Volume II. The Bronze Age and Early Iron Age in the Margiana Lowlands: Facts and methodological proposals for a redefinition of the research strategies 9781407302935, 9781407333007

Citation preview

BAR S1806 2008

Centro di Ricerca per le Tecnologie Multimediali Applicate all’Archeologia – Università di Bologna

THE BRONZE AGE AND EARLY IRON AGE IN THE MARGIANA LOWLANDS

Dipartimento di Archeologia – Università di Bologna

SALVATORI & TOSI (Eds)

Istituto Italiano per l’Africa e l’Oriente

B A R

The Archaeological Map of the Murghab Delta Studies and Reports

Series Editors: Annageldy Gubaev, Gennady A. Koshelenko and Maurizio Tosi

Volume II

The Bronze Age and Early Iron Age in the Margiana Lowlands Facts and methodological proposals for a redefinition of the research strategies Edited by

Sandro Salvatori and Maurizio Tosi with the editorial collaboration of

Barbara Cerasetti

BAR International Series 1806 2008

The Archaeological Map of the Murghab Delta Studies and Reports Series Editors: Annageldy Gubaev, Gennady A. Koshelenko and Maurizio Tosi

Volume II

The Bronze Age and Early Iron Age in the Margiana Lowlands Facts and methodological proposals for a redefinition of the research strategies Edited by

Sandro Salvatori and Maurizio Tosi with the editorial collaboration of

Barbara Cerasetti

BAR International Series 1806 2008

ISBN 9781407302935 paperback ISBN 9781407333007 e-format DOI https://doi.org/10.30861/9781407302935 A catalogue record for this book is available from the British Library

BAR

PUBLISHING

Editorial Note We have used for the transliteration from Cyrillic the system of the Library of Congress without diacritics, mainly to facilate the research on WEB. We would like to thank Gabriella Elina Imposti, Associate Professor of Russian Language and Literature of University of Bologna, for her suggestions to write this editorial note and to consult the book Albertazzi S. and D. Possamai (2002). Postmodernism and Postcolonialism. Proceedings of the Conference held in Bologna, October 5th, 2001. Padova. For local place-names we have used, when possible, the most accepted form in the English language archaeological literature. In other cases the Russian names were preferred instead of the Turkmen ones, because the local archaeological literature was produced in the Russian language .For abbreviations used in the text a list has been provided in the database of the archaelogical sites in the attached DVD.

Table of Contents Field Partecipants................................................................................................................................................ iv In Memoria of Iminjan Suleymanovich Masimov 1940-2006............................................................................ v Introduction.......................................................................................................................................................... x Sandro Salvatori and Maurizio Tosi Section A) Research Methods and Computer Aided Archaeology in Margiana 1. Transects and Other Techniques for Systematic Sampling.............................................................................. 1 Maurizio Cattani and Sandro Salvatori Introduction.................................................................................................................................................... 1 The Activities of the “Archaeological Map of the Murgab Delta” Project.................................................... 1 Appendix...................................................................................................................................................... 15 2. A GIS for the Archaeology of the Murgab Delta........................................................................................... 29 Barbara Cerasetti Introduction.................................................................................................................................................. 29 A First Setting of the Informative Archaeological Map: the GIS and the Remote Sensing Data................ 29 Main Research Applications......................................................................................................................... 31 Conclusions.................................................................................................................................................. 34 3. The Murgab Delta in Central Asia 1990-2001: the GIS from Research Resource to a Reasoning Tool for the Study of Settlement Change in Long-Term Fluctuations.......................................... 39 Maurizio Cattani, Barbara Cerasetti, Sandro Salvatori and Maurizio Tosi Introduction.................................................................................................................................................. 39 The Archaeological Map of the Murgab Delta............................................................................................. 40 Intrusive Camp‑site Scatters of Alien Pastoralists....................................................................................... 42 The GIS as a Tool for Model Building and Research Planning................................................................... 44 4. Non-graphic Information Systems and Diachronic Transformations in Margiana........................................ 47 Joshua Wright Introduction.................................................................................................................................................. 47 Environmental Change: Shifts in the Alluvium and Desert......................................................................... 48 Site Size Dynamics....................................................................................................................................... 48 Site Continuity across Periods..................................................................................................................... 49 Water Control and Access............................................................................................................................ 50 Self-Organised Criticality: the Nature of Intersite Connections in Margiana.............................................. 50 Method......................................................................................................................................................... 52 Discussion.................................................................................................................................................... 54 Section B) Middle and Late Bronze Age 5. The Margiana Settlement Pattern from the Middle Bronze Age to the ParthianSasanian: a Contribution to the Study of Complexity....................................................................................... 57 Sandro Salvatori Introduction.................................................................................................................................................. 57 Middle Bronze Age ..................................................................................................................................... 59 Late Bronze Age........................................................................................................................................... 63 Iron Age........................................................................................................................................................ 67 Conclusions.................................................................................................................................................. 71 6. Cultural Variability in the Bronze Age Oxus Civilisation and its Relations with the Surrounding Regions of Central Asia and Iran.................................................................................................. 75 Sandro Salvatori Introduction.................................................................................................................................................. 75 Middle Bronze Age...................................................................................................................................... 77 Architecture............................................................................................................................................ 77 Pottery.................................................................................................................................................... 78 

Anthropomorphic Terracotta Figurines.................................................................................................. 78 Compartmented Bronze Seals . .............................................................................................................. 79 Cylinder Seals (Margiana, Bactria, Indus Valley, Iran)......................................................................... 80 Metal Phials, Parade Axes, Palettes, Applicators for Cosmetics . ........................................................ 80 Chlorite Containers................................................................................................................................ 82 Chlorite Phials........................................................................................................................................ 84 Alabaster Vessels.................................................................................................................................... 87 Small Stone Columns and Rods.............................................................................................................. 88 Small Columns.................................................................................................................................. 88 Subcylindrical Stone Rods................................................................................................................ 90 Late Bronze Age........................................................................................................................................... 90 Architecture............................................................................................................................................ 91 Pottery.................................................................................................................................................... 91 Glyptic.................................................................................................................................................... 92 Conclusions.................................................................................................................................................. 93 7. Unpublished Stamp-seals from the North-western Murgab Delta................................................................. 99 Iminjan S. Masimov † and Sandro Salvatori 8. A New Cylinder Seal from Ancient Margiana: Cultural Exchange and Syncretism in a “World Wide Trade System” at the End of the 3rd Millennium BC...........................................................111 Sandro Salvatori Section C) Final Bronze 9. Excavations at Sites No. 1211 and No. 1219 (Final Bronze Age)............................................................... 119 Maurizio Cattani Palaeo-environmental Context of the Murgab Delta in the Bronze Age and the Definition of its Cultural Aspects............................................................................................................... 119 Surface Analysis and Topographical Notes within the Site....................................................................... 120 The Archaeological Excavation................................................................................................................. 121 Excavation in the Late Bronze Age Built-up Area. Site No. 1219............................................................. 126 10. The Final Phase of the Bronze Age and the “Andronovo Question” in Margiana.................................... 133 Maurizio Cattani Introduction................................................................................................................................................ 133 Method of the Research.............................................................................................................................. 133 The Environmental Context of Present-day Margiana............................................................................... 134 Types of Late Bronze Age Finds................................................................................................................ 137 Catalogue of ICW Sites.............................................................................................................................. 143 Chronology of the Finds in the Murgab Delta........................................................................................... 146 Cultural Setting: the Relationship between ICW and the Andronovo Horizon......................................... 146 Conclusions................................................................................................................................................ 148 Section D) The Iron Age

ii

11. An Aspect of the Early Iron Age (Yaz I) Period in Margiana: Ceramic Production at Site No. 999......... 153 Gian Luca Bonora* and Massimo Vidale** Introduction................................................................................................................................................ 153 Previous Research...................................................................................................................................... 153 Archaeological Back-ground...................................................................................................................... 155 The Early Iron Age Settlement Pattern in the Murgab Delta..................................................................... 156 Growth of Social Complexity: Was Wittfogel Right? . ............................................................................. 158 Chronology................................................................................................................................................. 159 Monitoring Craft Production: Ancient Kilns in the Soviet Archaeological Literature ............................. 159 Yaz I Pottery: Preliminary Descriptions..................................................................................................... 161 A Short Survey of Site No. 999.................................................................................................................. 162 The Site...................................................................................................................................................... 164 Surface Survey and Activity Areas ........................................................................................................... 164 Ceramics..................................................................................................................................................... 165 Catalogue of Pot-sherds Collected on the Surface..................................................................................... 166

Comparisons and Some Chronological Implication................................................................................... 176 Ceramic Technology ................................................................................................................................. 178 Excavation Trenches: a Trashing Ground and a Pottery-firing Kiln.......................................................... 181 Discussion ................................................................................................................................................. 186 The Kiln: Technical Features .................................................................................................................... 187 Conclusions ............................................................................................................................................... 189 12. Iron-working and Ceramic Recycling on the Surface of a Late Iron Age Fort at the North-eastern Fringe of the Murgab Delta....................................................................................................... 195 Massimo Vidale, Enrico Battistella and Giuseppe Guida Introduction................................................................................................................................................ 195 Fortifications and Craft Areas as Observed on Surface.............................................................................. 197 Iron-working Activity Areas....................................................................................................................... 198 The Archaeological Record of Iron Working: Spatial Patterning and General Remarks........................... 202 Analytical Tests.......................................................................................................................................... 204 Internal Structure of Smelting Slag, Slag Drops and Hammering Scales.................................................. 205 Quantitative Chemical Analysis................................................................................................................. 209 EDXRF Analysis........................................................................................................................................ 209 XRD Analysis............................................................................................................................................. 209 Iron Production at Site No. 172. Hypothetical Reconstruction.................................................................. 210 Recycling of Pottery by Chipping. The Evidence of Grog Production...................................................... 212 Pottery Collected on Surface...................................................................................................................... 216 Buff Ware................................................................................................................................................... 216 Coarse Chaff-tempered Containers and Walls........................................................................................... 219 Coarse Grog-tempered Ware...................................................................................................................... 219

iii

13. An Egyptian Vessel at Site No. 203........................................................................................................... 221 Sabina Malgora Introduction................................................................................................................................................ 221 Stone Vessel................................................................................................................................................ 221 Basalt.......................................................................................................................................................... 221 Literature.................................................................................................................................................... 223

Field PartcipanTs

Name Amanliev N. Annaev A. Babaev A. Bairamov A. Bakyeva O. Byashimova N. S. Bonora G. L. Carli M. Castellani V. Cattani M. Cerasetti B. Genchi F. Genito B. Gubaev A. G. Gundogdyev O. Khodzhaniyazov T. Koch P. Kurbanov A. Leoni E. Masimov I. S. † Menghi E. Muradova E. Pedrelli R. Putzolu C. Radovich A. Salvatori S. Tosi M. Udeumuradov B. N. Usai D. Vidale M. Zanfini M. Zheldakova A.

Nationality Turkmen Turkmen Turkmen Turkmen Turkmen Turkmen Italian Italian Italian Italian Italian Italian Italian Turkmen Turkmen Turkmen Italian Turkmen Italian Turkmen Italian Turkmen Italian Italian Turkmen Italian Italian Turkmen Italian Italian Italian Russian

Speciality Student Archaeologist Archaeologist Student Archaeologist Archaeologist Archaeologist Archaeologist Physicist Archaeologist Archaeologist Archaeologist Archaeologist Archaeologist Archaeologist Archaeologist Student Archaeologist Archaeologist Archaeologist Archaeologist Archaeologist Surveyor Surveyor Student Archaeologist Archaeologist Archaeologist Archaeologist Archaeologist Surveyor Student

iv

1997 1998 1999 2000 2001 2002 2005 X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X

X X X X X

X

X

X

X X

X X X

X

X X X

X

X

X X

X

X

X

IMINJAN SULEYMANOVICH MASIMOV 1940-2006

In mid-January 2006 we received sorrowful news: on January 11th, after painful and prolonged illness, Iminjan Suleymanovich Masimov died — Imil Masimov as he was called by everybody related with archaeology of Central Asia. It is difficult to express the feeling getting to know these tidings. We knew that Imil was seriously ill for a long time, but we never have heard his complaints. The main traits of Imil Masimov’s individuality — genuine fortitude, personal modesty and deep devotion to the science, obscured somehow by the critical state of his health. Indeed, he has been working as long as his last days. Now we still can fully estimate his fundamental contributions to the archaeology of Bronze Age of southern Turkmenistan. Native of the Uighur village of Chilik in Alma-Ata province of Kazakstan, after graduation in Tashkent State University he started to work in Turkmenistan where he spent all his life and scientific activities. Since 1967 he worked in the Institute of History, AS of Turkmenistan, and from a Laboratory Assistant he passed to a Senior Scientific Assistant at the Department of Archaeology. Having post-graduation from the Leningrad Department of the Institute of Archaeology (LOIA), AS USSR, in 1973 he defended his candidate dissertation. After the abrogation of the Academy of Sciences of Turkmenistan and the foundation of the Institute of History joined to the Ministry of Turkmenistan, Iminjan S. Masimov continued there his work as a Leading Scientific Assistant. In 2005, a few months before his decease, he returned to his historical mother-land. In no way this formal biography reflects Imil’s vivid personality and passion of a true discoverer. Thanks to the Chair of Archaeology in Tashkent University, Masimov was able to face the study of cronologically different sites and the archaeological systematization particularly of the most mass-produced ceramic evidence. During his post-graduated studies in the LOIA AS USSR and his researches at Altyn-depe, under the direction of Vadim M. Masson, Masimov was actively mastering and developing the techniques of archaeological excavation of extremely complicated objects from the Bronze Age sites of Central Asia with their multi-layered remains of mud-brick architecture. Joint studies with A. A. Marushchenko and A. F. Ganyalin in 1970s–1980s enabled Imil to accumulate such knowledge of particular sites in quite different nooks of Turkmenistan and of the archaeological evidence, collected in the depositories of Ashkabat that no other expert has ever possessed. Three major stages or research directions are distinguishable in the scholar biography of Masimov, and all the three are related with an extremely important problem of the origins and the development of the so-called Margiana Civilisation of the Bronze Age.



The results of his excavations of the crafts-men’s quarters at Altyn-depe and of potters’ kilns at other Bronze Age sites are reflected in his candidate’s dissertation and in the monograph “Keramicheskoe proizvodstvo epokhi bronzy v Yuzhnom Turkmenistane” (Ceramic Production in Southern Turkmenistan during the Bronze Age). On the basis of the data collected, he has demonstrated the role of pottery making as one of the most important manufactures in the structure of the economy of the early urban culture in Middle Bronze Age Turkmenistan of the late 3rd millennium BC.

