface and the other is from uncontaminated deposits close to the base of the sequence: BM 2007,6001.8217, 8501. 222  Whitehouse 2009: 9. 216  217 

223 

176

Piacentini 1992: 118.

Seth, M. N. Priestman: Bushehr, Dashtestan and Siraf suggest that the unique combination of factors at Siraf provided the impetus for the massive transformation of maritime activity seen in the late 8th century, one cannot escape the conclusions that the particular cocktail of a transformed political landscape under the Abbasid administration, the possible policing powers over maritime trade entrusted to forces operating from Siraf, and the freedom in which to operate created by Siraf ’s isolation, may have been significant in enabling those suitably poised to take advantage of a momentous opportunity which transformed the history of the region.

Following this scenario, it seems not unlikely that Siraf could have been founded not long after the destruction of Rishahr. Why a city would be newly founded in such a seemingly inhospitable and isolated location remains an enigma. One suggestion has been to view the fort and watch-post represented by the citadel as having been designed primarily to regulating mercantile activity within the region.224 Archaeologically, this interpretation seems reasonable and, if applicable, then the siting of an outpost in a particularly isolated position may have been a deliberate decision taken in order to place the settlement beyond the reach of potential interference. Perhaps somewhat conversely and against the initial intention, the regulatory powers operating from Siraf would have served to concentrate certain impounded or authorised commodities at the port and with them the constituent elements of an operational emporium. This would include vital components of infrastructure such as warehouses, quays and a labour force trained in the handling and movement of freight, together with some of the institutional mechanisms such as judiciary and banking essential to the smooth operation of trade.225

Scope remains for a much more specific understanding of the changes in the upper Persian Gulf trade network between the Sasanian and Islamic periods. What seems apparent now is that the fate of several large trade emporia within the Persian Gulf were affected, and that the growth of Siraf can be seen as a direct consequence of these processes, specifically the decline of the earlier major late pre-Islamic centre on Bushehr. With this shift, we see the beginning of a process of emporia migration which resulted finally in the main regional centre of maritime trade being situated in the mouth of the Persian Gulf at Hormuz. There were many processes involved in these events, and they are symptomatic of the gradual trend towards a growing entrepreneurial trade within the Indian Ocean of manufactured goods and exotics which contrasts with the model of stateregulated inter-Persian Gulf trade of relatively lowreturn staple produce during the Sasanian and, to a lesser extent, earlier Islamic periods. Similarly, the shift which one sees over nearly two millennia from the rise of Bushehr in the middle Persian Gulf to the dominance of Hormuz at its mouth may be connected with the transformation of the Persian Gulf as a significant distribution unit in its own right, to its status as a node within a much more extensive trading system which spanned much of the Indian Ocean and, with the arrival of the Portuguese, much of the globe.

Although the argument starts to enter the realm of ‘chicken and egg’ – which came first: Siraf or trade? – the nature of activity underway at Siraf in the early 9th century was clearly substantially different to that at Bushehr in the late Sasanian period. By that time military architecture had made way for impressive civil works. High status luxury imports were being brought back to the Persian Gulf from the furthest corners of the Indian Ocean and private enterprise offered the highest possible rewards. Such private risk-taking may not have been possible under the previous state-regulated system and yet it was this potential which would have driven forward innovation in navigation, cartography and boat-building, and allowed individuals to make the first direct voyages along previously uncharted routes. Although it may be overstating the situation to

224  225 

Piacentini 1992: 135. Chaudhury 1985: 197.

177

PART TWO: Environment and Resources St John Simpson

Department of the Middle East, The British Museum, London, WC1B 3DG.

‘Evidence on economic matters is as woefully lacking for the pre-Islamic period of Iranian history as it is for the early Islamic period. Studies of Sasanid Iran usually speak only in generalities about agriculture on the Iranian plateau. Scholars concur that the primary crops were wheat and barley and that various other foodstuffs were produced in small quantities for local consumption ... Within the general profile of the agricultural economy of the Sasanid Empire, Iran’s plateau region seems to have produced only a few exportable commodities, mostly nuts, dried fruits, and saffron. Control of the land lay in the hands of the great lords and lesser gentry, who are usually described as enjoying a rural lifestyle’. Bulliet 2009: 12.

Sasanian Archaeology (Archaeopress 2022): 178–191

St John Simpson: Part Two: Environment and Resources - Introduction

Figure 1. An agricultural scene in southern Iraq showing date palms and an unmaintained sluggish irrigation channel (photograph: author, 2016)

the integrity of government officials upon the reliability of the vizier, and the pinnacle of all these is the vigilance of the king in resisting his own inclinations, and his capability so to guide them that he rules them and they do not rule him’.4

Introduction Bulliet highlights a possible contrast between life on the Iranian plateau with that in the Mesopotamian lowlands, and it is the latter region which, until recently, has provided most of the evidence for particularly the agricultural economy of the Sasanian empire, whether textual or inferred from archaeological surface surveys. The question of continuity or change is also central to discussions of the evolution of the economy in Late Antiquity, whether Byzantine, Sasanian or early medieval.1 Adams repeatedly argues that the Mesopotamian hydraulic regimes were constructed by royal fiat and were unstable,2 and the early 7th century consequently has been viewed as a period of economic collapse which mirrors that of the Sasanian ruling class.3 The weaknesses of this thesis have been discussed in the previous section but the following section compares some of the textual and archaeological sources for the agricultural products, and is followed in the final section by evidence for the manufacturing economy.

This passage from al-Masudi’s history, The Meadows of Gold, makes it clear how, in the eyes of this 10th century author, the success of the Sasanian state lay in its effective integration of royal, military and civilian affairs, and how everything was underpinned by a successful agricultural economy. Early Islamic accounts provide information on crops and taxation during the preceding Sasanian period, Pahlavi texts offer references to fruits and foodstuffs, Classical and Syriac authors add other information but the most important textual source for the agricultural economy is the Babylonian Talmud (henceforth Bavli). Compiled in the 5th century but based on legal cases recorded from Babylonia from the 3rd century onwards, it is an invaluable source of information about that region and the following section draws heavily on it.

Textual reconstruction

Within Mesopotamia the staple crops were wheat, barley and dates, as they had been for millennia and their economic importance is reflected in the early Islamic tax registers:

‘Royal power rests upon the army, and army upon money, and money upon the land-tax (kharaj), and the land-tax upon agriculture, and agriculture upon just administration, and just administration upon the integrity of government officials, and

‘Al-Walid ibn-Salih from Mus‘ab ibn-Yazid abu-Zaid al-Ansari’s father: The latter said, “Ali-bn-abi-Talib sent me to the land irrigated by the Euphrates,

1  Jones 1964; Daryaee 1993; Foss 1995; Morony 1994; 2001/02; 2004; Banaji 2001; Decker 2009; 2011. 2  Adams 1965; 1981. 3  Rahimi-Lahidjani 1988; Christensen 1993.

4 

179

Mas’udi, quoted by Adams 2006: 22.

Sasanian Archaeology mentioning different cantons and villages, and naming Nahr al-Malik Kutha, Bahurasir, ar-Rumakan, Nahr Jaubar, Nahr Durkit and alBihkubadhat [Veh Kubad]. He ordered me to assess on every jarīb of wheat, if thickly sown, one dirham and a half and one sā‘; if thinly sown, two-thirds of a dirham and if not so thickly or thinly sown one dirham; and on barley, one-half of that. He also ordered me to assess on the gardens that include palm-trees and other kinds, 10 dirhams per jarīb; on one jarīb of vine-trees, if its trees had been planted for three full years and a part of the fourth, and if it bears fruit, ten dirhams, with nothing on palmtrees that are outside the villages and the fruits of which are eaten by the passers-by. On vegetables, including cucumbers, grains, sesame and cotton, he ordered me not to assess anything”’.5

times in the Bavli: one type referred to as a dequre is described as narrow, perforated and easily stackable, suggesting that it might refer to asphalt-lined torpedo jars which are so ubiquitous at southern Mesopotamian sites from the Parthian to early Islamic periods.14 There is a strong familiarity with agricultural practice and tradition in the Bavli, with regular references to fallowing, crop rotation, multiple cropping and cross ploughing and the temporary blockage of canals and irrigation ditches in order to allow cleaning (Fig. 1). Manuring may have been reserved for high-yield orchards and vegetable gardens rather than ordinary fields, and the late Tony Wilkinson interpreted extensive scatters of pottery in the region of Zibliyat and elsewhere as evidence of this last practice.15 Flax was a major crop in Babylonia and its cultivation is attested from Pumbadita / Falluja16 and near Nehar Abba, near Sura.17 The use of a water wheel as a lift for the cultivation of flax fields at Veh Ardashir is also mentioned.18 Cotton was another fibre crop which is mentioned in the post-Conquest tax registers for the Sawad region although one assessment gave it only half the levy for vines or clover, and another for ‘the land irrigated by the Euphrates’ excluded it, along with vegetables, altogether.19 This is a water-hungry summer crop which is now attested archaeobotanically from Sasanian Merv and providing the first reliable evidence for its cultivation at this period, also discussed below.

A daily grain market is referred to at Nehardea,6 and the twin towns of Hini and Shili, somewhere near Sura, were also well-known grain markets.7 The consumption of thick wheat or barley porridge with bread is mentioned in the Bavli in connection with the Jewish community at Husal, near the Euphrates between Nehardea and Sura.8 Barley and dates were used to make beer and wine, and date syrup continued to be one of the important traditional sources of sweetener until the import of processed sugar in the late 20th century; by contrast, cane sugar was a luxury at this period, as discussed below. Palms were normally cultivated in plantations and as cover crops.9 Other cultivated fruit trees included figs, olives and pomegranates. Viticulture is also mentioned in the vicinity of Ctesiphon, Nehardea, Pumbadita, Sura, Pum Nahara and Harpanya: with the exception of Ctesiphon, all lie close to the Euphrates.10 The Bavli quotes Rabbi Chisda, who presided over schools at Mata Mehasya and then Sura, as saying that ‘there are sixty types of wine, the most superior is “aromatic red”, the most inferior is white “tilia”’.11 This implies a sophisticated market and a major new industry as beer had been traditionally drunk in Babylonia and was still regarded as a staple, judging by other passages in the Bavli which refer to ‘storehouses of beer’ and the personal preference by another rabbi that ‘I have drunk Babylonian beer and it is more effective than [the water of palm trees]’.12 The ‘fermenting, sweetening and keeping’ of wine is mentioned on a Jewish Aramaic magic bowl from Mesopotamia,13 and the transport of wine in wooden barrels and different types of jars is mentioned several