The second direction was the analysis and the publication of the results of Alexander F. Ganyalin’s excavatio��� ns at Namazga-depe. This work not only opened for scientific studies the important archaeological collections and documents, otherwise under the danger of perishing, but also enabled Masimov to show the local peculiarities of the Middle Bronze Age culture of the central area in the piedmont zone of Turkmenistan. In particular, he distinguished an entire series of pottery forms and the peculiarities of Namazga-depe manufacture. Thanks to the publications we are at present able to judge not only the ceramic production and mass-produced categories of artefacts, but also the character and the peculiarity of house-building at Namazga-depe, comparing the small difference of the living room hearth‑shape of Altyn-depe and Namazga-depe. At Altyn-depe the hearths of nearly square adobe bases, located in the centre of a room, are characteristic of the Middle Bronze Age living houses. This peculiar structure continues the tradition of the Late Eneolithic period and Early Bronze Age. In the houses of Namazga-depe the hearths are predominantly located close the walls or inside them. These so-called fireplaces are one of the most important traits of Bactria-Margiana settlements, dated on the late 3rd–2nd millennium BC. Finally the third and the most significant direction of scientific activities of Iminjan S. Masimov became his “heroic” studies in Margiana. Under the conditions of water‑less desert, with miserable financing and almost without any technical means, the tiny group of enthusiast archaeologists, headed by Masimov, surveyed and mapped the most ancient oases in the Murgab River Delta: Kelleli, Egri Bogaz, Taip, Adzhi Kui and Adam Basan. The significance of these investigations is fundamental, because the major part of the lands are at present subjected to the agricultural exploitation and numerous sites are disturbed or destroyed. The discoveries and the excavations of the earliest sites in the Kelleli Oasis, together with the finds from the Egri Bogaz Oasis, constitute the basis of the modern ideas about the formation of the splendid Bronze Age Margiana Civilisation in the Murgab Delta, suggesting the beginning at the end of the 4th millennium BC. At Kelleli 4 Masimov investigated the so-called house-fort — an approximately square defensive structure with rectangular towers and inner dwellings. Later such forts, characterized by more elaborated architectural structures, became an indispensable component in the structure of central settlements in oases of Margiana and Bactria during the end of the Middle and the beginning of the Late Bronze Age. During the investigations of the Murgab settlements Iminjan S. Masimov found the first imprints of early eastern cylindrical seals. The list of the discoveries of this brilliant field archaeologist is full of striking intuition and can be continued by other scholarships. An important result of Masimov’s studies seems to be his hypothesis on the prolonged period of the establishment of the Murgab Civilisation inside Margiana on the basis of close contacts with settlements located along the piedmont area of the Kopetdagh mountains. This point of view has been proposed by Masimov in the last papers and notes (2000, 2005) and in the works of his disciple Berdy N. Udeumuradov. Based on detailed knowledge of mass evidence — pottery, statuettes and seals, and on his huge field experience, this hypothesis has been recently confirmed by some new discoveries. The first sites dated to the Late Eneolithic have been found in Uzbekistan along the routes of the ancient population of the southern Central Asia to the metal sources in the Zarafshan Valley. Iminjan S. Masimov is not only the author of over eighty scientific works, but also a talented guide of young archaeologists. Supervising graduation and term studies of students from the Turkmenian State University, he combined his exactingness with the ability to give them a huge knowledge and, above all, to captivate young people with scientific researches. Our foreign colleagues will recount us better about the international authority of Iminjan S. Masimov and his participation in the international project for the creation of the archaeological map of the Murgab Delta region. He was a modest worker, a deeply honest person, a sincere friend and an unpretentious man in every day life. Such he will ever remain in the memory o��������������� f who knew him. Nurgozel’ S. Byashimova, Liubov B. Kircho and Ruslan G. Muradov

This conclusion is confirmed by the evidence of B. A. Litvinskij’s 1952 excavations at Namazga‑depe



i

v

Scientific Works of

Iminjan Suleymanovich Masimov 1. Studies of the Bronze and Early Iron Age sites near the Railway Station of Baba-Durmaz // Kara-Kum Antiquities. Issue II. Ashkhabad, 1968, pp. 25–29. 2. Stratigraphic Exploration Pit at Kosha-Depe // Archaeological Discoveries of 1968. Moscow, 1969, pp. 440– 441. 3. Buddhist Statuette from Ak-Kala (the Middle Amu Darya River) // Soviet Archaeology. № 3. 1969, pp. 247–251. (In collaboration with V. N. Pilipko). 4. Bezymyanny (Nameless) Hill // Sites of Turkmenistan. № 1. Ashkhabad, 1970, p. 13. 5. The Craftsmen’s Quarters at the Site of Altyn-Depe // Archaeological Discoveries of 1969. Moscow, 1970, pp. 421–422. 6. Excavations of the Craftsmen’s Quarters of the Bronze Age at the Site of Altyn-Depe // Kara-Kum Antiquities. Issue III. Ashkhabad, 1970, pp. 51– 63. 7. Studies of Bronze Age Ceramic Production in Southern Turkmenistan // Papers of Sessional and Plenary Sittings of the All-Union Scientific Session devoted to the Results of Field Archaeological and Ethnographical Studies of 1970. Tbilisi, 1971, pp. 109–110. 8. Excavations of Bronze Age Ceramic Kilns in southern Turkmenistan // Archaeological Discoveries of 1970. Moscow, 1971, pp. 432–433. 9. Investigations of Bronze Age Ceramic Kilns at the Settlement-site of Ulug- Depe // Kara-Kum Antiquities. Issue IV. Ashkhabad, 1972, pp. 35–46. 10. New evidence on Bronze Age Ceramic Kilns // Proceedings of the Academy of Sciences of the Turkmenian Soviet Socialist Republic, Human Sciences Series. № 2. 1972, pp. 82–85. 11. Excavations in the Craftsmen’s Quarters at Altyn-Depe // Archaeological Discoveries of 1971. Moscow, 1972, pp. 528. 12. Excavations of the Potters’ Quarters of the Bronze Age at the Site of Altyn- Depe // Papers of Sessional and Plenary Sittings of the All-Union Scientific Session devoted to the Results of Field Archaeological and Ethnographical Studies of 1971. Moscow, 1972, pp. 281–282. 13. Excavations of Bronze Age Ceramic Kilns at Altyn-Depe // Archaeological Discoveries of 1972. Moscow, 1973, pp. 482–483. 14. Excavations of a Bronze Age Sanctuary at Altyn-Depe // Soviet Archaeology, № 2, 1973, pp. 139–145. 15. Pottery Making of a Bronze Age Early Urban Centre (according to the evidence from Altyn-Depe) // Ancient City in Central Asia. Theses of Conference Papers. Leningrad, 1973, pp. 12–13. 16. Ceramic Production in Southern Turkmenistan during the Bronze Age. Abstracts of the Dissertation for Competition for the Scientific Degree of Candidate of Historical Sciences. Leningrad, 1973, p. 16 ff. 17. Pottery Making in Bronze Age Early Urban Centre (on the basis of the Evidence from Excavations at AltynDepe) // Theses of Papers of the Session devoted to the Results of Archaeological Researches of 1972 in the USSR. Tashkent, 1973, pp. 86–88. 18. A New Excavation Area at Altyn-Depe // Archaeological Discoveries of 1973. Moscow, 1974, p. 509. 19. Surveys of Early Bronze Age sites // Archaeological Discoveries of 1973. Moscow, 1974, pp. 507–509. (In collaboration with A. A. Lyapin). 20. Ceramic Assemblage from Excavation “A1” at the Site of Namazga-Depe // Material Culture of Turkmenistan. Issue 2. Ashkhabad, 1975, pp. 39–55. 21. Finds from Gyaur-Kala in Old Merv // Sites of Turkmenistan. № 2 (20), 1975, pp. 16–17. 22. New Oasis of the Bronze Age in the Lower Reaches of the Murgab River // Archaeological Discoveries of 1975. Moscow, 1976, pp. 550–551. 23. Antiquities of the Kelleli Oasis // Sites of Turkmenistan. № 2. 1976, pp. 30–31. 24. Ceramic Production in Southern Turkmenistan during the Bronze Age. Ashkhabad, 1976, p. 112 ff. 25. Material Culture of Southern Turkmenia during the Developed Period and Late Bronze Age // Prehistoric Turkmenistan. Ashkhabad, 1976, pp. 123–152. 26. Archaeological and Surveying Works in the Lower Reaches of the Ancient Murghab River // Archaeological Discoveries of 1976. Moscow, 1977, p. 552. (In collaboration with A. A. Lyapin). 27. Excavations of the Potters’ Quarters at Altyn-Depe in 1971 // Kara-Kum Antiquities, Issue VI. 1977, pp. 6–21.

ii

v

28. The Lower Murghab in the 2nd millennium B.C. and its Development by Early Agricultural Tribes // International Symposium: Ethnic problems of the Ancient History of Central Asia. Abstracts of papers presented by Soviet Scholars. Moscow, 1977, pp. 25–27. 29. Studies of the Early Bronze Age Culture on the Lower Murghab // Archaeological Discoveries of 1977. Moscow, 1978, pp. 546–547. (In collaboration with R. Kurbanova, L. Rusakovskaya, V. Vdovin, B. Udeumuradov). 30. Finds from Shor-Depe // Sites of Turkmenistan. № 1. 1978, pp. 22–23. 31. Excavations of Living Complexes at Altyn-Depe // Kara-Kum Antiquities. Issue VII. Ashkhabad, 1978, pp. 3–13. 32. Archaeological Surveys of Sites in the Mary Province // Kara-Kum Antiquities. Issue VIII. Ashkhabad, 1979, pp. 55–62. 33. Studies of Bronze Age Sites on the Lower Murghab // Soviet Archaeology. № 1. 1979, pp. 111–131. 34. The Problem of the Development of the Lower Murghab by Early Agricultural Tribes // Proceedings of the Academy of Sciences of the Turkmenian Soviet Socialist Republic, Human Sciences Series. № 2, 1979, pp. 72–77. 35. Zoomorphic Statuettes from the Kelleli Oasis // Sites of Turkmenistan. № 1. 1979, pp. 4–6. 36. Studies of Bronze Age Sites in the Murghab delta // Advances of Central Asian Archaeology. Issue 4. Leningrad, 1979, pp. 82–83. 37. The Second Year of Excavations at the Settlement-site of Taip-depe // Archaeological Discoveries of 1978. Moscow, 1979, p. 568. (In collaboration with V. Yu. Vdovin and B. N. Udeumuradov). 38. Kelleli: a New Oasis of the Bronze Age Lower Murghab // New Researches on Archaeology of Turkmenistan. Ashkhabad, 1980, pp. 3–21. 39. Typological and Spatial Analysis of Ancient Settlements // Techniques of Archaeological Researches and Regularities in the Development of Ancient Societies. Theses of Papers, Ashkhabad, 1980, pp. 40–43. 40. Excavations at Taip-depe I // Archaeological Discoveries of 1979. Moscow, 1980, pp. 464–465. 41. New Finds of Bronze Age Seals from the Lower Murghab River // Soviet Archaeology. № 1. 1981, pp. 132– 150. 42. Seals from the Ancient Murghab Oasis // Sites of Turkmenistan. № 1 (31), 1981, pp. 18–20. 43. Excavations at the Kelleli Oasis // Archaeological discoveries of 1980. Moscow, 1981, pp. 466–467. 44. The Study of Bronze Age Sites in the Lower Murghab // The Bronze Age Civilization of Central Asia. Recent Soviet Discoveries, (ed. by Ph. L. Kohl). Armonk; New-York, 1981, p. 194–220. 45. The First Soviet-American Symposium on the Problems of Archaeology of Central Asia. Near East and Ancient India // Proceedings of the Academy of Sciences of the Turkmenian Soviet Socialist Republic, Human Sciences Series. № 5. 1982, pp. 94–96. 46. The Kelleli Oasis and Problems of the Cultural Genesis in Margiana during the Bronze Age // All-Union Symposium: Cultural Progress during the Bronze and Iron Ages. Theses of Papers. Yerevan, 1982, pp. 53–55. 47. Bronze Age Culture on the Lower Murghab and its Connections with Bactria // The Earliest Cultures of Bactria. The Medium, Development, Connections. Theses of Papers. Dushanbe, 1982, pp. 73–75. 48. Soviet-American Symposium // Sites of Turkmenistan. № 2. 1982, pp. 22–23. 49. New Sites of the Early Iron Age in the Murghab Oasis // Recent Archaeological Discoveries in Turkmenistan. 1982, pp. 20–33. 50. Types of Ancient Settlements and Social and Economical Reconstruction of Ancient Societies (by the Example of Sites in the Murghab Oasis) // Information Bulletin of the International Association for the Study of the Cultures of Central Asia (IASCCA). Issue 2. Moscow, 1982, pp. 76–78. 51. Recent Researches of Bronze Age Sites on the Murghab River // Archaeology of Central Asia and Near East. 2nd Soviet-American Symposium. Theses of Papers. Samarkand, September 6–12, 1983. Tashkent, 1983, pp. 72–74. 52. The Culture of the Lower Murghab River in the Bronze Age and its Connections with Bactria // Information bulletin of the International Association for the Study of the Cultures of Central Asia (IASCCA). Issue 4. Moscow, 1983, pp. 63–64. 53. Excavations in the Kelleli Oasis // Archaeological Discoveries of 1981. Moscow, 1983, pp. 476–477. 54. Living House on the Kelleli // Sites of Turkmenistan. № 2. 1984, pp. 16–18. 55. New Evidence on the Early Iron Age on the Lower Murghab // Turkmenistan in the Early Iron Age. Ashkhabad, 1984, pp. 12–27. (In collaboration with B. N. Udeumuradov). 56. Excavations in the Kelleli Oasis // Archaeological Discoveries of 1982. Moscow, 1984, pp. 488–489. 57. Excavations of Sites in the Ancient Murghab Delta // Archaeological Discoveries of 1983. Moscow, 1985, pp. 552–553. 58. Le Bassin du Murghab à l’Age du Bronze et ses relations avec la Bactriane // L’archéologie de la Bactriane ancienne. Paris, 1985, pp. 223–228.

iii

v

59. Excavations of the Fortified Settlement of Kelleli 3 // Archaeological Discoveries of 1984. Moscow, 1986, pp. 472–473. 60. Recent Investigations of Bronze Age Sites on the Murghab River // Ancient Civilisations of the East. Tashkent, 1986, pp. 171–181. 61. Innovations on the Lower Reaches of the Murghab River during the Bronze Age // Technological and Cultural Progress during the Early Agricultural Epoch: Theses of the Conference. Ashkhabad, 1987, pp. 59–60. 62. Excavations at Kelleli 3 (TSSR) // Archaeological Discoveries of 1985. Moscow, 1987, pp. 605–606. 63. New Evidence for the Problem of Connections of the Lower Murghab Area with India and Baluchistan in the Bronze Age // Proceedings of the Academy of Sciences of the Turkmenian Soviet Socialist Republic, Human Sciences Series. № 2. 1988, pp. 27–33. 64. Bronze Age Terracotta Sculpture from the Kelleli Oasis // Ancient and medieval archaeology of Central Asia. Tashkent, 1990. 65. The Caravanserai of Minar // Proceedings of the Academy of Sciences of the Turkmenian Soviet Socialist Republic, Human Sciences Series. № 6. 1990, pp. 38–42. (In collaboration with J.. Durdyev). 66. The Kelleli Oasis in the Bronze Age // Merv in the Ancient and Medieval History of the East. Theses of Papers. Ashkhabad, 1990, pp. 21–23. 67. Once again about the Medieval Caravan Route from Merv to Khwarezm // Sites of Turkmenistan. № 1. 1990, pp. 12–13. (In collaboration with J. Durdyev). 68. The Murghab Valley: Interaction of the Society and Nature in the Bronze Age // History of Interaction between Society and Nature: Facts and Conceptions. Theses of Papers. Moscow, 1990. 69. The Murghab Basin in the Bronze Age // Merv in the ancient and medieval history of the East. II. Theses of Papers. Ashkhabad, 1991, pp. 15–17. 70. Structuring of Oasis Systems and the Cultural and Historical Processes in Margiana during the Bronze Age // Social-space Structures in the Stage Characteristics of Cultural-historical Processes. Theses of Papers on the Interregional Conference, Moscow, 1992, pp. 80–84. 71. Metalworking in Southern Turkmenistan in the Bronze Age // Theses of Papers, Dushanbe, 1993. 72. Neolithic Hunters on the Murghab River // Merv in the ancient and medieval history of the East. Ancient Merv — Modern Mary. Mary, 1994 (co-author). 73. Cultural and Historical Processes in Margiana during the Bronze Age. Ibid. 74. Archaeology of Turkmenistan in 1990s // Collection of articles dedicated to M. E. Masson’s 100th birthday. Tashkent, 1997. 75. Preliminary Analysis of the Bronze Age Material Collected by the Margiana Archaeological Project and a First Chronological Assessment // The Archaeological Map of Murghab Delta. Preliminary Reports 1990–95. Rome, 1998, p. 35–46. (In collaboration with S. Salvatori and B. Udeumuradov). 76. Ancient Cultural Links between Turkmenistan and India According to Archaeological Evidence // India and Turkmenistan: “Living Heritage”. New- Delhi; Ashgabat, 1998. 77. Origins of the Ancient Murghab Civilisation // Cultural Heritage of Turkmenistan (Remote Origins and Modern Perspectives). Materials for International Scientific Conference. Ashgabat; Saint-Petersburg, 2000, pp. 25–26. 78. Seals from Excavations of the American Expedition in Anau // Neutral Turkmenistan, 09.10.2002. 79. The Silk Road and Problems of Interrelations between Ancient Civilisations // The Dialogue of Civilisations. Bishkek, 2002. 80. Traces of the Anau Culture in China // R. Pumpelly and the Development of the Archaeological Science of Turkmenistan. Theses of Papers. Ashgabat, 2003. 81. Metallic Seals from Ancient Anau // Neutral Turkmenistan, 2003. 82. History of Anau in the World Science // Neutral Turkmenistan, 11.09.2003 (co-author). 83. R. Pumpelly Expedition Researches at the Merv Oasis // Turkmenistan is a Nativeland of Anau Culture and White Wheat (Ak Bugday) (to the 100th Anniversary of Archaeological Expedition of the American Scientist Raphael Pumpelly to Turkmenistan). Abstracts of Reports of the International Scientific Conference (October 22–23, 2004). Ashgabat, 2004, pp. 87–88. 84. Prehistory of the Discovery of the Margush Civilisation // Renascence. № 7. 2005.