Rice cultivation seems to have been limited largely to Kaskar and Maysan, i.e. regions in lower Babylonia where water was more freely available for its cultivation. The origins and diffusion of rice are slightly controversial. Ghirshman claims that it was introduced into Mesopotamia during the Achaemenid period,20 and impressions in bricks recovered as part of the Susiana Survey were provisionally identified as belonging to rice and dated to the Parthian period (see below), as was a scatter of carbonised short-grained rice from a Parthian context at Susa.21 Rice became somewhat more widespread during early medieval times but still remained a luxury and an ingredient in sweet dishes rather than a staple, except in the marshes where it remained an important summer crop until recent times.22

Simpson 2003a: 354–55. Wilkinson 1989; 1990. 16  Gittin 27a: ‘a flax-house in Pumbadita’. 17  Shabbos 140b: ‘A Torah scholar who buys a linen garment should buy it from the area of Nehar Abba’, as it was considered of high quality. 18  Eruvin 104a–b. 19  al-Baladhuri 2002: 428–29. 20  Ghirshman 1954: 182. 21  Miller 1981. 22  Simpson 2003a: 357–58. 14  15 

al-Baladhuri 2002: 428–29. ‘Avodah Zarah 38b. 7  Beitzah 25b; Bava Metzia 63b, 72b. 8  Nedarim 49b. 9  e.g., Eruvin 15a. 10  Simpson 2003a: 354. 11  Gittin 70a. 12  Shabbos 110a; see Broshi 1984: 26–27; Geller 2004: 237–40. 13  Naveh and Shaked 1993: 134–35, bowl 24. 5  6 

180

St John Simpson: Part Two: Environment and Resources - Introduction

Figure 2. Camels bringing brushwood to market: this postcard view is of Aden in the 1960s but can apply to much of the Middle East in antiquity (British Museum, EPH-ME 6841)

at Ctesiphon and which had a lasting effect on Abbasid court cuisine.29

Fields or plots of cress, cucumbers, cuscuta, garlic, horse beans, leeks, lucerne (or clover), millet, oats, peas, pumpkins, rye, saffron, sesame, spelt and turnips are also mentioned in the Bavli and Arab Conquest tax registers: the former suggests that often they were on the edges of settlements and were fenced sometimes for their protection against thieves and goats.23 Abaye refers to ‘gourd, fenugreek, leek, beets and dates’.24 All are staples of Mesopotamian agriculture and it is no surprise that most of the plant names in the Bavli are Akkadian loan words,25 and many recur as ingredients in the 10th century ‘Baghdad Cookery Book’ compiled by ibn Sayyār al-Warrāq.26 Another feature of this later source is the wide range of local recipes and regional specialities. This was not something which was specific to Abbasid Iraq but probably reflects a long tradition of appreciative knowledge of different dishes: the Old Babylonian cuneiform culinary tablets in Yale list 35 types of stew, including one described as ‘Elamite broth’ and another as in ‘Assyrian style’,27 and the Mishnah remarks that the following ‘must be removed at Passover: Babylonian porridge, Median beer, Edomite vinegar and Egyptian barley-beer’.28 Kwasman has highlighted the prevalence of Iranian loan words in the Bavli for methods of food preparation and types of drinking vessel, doubtless reflecting the trickle-down influence of Iranian tastes and fashions from the court

The supply of fuel for domestic and industrial needs is a major part of human activity, whether in traditional or modern societies, and is sometimes overlooked when looking deeper into antiquity (Figure 2). Two Middle Persian documents in the 7th century Pahlavi Archive in Berkeley refer to firewood (ēzm) being measured by the donkey-load (xarvār), and the price of two such loads is given as a drachm.30 In the case of marginal environments fuel supply could be a major factor in determining the abandonment of a settlement or industry. In regions such as Mesopotamia, however, there was an abundance of agricultural by-product which could be used, whether for hearths, ovens or kilns. One source was manure, another was twigs and branches, and the Babylonian Talmud refers to both. One example is a case related by Rabbi Shmuel: ‘Concerning a person who swept debris (e.g. weeds and twigs) from a parcel of land, and later claimed that it was his intent to acquire the land thereby, the law is as follows: If he swept for the sake of improving the land, he has acquired the land. But if he swept for the sake of gathering wood for fuel, he has not acquired the land. What are the circumstances that reveal his intention? If he took both large and small [twigs], we assume that he removed them for the sake of improving the land. But if he took only the

e.g., Bava Basra 4b, 5a; see Cohen 1937: 174, 249–50; Newman 1932: 106–107. 24  Horayos 12a; repeated in Kereisos 6a. 25  Kwasman 2015. 26  Nasrallah 2007. 27  Bottéro 1995. 28  Pesachim 42a. 23 

29  30 

181

Kwasman 2015; see Simpson 2003a. Gignoux 2008a: 136.

Sasanian Archaeology Table 1. Sample incantation bowl texts which include references to householders’ livestock as subjects of magical protection Client

Script

Findspot

Publication

‘Farrukdad the son of Zebinta and of Jewish Qamoi the daughter Aramaic of Zaraq’

Babylonia

‘[names of clients] and their sons, their daughters, Gordon 1934a: 328– their slaves, their handmaidens – both male and 30, Gordon D = Isbell female – and camels, steeds, oxen, donkeys, all large or 1975: 112–13, text 49 small cattle, and everything there is or is going to be (Baghdad, IM 6519) on this threshold’

‘the lover of ‘Ibai the Jewish son of Mamai’ Aramaic

Nippur

‘Brik-Marya the son of Rishinduk, and for Jewish Dustai the daughter Aramaic of Dodai

?

Passage

Gordon 1937: 91, Gordon J = Isbell 1975: 126, text 55

Gordon 1941: 127– 28, pls 7–9, Gordon 7 = Isbell 1975: 100– 101, text 42 (Musée du Louvre, AO 2099)

‘the donkeys, the oxen’ ‘Brik-Marya the son of Rishinduk, and for Dustai the daughter of Dodai, his wife, from their sons, from their daughters, from their cattle, and from their possessions’

Gordon 1937: 104, Gordon O = Tell Ibrahim / Yamauchi 1967: 276– Mandaic Kutha 79, text 28 (British Museum 91724)

‘the house, the dwelling, the mansion, and the building, and the body, the wife, the sons, and the daughters, and the cattle, the property, the slaves, and the handmaids’

Mandaic Khouabir

Pognon 1898: 23 = Yamauchi 1967: 190–93, text 11

‘the building and the animals’

Mandaic Nippur

Montgomery 1913: 244–45, bowl 38 = Yamauchi 1967, 256–59, text 23

‘the house, the dwelling, mansion and building, the sons and daughters, the cattle and household vessels of Hinduita daughter of Doday and (of) Marada, her husband [Marabba son of Hinduita] and her sons [Zadoye, Dazaunoye] and daughters ... her house and her bed and from their [wives of her sons] and their sons and daughters and their cattle’

Mandaic Nippur

Montgomery 1913: 253, bowl 40 = Yamauchi 1967, 262–65, text 25

‘Shabha the son of Shirin’

Mandaic ?

‘Farrukzad, son of Ahai, and his wife Shahik, daughter of Rashnuiduk’

Gordon 1941: 344–45 ‘the house and to the dwelling and to the mansion and = Yamauchi 1967: to the building (?) and to the oxen possessed of horns 280–83, text 29 and to the asses possessed of teeth’ (Fitzwilliam)

Mandaic ?

Moriggi 2001: 215–20 ‘the dwelling and the mansion and the building and (Rome, IsIAO 5210) the cattle and the property and the food [plural]’

‘šyl’y son of ‘ym’d’bwh’

Syriac

this house of [client] and its inhabitants, the ones Moriggi 2014: 154– 60, bowl 32 (Musée entering the door and the ones going up the roofs, his du Louvre AO 17.284) grain, the field, the cattle and the property’

‘Shrula the son of Duktanuba’ and his wife Qaqay the daughter of KRPST ‘G[ukaye AdurYazda]ndar the son of Kzabiat’ -

‘Hinduita bath Doday’

‘Xaro son of Mehanosh’

Mandaic Nippur

‘brym’ son of ‘zdndwk’ Syriac

‘Bahroi bath BathSahde’, wife of Mihr- Syriac hormizd bar Mami

Hit area railway 700 m. north of Qadhimain

Nippur

Gordon 1941: 347 = Yamauchi 1967: 284– ‘this house, dwelling, mansion, building, and the cattle 87, text 30 (Hilprecht and property’ Sammlung)

Moriggi 2014: 120–23, bowl 24 (Belgrade, Jewish Historical Museum)

‘the body and the wife and the male sons and the female daughters and the house and the dwelling, the mansion and the building of the cattle, the ass, bull, goat and swine of [client’s name] ... the building of the ass, bull, goat, and swine and live(stock) of [the client]’

‘his house and for the people of his house, for his property and for his cattle and for his body’

Montgomery 1913: 232, bowl 34 = ‘his house, his wife and his sons and his daughters and Moriggi 2014: 47–51, his cattle and his property and in all his dwellings’ bowl 6 (Penn, CBS 9012)

182

St John Simpson: Part Two: Environment and Resources - Introduction

Client

Findspot

Publication

Syriac

Nippur

Montgomery 1913: 242, bowl 37 = ‘the house and threshold, the wife and the sons and Moriggi 2014: 60–63, the daughters, the cattle, [and all that] belong to him’ bowl 9 (Penn, CBS 2943)

‘’h’ son of Pr’hormuz’ Syriac

Nippur

‘Zaroi son of ...’

‘kwsrw son of q’qy’

-

Script

Syriac

Syriac

?

acquired in Basra

Passage

Moriggi 2014: 88–91, ‘his house, his wife, his sons and his daughters [...] his bowl 16 (Baghdad, IM cattle’ 60960) Moriggi 2014: 138–43, bowl 28 (Smithsonian, AO27064-0)

‘the house and the men entering his door and going up on his roof, his grain, his field and his cattle and his possessions’

Moriggi 2014: 27–31, bowl 2 (Catholic University of America, ‘the one who suffocates the animals’1 Washington, Semitics/ICOR Collections H156)

Moriggi (2014: 29) notes that the word used to refer to the suffocator of animals is based on a root meaning ‘to hold back, to hitch, to harness’, implying that the demon used their harness as a means of strangulation: this suggests that they were horses, donkeys or cattle kept under cover, rather than sheep, goats or pigs which were tended outdoors.