ix

Introduction Sandro Salvatori and Maurizio Tosi Sarianidi’s excavations of Gonur and Togolok have proven beyond doubt the great wealth and high standards of political complexity attained by the people who were settled across the irrigated lands of the Murghab Delta during the second half of the 3rd millennium BC. The parallel evidence gathered from the archaeological surveys, we have carried out since 1990, has indicated that the primary agents of that great cultural expansion were land, labour and water. The Murghab Delta was in no way different from all other cradles of civilization that blossomed throughout the Middle East and South Asia in the Early Bronze Age: both wealth and complexity were based on vast reserves of food provided by the combined exploitation of natural and human resources. The wealth of those alluvial settings by intersected steppe and swamps like the Nile or the Indus Valleys was in no way richly endowed by nature. We are all aware, since Adam Smith, that the wealth or poverty of nations bears little relationship to the abundance in minerals or fertile soils. The critical factor is rather the application of human labour to the natural potentials: more work more food, more food more people and more people more work, provided they learn the rules of sharing (e. g. Paterson 1976: 26). In fact this productive capacity required, then as now, the synergies of social alliances, hence the great significance of monumental buildings, shrines, ornaments and ritual performances in all spheres of communal life. The magnificent material life of Gonur and the other centres within the Murghab Delta stands as a direct witness of the elites’ political skills in harnessing the labour of many. In all arid lands human labour was the dynamic nexus to combine water and soil for subsistence and wealth, while the food produced was the essential reward that cemented society together. To irrigate thousands of hectares along the radiating sedimentary slopes of the Murghab Delta required immense manpower engagements: thousands of individuals had to be engineered in cohorts and brigades to direct the river’s flow into a web of channels, partly natural and partly artificial. The result was the creation of a continuous alluvial plain, a man-made landscape that during the Middle Bronze Age integrated within a single agrarian territory all towns and villages. We do not know yet whether they all belonged to a single political activity, but with a total extension of 25-30,000 sq. km. Margiana was comparable with the other agrarian homelands of civilisation. This landscape continuity was first indicated by the lack of sands in all geo-morphological sections discussed in our first report by Cremaschi (1998) and later confirmed by the systematic transects of intensive survey work presented here in Chapter 1. When emphasizing the critical importance of collective labour in the establishment of a vast irrigation system there is no need to quote K. Wittfogel’s debated propositions on early hydraulic societies. The archaeological evidence from all arid lands that housed the formative processes of urban civilization has repeatedly confirmed the primacy of irrigation works to the growth of nucleation and accumulation (e.g. Adams 1982). Wittfogel’s ideas on the critical priority of water management can hardly be regarded as a hypothetical proposition anymore. The Murghab Delta was certainly no exception to this rule and hydraulic engineering had to be the principle factor behind the expansion of agriculture in the Bronze Age and a principal decisional factor for the polities that shaped what we would like to call henceforth the Earliest Central Asian Civilization in Margiana in the lands around the southern edge of the Oxus Basin, postponing any debate whether they developed locally from Chalcolithic communities or had moved there in a massive migration from the west at the end of the 4th millennium BC.



By summing up the surface of all Middle Bronze Age sites we may reckon that between 60,000 and 100,000 people lived in the Murghab Delta during the second half of the 3rd millennium BC, not to include the more mobile pastoral segments that exploited the surrounding steppes for stock-breeding. It is a very large population, enriched by vast food resources that dynamically acted in all sectors of economic production and social exchanges. The archaeological record of crafts and industries generously indicates that not all labour in the Murghab cradle was exhausted by the many facets of irrigation agriculture. Wood, basic metals like copper or lead, and other minerals albeit scarce in the delta lowlands, were abundant at close distance to the south, in the mountains of Iran and Afghanistan, at few days of walk upstream, along the gentle slopes of water courses. Moreover a vast network of medium- and long-distance exchanges supplied the local craftsmen with materials from distant lands: lapis from Badakshan, tin from the Zeravshan, ivory from India, shells from the Arabian Sea. Along with the materials came objects for ornament, game pieces and the prestige displays most characteristic of those cultures: seals from Mesopotamia; again seals, dice, pawns, combs and beads from the Indus Civilization; shells, carved steatite vessels from the Magan’s shores between Oman and Baluchistan. So high is the quality of the craftworks emerged from Gonur and other towns in the delta, so elaborated their artistic expressions that we expect the Murghab heartland

would have exercised a deep and lasting cultural influence in all directions throughout Eurasia during and after its flourishing age. At the beginning of the 2nd millennium BC economic growth and expansion came to an end: the balance among people, land and water was upset by environmental and political changes. With the sand from the North penetrated the Tazabag’yab stock-breeders who pushed southwards the Andronovo techno-complex from the Eurasian steppes. The nature of their relations with the established farmers and their settlements across the shrinking alluvial plains remains a major target of our work in the Murghab Delta. The recent discovery of more permanent settlements of this period in the southern sections of the delta, with substantial deposits and buried structures, will probably disclose a better understanding of their cultural and political interconnections (see Chapter 9). This second volume of the Archaeological Map of the Murghab Delta represents a further step forward from the original work of the Russian-Turkmen-Italian Joint Project and it is an even representation of the work carried out from 1996 to 2005, when a new protocol has been signed by the “Istituto Italiano per l’Africa e l’Oriente” and the “State Institute of Cultural Heritage of the Peoples of Turkmenistan, Central Asia and the Orient under the President of Turkmenistan”. The presented papers somehow close the first part of our Project, originally designed in 1990-1991 to prepare an archaeological map that would include in a single digital frame all sites across the Merv Oasis, i.e. the eastern branch of the Murghab Delta, the ancient land of Margush. From 1995 onwards, the work of our Project has been chiefly directed to enlarge our knowledge in the northern sections of the delta where the largest number of Bronze and Early Iron Age sites is presently visible. While we carried out year after year the gargantuan work to explore this huge land piece of over 20,000 square kilometres fragmented by entrenching canals and moving sand dunes, two major excavation projects were taking place, carried out by other teams: to the south the British Archaeological Expedition at Merv headed by G. Herrmann and T. D. Williams, and to the north the joint RussianTurkmen Project at Gonur, under the direction of V. I. Sarianidi. Unfortunately, apart from courtesies and sporadic logistical assistance, little was exchanged in terms of critical and diagnostic information, particularly for what concerns the dating of the sites detected by the survey. While the work at Merv was concerned, with periods almost unrepresented in the lowlands north of “Antiochus Wall”, the emphasis on architecture exposure in Gonur’s excavations minimized the recovery of stratigraphical sequences for the relative dating of pottery variability. As a result we focussed on the exploration of sites and canals to develop a better definition of regional subdivisions within the sub-fossil alluvium. With few adjustments the survey work has been carried out along the same lines of the first 1990-1995 phase, considerably improving the technical equipment for site positioning and mapping. Systematic transects, a method of investigation first developed by T. J. Wilkinson for arid lands in the Near East (e.g. 1982; 2001), have proved quite effective to detail fluctuations in farmland extension and density of settlement. The use of CORONA and SAR radar satellite imagery, combined with early aerial photographs and maps, has greatly improved our capacity for detection and measurements of sites (see Chapters 2 and 3). A large number of transects (see Chapter 1) have been walked in areas previously randomly visited. To the eastern sector east-west oriented transects were walked across deserted sandy areas where a large number of Final Bronze Age campsites have been located, suggesting an increase of sharp pastoral activities well after 2000 BC. Particular attention has been directed to a group of Iron Age sites recorded in the previous publication as “site complexes” (see Chapters 1 and 5). This group of compounds was formerly described as huge shallow dispersions of pottery covering several hectares. Now we know that they represent series of small farm-houses scattered along the course of active river branches, possibly fed from dams and other complex water management systems built further south at the height of the delta’s hub in Iolotan (see Chapter 2). The Bronze Age settlement pattern has been studied in details (see Chapter 4 and 5) although we are well aware that the excavation methods applied so far in the Murghab Delta have missed to provide the detailed chrono-stratigraphy necessary to fully understand the complex dynamic between settlement, social organization and environmental fluctuations. In spite of these constrains we trust that new light has been thrown on critical periods as the Middle (see Chapters 5 and 6) and Final Bronze Age along with the massive evidence for the spreading of northern pastoralists across the Murghab Delta farmlands (see Chapter 10). Without a finer mesh of relative dating and the targeted use of geo-archaeological techniques for sedimentary analyses there are little, means to improve, unless changing the founding paradigms of the settlement survey. In fifteen years of continuous fieldwork we have learned from our own work that the relative stability of the Murghab Delta during the Holocene, particularly if compared with the erratic history of the Helmand in Sistan or the Tarim

i

x

in the Takla Makan, may have covered many sites under the coat of silt, in some points 4-5 meters thick. The Karakum sands moving from the north have hidden a vast land pieces, hindering direct scrutiny for prehistoric settlements along the earlier fan of delta branches radiating to the north of the continuous alluvial plains. In a number of cases over the years we attempted to develop ourselves more reliable templates of ceramic sequences for a better dating of the surface evidence. In 1992-94 excavations were carried out at the main mound of Takhirbai 1, under the direction of Maurizio Cattani, to cover the transition between the Late and Final Bronze Age. Limited excavations at Middle and Late Bronze Age sites (Adzhi Kui 1-9) were carried out in 1997 (Salvatori 2002), while from 2002 onwards the steppe-like Sites No. 1211 and 1219, marked by the characteristic Incised Coarse Ware (ICW), have been more exhaustively explored (see Chapter 9). Also Iron Age sites characterized by productive installations for ceramics and metal working have been intensively studied between 2000 and 2005 (see Chapters 11 and 12). New information on stamp and cylinder seals from the Murghab Delta are provided (see Chapters 7 and 8), along with a more general survey highlighting similarity and differences within the Margiana Bronze Age sequence and more generally inside the larger Middle Asia intercultural spheres incorporating the whole or parts of the Early Central Asian Civilizations (see Chapter 6). The constant upgrading of the GIS oriented study of the Murghab Delta has been one of the main efforts in order to comply with the original targets of the Project: the precise mapping of all archaeological sites in the Murghab Delta (see Chapters 2 and 3). Once accomplished the original aims of the Project, future research demands we move in different directions. We are aware that likewise elsewhere in the world the ultimate response will come from long-term excavation programs mobilizing the multidisciplinary spectrum of intersecting sciences. Stratigraphic controlled excavations at key sites like Gonur, Togolok and Taip would provide information on the transition from Middle to Late Bronze Age and a more credible pottery sequence for the Late Bronze Age than the actual one, based on grave goods later than the Late Bronze Age architectural features encapsulating them. By all means a more detailed knowledge of this period is needed to reconstruct the interactions between cultural transformations and the environmental changes, evidenced by the massive covering of sands on large portions of the delta. A further problem is that laid by the intriguing phenomenon of a well and fully developed Middle Bronze Age settlement pattern and the persisting lack of Early Bronze Age and Chalcolithic remains in the area preceding them. Although aware that “absence of evidence is no evidence”, even after many years of walking that sedimentary plain, we had to come across the scantiest remains of buried sites. Since the few indications presented by Kohl (1984: 146) and rejected by P’yankova (1994: 363), very little additional facts have been gathered to solve this crucial point. The crux remains between the extremes of a migration of an entire population in the area during the second half of the 3rd millennium BC and an indigenous development from Chalcolithic and Early Bronze Age groups. The answer ought to be somewhere along the continuum of variants between the propositions at loggerhead: the delta had to be inhabited at least since the Early Holocene, no matter whether at low density by mixed foragers alike the Kel’teminars of Khoresmia or by a network of nucleated farming settlements like Geoksyur in the Tedzhen Delta. The fact is since the beginning of explorations in the late sixties and more than 20 years after the publication of Kohl’s overview presentation we can still add very few new data in support of the hypothesis of a local development from Chalcolithic communities, comparable with the one that had occurred along the Etek piedmont plain and across the Tedzhen alluvium. The first piece of evidence was a single sherd of black burnished ware from a globular jar decorated with two applied parallel clay ribbons and a band of oblique parallel burnishing lines closely resembling the Hissar Tepe tradition, coming from Stratigraphic Unit 5 at Adzhi Kui 9 (Salvatori 2002: Figs. 68.10; 69). It points to a date in the first half of the 3rd millennium BC. The second one is a C-14 date from inside a storage pit dug into the basal layer at the same site (Ibidem: 119) which instead points to a late 4th millennium BC date. More significant, because it would be a direct evidence, is the presence of a Chalcolithic layer with traces of mudbrick architecture brought to light by Sarianidi below the external city-wall at Gonur North (Sarianidi 2005: 35). Although minimized by Sarianidi, it is the first piece of evidence of a Chalcolithic settlement in the Murghab Delta. However it urges us to further investigate the origins of the Middle Bronze Age civilization and the consistency ��������������������������������������������������������������������������������� We regret that the study of the materials found is still waiting for completion.