1 

It is likely that these animals were grazed on fallow fields and/or abandoned levees and the former practice is specifically cited in one passage;33 unripe grain is also said to have been used as animal feed in the Veh Ardashir area: ‘it is customary for farmers in that area to cut their grain while it is still unripe’.34 Unsurprisingly, the unrestricted foraging by goats either within the countryside or settlements was frowned upon and considered sufficiently irritating to cause regular litigation. Several cases appear to refer to Veh Ardashir:

large ones and left the small ones, it is evident that he removed them for the sake of gathering wood’.31 Arable agriculture dominated the landscape of the alluvial plains of Mesopotamia, Margiana and probably many other regions, but the significance of other economies should not be overlooked. The population was not vegetarian, cities and villages, civilians and soldiers, needed to be fully provisioned, and a complex pattern of land-use was developed. References within the Bavli and occasionally on incantation bowls indicate that many households, even in towns, owned cattle, sheep and goats (Table 1): great care was therefore taken in the management of their grazing in order to ensure minimal complaint and maximum economic benefit (and also ensure their protection through inclusion in the magical spells). Shepherds were responsible for the daily management of sheep and goats which were returned to individual households at night, at least in the case of small settlements:

‘There was a goat that saw a turnip on the top of a barrel. It pawed its way and climbed up to the top of the barrel, and ate the turnip and broke the barrel. Rava obligated [the goat’s owner] full damages for the turnip and the barrel’.35 ‘If those goats of the meat market cause damage, we warn their owners about this two or three times. If [the owner] heeds the warning and restrains his goats, fine. And if he does not heed the warning, then we say to him: “Sit at the butcher’s place and take your money!”’.36

‘Goats are different than other animals, for they are customarily entrusted to the care of a shepherd and therefore do not roam about freely … But there are periods in the morning and evening when the goats are not supervised … In Nehardea thieves abound and [the goats] are therefore delivered from hand to hand’.32

‘There was a certain woman who entered a certain house with the homeowner’s permission to bake bread there. A goat belonging to the homeowner came, ate the dough, became overheated and died. Bava Basra 29b. Bava Basra 36a. 35  Bava Kamma 20a. 36  Bava Kamma 23b. 33 

31  32 

34 

Bava Basra 54a. Bava Basra 36a.

183

Sasanian Archaeology meat are mentioned,42 and the consumption of sheeptail fat is given in another passage.43 Meat was bought for cash or credit and by weight or size of portion, for which there were different standards in different towns.44 Pan balances are depicted on slightly later Sogdian ossuaries from central Asia and small flat copper alloy weights resembling those found in early Islamic contexts at city-sites such as Siraf and Ayla have been excavated at Veh Ardashir.45 However, there is a telling reference in the Bavli to preferences for weights of stone or glass over metal which implies that there were other types which have not yet been recognised archaeologically and/or perhaps may have been mistaken for gaming pieces.46 Choice cuts of meat were reserved by wealthy customers who used servants to carry them home. Animal heads were dressed and sold, perhaps as the centrepiece of a meal, an ingredient of broth or a source of jellied meat, and a supplier of calves’ heads is mentioned.47 A ‘market of the leather workers’ is mentioned in association with Rava;48 in another passage, black oxen were said to provide the best hides: ‘A black ox for its skin, a red ox for its meat, a white ox for ploughing’.49 These industries probably facilitated the co-location of related crafts such as tanneries or horn workshops.

Figure 3. Sheeps’ heads for sale at a butchers in Kermanshah / Bakhtaran (photograph: author, 2000)

Rava obligated [the woman] to pay the homeowner the value of his goat’.37 Professional fatteners and breeders of animals also feature: ‘Rabbi Elazar of Hagronya [probably Agranum near Nehardea on the Euphrates] bought an animal and consigned it to his sharecropper to fatten, for half the profit and half the loss. [He] fattened it, and [R. Elazar] gave him the head as his wage, and also gave him half of the profit’.38 Some places, such as Mata Mehasya, appear to have specialised in the rearing of livestock. The name of another settlement, Dura dhe-Ra‘awata, translates as ‘settlement of shepherds’:39 although its location is unknown, the fact that the four legal cases mentioning this place were referred to rabbis probably based in the academy at Pumbadita suggests that it was situated near that town and therefore lay in the middle Euphrates area.

Moreover, regular references in the Bavli and on incantation bowls attest the keeping of chickens; ducks and geese are also mentioned as sources of food and roast goose was a popular dish.50 Finally, pigs are occasionally mentioned in the Bavli although their upkeep was of course prohibited within Jewish communities: one passage refers to how ‘the condition of Deruqart [probably near Sura] is like that of a town in which there are pigs’,51 suggesting that they were allowed to roam free as scavengers, as in earlier periods,52 and pork was certainly consumed at Merv during this period (see below). The impression from the Bavli is that sheep and goat were the main types of livestock and the reduction of cattle may, as Adams remarked, reflect a situation whereby large-scale increases in arable agriculture was at the expense of former grazing areas within the Sawad and this led to a reduction in livestock herds in this region.53 However, the role of pastoralism within the economy was probably more significant in

Cattle were an important source of meat and leather, but they were also used for ploughing, threshing, and were of course a valuable source of milk. According to one passage in the Bavli, large white oxen were used in ploughing,40 but cows were preferred for threshing as they were considered more docile. Places mentioned in connection with butchers or slaughterhouses include Husal, Pumbadita and Sura,41 and it appears that meat was commonly purchased from stalls and shops but slaughterhouses, specialist butchers and direct provisioning of fattened animals to households are also attested (Figure 3). Shops selling thighs of

Bava Basra 88a. Pesachim 3b. 44  Bava Basra 83b, 89a, 90a. 45  See Whitehouse 1968: 19; Whitcomb 1994: 9. 46  Bava Basra 89b. 47  Bava Metzia 69a. 48  Bava Metzia 24b. 49  Nazir 31b. 50  Pesachim 74b; Beitzah 33a. 51  Niddah 58b; see Oppenheimer 1983: 110. 52  e.g., Jursa 2016: 190, 193. 53  Adams 1981: 213–14. 42  43 

Bava Kamma 48a. Bava Metzia 69a. 39  Bava Basra 54b, 82b–83a, ‘Eruvin 7b, 12a. 40  Nazir 31b. 41  Chullin 110a, 132b. 37  38 

184

St John Simpson: Part Two: Environment and Resources - Introduction

Figure 4. Using a Siraf-type flotation machine at Merv radically improved the recovery of plant remains from Sasanian contexts (photograph: author, October 1995)

silken cloaks, woollen rugs and woven carpets – a great quantity of them and very beautiful, but on account of their weight they burnt them all’.58

other regions, particularly in northern Mesopotamia. Written sources suggest that stock-breeding and the grazing of sheep and goat were important factors in the economy of the north Jazira and upper Khabur where a reduction in settlement density during the Sasanian period suggests a shift to non-sedentary pastoralism,54 probably reflecting the gradual aridification of the region as a result of the northward shift of the dryfarming isohyets.55 During the late 5th century, this region was occupied by members of the Banu Taghlib who are believed to have been driven out of Arabia by drought and the expansion of the Kinda.56 Cattle herds belonging to the Sasanian garrison at Nisibis are recorded as having grazing rights as far as the northern slopes of Jabal Sinjar, and cattle, horses and goats feature heavily among the livestock of the Jazira as late as the 760s, according to the Syriac Chronicle of Zuqnīn, which also describes some of the social tensions arising from greater competition over agricultural resources.57 Moreover, Theophanes’ account of Heraclius’ seizure of Khusrau II’s palace and residence at Dastagerd refers to the parks containing ‘an infinite number of ostriches, gazelles, wild asses, peacocks and pheasant’ and how the the storerooms contained

Surveys carried out in the Kangavar valley of the western central Zagros suggest that the density and distribution of Parthian settlement there was similar to that of the 1960s,59 and led to the hypothesis: ‘with such intense settlement densities in the Parthian and Sasanian periods, in both the Zagros uplands and Mesopotamian lowlands, it would then be logical that some form of transhumance or pastoral nomadism was occurring the uplands. These conjectured migrations should be considered as the consequences of population growth and land pressure, which are currently the factors that determine the transhumant cycle in the Zardeh basin’.60 Moreover, an independent survey of the Izeh plain in north-east Khuzestan carried out in 1976 revealed ‘an increased number of campsites’ dating to the Parthian and/or Sasanian period, when rural settlement appears to have contracted but the settlement of Čoqa Qal ‘grew to about 2400 people and a village in the south-east part of the plain became a centre with about 1000 people’.61 The importance of pastoral nomadism in the preIslamic economy of Iran is a hotly debated subject and

‘a great quantity of aloes and big pieces of aloes wood, each weighing 70 or 80 lbs, much silk and pepper, linen garments beyond counting, sugar, ginger and many other goods. Others found silver bullion, Wilkinson and Tucker 1995: 69–71; Lyonnet 1996: 353–53; de Aloe 2008; Ur et al. 2010: 119–22. 55  Rösner and Schäbitz 1991. 56  Morony 1984: 217. 57  Harrak (ed.) 1999; see Morony 2000: 8. 54 

Quoted by Greatrex and Lieu (eds) 2002: 215. Young 1974: 27. 60  Evans 1980: 4. 61  Eqbal 1979: 116. 58  59 

185

Sasanian Archaeology

Figure 6. A typical Sasanian domestic hearth at Merv (photograph: author, 1999)

Figure 5. Cotton bushes at Chilburj in the Merv oasis (photograph: author, October 1992)

in barrels of brine and the water transport of tzachanta fish to Sikra near Mahoza on the Nahr al-Malik.64

more measures are needed to assess the archaeological evidence.62

Archaeological evidence ‘In these regions there are many fields, planted with vineyards and various kinds of fruits. Here too palm trees are wont to grow, extending over a wide expanse as far as Mesene and the great sea [Persian Gulf], in mighty groves. And wherever anyone goes, one constantly sees palm branches with and without fruit, and from their yield an abundance of honey [syrup] and wine is made’.65

Finally, in riverine areas and in the far south of Iraq and along the Persian Gulf, fishing must have been an important part of the food economy, as it had been for millennia before, and various types of fish are mentioned as food in the Bavli. One passage comments on the merits of salted meat over fish,63 and a ruling attributed to R. Pappa that ‘all things that are aged are better, except for dates, beer and small fish’ suggests that large fish may have been salted but there were sufficiently large quantities and variety of other types for these to be bought fresh. Other passages allude to fishing in the Badita canal by inhabitants of a place called Levai and the salting of the catch for consumption on alternate days of a festival, a ruling at Parziqiya near Sura over the eating of roasted binita fish, the sale of fish

This passage from Ammianus’ description of Babylonia evokes a rich picture of agriculture in lower Mesopotamia but how far can such textual sources be supported or amended by the archaeological evidence? Across the Middle East, environmental analyses have been regarded as a fundamental tool for reconstructing Moed Katan 11a, Pesachim 76b, ‘Avodah Zarah 40a; see Oppenheimer 1983: 178–79, 345, 402. 65  Ammianus, History 24.3.12. 64 

See Potts 2014. 63  Bava Basra 74a. 62 

186

St John Simpson: Part Two: Environment and Resources - Introduction

Figure 7. Harvesting camel thorn as a source of fodder at Merv (photograph: author, October 1996)

terminal Parthian’ fired bricks found on the Susiana Survey have been reportedly identified as belonging to rice although these identifications have since been challenged.72

the economy at early prehistoric sites but have rarely been adopted at late historic sites on the grounds that they are unnecessary in periods where there are texts. The fallacy of this assumption will become clear below. Archaeobotanical studies have been slow in testing the literary sources but there is growing awareness of this and important studies published from Late Antique sites in Egypt,66 as well as much later contexts in Turkey.67