ii

x

of the Chalcolithic and Early Bronze Age population in the area. This remains one of the limits of a project based on surface evidence, no matter how careful, systematic and intensive. Nevertheless geophysical prospections by magnetograms and geo-radar might disclose greater opportunities for further non-destructive explorations around the main sites and across the northern reaches of the delta hidden by the coat of shifting sand. In order to establish the potentials of these means in October 2007, we have carried out a first run of magnetometric measurements around some of the main Bronze Age sites with the aid of a leading geophysicist like Helmut Becker. We ought to begin moving toward a different paradigm that require new means of investigation. Middle and Late Bronze Age sequences are not yet properly known. The most serious attempt to establish a sequence of the Late Bronze Age of the area has been carried out by L. P’yankova (1989; 1993) using the pottery from the graveyards excavated by Sarianidi at sites like Togolok 1 and 21. Unfortunately these graveyards cut into the structures and deposits of earlier buildings and represent only post-settlement phases. They are later than the impressive palace-like structures exposed at those sites. Incidentally the same has happened with the excavated Surkhandar’ya sites in Northern Bactria, Sapalli and Dzharkutan (Askarov 1973; 1977; Askarov and Abdullaev 1983): we actually know important architectonic features of the Late Bronze Age Margiana settlements, but we can hardly recognise the correlated material assemblages (Sarianidi 1998). This situation strongly hinders the analysis that could be otherwise conducted on survey data and does not allow a proper understanding of the process which leads to the demographic collapse recorded between the Late and Final Bronze Age. We deeply regret that the great chance we still had at the beginning of the nineties of the last century to study in detail the cultural evolution of an urban complex society of the Bronze Age in the Murghab Delta has been greatly compromised. The present day impressive expansion of agricultural exploitation across the Murghab Delta, with the destruction of hundreds of ancient settlements, commands new research strategies eventually oriented to the excavation of some of the largest multi-stratified sites like Togolok 1 and Taip 1 using more attentive excavation techniques targeted to the chronological seriation of material cultures. The task seems to be very difficult and we think, as suggested above, that only a large international investment for long-term explorations would possibly help to complete the picture of the early history of mankind in this critical hub in southern Central Asia. Finally, we like to offer this book to the memory of Iminjan S. Masimov that shared with us, during the many years of the project, his invaluable knowledge of the region. We have not to recall his well known role in Central Asian archaeological research, but his human kindness and generosity that will be engraved forever in our memory. Emil was a champion of Adamcilik, the beautiful Turkic expression we learn as soon as our hearts and minds were directed to the great Central Asian heritage. Acknowledgments The Russian-Turkmen-Italian Archaeological Project has been financed by the Italian Ministry for Foreign Affairs, the Italian Institute for Africa and the Orient (Is.I.A.O., former Is.M.E.O.), the University of Naples “L’Orientale” (former I.U.O.N.: Istituto Universitario Orientale di Napoli), the University of Bologna (UNIBO) and through the research funds of the Italian Ministry of University and Research (Co. Fin. ex 40% quota). We would like to thank the President of the Is.I.A.O., Prof. Gherardo Gnoli, for his constant support and the Turkmen Authorities and Institutions for their precious help throughout this Project. The list of the people to be thanked is really rather long, and to be followed by an acknowledgement of numerous debts contracted by the Mission in the course of many seasons of work. Furthermore the Editors wish to thank all the members of the Project for the contribution they have offered under various titles during the course of these years. In particular we like to acknowledge the precious help provided by Aydogdy Kurbanov in the spelling of Russian and Turkmen place-names and by Paolo Mozzi for his suggestions about ancient irrigation evidences. A big debt has been contracted with Barbara Cerasetti for her assistance in following all the stages in the making of the present book: without her dedication it would have been an even longer process. Extensive excavations of key sites have been conducted only by V. I. Sarianidi. Admittedly a lot of information and vast collections of materials have been recovered in few years, but the methods used can hardly be considered accurate. The excavations of settlement remains at Gonur North have been carried out by artificial cuts of approximately 35-40 cm and different building phases could not be recorded, along with the relative pottery assemblages. The same can be said about Togolok 1, Togolok 21 and the Adzhi Kui 8 excavations. No information has been published about the large Late Bronze Age graveyard excavated on top of the Gonur North mound, evidently related with the Gonur South fortified settlement (See for a critical evaluation C. C. Lamberg Karlovsky 2003. A reply by V. I. Sarianidi has been published in Sarianidi 2007: 186-195) 

iii

x

We have to acknowledge the SEPS (Segretariato Europeo per le Pubblicazioni Scientifiche) which has generously funded the translation into English and the language revision for most of the contributions. Finally, we would like to thank Joshua Wright for his kindness to review and correct parts of the texts and Gian Luca Bonora for reviewing the Russian bibliography. References Cited Adams, R. McC. 1982. “Die Rolle des Bewässarungbodenbaus bei der Entwicklung von Institutionen in der altmesopotamische Gesellschaft”, in Hermann and Sellnow (eds) 1982: 119-140. Askarov, A. A. 1973. Sapallitepa. Tashkent. Askarov, A. A. 1977. Drevnezemledel’cheskaya kultura epokhi bronzi yuga Uzbekistana, Tashkent. Askarov, A. A. and B. N. Abdullaev 1983. Dzharkutan. Tashkent. Brothwell, D. R. and A. M. Pollard (eds) 2001. Introduction to Archaeological Science. London. Cattani, M., B. Cerasetti, S. Salvatori and M. Tosi 2003. “The Murghab Delta in Central Asia 1990-2001: GIS from a Research Resource to a Reasoning Tool for the Study of Settlement Change in Long-Term Fluctuations”, in Doerr and Sarris (eds) 2003: 125-131. Cremaschi, M. 1998. “Palaeohydrography and Middle Holocene Desertification in the Northern Fringe of the Murghab Delta”, in Gubaev, Koshelenko and Tosi (eds) (1998): 15-25 Doerr, M. and A. Sarris (eds) CAA 2002, The Digital Heritage of Archaeology. Computer Applications and Quantitative Methods in Archaeology. Proceedings of the 30th Conference. Heraklion, Crete (Greece), April 2002. Heraklion. Gubaev, A., G. Koshelenko and M. Tosi (eds) (1998). The Archaeological Map of the Murghab Delta. Preliminary Reports 1990-95. Reports and Memoirs. Series Minor, III. IsIAO. Roma. Hermann J. and I. Sellnow (eds), Produktivkräfte und Gesellschaftsformationen in vorkapitalistischer Zeit, Berlin. Lamberg Karlovsky, C. C. 2003. “Civilization, State, or Tribe? Bactria and Margiana in the Bronze Age”, The Review of Archaeology 24 (1): 11-19. Kohl, Ph. L. 1984. Central Asia. Palaeolithic Beginnings to Early Iron Age. Paris. Paterson, J. H. 1976. Land, Water and Resources. An Introduction to Economic Geography. London. Pyankova, L. 1989. “Pottery Complexes of Bronze-Age Margiana (Gonur and Togolok 21)”, IASCCA Information Bulletin 16: 27-54. P’iankova, L. 1993. “Pottery of Margiana and Bactria in the Bronze Age”, IASCCA Information Bulletin 19: 109-127. P’yankova, L. 1994. “Central Asia in the Bronze Age: Sedentary and Nomadic Cultures”, Antiquity 68 (259): 372-387. Salvatori, S. 2002. “Project “Archaeological Map of the Murghab Delta”(Turkmenistan): Test trenches at the Bronze Age Sites of Adzhi Kui 1 and 9”, Ancient Civilizations from Scythia to Siberia 8 (1-2): 107-178. Sarianidi, V. I. 1998. Margiana and Protozoroastrism. Athens. Sarianidi, V. 2007. Necropolis of Gonur. Athens. Sarianidi, V. I. 2005. Gonurdepe. City of Kings and Gods. Ashgabat. Wilkinson, T. J. 1982. “The Definition of Ancient Manured Zones by Means of Extensive Sherd-Sampling Techniques”, Journal of Field Archaeology 9: 323-333. Wilkinson, T. J. 2001. “Surface Collection, Field Walking, Theory and Practice, Sampling Theories”, in Brothwell and Pollard (eds) 2001: 529-541.

xiv

Chapter 1

Transects and Other Techniques for Systematic Sampling Maurizio Cattani and Sandro Salvatori

Introduction

In Margiana, besides the previously mentioned ItalianRussian-Turkmen undertaking, in 1991 a collaboration was started between the Centro Studi e Ricerche Ligabue (CSRL) of Venice, the Ministry of Culture of Turkmenistan and the Institute of Archaeology of the Academy of Sciences of the Soviet Union (later of Russia), within V. I. Sarianidi’s archaeological mission in the archaeological district of Gonur. The mission of the CSRL, led by S. Salvatori, was oriented to the identification and excavation of the Middle Bronze Age graveyard related to the large settlement of Gonur North (Salvatori 1993, 1994, 1995). Subsequently the excavation of the burial ground went on, up to the present time, under the leadership of V. I. Sarianidi and B. N. Udeumuradov, with the co-funding of the CSRL and above all of the Greek Ministry of Culture and of other Greek private institutions.

The prehistoric research in the Murgab Delta, the Margiana of the historical era, has seen three great seasons of research whose characteristics were very different, both because of their methods of investigation and because, during the third stage, the area was opened to European and North-American researchers. The first chapter of the pre-protohistoric research into the Murgab Delta is connected with the work carried out, in the second half of the nineteen-fifties, by V. M. Masson (1959): though focused on the southern part of the delta and on the Iron Age sequence (excavations in the large, pluristratified settlement of Yaz-depe), this investigation brought to light more ancient settlements dating back to various phases of the Bronze Age, in the southernmost parts of the delta, such as the sites of the area of Adam Basam and Takhirbai.

In the Medieval capital of the region, the ancient Merv, which was established around the Achaemenid fortress of Erk-kala and the Greek-Hellenistic city of Gyaur-kala, from 1992 there began a British archaeological mission of the University College of London, led by G. Herrmann (Iran, 1993-2000) and by T. D. Williams from 2001 up to now.

A second campaign carried out in the nineteen-seventies and eighties included more systematic surveys in the northern parts of the delta, the mapping of hundreds of Bronze Age sites, and their excavation both by Russian scholars (V. I. Sarianidi) and by Turkmen ones (I. S. Masimov). At the end of the nineteen-eighties, there came to a head the work of some American, French and Italian scholars who had been holding meetings and exchanges with their Soviet colleagues for several years (Deshayes 1977; Kohl 1981). In this new climate of exchanges between the Soviet archaeological institutions and the Western world, in 1990 an international project was started: the “Archaeological Map of the Murgab Delta”, under the leadership of A. G. Gubaev, G. A. Koshelenko and M. Tosi, respectively representing the Turkmenistan State University, the present Institute of Archaeology of the Russian Academy of Sciences (IARAS) and the Is.M.E.O. (now Is.I.A.O.).

The Activities of the “Archaeological Map of the Murgab Delta” Project The Murgab Delta has become a completely unusual archaeological laboratory, as a result of the increasing interest of the Turkmen administration in extending the agricultural areas of the country wherever possible. The researches carried out during more than a decade, from 1990 to the present time, have coexisted with, and had to adapt to, the deep environmental change brought about by the expansion of cultivated land: the strategies and methods of archaeological field work had to change gradually. During the very first years of work, only the southern part of the delta had been put to crops, basically copying the agricultural areas of the Parthian-Sasanian Era and then of Medieval times, around the large centre of Merv.

In the meantime, several other Western archaeological enterprises were started in the Central Asian Republics; the activities of the Soviet researchers were going on, in particular V. I. Sarianidi’s excavations at Gonur South and, from 1992, also in the largest northern depe. The break-up of the Soviet Union, though it produced a sort of institutional revolution, essentially did not modify the cultural institutions or suppress the new trend towards an open-minded attitude to international collaboration.

In those years, the central-northern part of the delta was still a vast desert region with wind-borne sand accumulations from the Karakum Desert. This situation had helped both 

M. Cattani and S. Salvatori

Figure 1.1 Transect lines carried out from 1995 to 2005.

the surveys and excavations of the Soviet archaeologists in the early nineteen-fifties, and the first stages of the ItalianRussian-Turkmen Project.

In 1995 the new territorial situation, which was undergoing an huge transformation, combined with the need to further develop our knowledge of the ancient settlement system and to understand to what degree the desertification of the environment had influenced the settlement choices on a regional level, gave rise to the decision to realign the operations relative to the archaeological survey with different operating strategies, no longer based on a systematic coverage of the territory. In consideration of, and thanks to, the progresses in the study of the data collected up to then, the decision was made to organise a series of survey transects (Figure 1.1), covering the areas interposed between the larger settlement groups of the system, as defined by the researches of the Soviet archaeologists. The data collected in the central-northern part of the delta, as a matter of fact, when combined with those previously collected and published by V. I. Sarianidi, clearly revealed a hierarchical-type territorial organisation of the Bronze Age settlements; this organisation was clearly contrary to the hypothesis of discrete concentrations (oases) and suggested a spatial continuum whose settlement rules were dictated by political-geography principles rather than by environmental limitations. The same conclusions were attained also by the thorough geo‑morphological analysis carried out, during the first years of the development of the project, by M. Cremaschi, who utilised the aerial-



The archaeological survey activities of the project were initially concentrated in the southern part of the delta and characterised by a great intensity (Bondioli and Tosi 1998). Only later on, the activities were extended to the rest of the delta, so as to overlap with the areas that had seen the pioneering surveys of the Soviet archaeologists. This new stage coincided with the first great drive for the recovery of agricultural land that led to the northward enlargement of a close network of irrigation canals. It was then possible to significantly enrich the catalogue of the archaeological sites, particularly of the Bronze and Iron Ages, in the delta region, and, with the help of aerial‑photograph coverage and satellite photographs, to produce some reconstructions of the ancient delta system and of the old branches of the Murgab that were active during the third, second and first millennia BC, and to establish a relationship between them and the peopling dynamics in that area (Cremaschi 1988; Marcolongo and Mozzi 1998). The advancing surge of the agricultural world was already bearing down on archaeological research, making it impossible to maintain the previous survey procedures on increasingly extensive parts of the delta.

Transects and Other Techniques for Systematic Sampling

Figure 1.2 Surface sherd counting on a line from Site No. 193 to Togolok 1.

photograph coverage of the area and the data collected during intensive survey campaigns on the ground (Cremaschi 1998).

It is common knowledge that the transect technique, one of the various territorial sampling methods used in archaeology (Redman 1974; Mueller 1975; Ammermann 1981, 1985), has been applied within survey strategies based on concepts of statistical sampling, such as statistically calibrated projections for the production of a view of ancient peopling on the basis of a territorial sample proportional to the entire territorial system under examination. Other transect survey techniques have been developed in the past in order to investigate the spatial distribution of the archaeological material (Blintliff and Snodgrass 1985; Blake et al. 1988): they were based on rectilinear, uni‑directional routes that started from areas with a high density of archaeological material (chiefly pottery fragments), which usually indicates the presence of dwellings or artisanal activities, proceeded towards areas where the presence is quantitatively less dense at the periphery of these centres, then towards areas where the presence of archaeological material is extremely scanty but constant, indicating agricultural land with manuring activity, and finally reached areas that were absolutely devoid of archaeological material, i.e. had not been systematically exploited and were almost a no man’s land where agricultural activities did not or could not arrive. These techniques were successfully used for determining

A decidedly new picture of the environmental configuration of the delta during the second half of the 3rd millennium BC and the beginning of the 2nd millennium was thus emerging. The delta area had not yet been encroached on by the Karakum sand, and appeared as a vast irrigated plain, furrowed by the deltaic branches of the Murgab. Only in a subsequent period, during a late stage of the Late Bronze Age, the Karakum sand advanced and invaded the northern front of the delta with linear north-south dunes, penetrating deeply southwards. The need to document (also from an archaeological point of view) this situation, which was quite different from the one that characterised the region in the 2nd millennium BC, led us to organise a program of transect surveys along the axes that interconnect the major sites of the Bronze Age settlement system.

On the causes of sand encroachment, an interesting conjecture has been proposed recently in Lecomte and Francfort 2002.