Subsequent archaeological research projects at Merv and Kush were partly designed to redress this shortage of reliable archaeobotanical data. In the case of Merv, winter staple food crops included free threshing wheat, six-row hulled barley and lentil, other crops included broomcorn millet, lentil, pea and melon/cucumber, and a variety of local and imported fruits and nuts were also available including grape (Vitis sp.), apple (Malus and Sorbus sp.), apricot (Armeniaca sp.), peach (Persica sp.), almond (Amygdalus sp.), pistachio and jujube.73 In addition, accidentally carbonised seeds found in mid-4th century and later contexts prove that cotton (Gossypium sp.) was an important summer crop (Figure 5), as it was in Khorezm in the same period.74 This implies a Sasanian precursor for the mixed cotton/silk mulham/malham (and indirectly therefore silk) industries attested from this city during the 9th century and later.75 Further archaeobotanical finds published below for the first time from Kirpichli depe

Archaeobotanical analyses for Sasanian sites are at their beginning but already offer some important insights (Figure 4). Sieving and flotation were carried out during the excavation of late Sasanian contexts at Tell Baruda at Veh Ardashir but the results remain unpublished apart from a reference to the discovery in one room that ‘the whole layer seems to have been occupied by insects after the site was abandoned by human beings’ in the 7th century.68 The remains of pomegranate, almonds and woven cotton are reported from a late Sasanian funerary repository at Shahr-i Qumis,69 ‘burnt almonds and carbonized figs’ are reported from a 7th century room on the citadel at Qasr-i Abu Nasr, although the source of the identifications is unknown,70 and barley impressions noted in fired bricks at the Diyala site of Tell Bismaya.71 Impressions in so-called ‘middle– Smith 2003; Cappers 2006. Kennedy 2000. 68  Costantini, Tosi and Vigna Taglianti 1975/77: 248; see Venco Ricciardi 1977. 69  Hansman and Stronach 1970b: 148–50. 70  Whitcomb 1985: 168. 71  Jacobsen 1982: 16, 23, 25. 66 

72  Wenke 1975/76: 41, 87–88, 106–108, 120, 144–46; Nesbitt pers. comm. 73  Nesbitt 1993; 1994; Boardman 1995; 1996; 1997; 1999; Simpson 2019d. 74  Brite and Marston 2013. 75  Simpson 2008b: 248–49; see Lamm 1937: 95–99, 199–200.

67 

187

Sasanian Archaeology according to the local ecological zone. Tamarisk and ziziphus were common but there was also evidence for various chenopods, maple/sycamore (Acer sp.), hornbeam/alder (Carpinus sp./Alnus sp.), deciduous oak (Quercus sp.), sweetgum/liquidambar (Liquidambar sp.), alder buckthorn (Frangula sp.), beech (Fagus sp.)/plane (Platanus sp.), willow (Salix sp.) and/or poplar (Populus sp.). However, some fuel – notably Ulmus sp./Zelkova sp. (elm/zelkova) – may have been imported as these arboreal species were found in kilns situated within treeless steppe; moreover, the presence of Pomoideae, Prunoideae and plum (Prunus sp.) hint at the cultivation of fruit trees.81

provide additional evidence for the 8th century from Dehistan.76 During the excavations at Merv, domestic hearths within Sasanian houses were typically filled with pale powdery ash containing almost no visible charcoal but including large numbers of accidentally carbonised cotton seeds (Figure 6). The source of this is either cotton sticks or dung fuel containing cotton seeds, in either case offering indirect evidence for the cultivation of cotton as a fibre crop, for which Merv was famous in the early medieval period.77 In other contexts, evidence for wood charcoal was found which indicated the burning of ad hoc fragments derived from coppicing.78 The environmental analyses therefore permitted the reconstruction of fuel supply at a major urban centre with recycled agricultural by-product as a major source.

The faunal assemblages recovered from forts on the Gorgan wall, a possible campaign base at Qal‘eh Kharabeh and the city-site of Dasht Qal‘eh suggest an overall pattern which is rather similar to Merv but with some significant variation between the different types of site on the Gorgan plain.82 Sheep and goat were found throughout, with sheep outnumbering goat in all cases where the data was available, and almost all were slaughtered between the ages of 2–4 and 6–8 years, implying that they were primarily valued for their wool and hair. Cattle and pig were again represented at almost all sites, the former primarily used as draught animals whereas the latter were slaughtered young: however, whereas sheep/goat dominated the assemblages at Qal‘eh Kharabeh and Dasht Qal‘eh, cattle and pig were much more common at Fort 4. Single bones belonging to water buffalo and camel, both potentially draught animals as well as sources of nutritious milk and hair, were found at Dasht Qal‘eh. Chicken bones were common at many sites but there was also evidence for butchered ducks, swans and migratory birds at Qal‘eh Kharabeh. Wild boar was present at Fort 4, and red deer (including a piece of sawn shed antler) was present at Dasht Qal‘eh: whereas the former represents hunting in a marshy environment, the latter probably derives from the forested zone of the Elburz. Fish-bones recovered from Qal‘eh Kharabeh and Dasht Qal‘eh suggest that carp and roach may have been traded as whole fish, whereas pikeperch, sturgeon and shad were mainly represented by their meaty portions; by contrast, fish is absent or rare at any of the forts investigated to date.83

The following wild species were also represented at Merv: Galium sp., Suaeda type and Boraginaceae. Moreover, the seeds and pods of small seeded legumes, including camel thorn (Alhagi sp.), were very common in samples from 4th–5th century Sasanian contexts in the lower city. Camel thorn grows wild throughout the oasis today, providing an effective natural balance to progressive salinisation and an important winter reserve fodder crop for cattle and other animals (Figure 7). A similar pattern of distribution and use is presumed in antiquity although the excavated seed remains may also derive from deliberate use of dried camel thorn fuel, particularly in bread ovens where it is a preferred fuel.79 These legumes nevertheless hint at the importance of the herding of livestock and suggest that a range of environmental techniques are necessary to help reconstruct the surrounding landscape. This was strongly confirmed by the results of the analyses of the charcoal now completed from all of the excavation areas carried out at the site from the Sasanian through to the post-medieval periods. A preliminary report on results from excavations directed by Jebrael Nokandeh at the multi-period site of Qizlar Qal‘eh, on the western section of the Gorgan wall, indicate the cultivation of barley (Hordeum vulgare), six-row barley, compact wheat (Triticum compactum) and grape (Vitis vinifera), with a wide variety of weed species represented, including Linum sp., Valerianella dentata, Trifolium sp., Melilotus sp., Phalaris sp., Aizoon sp., Bromus sp. and seeds of the Malvaceae family, plus tamarisk charcoal.80 Analyses of charcoal residues found within kilns used to construct the Gorgan wall, as well as domestic contexts at Qal‘eh Kharabeh, suggest that most fuel was locally sourced as it varied

There are relatively few other zooarchaeological analyses for this period but the range and significance of their findings is reviewed below. There are brief species lists of excavated animal bone recovered by hand from Bastam (post-Sasanian), Noruz-Mahale (late Parthian or early Sasanian) and Shahr-i Qumis (mainly Parthian), although the principal focus of these excavations in northern Iran was on remains of earlier

Atagarryev, this volume. Gale, this volume; Boardman 1996: 20. 78  Gale, this volume. 79  Jacobsen 1982: 68, appendix 24; A’lam 1990. 80  Kimiaie 2006. 76  77 

Poole and Gale 2013. Mashkour 2013. 83  Radu 2013; see Mashkour 2013: 563. 81  82 

188

St John Simpson: Part Two: Environment and Resources - Introduction baited basket traps (which were traditionally made in the Persian Gulf from plaited palm fronds), drifting gill nets, weighted beach seines or hook and line. However, the pelagic tunas and jacks feed in open water and may have been caught in gill nets, by hook and line or using spears.89 The latter species are today mostly fished when optimum sea conditions and steady weather combine from October–December and March–May whereas the optimum time for catching the shallow water species is during spawning during April–June. However, despite extensive fine-sieving, very few bones of small fish were found at Kush. The exceptions include sardine and silverbiddies which are shoaling species and therefore may be caught in beach seine nets. Despite variations in recovery techniques and sieve mesh size, the range of species and fish sizes broadly matches the pattern previously noted from other coastal sites in the Persian Gulf and suggests all-year round fishing but giving preference to species which yield large quantities of meat as well as possibly exploiting different marine environments – and thus different species – according to the seasons. This pattern differs to that reconstructed for the 8th century Christian communities on Sir Bani Yas where most of the consumed fish were small in size and caught in shallow sand and reef areas around the island whereas there were relatively low numbers of larger jacks or pelagic species such as tuna.90

periods and the hand recovery probably biased the sample in favour of larger mammals.84 Another small assemblage excavated at Jalaliye in the Kaluraz valley in Gilan province, and broadly attributed to the ParthoSasanian period by the excavators, was dominated by sheep/goat with smaller numbers of cattle bones but the very low numbers of red deer hint at hunting in the forests which then dominated the landscape.85 A very small and highly fragmented assemblage recovered from limited excavations by the author at the rural late Sasanian site of Qara Dere, in the upper Tigris valley, suggested the herding of sheep/goat with additional consumption of pork and possibly gazelle, which was still hunted on the nearby steppe as late as the 19th century.86 A small quantity of pig was also represented in the small zooarchaeological assemblage recovered from a late Sasanian domestic context at Tell al-Sarāh in the Hamrīn basin of east-central Iraq but the assemblage, both there and at a contemporary late Sasanian farmhouse excavated at nearby Tell Mahmūd, were again dominated by sheep/goat.87 The presence of gazelle and possibly wild pig from excavated contexts of the same period at Merv hint at provisioning through hunting: the subject of the royal hunt is a familiar one in Sasanian art and courtly literature, but the employment of professional hunters ‘supplying wild animals and partridge and fish’ is attested from Amida in 503 and we might expect this to be catering for the middle classes rather than royalty.88 Ian Smith’s ongoing analysis of the animal remains excavated at Merv should provide one of the most detailed glimpses into provisioning of meat at a major historic urban centre in south-west central Asia, and in this volume he presents a preliminary view of evidence for butchery practices. The analysis of the remains from the late Sasanian citadel suggests the herding of sheep/goat as the most important forms of livestock and that these were kept not only for their meat but also for their wool.

The zooarchaeological results from these few and widely separated sites suggest different economies for different eco-zones but are consistent in illustrating the importance of the herding of ovicaprids, whereas cattle appear to have been less important. One of the causes for this may have been the increasingly dry climatic phase which particularly affected more marginal zones, and palynological and sedimentological evidence from cores in Lake Khatouniye in north-east Syria confirm that this region underwent a more arid phase during this period.91 The correlation of oxygen isotope curves with pollen cores from lakes Zeribar and Mirand in the Zagros mountains supports this evidence for a dryer phase centred around 1500 cal. BP, i.e. the beginning of the 6th century.92 Pollen diagrams from cores taken from Lake Almalou in north-west Iran and Lake Maharlou in Fars province provide a complementary picture from the Iranian highlands as they indicate the beginnings of large-scale arboriculture in the 3rd millennium BC, with cultivation of Persian Walnut (Juglans regia L.), Oriental Plane Trees (Platanus orientalis L.) and probably grapevines (Vitis) during the Sasanian period.93 The grapes of Fars were famous during the early medieval period,94 the forms of and imagery

Fish, on the other hand, was barely consumed at Merv but unsurprisingly was more important in the Persian Gulf where there is evidence from this period (and later) at Kush. A limited variety of shallow and deep-water species of large fish, including Scombridae (tunas and mackerel), Sparidae (seabreams), Carangidae (especially jacks but also scad and trevallies) and Lethrinidae (emperors), were consumed in varying quantities throughout the sequence at this site. The seabreams, jacks and emperors prefer reef and shallow inshore habitats and analogy with traditional fishing practices suggests that they may have been caught throughout the year using a combination of tidal barrier traps,

Beech 2004: 44–49. Beech 2004: 110–21, 182. Rösner and Schäbitz 1991. 92  Stevens et al. 2006. 93  Djamali et al. 2009; 2010; 2015: fig. 3. 94  e.g., Le Strange 1905: 262.

Boessneck and Kokabi 1988: 215–20; Takai 1966; Hansman and Stronach 1970b: 149–50. 85  Mashkour 2005. 86  Dobney forthcoming. 87  Fedele, this volume. 88  Simpson 2003a: 352, 358–59.

89 

84 

90  91 

189

Sasanian Archaeology on late Sasanian silverwares imply that wine was a popular beverage,95 and asphalt-coated torpedo jars found at sites on both sides of the Persian Gulf probably represent the discarded packaging of wine which was either consumed locally or moved to/from the coast in wineskins.96 During this period an increase in sporormiella, a coprophilous fungus associated with domestic livestock, was noted in another core taken from Lake Parishan, also in Fars, whereas a core sample from the Tuska Tchal peat bog, in the eastern part of Gorgan, not only shows a peak in Persian Walnut in the early Sasanian period but the appearance of Oriental Plane Trees, which are typically cultivated as a shade crop in Iran, during the late Sasanian period.97 Finally, an additional core taken from Lake Neor in the Talysh region of north-west Iran indicates an open grassy steppe landscape suitable for grazing, and the fossilised remains of beetles associated with dung points strongly to the area being dominated by sheep/goat herds in a high region where transhumance is still practised by specialised herders.98

volume) has argued that date palm cultivation played a more important role than previously suggested in the northern Persian Gulf region. Archaeobotanical results and excavated charcoal remains from Merv offer some of the best archaeological environmental evidence for the Sasanian period. They prove the existence of fruticulture and the cultivation of summer cotton, as well as a range of cereals which were harvested in the autumn. This implies that there must have been a complex agricultural regime with careful management of land and water resources. It also strongly suggests a Sasanian precursor for the elaborate water management practices described in the 10th century, whereby ‘a commander ... with more jurisdiction than the commander of the garrison [in Merv itself]’ and responsible for thousands of workers, watchmen and divers maintained the main dam on the Murghab, allocated water depending on the level which was regularly measured, and ‘when the superintendant gauges the water, he sends the message at high speed specifically to the office of the river administration; then they send messengers to all the supervisors of the branch canals, and the water is distributed on the basis of the appraisal’.101 At Merv, the addition of fruits such as grapes, almonds and peaches to the crop repertoire known from earlier periods in the oasis is a significant change to the local economy, and contradicts the suggestion that crops such as apricot were a late addition to Sasanian agriculture.102 The fact that several of these crops (almond, peach, melon and grape) are sensitive to soil salinity, and ‘that the full suite of crops is present in broadly similar quantities throughout the sequence, strongly suggest that irrigation agriculture was sustained through this period without occurrence of catastrophic salinization’.103 Moreover, they imply a stable long-term economy which undermines the hypothesis of cotton being an innovation of the socalled ‘green Abbasid revolution’104 or part of a sudden cotton boom in Khurasan during the 9th and 10th centuries.105 A similar story applies to rice for which the archaeobotanical and textual evidence combine to indicate a longer history and more complex pattern of consumption than previously recognized.106 The same result has been argued by Decker using combined textual and archaeobotanical evidence, and who focused on four of the crops claimed as part of the ‘Islamic Agricultural Revolution’ – durum wheat, rice, cotton and artichoke – to conclude that ‘Islamic farming structures were built atop earlier Roman and

Conclusion Archaeological surveys in the Mesopotamian lowlands, Susiana, Margiana and the Mughan / Mil plains prove that irrigation agriculture was a fundamental part of the economy, and references in the Bavli suggest a complex range of crops being grown in central Mesopotamia. In other areas dry farming was practised, with the grazing of large mixed herds on the steppe. There is also evidence for qanats from textual sources: although there is still no independent evidence to date these, they surely played an important role by the Sasanian period in arid regions such as central and southern Iran. It was probably in this latter period that the hydraulic technology changed from one which used underground water channels or tunnels to divert water from groundwater underflow in wadi beds or depressions where runoff accumulated, the type widely attested from south-east Arabia during the Iron II phase (c. 1200/1000–600 BC), to one where deep channels were dug which tapped aquifers and which are true qanats.99 Following the completion of the Kur river basin survey, Hartnell argues that industrialised agriculture, including water-milling, began to be important in northern Fars during the Sasanian period, and that this marks the beginning of a trend which continued into the early medieval period.100 Moreover, Asadi (this Gunter 1988. Simpson 2003a: 355; this volume: 255. Djamali pers. comm., November 2013; see Potts 2018. 98  Ponel et al. 2013. 99  This distinction is made by Boucharlat (2001; 2003), where the south-east Arabian evidence is discussed. Mouton and Schiettecatte (2013: 40–45) also observe how increasing aridity probably led to the abandonment of the south-east Arabian versions and the abandonment of the settlements they sustained. 100  Hartnell 2014. 95  96  97 

al-Muqaddasī 2001: 269. Gyselen 1998: 104. Nesbitt and O’Hara 2000: 112. 104  Watson 1983. 105  Bulliet 2009. 106  Nesbitt, Simpson and Svanberg 2010; see Watson 1983. 101  102  103 

190

St John Simpson: Part Two: Environment and Resources - Introduction is yellow’,119 and colours were visible on stuccoes from Dezh-e Guri immediately after discovery.120 Doubtless other such buildings will be discovered in the future but it is perhaps telling that, so far, their distribution is within fertile agricultural lands, often situated away from large centres of population. Gibson remarked following his survey of the Kish region in 1966/67, as well as ‘in excursions in other parts of Iraq, many small sites could be seen to have stucco decoration. Stuccoed buildings indicate more than private houses, but not palaces nor necessarily the residences of local rulers’.121 He added a few details on one of these sites called Abu Taraichiyah, situated 22 km east of Kish: ‘Four mounds covering an area 700 x 400 x 3. Largest, central mound shows remains of large baked brick (30 x 30 x 7 cm.) villa with stucco decoration, including trefoil and vine motifs, as well as half columns imitating palm trunks’.122 This hypothesis is supported by inscriptions on the walls of the building at Bandian which appear to refer to the marzbān Veh-Mihr-Shapur who is recorded in Armenian sources during the reign of Varahran V (420– 438).123 It is one of these complexes that Ammianus probably refers in his account of the Ctesiphon area during Julian’s campaign:

Persian landscapes: these were usurped rather than swept away’.107 The implication of this evidence is that agriculture was a crucial part of the economy. Morony has argued that self-sufficient village estates using chattel labour and managed by agents on behalf of a landed aristocracy of absentee landlords were typical of Sasanian Iraq, and that this pattern was not only common to Roman Egypt but also continued in Iraq after the Islamic conquest.108 Without ostraca – although it would be surprising if more of these did not turn up in future excavations – this is more difficult to test archaeologically, as is the entire question of the administration of the agricultural economy. However, one observation might be ventured and that concerns the so-called ‘palaces’ found first at Kish,109 Chal Tarkhan,110 Damghan,111 Sarvistan112 and, more recently, at Hajiabad,113 Bandian,114 and now Dezh-e Gri in Luristan.115 These complexes all appear to be modest in size and most consist of an iwan-plan reception hall, either roofed with a barrel vault or supported by four pillars and decorated with stuccoes, with various rooms behind and, at least in some cases, an enclosed but open courtyard in front. In the case of Bandian, one of the back rooms contained a fire-altar and a similar feature characterises the less well preserved building at Mele Heiram which the excavator therefore interprets as a ‘fire temple’.116 The style of the stuccoes varies. At Bandian they were made by cutting into the wet plaster and include scenes of combat, hunting, fire-worship and banqueting, and thus resemble murals or rock reliefs in terms of composition although no traces of pigment were observed by the excavator.117 By contrast, those from other sites in northern and western Iran, eastern Arabia and Mesopotamia were prefabricated using moulds and the decorative effect is dominated by the repetition of single motifs or portraits enclosed within heavy borders; moreover, many, if not all, of these were originally painted: at Kish, it was noted that ‘some were coloured in yellow and red’,118 and ‘on many of them the brilliant colours remain. The lion plaque, for instance, has a blue background and the beast itself

‘we encamped in a rich territory, abounding in orchards, vineyards, and green cypress groves. In its midst is a pleasant and shady dwelling, displaying in every part of the house, after the custom of that nation, paintings representing the king killing wild beasts in various kinds of hunting; for nothing in their country is painted or sculptured except slaughter in divers forms and scenes of war’.124 Archaeological sites, whether ‘palaces’ or cities, are part of a wider landscape and the study of their relationship and impact on the environment are rich subjects for research. The final section which concludes this volume looks at the evidence for material culture and what the range of evidence is for its production and circulation across the Sasanian empire.

Decker 2009: 206. Morony 1981: 166; 2004: 170. 109  Moorey 1978: 124–41. 110  Thompson 1976. 111  Kimball 1937. 112  Bier 1986. 113  Azarnoush 1994. 114  Rahbar 1998; 2004; 2007a; 2007b. 115  CAIS 2013. 116  Kaim, this volume. 117  Rahbar, pers. comm., 2001. 118  Langdon 1931b. 107  108 

Langdon 1931a. CAIS 2013. Gibson 1972: 113. 122  Gibson 1972: 138 = Kish Survey Site 84. 123  Gignoux 1998; 2008b. 124  Ammianus, History 24.6.3. 119  120  121 

191

Charcoal and Wood in Fuel Residues in Sasanian and Early Islamic Merv: Wood Resources and Environmental Implications Rowena Gale

Bachefield House, Kimbolton, Leominster, Herefordshire, HR6 0EP.