M. Cattani and S. Salvatori the limits of the “site catchment areas” , i.e. for evaluating the territories of ancient centres and the boundaries between the areas of influence of the various settlements present in a territory (Wilkinson 1982, 1989, 1990), thus obtaining a measure, so to speak, of the cultural, administrative, political or environmental discontinuity.

researches aimed at explaining the function of the mound areas and of the areas between them. The transects carried out between 1995 and 2002 used a different method. First of all, they were traced chiefly with an east-west course and on lines that tended to interconnect the main centres of the settlement groups of the delta, as they had been defined by the Russian archaeologists’ surveys. They can be represented as strips of various widths, along which the survey was carried out systematically; their purpose was not to collect statistical or areal data, but rather to evaluate the distribution of the archaeological material and settlements along the axis of the spatial interconnection between the settlement groups previously defined as oases.

Methods of this sort were used also within our project, above all in relation to a particular problem represented by the so-called “site complexes” that were perceived during the surveys of the first years as continuous scatters of archaeological material not characterised, morphologically speaking, by settlement mounds. This experiment was carried out by S. Cleuziou in the area of Togolok and published in the previous volume, where our researches in the Murgab Delta are summarised (Cleuziou et al. 1998).

Along these linear routes, GPS-aided georeferences of the presence of archaeological materials and of their concentration or dispersion were acquired, in quantitative terms (low, medium, high). Where there were settlement mounds or isolated industrial areas (pottery kilns), their contours, extension and structural characteristics were recorded. The geo‑morphological characterisation of the context (takyr, sand, palaeochannel, etc.) was recorded as well.

Undoubtedly this technique should be adopted to investigate a series of sample areas in various parts of the delta, and should be applied to situations having different chronological and cultural characteristics. In the specific case, the work carried out by S. Cleuziou made it possible to analyse the phenomenon of the pottery scatter produced by a settlement with a complex, diversified structure, and to reveal its polycentrism (Figure 1.2). It is true, on the other hand, that the situation revealed by the analysis is the outcome of the transformation of the settlement pattern, presumably over a span of about 700 years (2200-1500 BC). For the time being, it is impossible to break down this archaeological district into its subspecific chronological components, i.e. into the various phases and sub-phases of the Bronze Age sequence. However, settlement continuity is in itself a strong indication of the importance of the centre of Togolok 1 in the region, particularly during the Late Bronze Age. The data obtained with the transect performed in the area of Togolok 1, when processed and displayed in the territorial context, made it possible to achieve a different interpretation of the interrelationships between a certain number of pluri‑stratified settlement mounds (Sites No. 190, 200, 199, 145, 192 e 197). This series of discrete entities (low mounds) seems actually to belong to a single large settlement, presumably no less than 60 hectares, which perhaps reveals the presence of a real, large-sized central place in the Margiana of the Late Bronze Age. A similar situation, though on a smaller scale, may be surmised for Taip (in this perspective, the area of Taip is presumably about 25 hectares). At present, this interpretation can be proposed only as a conjecture, pending more targeted

This strategy was adopted, as we have already pointed out (Salvatori 1998), in order to bridge some evident gaps in the data published by the Russian archaeologists (Sarianidi 1990), undoubtedly resulting from the exploration method adopted by them, along north-south penetration axes. The first transect carried out in 1995 between Gonur and Adzhi Kui made it possible to identify not only some new settlements, from both the Middle and Late Bronze Ages, but also several pottery scatters on the surface of the takyr. The findings of this first transect led us to intensify our efforts, and in the subsequent years other fifteen transects were carried out, some of them only partial because of the obstacles represented by new irrigation canals and by the practice of flooding the areas put to crops. This selective, targeted survey activity on linear routes confirmed that the delta (even in the areas that were partly covered with sand belts, both hardened and still active) must have appeared, during the 3rd millennium BC and the first few centuries of the second, as a vast alluvial plain, furrowed by the meandering branches of the river and basically devoid of sand. A further confirmation of the new picture that was beginning to appear came from the identification of a vast settlement that bridged a significant gap in the distribution of the medium-sized sites within the Bronze Age territorial system of the delta.

Unfortunately V. Sarianidi’s methods of excavation and publication do not make it possible to clearly understand the development, with time, of the various archaeological entities investigated in the area of Togolok. L. P’yankova’s attempts (1989, 1993, 1994) to produce a viable ceramic sequence for the Late Bronze Age are only an initial approach to the problem, because she was able to work almost exclusively on items coming from graveyard contexts that occupied the areas of the architectural complexes after they had been abandoned. 

The data relative to the transects are supplied in the Appendix to this book.





Transects and Other Techniques for Systematic Sampling The analysis of the settlement pattern of the delta during the Bronze Age carried out by means of the technique of Thiessen Polygons revealed some abnormal empty spaces between the south-eastern part of the ancient delta and the north-western part. The identification of Site No. 723, which was located at the centre of one of the previously empty polygons, confirms that the settlement pattern of the delta had not been determined by environmental constraints such as the presence of sandy formations, but by rules that were inherent in the social and political dynamics of the human groups involved.

The settlement network that was thus formed complies, however, with precise rules of political geography, as demonstrated by the application of Thiessen Polygons to the settlement system of Margiana, which, during the Middle Bronze Age, was based on a hierarchical system with at least three orders of magnitude (Salvatori 1998), pivoting on the central site of Gonur 1 North. What we have described above refers specifically to the problems of the territorial organization during the Middle and Late Bronze Ages; but the settlement pattern during the Iron Age presented equally impelling problems.

The only real environmental determinant in the Murgab Delta of the Bronze Age, as everywhere else and at any latitude in the planet, was water. It is true that in other places, in different climatic conditions, the location of sites near water‑courses was determined chiefly (though not exclusively) by reasons connected with the transportation of materials, goods and people. In areas with a low yearly rainfall, the fact that the vicinity of water‑courses was preferred for the location of urban concentrations was dictated above all by the need for a constant and easilyobtained water supply for farming activities (abduction of canals for irrigation).

Other operating methods were subsequently designed in order to make it possible to describe and explain archaeological entities that belonged to a later period (chiefly the Iron Age) and therefore referred to environmental and geo‑morphological conditions that had been deeply changed by the desertification of the delta resulting, on the one hand, from the progressive reduction of the flow of the Murgab River and, on the other hand, from the consequently relentless encroachment of the Karakum sand carried by the dominant northern winds. During the surveys of the first years of the project, vast flat‑land areas where pottery was scattered in a high and medium concentration were found. It was precisely the concentration rate and quantity of pottery that led to the reasonable conclusion that these scatters were not the result of manuring activities; the areas therefore were called “site complexes”, borrowing a term used in the archaeology of Central America (Bondioli and Tosi 1998: XVIII). This, however, was not an accurate choice of words, because

Similarly to what happened in Mesopotamia, the Indus Valley, Egypt, China, Elam and the Helmand Civilization, in Margiana, during the Bronze Age, the main settlements were located along the branches of the broad delta fan; from these branches, with time, artificial canals were deviated in order to bring water also to the areas that were more outlying with respect to the natural water‑courses.



Figure 1.3 Three dimensional contour map of the area of Sites No.173-174 with a slightly exaggerated scale.

M. Cattani and S. Salvatori “site complex” defines complex realities characterised by the distribution, over a vast archaeological territory, of various discrete units that can be individually described as dwellings, administrative complexes, religious complexes, areas of craftsmanship activity, communal meeting areas, game areas, etc. In the case of protohistoric Margiana, it would be more consistent to apply this term, for instance, to the district of Togolok 1, mentioned above, and to that of Taip 1, but decidedly not to these vast and apparently undifferentiated expanses of archaeological material, chiefly ceramics. On the basis of these considerations, we felt the need to endeavour, where it was still possible, to break down the areas called “site complexes”, selecting sample areas in various parts of the delta. One of the areas thus selected (Sites No. 211, 212 and 213), which meanders for at least 7 km, is located in the eastern part of the delta, 5 km north of the Achaemenid, Parthian and Sasanian fortress of Dingli Kishman. Another area, which turned out to be heavily altered by the expansion of agricultural activities, is in the central part of the delta (Sites No. 203, 204, 206 and 119).

an indistinct area of pottery scatter was the result of the presence of a settlement system that had developed on the banks of some branches of the delta, following their course for several kilometres. This new perspective made it possible, in the case of Sites No. 211, 212 and 213, to reconnect this widespread distribution (whose course was forced, because it was determined by the route of the river) with other contemporary archaeological entities located slightly to the north along the same river branch, such as Sites No. 173 and 174, indicated in the first survey stage as two small mounds. During the campaign of 2000 these two mounds, only about 150 m apart, were re-examined, and we were surprised to find that both were at the centre of a vast settlement area characterised by a substantial pottery scatter (Figure 1.3). Greater densities of material could easily be noticed in coincidence with even very limited altimetric variations. It was increasingly evident that these altimetric variations, which were associated with high densities of archaeological material (not only pottery) and which consisted of sub-rectangular areas slightly raised above the alluvial plain and separated by alluvium areas with a lower density of material, were nothing other than dwelling units, probably monophase ones. This large settlement (approximately 27 hectares), with two central mounds (Sites No. 173 and 174) that probably were the administrative, business or political centres, seems to be the place where the vast area of agricultural development

The new investigation showed that, under a more careful observation, these extremely vast scatters of material coincided with an uninterrupted series of slight altimetric variations (20 to 40 cm), more or less regularly spaced out and distributed on the margin of meandering palaeo‑channels. It became immediately evident that what had previously been understood and described as



Figure 1.4a Pottery frequency along a discontinuously oriented line across Sites No. 212-213.

Transects and Other Techniques for Systematic Sampling

Figure 1.4b Pottery frequency along a discontinuously oriented line across Sites No. 212-213.



Figure 1.4c Pottery frequency along a discontinuously oriented line across Sites No. 212-213.

M. Cattani and S. Salvatori

Figure 1.4d Pottery frequency along a discontinuously oriented line across Sites No. 212-213.



Figure 1.4e Pottery frequency along a discontinuously oriented line across Sites No. 212-213.

Transects and Other Techniques for Systematic Sampling

Figure 1.4f Pottery frequency along a discontinuously oriented line across Sites No. 212-213.

houses in all Turkmenistan (Figure 1.6). Immediately to the east and west of the two settlement lines at the margins of the palaeo‑channel, there are vast stretches of takyr on which we found a widespread, low- and very-lowconcentration scatter of potsherds belonging to the Yaz III phase of the Iron Age Central Asian sequence. The low density of material on easily irrigated areas beyond the settlement line suggests that they were real agricultural areas and that the low-density distribution of ceramic material was the result of manuring procedures. The model that is deduced from all this is therefore structurally acceptable and can be represented as in Figure 1.7.

that meandered southwards along the river branch was coordinated. In order to be able to document this settlement reality, quite unusual because of its extension and axial development, with a graphically representable method, we decided to apply an observation strategy based on the coverage of linear routes. In other words, we proceeded to trace a discontinuously oriented line that crossed the entire area of the “site complex” (Sites No. 212-213) from south to north, along the eastern bank of the palaeo‑channel, interconnecting the altimetric anomalies that were present. Every 20 m, a fixed area of 16 sq m was sampled for a counting of the ceramic material. Finally, a deviated stretch of the route interconnected the two banks of the palaeo‑channel. The result of this anomalous procedure, in which we played it as it came on the altimetric discontinuities of the landscape, can be plotted on a chart that clearly shows the sinusoidal course of the pottery scatter, in direct relationship with the increases in altimetric value (Figures 1.4a-f). Other observations we were simultaneously able to make along the route covered by us within this vast area are also interesting, because they confirm our overall interpretation. In coincidence with some of the surveyed areas, we noticed some traces of unbaked-brick wall structures, while in the alluvial areas we had the chance to document the presence of fireplace structures (Figure 1.5), similar to the present-day ovens, truncated-cone-shaped or dome-shaped, that are typical of the courtyard area of rural

The decoding of the archaeological reality underlying the unusually vast scatter of chronologically and culturally homogeneous archaeological material (we refer to areas Sites No. 211, 212 and 213, which form a sequence along the same canal) allows an integrated interpretation of the settlement type along the eastern branch of the Murgab Delta: a succession of dwelling units, probably of an agricultural nature, that follows the banks of the easternmost branch of the delta and has a more compact and formally more consistent concentration to its north. This concentration has developed around two artificial hill formations, at least one of which (Site No. 173) may be the centre of the system; the latter may have included, among other things, an entire series of defensive installations (frontier fortresses protecting the western and eastern fronts of the delta). This series of fortresses is located in a good strategic position, on two south-east to north-west lines (Figure 1.8). If, as we believe (also on the basis of the contemporaneity of the archaeological entities),

The numerical data are reported in the Appendix (On-the-line Transect 2000).





M. Cattani and S. Salvatori the three transects that were carried out now is almost entirely occupied by sand, and in any case, in the few takyr we met with, we did not find any archaeological material. To find traces of human settlement we had to arrive near the Takhirbai Canal, where there is evidence of both Iron and Bronze Age settlements, including some important dwelling remains ascribable to the steppe culture (Site No. 1211). To return to the problem of the structural pattern of these settlements, we must add that in other places, but still in relation to the settlements of the Iron Age, we had a chance to observe a second pattern as well (Figure 1.9). Another “site complex” we were able to study, in a central area of the middle delta, about 10 km west of Takhirbai 3, is Site No. 203. In this area, during the first stages of the survey on the Murgab, there had been many reports of vast areas covered with potsherds, significantly dated to the Iron Age, chiefly to its Yaz III phase, in concentrations or high-density scatters. At present, as a result of the expansion of agricultural activity, a great number of these areas have undergone changes that are so deep as to delete most of the surface archaeological evidence. In many cases, therefore, during the campaign of 2002, we were only able to record the complete disappearance of the previously plotted archaeological sites. Elsewhere, in the very extensive formations, we were able to recognise only some very limited residual portions that were still unaltered (e.g. Sites No. 50, 41, 160 and many others).

Figure 1.5 Fireplace structures on a flat area between small Yaz III farm‑houses.

it represents a harmonized system, than it must be the product of a great remodelling operation for the protection of a predominantly agricultural productive unit supported by hydraulic-control operations. It may be connected to the defensive lines formed of series of fortresses that delimit the eastern margin of the area, the administrative centres (Sites No. 173-174) and the production centres (Site No. 172). The system appears to exploit a palaeo‑channel that perhaps was activated artificially. It is undoubtedly remarkable that very few archaeological pieces of evidence ascribable to periods other than Yaz III were found. The fact that there are no settlements from a previous period in the area seems to suggest that the deltaic branch was not active during the Bronze Age.

Unlike most of the Yaz period sites in this central part of the middle delta, Site No. 203 had been impaired by cultivations only in its southern part. Here too it was possible to recognise, in the preserved portion, a series of dwelling units, aligned and regularly spaced. Once more we applied the interrupted-linear-transect technique, counting the items of archaeological material on 4 x 4 m squares 20 m apart: the result is represented in Figure 1.10. As in the area of Sites No. 211, 212 and 213, to the east, the dwelling units are clearly recognisable in coincidence with the numerical-value peaks in the frequency of potsherds, and the free areas between the discrete dwelling units are clearly recognisable as well.

In order to verify the general situation in which this remarkable sequence of Iron Age archaeological entities is set, some transect surveys were carried out between this eastern canal and the nearest canal to the west, the Takhirbai Canal. The area crossed by us along the axis of

Immediately to the east of this alignment of houses or small farms, we detected the presence of a second line of dwellings, separated only by a narrow strip of sand. Further eastwards we noticed a flat takyr belt characterised by a medium-to-high pottery density, resulting from the radiant scattering from the actual dwelling area. Beyond this, See the notes and data relative to Transects 9 and 10 in the Appendix. 