Introduction The aim of this paper is to illustrate the use of fuel woods, establish evidence of cultivated species and assess the environmental implications of irrigation farming at Merv in the Sasanian and early Islamic periods (Figure 1). It also provides an overview of the results of analyses of charcoal and wood recovered from excavations in three areas of the ancient city-site of Merv: a 4th–5th century (mid Sasanian) residential quarter in the north-east portion of the lower city of Gyaur-Kala (henceforth: MGK5) (Figure 2), a late 6th–7th century (late Sasanian) residential part of the citadel at Erk-Kala (henceforth: MEK1) (Figure 3), and a 9th century industrial workshop in Gyaur-Kala (henceforth: MGK4) (Figure 4).1 The detailed analyses will appear in the forthcoming final publications of these excavations.2 Evidence of the advance from dry to irrigated agriculture in Turkmenistan has been traced from the 6th millennium BC and suggests that, in the south, early farming was restricted to the piedmont zone close to the northern foothills of the Kopet Dagh mountain range which form the present border with highland Iran (Figures 5–6).3 Soviet scholars have argued that during the late 3rd millennium BC the Bronze Age inhabitants began to develop settlements and early irrigation systems around the edges of the deltaic fans of major rivers such as the Geoksyur and Murghab, although the date of the first settlements in these areas is still a subject of academic discussion and the large amount of sedimentological overburden in the central and upstream areas of these fluvial systems complicates matters.4 The floristic diversity of the natural vegetation prior to these ancient engineering works is unknown but the low ratio of charcoal to dung in fuel residues This paper arose from research on the excavated charcoal remains from Merv and was funded by the Wenner-Gren Foundation for Archaeological Research; it concentrates on residues obtained from the first three seasons of excavations at Merv conducted between 1992 and 1994. 2  e.g., Simpson et al. forthcoming a. 3  Lisitsina 1968; 1981; Harris and Masson et al. 1993. 4  Kohl 1984: 143–50; Hiebert 1992; Gubaev, Koshelenko and Tosi (eds) 1998. 1 

Sasanian Archaeology (Archaeopress 2022): 192–211

from excavations of Bronze and Iron Age contexts at Tahirbaj and Gonur depe suggest a paucity of wood resources.5 By the Sasanian period irrigation systems had evolved into highly complex structures, and the rich charcoal residues recovered from these periods at Merv suggest a greater availability of wood. Moreover, the charcoal analysis indicates that the artificially induced environment in the oasis supported a range of mesic arboreal/shrubby species but since their survival was linked inherently to contemporary farming practices which later fell into disuse, scant evidence of their existence remains today. Halophytic shrubs (e.g., members of the Chenopodiaceae) were identified more frequently in late Sasanian contexts and may indicate changes in soil salinity or a more widespread trade in imported wood/charcoal fuel. A survey of the present day vegetation undertaken within the oasis by Nesbitt in 1992 and 1993 outlined the extensive growth of xerophytic shrubs, particularly in areas affected by recent changes in the water table following the construction of the new Karakum canal.6 Trees are small and sparse, and the extremes of climate, drought and increasing salinity of the soil deter all but the toughest halophytes. The densely populated settlements in the oasis during the Sasanian (3rd–7th centuries) and the early Islamic (8th–9th centuries) periods probably saw the most intensive period of irrigation farming.7 The remains of almond and peach stones, as well as grape pips,8 suggest orchards, perhaps similar to those supplying the exotic fruits (including pomegranates and walnuts) which contemporary Chinese sources refer to being imported from central Asia.9 Charcoal identified as almond/ peach/apricot, grape and possibly apple, pear or quince is consistent with evidence of horticulture. In addition, a number of other mesic species were identified, including maple, ash, plane, pistachio, willow/poplar, elm, juniper and hawthorn/apple/pear/quince/Sorbus Miller 1993; Nesbitt, pers. comm. Nesbitt 1993; 1994. 7  Hiebert 1992; Simpson 2014b. 8  Nesbitt 1993. 9  Schafer 1963. 5  6 

Rowena Gale: Charcoal and Wood in Fuel Residues

Figure 1. Air view of the ancient city-site of Merv, looking east with Erk-Kala on the right (photograph: St J. Simpson, 1994)

of the charcoal examined; a few samples were selected from heavy residues for comparative purposes but provided no additional information. Two charred buds were also examined and although multiple budscales were noted on each, they were too degraded for identification.

and tamarisk. Indigenous species of these taxa grow in the mountains and foothills of the Kopet Dagh (some distance from the oasis), or in tugai. Seeking the origin of the fuel supplies therefore questions the status (i.e., natural or cultivated) of the species used. Wood and timber requirements of the populace for construction, artefactual uses (production of tools etc.), fuel for domestic hearths and, probably more significantly, for pyro-industries such as metalworking, pottery or brick kilns and other industries must have been proportionately high. The abundance of charcoal in the fuel residues suggests that wood was widely available and was probably obtained both from within and beyond the oasis; moreover, wood fuel appears to have been used in conjunction with dung cakes (Figure 7).

The charcoal was mostly firm and well preserved although some samples (e.g. (1) and (6) ) from the furnace in MGK4 were extremely hard, cinder-like and difficult to process. The fragments of uncharred wood from MEK1 (ctx 114, 238.1, 245.1) and MGK5 (ctx 12) related to different phases of occupation, separated by up to three centuries. Although very fragile, the wood was well preserved with little evidence of cellular collapse, probably due to soil salinity/desiccation. Costantini describes well preserved wooden artefacts from Bronze Age contexts at Shahr-i Sokhta in southeast Iran and refers to the sealing of contexts by a saline ‘crust’ as the major factor of preservation at that site.11

Materials and methods Charcoal was extracted from bulk soil samples by sieving and flotation (Figure 8).10 The charcoal examined was recovered mainly from floors, pits, burnt areas, industrial waste, hearths and refuse deposits. A few fragments of uncharred wood were present in refuse deposits in MEK1 and MGK5. Priority was given to samples confidently assigned to known contexts in securely stratified deposits, and particularly those suitable for comparative assessment with seeds and other plant remains. Flot samples provided the bulk 10 

Charcoal fragments measuring >2mm square in crosssection were examined. Fragments from each sample were fractured to expose fresh transverse surfaces and sorted into groups based on the anatomical features observed using a X20 hand lens. Representative fragments from each group were prepared for further examination. These were further fractured to expose tangential and radial surfaces and supported in sand. The charcoal was examined using an incident-light

Nesbitt 1993; 1994.

11 

193

Costantini 1977.

Sasanian Archaeology

Figure 2. Air view of the mid Sasanian residential quarter in Gyaur-Kala at the beginning of excavation (photograph: St J. Simpson, 1995)

Figure 3. Air view of part of a late Sasanian residence in the citadel of Erk-Kala at an early stage of excavation (photograph: St J. Simpson, 1995)

microscope at magnifications of up to X400. The same method was used to prepare the wood. Diagnostic features were matched to reference material obtained by Mark Nesbitt from the environs of the site and the Repetek reserve, and also microscope slides from the reference collection in the Jodrell Laboratory, Royal

Botanic Gardens, Kew. Atlases of wood structure by Fahn, Werker and Baas,12 Jagiella and Kürschner,13 and Parsa pajouh and Schweingruber14 were also used. Fahn, Werker and Baas 1986. Jagiella and Kürschner 1987. 14  Parsa pajouh and Schweingruber 1985. 12  13 

194

Rowena Gale: Charcoal and Wood in Fuel Residues

Figure 4. Air view of an early medieval metal workshop in Gyaur-Kala during excavation (photograph: St J. Simpson, 1995)

Figure 5. View of the lower northern slopes of the Kopet Dagh at Nysa, looking south (photograph: St J. Simpson, October 1996)

The charcoal mainly consisted of fragmented roundwood; where possible diameter widths were recorded. Heartwood, when present, was noted on tables (to be included in individual site reports) as (h); similarly, narrow roundwood (diameter 20 mm) as (s).

Measurements from charred material should be taken only as a rough guide since the volume and weight of wood from a living tree is considerably reduced during carbonisation (see Fuel) and varies according to the moisture content, cellular structure and

195

Sasanian Archaeology

Figure 6. View of the piedmont zone near Anau, looking south towards the northern edge of the Kopet Dagh and the Iranian border (photograph: St J. Simpson, April 1991)

Figure 7. Drying dung cakes as a source of fuel: this scene is in a village in Azerbaijan (photograph: St J. Simpson, 2001)

temperature when burnt,15 facts which are unknown in archaeological samples.

although in some instances, e.g., Pistacia, anatomical differences have been recorded at species level, but these can be difficult to observe and may be unstable in juvenile wood.16 Intergeneric similarities occur in some families, e.g., Salicaceae (Salix and Populus) and in some subfamilies, e.g., Pomoideae (Rosaceae) which includes Crataegus, Cydonia, Malus, Pyrus, and Sorbus. Unrelated genera can sometimes appear superficially similar, particularly when examining poorly preserved or small

Taxa identified with associated environmental and economic data A wide range of trees and shrubs were identified from the charcoal deposits and are discussed below. In general, wood can be identified reliably to genus 15 

Prior and Alvin 1983.

16 

196

Fahn, Werker and Baas 1986.

Rowena Gale: Charcoal and Wood in Fuel Residues

Figure 8. Collecting whole earth samples from secure contexts in MGK5 for flotation (photograph: St J. Simpson, 1996)

regrowth from stools in the south-west and central Kopet Dagh associated with Juniperus. A. pubescens occurs in the shibylak (a mixed shrub/small tree layer on foothills and rocky slopes) of Kugitang in the extreme south-east of the country. Maple wood is evengrained, strong and has many artefactual applications; it is also an efficient fuel. In the present day, maple (Turkmen kerkaw, Russian klyon) is grown next to the site at Collective Farm ‘Merv’, as well as elsewhere in the oasis, to provide shade on the roads and property boundaries but the wood itself is not apparently used.19

pieces of archaeological material, e.g., Ulmus and Morus. These problems are discussed in full below. Most of the taxa identified are indigenous in Turkmenistan but, since the charcoal was mainly derived from fuel residues, both indigenous species and/or cultivars may be implicated (see discussion below). The indigenous species named in the following section are considered to be those most likely to have been available.17 Broadleaf taxa:

Anacardiaceae

Aceraceae Acer L., maple. Main diagnostic features: vessel distribution diffuse porous, solitary, radial pairs or multiples; perforation plates simple; spiral thickening on vessels; intervascular pitting coarse, polygonal. In TLS rays mainly homocellular, some end cells triangular (upright in RLS).

Pistacia L., pistachio Main diagnostic features: vessel distribution ring- to semi-ring porous; vessels with tyloses, tertiary spiral thickening on narrower vascular elements; vascular tissue and axial parenchyma storied; radial resin canals present in some rays. Indigenous species include P. vera L. (height 5–7 (10) m) and P. badghysi K. Pop. (height 4–6 m); the latter is possibly a hybrid or descendent of P. vera and occurs in a single colony in the south east.20 P. vera is longliving, possibly up to 400 years, usually develops several trunks and a dense, spreading crown.21 It can withstand extremes of temperature and severe drought: its roots

Indigenous species include A. turcomanicum Pojark. (height 3–12 m) and A. pubescens Franch. (height 2–3 m). Dense forests of A. turcomanicum once clothed the slopes and valleys of the Kopet Dagh but these have been decimated by wood-cutting.18 Fragmentary stands still occur on the north-facing slopes in the west of the range (altitude 800–1700 m), and some shrubby

Akmuhammat Annayew, pers. comm. Zohary 1952; Nikitin and Gel’dikhanov 1988. 21  Browicz 1988. 19 

The classification follows Nikitin and Gel’dikhanov 1988. 18  Popov 1994. 17 

20 

197

Sasanian Archaeology penetrate the soil to a depth of 5–6 m enabling it to tap lower water levels. Young trees can survive fires by regenerating from the base of the trunk.22 P. vera is characteristic of steppe-forest, steppes or semi-desert.23 Pistachio groves were probably relatively common in subtropical semi-savanna on the low foothills in southern Turkmenistan, as shown, for example, by the remnants of large woodlands in the Badghyz region.24 Badghyz was only recently populated by humans and, as a result, the woodlands have suffered from stock grazing, wood-cutting and nut harvesting.25 Within living memory (1950/60) pistachio (Turkmen tithe) grew in the southern part of the Merv oasis at Iolatan but it is now probably rare.26

as a solitary deep-rooted (to 5 m) shrub or tree, with several layers of branches, in shallow wadis or sandy dunes.31 H. persicum communities are the most well developed of all plant associations in the desert; the understorey may include tall shrubs of Salsola richteri, Calligonum spp. and Ephedra strobilacea, smaller shrubs and subshrubs of Ephedra intermedia, Artemisia kelleri and Astragalus spp., and a herbaceous layer.32 While travelling in central Asia in the early 1920s, Krist refers to white saxaul trees growing up to 10 m in length with trunks up to 35 cm in diameter (Figure 9).33 These appeared to lie mainly buried under the sand and were particularly sought for fuel and, when burnt, the wood gave off a spicy fragrance. Since the wood burns to ash it is not possible to produce charcoal from white sexual,34 but the white wood ash was used medicinally and as a disinfectant.35 H. aphyllum is more water dependent and grows more commonly in areas with a high water table and a source of surface water.36 It can form dense forests but, once felled, these remain as low brushwood scrub up to 2 m tall.37 In areas where H. aphyllum grows sparsely it sometimes forms a mixed community with H. persicum. Black and white saxaul provide fodder for camels and sheep. Black saxaul wood is hard, heavy, difficult to cut and too brittle for most uses except fuel. It is, however, the principal fuel wood in desert regions (Figure 10) and produces excellent charcoal, the size and quality of which made it particularly desirable by blacksmiths in the late 19th century and, by inference, probably for many preceding centuries.38 When cut, the stock regenerates rapidly, e.g., over 2 m in four years. In the medieval period it was used for firing bricks.39 Suaeda arcuata makes good fuel which burns well in damp weather and is used in the oasis today, particularly in winter.40

P. vera also provides tannins (from the leaves, galls and bark), a resin used in dyeing, a heavy, durable used wood for carpentry, excellent wood fuel and high calory charcoal, and nutritious nuts.27 During the 18th century pistachio charcoal was imported from the Badghyz to the Merv oasis (Abdulla Khan Kala) to make copper items, cannon-balls, cast iron and tin.28 Today pistachio is used in the Merv oasis as a source of autumn and winter fuel.29 Chenopodiaceae Main diagnostic features: included phloem; perforation plates on vessels simple; axial elements storied. Many genera are anatomically similar but differences in the distribution of the vessels and phloem tissue allows some separation into groups. The wood is very hard and dense and with charred material it is difficult to prepare suitably flat surfaces for examination. The samples examined suggested that three or more taxa were present including Haloxylon, Salsola and/or Suaeda. The Salsola and Suaeda types were difficult to distinguish. A few fragments were clearly members of the Chenopodiaceae but did not match the available reference material.

Malvaceae Gossypium L., cotton Main diagnostic features: vessel distribution diffuse porous, solitary, in radial pairs and multiples; perforation plates simple, vessel pitting small; rays tall, 1–3 cells wide, heterocellular with long uniseriate tails. One piece of charcoal was tentatively identified as Gossypium.

There are numerous subshrubby and shrubby members of this family and many of the Turkmen species are xerophytic halophytes. Woody members tend to be less salt-tolerant than the succulent herbaceous species. For example, H. persicum Bunge ex Buhse, white saxaul (height 2–5 m) and H. aphyllum (Minke.) Iljin, black saxaul (height 7–8 m) which form independent or mixed shrub or forest communities on sandy or clay desert, or in dried-out river channels.30 H. persicum typically grows

Old World species of cotton, G. arboreum L. and G. herbaceum L., were cultivated in the Near East by Zohary 1938/39; Rustamov 1994. Rustamov 1994. Krist 1939. 34  Annayew, pers. comm. 35  Krist 1939. 36  Rustamov 1994. 37  Suslov 1961. 38  Annayew, pers. comm. 39  Nesbitt, pers. comm. 40  Nesbitt, pers. comm. 31 

Browicz 1988; Popov 1994. 23  Browicz 1988. 24  Popov 1994. 25  Popov 1994; Zlotin 1994. 26  Annayew, pers. comm. 27  Tseplyaev 1965; Costantini 1977; Browicz 1988. 28  Obelchenko 1963. 29  Annayew, pers. comm. 30  Suslov 1961; Nikitin and Gel’dikhanov 1988. 22 

32  33 

198

Rowena Gale: Charcoal and Wood in Fuel Residues

Figure 9. Saxaul trees near the tomb of Muhammad ibn Zaid at Merv (photograph: St J. Simpson, 1991)

Hellenistic and Roman times.41 These species are perennial but do not tolerate freezing temperatures. Both can be grown as annuals42 and, as such, plants attain large dimensions with woody stems. At Merv cotton could only have been grown as a summer crop and it is the main crop today (Figure 11). Oral evidence cited by Akmuhammat Annayew, describing his own childhood in the oasis at Merv, records the widespread use of cotton sticks as fuel.43 In his own house they were used for heating the sitting room, cooking food and in bread ovens. Two trailer loads collected after the cotton harvest was sufficient to fuel the family’s bread oven for the whole winter. When burnt the sticks left only a fine ash, unless the fire was small or aborted when small pieces of charcoal remained. The upper parts of the bushes were also burnt. From the 1970s, however, those who could afford to, preferred to use charcoal and, later still, charcoal was superceded by gas. Oleaceae Fraxinus L., ash Main diagnostic features: vessel distribution ring porous; latewood vessels small, solitary or in pairs, with vasicentric parenchyma; vessel pitting small; rays 1–2 cells wide, homocellular. F. syriaca Boiss., ash (height 20 m) grows in riparian forests on the lower and middle levels of the south-

Figure 10. A saxaul tree in the Karakum desert east of Merv (photograph: St J. Simpson, 1991)

Zohary and Hopf 1993; see Muthukumaran 2016. Watson 1983. 43  Annayew, pers. comm. 41  42 

199

Sasanian Archaeology Platanus orientalis L., plane or eastern sycamore (height 25–30 m) is a riparian forest tree which grows in the south-west Kopet Dagh. When mature it can form a massive canopy with wide sweeping branches. Since antiquity it has been esteemed in the East as a sacred tree,47 and it was widely planted in the Mediterranean and Near East in Greek and Roman times for its decorative value and as a shade tree.48 The wood is fairly strong but rather perishable. The decorative figuring, particularly on burr wood, is used for furniture making.49 Sargon II obtained wood from the Zagros mountains;50 plane wood was used at Tell Brak for construction and for a small carved object found at Hasanlu.51 In the Merv oasis the wood is known as platan but it is rarely used for fuel.52 Rosaceae Pomoideae: Crateagus L., hawthorn; Cydonia Miller, quince; Malus Miller, apple; Pyrus L., pear; Sorbus L. Main diagnostic features: vessel distribution diffuse porous, mainly solitary, perforation plates simple; rays 1–3 cells wide, mainly homocellular. The genera listed above are difficult to separate using anatomical methods. Prunoideae: Prunus L., plum; Cerasus Mill., mahaleb cherry; Armeniaca Scop., apricot; Persica Mill., peach; and Amygdalus L., almond. Figure 11. Detail of a cotton bush in the Merv oasis (photograph: St J. Simpson, October 1992)

Slight differences in the wood structure of some species (vessel distribution, and width and cellular composition of the rays) allow (cautious) separation into groups. The samples examined were similar in structure to the group including Armeniaca, Amygdalus and Persica. Some differences in vessel distribution and ray width were sometimes apparent, suggesting that more than one species was present, but these were inconclusive and may have been indicative of variable edaphic or climatic elements.

west Kopet Dagh.44 Archaeological evidence indicates that until the Bronze Age, F. syriaca was the dominant tree in tugai communities in the northern foothills of the Kopet Dagh.45 The wood (Turkmen dagdan) is strong, resilient and shock resistant, being excellent for tool handles, but is rather perishable. It is regarded as having talismanic properties and is reported as having been used in the construction of the dome of Sultan Sanjar’s mausoleum.46 Ash wood also provides high calory fuel and can be used green. Coppiced stumps regenerate quickly.

Main diagnostic features of Armeniaca, Amygdalus and Persica: vessel distribution semi-ring porous, vessels solitary or in multiples, spiral thickening present; rays 4–6 cells wide, heterocellular, often with sheath cells. Similarities between Eleagnus (an unrelated taxon) and species of Prunus with narrower, homocellular rays were not relevant in this instance.

Platanaceae Platanus L., plane Main diagnostic features: vessel distribution diffuse porous, perforation plates simple and scalariform, vessel pitting opposite to alternate, circular to elongated; rays mainly multiseriate, homocellular.