See Cattani, Chapter 9.



In actual fact the area was quite vast and decidedly continuous, and consisted of Sites No. 203, 204, 206 and 219, covering several hectares. 

Figure 1.6 Present-day truncated-cone-shaped or dome-shaped ovens

The data are reported in the On-the-line Transect.



10

Transects and Other Techniques for Systematic Sampling

Figure 1.7 Synthetic model of Yaz III farm‑houses pattern along a river branch.

the takyr, which was basically open, though occasionally interrupted by more or less extensive sand strips, showed only the constant presence of low-density scatters, and at about 200 m from the dwelling area there was a meandering, north-south depression, a few tens of metres wide, that corresponded to a segment of a palaeochannel mostly filled up with recent and/or hardened sand. Beyond the palaeo‑channel, the flat takyr area dotted with potsherds started again. About a hundred metres further to the east, the density of material suddenly increased, and a little further on there appeared the usual small altimetric discontinuities, covered with substantial concentrations of material, chiefly pottery, indicating the presence of dwelling units. This new settlement area stretches along the axis of the palaeo‑channel for at least 600 m and covers a surface of about 6 hectares (to the south, the settlement has been deleted by cultivations); we called it Site No. 415. The pattern that presents itself in this area (Sites No. 204, 206, 415, 203 and 119) seems to be formed of a palaeo‑channel, on both sides of which there are cultivated fields that are bounded, to the east and to the west, by linearly developed settlements. Considering a more extensive area, the system defined by the sites mentioned above seems to stretch along the same river branch on which there depends, slightly further southwards, another large concentration of Yaz presences (Sites No. 205, 163, 162, 161, 57, 143, 54, 142, 49, 160, 41, 40, 39, 50, 51, 158 and 159). This area was mapped during the first years of the project, and now has been almost entirely taken over by agriculture. During our visit in this area in 2002, we

realised that it had been almost completely destroyed, as regards its surface archaeological traces. Only in the area of Site No. 50 we were able to locate the residual remains of a Yaz III farm. We found the same situation in Sites No. 160 and 41. The present conditions of the territory do not allow us to advance conjectures about the pattern of this vast concentration, though it is clear that, like the others, it probably developed along the course of a delta branch. Equally critical is the situation of other three clusters of presences that are chronologically ascribable to the Iron Age and seem to follow each other along other fluvial branches in the same area. Many of these presences have been almost completely deleted by the development of cultivations, and only occasionally we were able to document the survival (up to the autumn of 2002) of (by then isolated) remains of Yaz farms; an exception was a new Yaz III settlement (Site No. 125), which was detected in the northern portion of Cluster 3 (see note 10) and still contained the remains of at least six dwelling units. In conclusion, it is reasonable to assert that the settlement pattern in the Murgab Delta underwent, between the Middle Bronze Age and the Iron Age, dramatic transformations that were significantly correlated with geo‑morphological Cluster 1: Sites No. 128, 117, 94, 96, 98, 89, 91: Cluster 2: Sites No. 122, 123, 1409, 126, 55, 1410, 208, 146; Cluster 3: Sites No. 168, 77, 76, 78, 72, 75, 73, 71, 72, 70, 45, 44, 43, 42, 1023, 37, 38, 1009, 664, 665, 52, 53, 48, 1010, 56, 124, 47, 1024, 1025, 63, 124, 60, 61, 59, 46, 1011, 230, 162. 

11

M. Cattani and S. Salvatori

Figure 1.8 Fortress lines bordering the Yaz III settlement pattern.

transformations. Apparently, during the Middle Bronze Age and the early stages of the Late Bronze Age, the Murgab Delta was not considerably encroached on by sand formations: the entire delta was continuously settled. During the final stages of the Late Bronze Age, whatever the causes of this were, sand seemed to move forward and invade the extensive alluvial takyr, and together with the advance of sand we see the entrance into the delta of small groups of nomadic pastoralists who were bearers of a material culture characterised by hand-made pottery decorated with incisions. A great number of small camps of these intrusive groups were found, always on sand and very often within depressions in the sandy areas. Only during the Final Bronze Age (Takhirbai 3 Phase) these nomadic groups seemed to acquire permanent settlements and to achieve farming activities, besides a stable relationship, and perhaps integration, with the local farming population. The settlements of this phase, and not only in Margiana10, reveal the co-presence of pottery typical of the population of the steppes of Central Asia and wheel-made pottery typical of the Takhirbai 3 Phase. For the time being,

however, this peopling appears to have been quite limited, and chiefly located in the area of Takhirbai 3. With the subsequent early phase of the Iron Age (Yaz I), the peopling seemed to become more extensively redistributed in the delta, probably along river branches that were still active. This new configuration, which was strongly conditioned by an even more massive presence of sand in the ancient delta, seemed to be reinforced during the subsequent Yaz II e III phases: in this period we notice an increase in the concentration of the population on the banks of active fluvial branches. The agricultural potentiality of the region now seemed to be limited almost exclusively to the areas by the sides of water‑courses, and this produced the archaeological phenomenon we have analysed above: it represents a significant adjustment to precise environmental constraints through an exacting effort to control the natural hydrographical regimen. References Cited Ammerman, A. J. 1981. “Surveys and Archaeological Research”, Annual Review of Anthropology 10: 63-88.

In northern Bactria, too, the arrival of nomadic populations that were bearers of the typical pottery with incised decoration is documented during the final phases of the local Bronze Age sequence (Avanesova 1995, 1996). 10

12

Transects and Other Techniques for Systematic Sampling

Figure 1.9 Alternative model of Yaz III farm houses pattern observed along a river branch.

Figure 1.10 Pottery frequency along a discontinuously oriented line across Site No. 203.

13

M. Cattani and S. Salvatori Ammerman, A. J. 1985. “Plow-Zone �������������������������� Experiments in Calabria, Italy”, Journal of Field Archaeology 12: 3340. Avanesova, N. 1995. “Bustan VI, une nécropole de l’âge du Bronze dans l’ancienne Bactriane (Ouzbékistan méridional): témoignages de cultes du feu”, Arts Asiatique L: 31-45. Avanesova, N. 1996. “Buston VI - the Peculiarities of the Proto-bactrian Complex of Uzbekistan”, Information Bulletin IASCCA ������� 20: 141-155. Blake, H., A. Bondesan, V. Favero, E. Finzi and S. Salvatori 1988. “Cittanova-Heraclia 1987: risultati preliminari delle indagini geomorfologiche e paleogeografiche”, Quaderni di Archeologia del Veneto IV: 112-136. Bintliff, J. L. and A. M. Snodgrass 1985. “The Cambridge/ Bradford Boeotian Expedition: the First Four Years”, Journal of Field Archaeology 12: 123-161. Bondioli, L. and M. Tosi 1998. “Introduction”, in Gubaev, Koshelenko and Tosi (eds.) 1998: IX-XIX. Cleuziou, S., V. Gaibov and A. Annaev 1998. “Off-site Archaeological Transects in Northern Margiana”, in Gubaev, Koshelenko and Tosi (eds.) 1998: 27-33. Cremaschi, M. 1998. “Palaeohydrography and Middle Holocene Desertification in the Northern Fringe of the Murgab Delta”, in Gubaev, Koshelenko and Tosi (eds.) 1998: 15-25. Deshayes, J. (ed.) 1977. Le Plateau Iranien et l’Asie centrale des Origines a la Conquête Islamique. ������ Paris. Gubaev, A., G. A. Koshelenko and M. Tosi (eds.) 1998. The Archaeological Map of the Murgab Delta. Preliminary Reports 1990-95. Reports and Memoirs. Series Minor, III 1st AD. Rome. Hiebert, F. T. 1994. Origins of the Bronze Age Oasis Civilization in Central Asia. (American School of Prehistoric Research Bulletin 42). Cambridge. ���������� Kohl, Ph. L. (ed.) 1981. The Bronze Age Civilization of Central Asia. New York. Lecomte, O. and H-P. Francfort ��������������������������������������� 2000. “Irrigation et société en Asie centrale des origines à l’époque achéménide”, Annales 57 (3): 625-663. Marcolongo, B. and P. Mozzi 1998. “Outline of Recent Geological History of the Kopet-Dagh Mountains and the Southern Kara-Kum”, in Gubaev, Koshelenko and Tosi (eds.) 1998: 1-13. Masimov, I. S. 1980. “Kelleli – novyj oazis epokhi bronzy nizovij Murgaba”, in Novye issledovaniya po arkheologii Turkmenistana, Ashkhabad: 3-21. Masimov, I. S. 1981. ���������������������������������� “The Study of Bronze Age Sites in the Lower Murgab”, in Kohl (ed.) 1981: 194-220. Masson, V. M. 1959. Drevnezemledel’cheskaya kul’tura Margiany. (����������������������������������������� Materialy i issledovaniya po arkheologii SSSR,����������������� no. 73). Moskva. Mueller, J. W. (ed.) 1975. Sampling in Archaeology. Tucson. Parpola, A. and P. Koskikallio (eds.) 1994. South Asian Archaeology 1993. (Annales Academiae Scientiarum Fennicae, Series B, Vol. 271). Helsinki.

Pyankova, L. 1989. “Pottery Complexes of Bronze-Age Margiana (Gonur and Togolok 21)”, Information Bulletin IASCCA ������� 16: 27-54 . P’iankova, L. 1993. “Pottery of Margiana and Bactria in the Bronze Age”, Information Bulletin ������� IASCCA 19: 591-595. P’yankova, L. 1994. “Central Asia in the Bronze Age: Sedentary and Nomadic Cultures”, Antiquity 68 (259): 372-387. Redman, C. L. 1974. Archaeological Sampling Strategies. (An Addison-Wesley Module in Anthropology 55). Cambridge. Salvatori, S. 1993. ����������������������������������� “The Discovery of the Graveyard of Gonur-depe 1 (Murgab Delta, Turkmenistan): 1992 Campaign Preliminary Report”, Rivista di Archeologia XVII: 5-13. Salvatori, S. 1994a. “A ��������� Late 3rd Millennium Graveyard at Gonur-depe 1 (Murgab Delta, Turkmenistan)”, in Parpola and Koskikallio (eds.) 1994: 657-666. Salvatori, S. 1994b. “Excavations ����������������������������������� at the Namazga V Late Graveyard of Gonur 1 (Murgab Delta, Turkmenistan). Preliminary Report on the 1993 Field Season”, Rivista di Archeologia XVIII: 14-39. Salvatori, S. 1995. “Gonur-depe 1 (Margiana, Turkmenistan): the Middle Bronze Age Graveyard. Preliminary Report on the 1994 Excavation Campaign”, Rivista di Archeologia XIX: 5-37. Salvatori, S. 1998. “Margiana Archaeological Map: the Bronze Age Settlement Pattern”, in Gubaev, Koshelenko and Tosi (eds.) 1998: 57-65. Sarianidi, V. I. 1990. Drevnosti strany Margush. Ashkhabad. Sarianidi, V. I. 1993. “Excavations ��������������������������������� at Southern Gonur”, Iran XXXI: 25-37. Sarianidi, V. I. 1998. Margiana and Protozoroastrism. Athens. Wilkinson, T. J. 1982. “The Definition of Ancient Manured Zones by Means of Extensive Sherd-sampling Techniques”, Journal of Field Archaeology 9: 323-333. Wilkinson, T. J. 1989. “Extensive Sherd Scatters and Land-Use Intensity: Some Recent Results”, Journal of Field Archaeology 16: 31-46. Wilkinson, T. J. 1990. “Off-site Archaeology”, National Geographic Research & Explorations 8 (2): 196-207.

14

Transects and Other Techniques for Systematic Sampling Appendix

ICW material (Incised Coarse Ware, the ceramics, typical of the nomadic populations, that can apparently be ascribed to the Tazabag’yab phase of the Andronovo culture, of the steppes of Central Asia) was detected both in the takyr areas and associated with sand-covered sites (Site No. 701), and also in areas totally covered with sand (Site No. 703)11.

(Figure 1.1) Transect 1 (1995) (approximately 7 km) (No field notes preserved) Points of M. Tosi’s 1995 transect 403370.91, 4231851.58 tr.95-1 402555.03, 4231829.91 tr.95-2 402525.94, 4231837.65 tr.95-3 402815.39, 4231762.04 tr.95-4 403162.31, 4231733.84 tr.95-5 403354.71, 4231716.75 tr.95-6 403489.94, 4231678.14 tr.95-7 403508.73, 4231785.20 tr.95-7 403674.52, 4231616.76 tr.95-8 403868.07, 4231573.77 tr.95-9 404075.13, 4231562.06 tr.95-10 404249.91, 4231535.95 tr.95-11 404418.38, 4231469.23 tr.95-12 404606.30, 4231442.97 tr.95-13 404797.48, 4231446.27 tr.95-14 404927.85, 4231365.20 tr.95-15 405046.95, 4231317.56 tr.95-16 405137.68, 4231340.55 tr.95-17 405357.38, 4231286.19 tr.95-18 405513.70, 4231302.86 tr.95-19 405692.31, 4231228.64 tr.95-20 405869.99, 4231198.83 tr.95-21 406032.25, 4231098.92 tr.95-22 406215.29, 4231028.36 tr.95-23 406418.07, 4231024.17 tr.95-24 406558.03, 4231013.31 tr.95-25 407793.38, 4230962.17 tr.95-26 406913.73, 4230983.32 tr.95-27 407051.02, 4230867.07 tr.95-28 407913.95, 4231299.28 tr.95-29 408818.32, 4231274.26 tr.95-30 408275.45, 4231267.44 tr.95-31 408486.52, 4231222.51 tr.95-32 408652.21, 4231165.15 tr.95-33 408790.40, 4231126.60 tr.95-34 408973.83, 4231089.40 tr.95-35 409155.98, 4231067.01 tr.95-36 409336.50, 4231029.84 tr.95-37

Finally we must report the presence, in point 14, of a palaeochannel alng whose banks we detected the presence of limited LBA pottery scatters. UTM Coordinates 389929,4234594

390393, 4234365 390623, 4234496 390703, 4234370

391367, 4233848 390881, 4234178 392568, 4233632 393209, 4233794 393235, 4233939 393963, 4233896 395217, 4233438

Transect 2 (1996) from Taip-1 to Adzhi Kui 9 (11.5 km)

395531, 4233414

Along the entire route, in the takyr areas, scatters or concentrations of pottery chiefly datable to the Late Bronze Age, but sometimes also to the Middle Bronze Age, were found, starting from point 3.

397538, 4233160

Entities that can be defined as actual dwellings are Site No. 700 (= Taip 10), sand-covered Site No. 701, where only the marginal scatters were detected; and also Site No. 709. Site No. 702 is a case apart, and contains material dating from the Middle and Late Bronze Ages. The pottery material, which is present in considerable concentrations, is located in an area that has already been disrupted by ploughing.