Several species of Crataegus form montane communities on the Kopet Dagh, mainly in association with Amygdalus. Other indigenous taxa growing in riparian Popov 1994. Pliny Natural History, 7.3. Lincoln 1986. 50  Postgate 1992. 51  Harris 1989. 52  Annayew, pers. comm. 47  48  49 

Nikitin and Gel’dikhanov 1988; Popov 1994. Lisitsina and Popov 1988. 46  Annayew, pers. comm. 44  45 

200

Rowena Gale: Charcoal and Wood in Fuel Residues grown along canals to stabilize the banks.64 The wood is light and soft and was used in ancient Babylonia for making beds, chairs, parts of doors,65 and at Shahr-i Sokhta for firewood and charcoal making,66 although other sources say the wood makes a poor fuel.67

forests in the Kopet Dagh include Cydonia oblonga Miller, Malus spp., Pyrus spp. and Sorbus spp., Prunus spp. and Cerasus spp.. Plant remains from prehistoric sites indicate that the edible fruits from many members of this family were gathered from the wild.53 Cultivation in the Old World was well advanced by the Christian era, and Islamic and 8th century Chinese texts refer to a wide range of fruits of commercial importance.54 The wood is close grained, hard and decorative. Traditional uses include cabinet-making, small artefacts (cogs etc) and turning. In southern Babylonia during the 3rd millennium BC apple wood was used for making beds, tables and chairs.55 Fruit wood burns evenly and provides a high energy fuel.

Several species of Salix grow in river valleys in the Kopet Dagh and, in Iraq, willow was often planted along canals.68 The wood is light and soft. Records of its use in central Asia and the Near East are sparse but include basket making in Iraq69 and roofing in Mesopotamia.70 Willow (Turkmen torangi) is very common in the Murghab delta and along the Amu dar’ya and reaches a height of 5 or 6 m; local uses in include roofing and tool handles; the tree is also used to line graves.71 Tamaricaceae

Salicaceae

Tamarix L., Tamarisk Main diagnostic features: vessel distribution ring- to semi-ring or diffuse porous; vessels solitary or in small multiples, perforation plates simple, vessel pitting small. Vascular elements and axial parenchyma storied; rays 2–20 cells wide. Anatomical differences between some species have been noted but since these are overlapping their use for identification is unreliable.72

Populus L., poplar Salix L., willow Main diagnostic features: vessel distribution diffuse porous; vessels solitary, in radial pairs or short radial multiples, perforation plates simple; intervascular pitting large, polygonal; vessel to ray pitting simple, tending to be angular in outline; rays uniseriate, usually homocellular in Populus and heterocellular in Salix but can be variable, particularly in juvenile wood.

A large number of woody species are indigenous and grow in a range of habitats: damp thickets with Populus, tugai, riverain forests, montane woodlands, low foothills, and flanking rivers.73 Many species are halophytic, deep rooted shrubs growing from 1–4 (5) m in height. In Iraq, tamarisk is planted on dikes and canal banks as windbreaks and to stabilize the soil.74 The light knotty wood is unsuited for most artefactual uses except crude tool handles and implements,75 and roofing.76 It is durable and has been used in recent times in Turkmenistan for sheep pens and foundations; it also makes good fuel.77 Tamarisk was identified in charcoal from hearth residues at Shahr-i Sokhta.78

Populus euphratica Olivier, euphrates poplar (height 4–8 m) and P. pruinosa Schrenk, (height 3–5 m) are riparian trees which form the principal components of the tugai56 or populetea (a riverain community formed with Tamarix, Ulmus and Fraxinus where flooding is frequent.57 P. euphratica grows as a rank smelling, spreading tree with a wide crown.58 It suckers readily and quickly forms dense thickets, often in shallow water in association with Tamarix, which can be coppiced every few years to produce fast-growing rods or poles.59 Poplar (Turkmen derek) is very common in the Tashauz region with coppiced trunks growing to a height of about 9 m.60 The trunks provide straight timbers which can be seasoned quickly without splitting once the bark has been removed.61 Since antiquity, poplar wood has been important throughout the Middle East for construction, particularly for roof beams.62 P. euphratica tolerates saline soils,63 and in Iraq poplar has, by tradition, been

Moorey 1999: 355. Mieroop 1992. 66  Costantini 1977. 67  Guest 1933; Annayew, pers. comm. 68  Guest 1933. 69  Guest 1933. 70  Moorey 1999: 348, 360–61. 71  Annayew, pers. comm. 72  Fahn, Werker and Baas 1986. 73  Nikitin and Gel’dikhanov 1988. 74  Powell 1992. 75  Costantini 1977. 76  Moorey 1999: 348, 360. 77  Annayew, pers. comm. 78  Costantini 1977. 64  65 

Zohary and Hopf 1993. Schafer 1963; Watson 1983; Zohary and Hopf 1993. 55  van de Mieroop 1992. 56  Popov 1994. 57  Costantini 1977. 58  Guest 1933; Powell 1992. 59  Costantini 1977. 60  Annayew, pers. comm. 61  Moorey 1999: 348. 62  Costantini 1977; Moorey 1999: 348, 355, 361. 63  Costantini 1977; Breckle 1986. 53  54 

201

Sasanian Archaeology Main diagnostic features: growth rings distinct; tracheids with uniseriate bordered pits; tracheid to ray pitting cupressoid; axial parenchyma present with smooth to pitted end walls. In radial view, crossfield pits cupressoid, 2–4 per field; ray cells with thick horizontal walls and smooth to nodular end walls. Resin canals absent. The anatomical features of Cupressus and Juniperus are very similar, particularly in juvenile wood.86 The geographical distribution of these genera, however, suggests that the charcoal and wood is more likely to be Juniperus than the Mediterranean Cupressus.

Ulmaceae Ulmus L, elm Main diagnostic features: vessel distribution ring porous; wide earlywood vessels with tyloses present in heartwood, narrow latewood vessels in multiples forming conspicuous tangential swags; perforation plates simple, intervascular pitting coarse, tertiary spiral thickening on narrow vessels; vascular elements and axial parenchyma storied; rays 1–6 cells wide, tall, homocellular. The structure is similar to that of Celtis (hackberry) but can be distinguished by the heterocellular rays of Celtis. Similarities occur between Morus (mulberry) and Ulmus but in Morus the rays are heterocellular with elongated, pointed end cells (best seen in tangential view).

J. turcomanica B. Fedtsch., turkman juniper (height 2–10 m) grows predominantly in the lower and mid range of the Kopet Dagh87 but can be found on the peaks and lower foothills.88 It is the most drought-tolerant of all the junipers and is long-lived (up to 1000 years). In the 19th century vast tracts of the mountain were densely populated by J. turcomanica but in recent decades these stands have been decimated by wood-cutters procuring fuel and building timbers, and by stock grazing.89 Despite its ability to regenerate from stump growth, frequent grazing or cutting of new shoots eventually kills the tree. The heartwood is reddish brown, close grained, easy to work, aromatic and very durable. The trunks from large trees provide strong timber for construction. In antiquity, juniper was considered a precious and prestigious timber.90 In present-day central Asia the wood is valued for fuel and is considered to make the best quality charcoal.91 Such was the value of juniper for charcoal-making in Uzbekistan, that in the 1920s 21 forest officers were appointed to protect the few remaining stands of trees, for timber, and they had the right to shoot unauthorised woodsmen.

U. carpinifolia Rupp. ex Suckow, elm (height to 20 m) is one of the taller constituents of riparian forests in the Kopet Dagh. It is cultivated in oases.79 Suslov states that, nowadays, oases are separated from deserts by walls of poplars and elms.80 The wood is tough, durable and difficult to work but it is, nevertheless, useful for a wide range of purposes. Elm wood burns slowly and tends to smoulder. At Merv the elm tree is known as garaagac (‘black tree’) and its only use is as fuel.81 Vitaceae Vitis L., vine Main diagnostic features: vessel distribution ring- to semi-ring porous; vessels in radial multiples, earlywood vessels wide, latewood vessels narrow; intervascular pitting scalariform, perforation plates simple and scalariform; rays multiseriate, tall, heterocellular.

Monocotyledons:

The wild vine, V. sylvestris C.C. Gmelin, clambers into the canopy of humid riparian forests of the Kopet Dagh and tugai.82 The cultivated form, V. vinifera L., was developed in the Mediterranean probably as early as the Bronze Age.83 Pliny noted that the wood was durable and used for columns and statues.84 In the present day the wood is used in the Merv oasis for heating and cooking although it is necessary to chop the wood into small pieces and season them well before use; thin branches are used in bread ovens.85

Small fragments of woody culm from monocotyledonous plants proved difficult to identify but most closely matched Phragmites australis (Cav.) Trin. ex Steud., common reed. These aquatic reeds are common in waterways throughout the Near East and central Asia growing to height of 4 m. They have been used extensively for buildings, screens, skiffs and a range of small artefacts, and as fuel and fodder.92 Unidentified samples

Conifers:

A few fragments remain unidentified and for reference their main characteristics are given:

Cupressaceae

Nikitin and Gel’dikhanov 1988. Suslov 1961. 81  Annayew, pers. comm. 82  Costantini 1977; Popov 1994. 83  Zohary and Hopf 1993. 84  Pliny Natural History, 14.1. 85  Annayew, pers. comm.

Fahn, Werker and Baas 1986. Nikitin and Gel’dikhanov 1988. 88  Popov 1994. 89  Popov 1994. 90  Meiggs 1982. 91  Krist 1939; Tseplyaev 1965. 92  Van de Hepper 1992; van de Mieroop 1992; Ochsenschlager 1992.

79 

86 

80 

87 

202

Rowena Gale: Charcoal and Wood in Fuel Residues here, Pistacia, Juniperus and Chenopodiaceae were more in evidence than in the earlier phase at Gyaur-Kala. At the time of writing comparatively few contexts have been examined from MGK4 (the site of a crucible steel production workshop) and subsequent excavation at the site has produced more charcoal. Compared with MEK1 and MGK5 a reduced range of possible tugai species was identified from MGK4. These included Poplar/Salix and Tamarix species which could have been obtained from plantations or canal banks rather than from true tugai. Wood from fruit trees, Chenopodiaceous shrubs and Pistacia was also present. Comparative analysis of fuel used in the Sasanian and Islamic periods was inconclusive owing to the relative paucity of charcoal from Islamic contexts and the disparity in use between domestic hearths of the Sasanian periods and the industrial furnace at the Islamic site. Nevertheless, some interesting issues arise as to the preferential selection of species, the origins of the fuel, the economics of fuel production, and the implications for the environment. Quantitative differences in species identified from the different periods may be attributable to preferential selection of fuel or, more probably, environmental changes following depletion of primary wood resources (see Environment).

Vessel distribution diffuse porous; vessels rather sparse, mainly solitary, some radial pairs and multiples; solitary vessels very wide; perforation plates simple; intervascular pitting small to medium, oval to elongated, crowded; axial parenchyma abundant, vasicentric and diffuse, ?storied; fibres thick-walled with narrow lumina; rays 1–?2 cells wide; probably chains of solitary prismatic crystals in strands of axial parenchyma. These fragments were present in samples (431) , (431C–D) and (437) from MEK1. The absence of a central core of pith suggests root rather than stem; diameters measured