397542, 4232645

397344, 4233178

397540, 4233155

397974, 4232247 398243, 4232144

Context Taip 1 MBA-LBA Large concentration of MBA to LBA pottery sherds. To the east we can observe the presence of a very small open takyr in a sandy dune landscape. Site No. 700 (=Taip-10). Takyr with low-density LBA pottery dispersion. Takyr with low-density MBA to LBA pottery dispersion. Edge of site covered with sandy dunes. LBA and ICW pottery sherds and furnace bricks. Site No. 701. To the east of the site, large area of agricultural exploitation. Concentration of MBA to LBA pottery in the agricultural fields. Site No. 702. Sandy area with ICW pottery concentration. Site No. 703. Takyr with medium-density dispersion of LBA and ICW pottery sherds. Site No. 705. Takyr with low-density LBA pottery dispersion (50 x 50m). Site No. 706. Takyr with low-density LBA pottery dispersion. Takyr with low-density LBA pottery dispersion. Takyr with low-density BA (?) pottery dispersion. Takyr with low-density LBA pottery dispersion. Takyr with low-density LBA pottery dispersion. Low-density LBA pottery dispersion along the silty bank of a palaeo‑channel. Takyr with low-density LBA pottery dispersion. Takyr with low-density LBA pottery dispersion. Site No. 707. Takyr with medium-density LBA pottery dispersion. Site No. 708. Takyr with medium-density LBA pottery dispersion.

This type of material has been found frequently in sandy depressions, within areas largely covered with sand. 11

15

M. Cattani and S. Salvatori UTM Coordinates 398448, 4232010

400740, 4231534 401531, 4231905

alternating with more or less extensive areas of bare takyr in which the presence of MBA, LBA and ICW pottery was detected. Some settlement mounds were identified, but they are chiefly covered with dunes (Sites No. 715, 717, 718 and 720). Only Site No. 753 had not been covered with sand. As regards point 25, the high density of pottery material in the bare takyr seemed to suggest the presence of a real settlement in the environs, but we were not able to locate it.

Context Low LBA mound partially covered by a sand dune. Site No. 709. From this site, looking towards Adzhi Kui 9, we could see a large agriculturally exploited area. Takyr with medium-density LBA pottery concentration. Adzhi Kui 9. MBA ��������������� settlement.

Transect 3 (1996) from Gonur 1 North to Auchin 1. (6 km)

Medieval pottery material, in a rather high-density concentration, was detected in a sandy area (Site No. 719).

The transect was interrupted by the interposed obstacle of a broad, modern irrigation canal and by the flooding of vast areas, in preparation to putting new fields to crops. Along the route, characterised by sand interspersed with open takyr areas, we found some limited pottery scatters in the takyr areas. In the sand, always within a sandy depression, we noticed the presence of a camp‑site (Site No. 711) containing ICW pottery together with wheel-made pottery datable both to the LBA and to the FBA (Final Bronze Age = Takhirbai 3 horizon). A pottery kiln, apparently isolated and datable to the LBA, was identified in point 2 (Site No. 713). UTM Coordinates 417900, 4231449

418147, 4231380

418200, 4231492 418584, 4231588

418903, 4231646 418862, 4231158 419268, 4231857 419638, 4232038 420113, 4232038 423373, 4231591

UTM Coordinates 385533, 4251690 385654, 4251814 385852, 4251850 386024, 4251877 386146, 4251866 386311, 4251877 386411, 4251875

Context Takyr with low-density LBA pottery dispersion. Traces of old pits which may suggest a cemetery area. Site No. 704. Takyr with low-density LBA pottery dispersion. Traces of old pits which may suggest a cemetery area. Site No. 710. Pottery‑kiln and LBA pottery sherds. Site No. 713. Sandy depression with medium-tohigh-density concentrations of ICW and LBA and FBA pottery sherds. Site No. 711. Takyr with medium-density LBA pottery dispersion. Large mound-like LBA settlement and pottery kilns. Site No. 712. Takyr with low-density LBA pottery dispersion. Takyr with low-density LBA pottery dispersion. Takyr with low-density LBA pottery dispersion. End point because of the presence of a large irrigation canal and irrigated fields.

386420, 4251869 386425, 4251860 386510, 4251858 386520, 4251846 386514, 4251756 386556, 4251720 386597, 4251578 391053, 4251152 391066, 4251140 391453, 4251120 391467, 4251115 391498, 4251078 393455, 4250908

Transect 4 (1996) from Kelleli 1 to Egri Bogaz 4. (approximately 11 km)

392046, 4250765

The area that was crossed is characterised (as almost everywhere else) by a great number of sand strips 16

Context Takyr with low-density MBA, LBA and ICW pottery dispersion. Site No. 713. Isolated LBA sherd in a sandy area. Takyr with low-density LBA pottery dispersion. Takyr with low-density LBA pottery dispersion. Takyr with low-density LBA pottery dispersion. Takyr with low-density LBA pottery dispersion. LBA pottery along a dune edge. Site ����� No. 714. Takyr with low- density LBA pottery dispersion. Takyr with low-density LBA pottery dispersion. Takyr with low-density LBA pottery dispersion. Takyr with low-density LBA pottery dispersion. Takyr with low-density LBA pottery dispersion. Two isolated LBA sherds in a sandy area. Takyr with low-density LBA pottery dispersion. Takyr with low-density LBA pottery dispersion and high-density dispersion along a dune edge. Site ������������� No. 715. Takyr with low-density MBA pottery dispersion. Takyr with low-density MBA pottery dispersion. Takyr with low-density LBA pottery dispersion up to the dune edge. Low-density unidentified pottery dispersion on a silty area in the midst of sands. Low-density BA pottery dispersion (kitchen‑ware) in a sandy depression. Site No. 716. Takyr with low-density unidentified pottery dispersion.

Transects and Other Techniques for Systematic Sampling UTM Coordinates 392089, 4250641

392122, 4249995

390842, 4250771 395939, 4250663 395867, 4250426 396263, 4250318

Context Takyr with high-density MBA and LBA pottery dispersion. The density increases near a barkhan dune that covers the actual settlement area. Site No. 717. Takyr with high-density MBA and LBA pottery dispersion. The density increases near a barkhan dune that covers the actual settlement area. Site ����� No. 718. Medium-density Islamic pottery dispersion on a sandy area. Site No. 719. Takyr with high-density MBA and LBA pottery dispersion. Site ������������� No. 720. Takyr with high-density unidentified (kitchen‑ware) pottery concentration. Mound-like settlement area with high concentration of MBA and LBA pottery. Site No. 753.

406960 407859 408095

4224637 4224518 4223683

1997 1997 1997

402665 403319 403085 405227 405581

4231607 4230789 4230557 4228032 4228098

1997 1997 1997 1997 1997

UTM Coordinates 402152, 4231879 406440, 4245032 412979, 4236852 410645, 4234535 404088, 4229557

Transect 5 (1996) from Adzhi Kui 1 to Egri Bogaz 2. (approximately 19 km)

406304, 4252124

The area turned out to be heavily covered with sand. In this bleak landscape we identified, in the midst of the sands, an ICW camp‑site (Site No. 745) and two mound-like settlements (Sites No. 746 and 747), containing Middle and Late Bronze Age pottery. UTM Coordinates 413290, 4234314 400643, 4234816 400593, 4235007

409759, 4246628 405607, 4227227 405148, 4226829

Context ICW pottery concentration in a sandy depression. Site No. 745. Low mound-like settlement area with MBA and LBA pottery concentration. Site No. 746. Low mound-like settlement area with MBA and LBA pottery concentration. Site No. 747.

406433, ������� 4226615 410572, 4225719 419759, 4244272 407707, 4223962

Transect 6 (1997) from Adzhi Kui 1 to Togolok 1. (approximately 16 km)

408974, 4232981 408677, 4223206

Along this transect, we recorded many settlements (mound-like settlements: Sites No. 957, 958, 959, 967 and 960), together with sand-covered sites of which we were able to see only the marginal areas (points 6 and 7) and high-density concentrations pointing to the presence of settlements in the environs; but we were not able to locate them (points 3, 4 and 5).

takyr takyr takyr

takyr takyr

LBA LBA LBA

LBA LBA LBA LBA

Context Adzhi Kui 1 (= Site No. 406). Concentration of very small unidentified pottery sherds along a dune edge. Takyr with high-density LBA pottery dispersion. Takyr with high-density LBA pottery dispersion. Takyr with high-density LBA pottery concentration. Site ��������� No. 408. LBA pottery concentration along a dune edge. LBA pottery concentration along a dune edge. Large mound-like settlement area with LBA pottery concentrations. Site No. 957. Mound-like settlement area with LBA pottery concentrations. Site ����� No. 958. Mound-like settlement area with LBA pottery concentrations. Site ����� No. 959. Takyr with low-density LBA pottery dispersion. Takyr with low-density LBA pottery dispersion. Mound-like settlement area with LBA pottery concentrations. Site ����� No. 967. Takyr with low-density LBA pottery dispersion. Mound-like settlement area with LBA pottery concentrations. Site No. 960. From this point to Togolok 1 only agricultural fields.

Transect 7 (1997) from Egri Bogaz 4 to Taip 1. (approximately 17 km) Along the line that interconnects the centres of Egri Bogaz-4 and Taip 1, in an area characterised by a strong presence of sand, we encountered some takyr with LBA ceramic material; two settlements covered with dunes, of which we were able to see only the edges (points 4 and 7); and two small, low mound-like settlements or camp‑sites with LBA and ICW material (Sites No. 961 and 962).

Starting from point 15 of the transect (Site No. 960), there begins a vast agricultural area that extends up to Togolok 1 and has disrupted the original landscape12.

I.e. the landscape of the stage that immediately preceded the operations for putting the desert or sub-desert areas of the delta to crops. 12

17

M. Cattani and S. Salvatori UTM Coordinates 405821, 4252346 406330, 4254436 393957, 4242844 391900, 4239208

391898, ������� 4239057

389713, ������� 4235710 425496, 4207693

takyr; a site datable to Yaz III, partly covered with sand (Site No. 1209); and, after an uninterrupted sand bed, the margin of another settlement area containing LBA material and completely covered by a dune (Site No. 1210). Near the modern, broad Takhirbai canal, which barred the way to our target, Site No. 357, we located a vast settlement which contained high concentrations of FBA and ICW ceramic material and was later divided into two apparently discrete areas (Sites No. 1211 and 1219)13.

Context Egri Bogaz-4 (=Site No. 288) Takyr with low-density LBA pottery dispersion. Takyr with low-density LBA pottery dispersion. High density of LBA pottery splinters on a dune edge. Very low circular mound (diameter about 12 m.) with high concentration of LBA and mostly ICW pottery sherds. �������������� Site No. 961. Low circular mound (diameter about 20 m.) with high concentration of LBA and mostly ICW pottery sherds. Site No. 962. High density of LBA pottery splinters on a dune edge. Site ������������� No. 966. Taip-1 (=Site No. 179)

UTM Coordinates 429279, 4206880 428686, 4206697 428497, 4206611 428165, 4206294 427562, 4205860 427029, 4205612

Transect 8 (1997) from Site No. 725 to Taip 1. (approximately 19 km)

424076, 4208897 424016, 4208958

Most of the linear route (approximately 19 km) that leads from Site No. 725 to Taip 1 is characterised by an uninterrupted sand coverage. Only at a distance of about 12 km from Site No. 725 we started finding settlement evidence again: three mound-like settlements, of which one, the largest (Site No. 963: approximately 2 hectares) is characterised by the presence of both MBA and LBA material; the other two, smaller settlements (Sites No. 964 and 965, respectively 0.28 e 0.1 hectares) yielded only LBA pottery. UTM Coordinates 406840, 4241419 395416, ������� 4236852

395591, ������� 4236250 395146, 4237199 425496, 4207693

422979, 4210430

Context Site No. 212. End point of pottery dispersion of Site No. 212. Takyr with low-density dispersion of IA pottery. Yaz III Site No. 1209. Partially covered with sand. LBA pottery along a dune edge (Site No. 1210). High-density concentration of FBA and ICW pottery (Site No. 1211). Takyr with LBA and ICW pottery dispersion (Site No. 980). LBA pottery concentration along the edge of a sandy area (Site No. 980 ?). Site No. 638 (Takhirbai 14). LBA pottery concentration.

Transect 10 (2000) from Site No. 638 to Site No. 172. (approximately 3.7 km) Along this route, characterised by a continuous bed of dunes and hardened sand, we did not detect any takyr areas or ancient anthropic evidence.

Context Site No. 725 Mound-like settlement with concentrations of MBA and LBA pottery sherds. The site is located in a takyr, 1.5 x 1.0 km, covered with a high-density dispersion of pottery. Site ����� No. 963 (MBA). Very low circular mound (diameter about 50 m) with concentrations of LBA pottery sherds. Site �������������� No. 964. Very low rectangular mound (m 40 x 20) with concentrations of LBA pottery sherds. Site �������������� No. 965. Taip 1 (=Site No. 179).

UTM Coordinates 422979, 4210430

426492, 4211601

Context Site No. 638 No open takyr exists between the two sites. Old sand dunes dominate the landscape. Site No. 172

Transect 11 (2002) from Site No. 1412 to Site No. 1411. (approximately 2.1 km) Area characterised by an alternation of dunes and takyr, together with extensive stretches covered with hardened sand. In the small takyr we noticed the presence of LBA pottery and, more rarely, of ICW. Along this route we detected some segments of palaeo‑channels (points 11, 19-20 and 22-23).

Transect 9 (2000) from Site No. 212 to Site No. 357. (approximately 2.2 km) Along the axis that connects the north-western limit of Area 212, in the direction of Site No. 357, we found, within an environment dominated by hardened sand, a scatter of Yaz III ceramic material (point 3), near a small

On these two sites, which are very important for understanding the dynamics of the integration of the nomadic population encroaching on the delta with the indigenous agricultural population, see Cattani, Chapter 10. 13

18

Transects and Other Techniques for Systematic Sampling UTM Coordinates Context

UTM Coordinates 415747, 4219005 415771, 4219021 415813, 4219057 415945, 4219095

416484, 4221075 416372, 4221018

416081, 4219093 416188, 4219102

416699, 4221335 416665, 4221334 416597, 4221235 416553, 4221168

416101, 4220754 416060, 4220691 416034, 4220656 415987, 4220542 415966, 4220510 415847, 4220465 415782, 4220437 415745, 4220401 415721, 4220382 415601, 4220202 415536, 4220155 415493, 4220115 415490, 4220129 415408, 4220089 415348, 4220010 415342, 4220009 415294, 4219954 415235, 4219884

Takyr with LBA pottery dispersion. Takyr with LBA pottery dispersion. Takyr with LBA pottery dispersion. Takyr with LBA pottery concentration. Takyr with LBA pottery dispersion. Takyr with LBA and ICW pottery dispersion (from this point to the next on sandy area). Takyr with LBA pottery and a single fragment of ICW pottery. Takyr with LBA pottery dispersion. Takyr with LBA pottery dispersion. Takyr with LBA pottery dispersion. Palaeo‑channel bed (centre point). Takyr with LBA pottery dispersion. Takyr with LBA pottery dispersion. Takyr with LBA pottery dispersion. Takyr with LBA pottery dispersion. Takyr with LBA pottery dispersion. Takyr with LBA pottery dispersion. Western palaeo‑channel bank. Eastern palaeo‑channel bank. Takyr with LBA pottery dispersion. Eastern palaeo‑channel bank. Western palaeo‑channel bank. Takyr with LBA pottery dispersion. Takyr with LBA pottery dispersion.

416339, 4219085 416453, 4219191 416540, 4219194 416653, 4219298 416690, 4219313 416748, 4219365 416778, 4219414 416845, 4219486 416926, 4219565 417160, 4219675 417189, 4219673 417439, 4219852 417570, 4219867 417649, 4219714

Transect 13 (2002) from Site No. 232 to Site No. 144. (approximately 2.9 km)

Transect 12 (2002) from Site No. 156 to Site No. 232. (approximately 4.2 km)

This area is characterised by uninterrupted sand, hardened dunes and shifting sand; but in the rare, small takyr there are scatters of LBA pottery.

This route too, like the previous one (from which it is only 1 km away, towards the south) turned out to be characterised by an alternation of dunes and takyr, together with extensive stretches covered with hardened sand. In the small takyr we detected the presence of LBA and ICW pottery. Along this route, too, we were able to detect some segments of palaeo‑channels (points 3-4 and 8-10), almost certainly connected to the ones that had been found slightly to the north. At point 22 we found a high-density concentration of LBA pottery that suggests the probable presence of a settlement in the environs, though we were not able to locate it. At point 17 (Site No. 1482), we found an eroded grave that yielded, besides human remains reduced to splinters of a few centimetres, a hand-made coarse ceramic beaker. UTM Coordinates 414917, 4218379 415404, 4218788 415416, 4218782 415446, 4218781 415640, 4218896 415670, 4218871

Context Western palaeo‑channel bank. Palaeo‑channel bed with LBA pottery. Eastern palaeo‑channel bank. Wide takyr with LBA pottery dispersion. Takyr with LBA pottery dispersion. Takyr with LBA and ICW pottery dispersion. Takyr with LBA and ICW pottery dispersion. Takyr with LBA pottery dispersion. Takyr with LBA pottery dispersion Hand-made pottery beaker associated with high-density concentration of human bone splinters (Site No. 1482). Takyr with LBA pottery dispersion. Takyr with LBA pottery dispersion. Takyr with LBA pottery dispersion. Takyr with LBA pottery dispersion. Takyr with high-density concentration of LBA pottery. Takyr with LBA pottery dispersion. Takyr with LBA pottery dispersion. Takyr with LBA and ICW pottery dispersion. Takyr with LBA pottery dispersion. Takyr with LBA pottery dispersion.

UTM Coordinates 417209, 4219399 417063, 4219025 416830, 4218785 416817, 4218742 416746, 4218611 416682, 4218510 416689, 4218305

Context Takyr with LBA pottery dispersion. Takyr with LBA pottery dispersion. Takyr with LBA pottery dispersion. Takyr with LBA pottery dispersion. Takyr with LBA pottery dispersion. Takyr with LBA pottery dispersion. Takyr with LBA pottery dispersion.

Transect 14 (2002) from the road to Auchin towards Site No. 340. (approximately 7 km)

Context Takyr with LBA pottery dispersion. Western palaeo‑channel bank. Eastern palaeo‑channel bank. Takyr with LBA pottery dispersion (from this point on sandy area). Takyr with LBA pottery dispersion. Takyr with LBA pottery dispersion.

This transect was not completed up to its target site (Site No. 340), because about 7 km from its origin a series of irrigation canals barred the way. A further attempt to reach Site No. 340 from Site No. 223, about 3.6 km north of the last point of the transect (Site No. 972), also failed because of the interposition of irrigation canals.

19

M. Cattani and S. Salvatori Most of the area we crossed was covered with sand, occasionally interrupted by small open takyr, where we recorded the presence of low-density dispersions of Iron Age pottery. The last stretch contained a series of Yaz III and IV farm‑houses, located along the eastern bank of a meandering palaeo‑channel. UTM Coordinates 429638, 4209824 429438, 4210042 429372, 4210187 429273, 4210310 429101, 4210428 428880, 4210503 428632, 4210639 428218, 4210727 428143, 4210783 427983, 4210891 427804, 4210986 427640, 4211067 427483, 4211170 427352, 4211225 427037, 4211397 426860, 4211488

UTM Coordinates 426643, 4211690

Context Takyr with low-density Yaz III pottery dispersion. Takyr with low-density Yaz III pottery dispersion. Takyr with low-density Yaz III pottery dispersion. Takyr with low-density Yaz III pottery dispersion. Yaz III farm‑house (Site No. 970). Yaz III-IV farm‑house (Site No. 971). Yaz III-IV site with farmhouses (EDA) (Site No. 972).

425299, 4213767 425141, 4213894 424691, 4214111

Context Starting point on the Auchin road. Takyr with low-density Yaz III pottery dispersion. Takyr with low-density Yaz III pottery dispersion. Takyr with low density Yaz III pottery dispersion Takyr with low-density Yaz III pottery dispersion. Takyr with low-density Yaz III pottery dispersion. Takyr with low-density Yaz III pottery dispersion. Takyr with low-density Yaz III pottery dispersion. Takyr with low-density Yaz III pottery dispersion. Takyr with low-density Yaz III pottery dispersion. Takyr with low-density Yaz III pottery dispersion. Takyr with low-density Yaz III pottery dispersion. Takyr with low-density Yaz III pottery dispersion. Takyr with low-density Yaz III pottery dispersion. Takyr with low-density Yaz III pottery dispersion. Takyr with low-density Yaz III pottery dispersion.

424653, 4214142 424395, 4214321 423702, 4217838

Transect 15 (2002) from Site No. 234 to Site No. 280. (approximately 9 km) The area we crossed is covered with an unbroken formation of hardened sand. Along our route, we met only with two small takyr with LBA pottery dispersion, and in the sandy part we detected a camp‑site with a high density of ICW pottery in a sandy depression. Much further on, near Site No. 280, we found a takyr area with a significant density of LBA pottery material, probably attesting to the presence of a monophase farm (Site No. 1483). UTM Coordinates 422881, 4229237 420525, 4227482 420061, 4227149 420039, 4227126 418603, 4221834

Context Site No. 234. Takyr with low-density LBA pottery dispersion. Takyr with low-density LBA pottery dispersion. ICW site in a sandy area (30 x 30 m). Takyr with medium-density LBA pottery concentration (20 x 20 m) (Site No. 1483).

On-the-line Transects with sherd counting Transect 1 (1994) COORD_E 418997 418882 418751 418676 418486 418376 418276 418140 418049 417942 417853 417711 417655 417526

COORD_N 4212898 4212862 4212837 4212823 4212784 4212767 4212750 4212725 4212698 4212679 4212656 4212607 4212582 4212552

FIELD_YEAR 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994

TYPE

SHERD_N° PERIOD

takyr takyr sandy area takyr takyr channel takyr takyr silty area

97 IA 1 5 0 1 8 0 1 0 260

barkhan takyr silty area

49 80

20

Transects and Other Techniques for Systematic Sampling COORD_E 417397 417215 417077 416900 416729 416609 416484 416313 416152 416021 415907 415807 415721 415536 415328 415195 415179 415096 415024 414943 414876 414750 414650 414551 414482 414417 414299 414224 414161 414096 414016 413957 413896 413742 425617 425644 425694 425743 425794 425842 425888 425955 426015 426110 426213 426265 426296 426398 426480 426540

COORD_N 4212527 4212509 4212444 4212394 4212318 4212297 4212270 4212263 4212226 4212175 4212149 4212118 4212078 4212071 4212069 4212050 4212060 4212070 4211947 4211844 4211725 4211622 4211524 4211397 4211298 4211212 4211076 4211053 4210915 4210825 4210720 4210606 4210490 4210327 4213508 4213389 4213250 4213153 4213061 4212981 4212858 4212769 4212692 4212505 4212308 4212176 4212083 4211888 4211799 4211589

FIELD_YEAR 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994

TYPE silty area silty area silty area flat area takyr silty area silty area takyr barkhan takyr takyr

1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994

takyr silty area silty area takyr takyr takyr sandy area sandy area sandy area sandy area takyr takyr takyr takyr takyr takyr

SHERD_N° PERIOD 15 0 5 26 210 136 16 0 0 9 12 0 2 14 69

takyr silty area

272 1141 155 73 IA 3; SAS; EISL 108 535 0 4 11 14 198 20 20 150 638 2041 IA 1 1999

takyr takyr barkhan takyr takyr takyr takyr takyr takyr takyr takyr takyr takyr takyr fortress

12 17 0 2 10 2 9 0 4 4 0 0 37 194

Transect 2 (1994) COORD_E 414659

COORD_N 4211795

FIELD_YEAR 1994

TYPE takyr

SHERD_N° PERIOD 69

21

M. Cattani and S. Salvatori COORD_E 414637 414603 414575 414531 414540 414528 414455 414405 414381 414336 414272 414166 414166 414168 414118 414072 413932 413778 413756 413766 413724 413671 413622 413498 413394 413349 413322 413182 413179 413079 413039 412977 412942 412878 412835 412798 412677 412642 412605 412530 412501 412464 412441 412404 412354 412217

COORD_N 4211919 4212033 4212149 4212276 4212479 4212610 4212635 4212893 4213024 4213217 4213405 4213496 4213583 4213713 4213904 4214010 4214194 4214362 4214498 4214679 4214833 4214970 4215243 4215481 4215722 4215880 4216052 4216250 4216392 4216547 4216680 4216879 4217033 4217224 4217369 4217632 4217813 4217992 4218139 4218302 4218427 4218532 4218651 4218762 4218857 4219079

FIELD_YEAR 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994 1994

TYPE takyr takyr takyr takyr sandy takyr sandy takyr takyr sandy takyr takyr takyr takyr takyr takyr takyr takyr takyr takyr takyr takyr

takyr takyr takyr

takyr takyr takyr

takyr takyr takyr takyr takyr takyr takyr takyr takyr takyr takyr

SHERD_N° PERIOD 158 6 20 66 4 6 14 7 5 473 148 LBA ; IA 3 34 18 18 37 47 36 10 4 6 10 24 3 24 20 8 14 56 LBA 26 2 13 0 0 2 105 207 141 149 1968 1999 1999 1999 1999 1999

Transects on broken lines On-the-line Transects (2000) UTM Coordinates

Distance from origin (m)

423499, 4214148

First line; first series 0-20 20-40

No. of sherds in the 16 sq m Context counting unit Takyr Takyr

15 19

22

Transects and Other Techniques for Systematic Sampling UTM Coordinates

Distance from origin (m)

Context

40-60 60-80 80

No. of sherds in the 16 sq m counting unit 24 11 -------

429892, 4205397 429898, 4205370

0-20 20-40 40-60

12 20 14

Takyr Takyr Sand dune edge

0-20 20-40 40-60 60-80 80-100 100-120 120-140 140-160 160-180 180-200 200-220 220-240 240-260

11 8 15 24 18 16 31 23 8 11 31 30 22

Takyr Takyr Takyr Takyr Takyr Takyr Takyr (ovens) Takyr (ovens) Sandy area Takyr (ovens) Takyr Takyr Sand dune edge

0-20 20-40 40-60 60-80 80-100 100-120 120-140 140-160 160-180 180-200

50 31 10 5 13 31 27 26 42 50

Takyr Takyr Sandy area (low sand) Sandy area (low sand) Sandy area (low sand) Farm-house edge Farm-house edge Farm-house edge Farm-house edge Farm-house

430120, 4205964

0-20 20-40 40-60 60-80 80-100 100-120 120-140 140-160

15 31 11 42 56 33 23 21

Farm-house Farm-house Farm-house Farm-house Farm-house Farm-house Farm-house Farm-house edge

430112, 4205930 430073, 4205915

0-20 20-40

50 0

Farm-house Channel bed

430048, 4205919

0-20 20-40

16 20

Dune edge

0-20 20-40 40-60

12 48 47

Channel bed Channel bank Farm-house

First line; second series 0-20 20-40

17 6

Takyr Sandy area

429924, 4205473 429972, 4205536 430007, 4205548

430103, 4205561

430128, 4205530 430130, 4205558

430043, 4205617

430132, 4205925

429992, 4205904 429939, 4205874

430088, 4205982

23

Takyr Takyr on sand dune Sand dune edge

M. Cattani and S. Salvatori UTM Coordinates

430200, 4206030

430209, 4206101

430239, 4206142

430218, 4206257 430135, 4206271

429985, 4206268 429959, 4206262 429960, 4206227 429871, 4206261 429845, 4206238 429829, 4206233

429756, 4206264

429733, 4206335 429757, 4206407

429730, 4206560

429776, 4206604

429806, 4206662 429811, 4206678

Distance from origin (m)

Context

40-60 60-80 80-100 100-120 120-140 140-160 160-180 180-200 200-220 220-240 240-260 260-280 280-300 300-320 320-340

No. of sherds in the 16 sq m counting unit 0 0 0 0 0 52 37 22 9 4 6 20 28 13 17

0-20 20-40 40-60 60-80 80-100 100-120 120-140 140-160 160-180 180-200 200-220

8 11 11 4 12 5 10 5 78 43 25

0-20 20-40 40-60 60-80 80-100 100-120 120-140 140-160 160-180 180-200 200-220

20 0 25 52 42 31 87 36 17 13 8

Takyr Takyr Takyr with sand Takyr with sand Takyr with sand Takyr Takyr Takyr Farm-house Farm-house Sandy area Sand dune Takyr Sandy area Takyr Farm-house Farm-house (sand-covered) Farm-house (sand-covered) Farm-house Farm-house (sand-covered) Sandy area Sandy area Sandy area

0-20 20-40 40-60 60-80 80-100 100-120 120-140 140-160 160-180 180-200 200-220 220-240 240-260 260-280 280 280-300

7 5 12 42 11 13 118 33 3 0 41 6 50 133 Oven 50

Takyr with sand Sandy area Takyr with sand Takyr Takyr with sand Takyr with sand Farm-house Takyr Takyr Sandy area Takyr Takyr with sand Takyr Farm-house Takyr Takyr

24

Sandy area Sandy area Sandy area Sandy area Sandy area Farm-house Farm-house Farm-house Ovens on takyr Sandy area (low sand) Sandy area (low sand) Takyr Takyr Takyr with sand Takyr

Transects and Other Techniques for Systematic Sampling UTM Coordinates

Distance from origin (m)

429820, 4206736 429871, 4206796 429908, 4206824 429996, 4206854 430057, 4206857

Context

300-320 320-340 340-360

No. of sherds in the 16 sq m counting unit 10 4 15

0-20 20-40 40-60

9 90 35

Takyr with sand Farm-house Takyr with sand

0-20 20-40 40-60

13 8 192

Takyr Takyr Farm-house

0-20 20-40 40-60 60-80 80-100 100-120 120-140 140-160 160-180 180-200 200-220 220-240 240-260 260-280 280-300 300-320 320-340 340-360 360-380 380-400 400-420 420-440 440-460 460-480 480-500 500-520 520-540 540-560 560-580 580-600 600-620 620-640

28 13 54 35 49 49 17 26 33 62 52 57 66 69 267 25 58 0 93 67 4 2 72 53 16 74 89 66 0 29 112 21

Dune edge Takyr Farm-house Takyr with sand Farm-house Farm-house Takyr with sand Takyr with sand Takyr with sand Farm-house Farm-house Farm-house Farm-house Farm-house Farm-house Takyr with sand Sandy area (low sand) Sandy area Farm-house Farm-house Takyr with sand Takyr with sand Farm-house Farm-house Sandy area (low sand) Farm-house Farm-house Farm-house Sandy area Sandy area Farm-house Sand

0-20 20-40 40-60 60-80 80-100 100-120 120-140 140-160 160-180

46 35 29 36 28 40 18 20 5

Takyr with sand Takyr with sand Takyr with sand Takyr with sand Takyr with sand Takyr Takyr Takyr Takyr

Takyr with sand Takyr with sand Takyr with sand

Second line 428704, 4208562

428586, 4208653 428568, 4208624

428529, 4208664 428469, 4208762

428449, 4208862

428401, 4208977

428403, 4209028 428426, 4209085

428385, 4209235

25

M. Cattani and S. Salvatori UTM Coordinates

428375, 4209317

428504, 4209423

428583, 4209360 428414, 4209391

428244, 4209544

428315, 4209620

428384, 4209620

428299, 4209679

428332, 4209745

Distance from origin (m)

Context

180-200 200-220 220-240 240-260 260-280 280-300 300-320 320-340 340-360 360-380 380-400 400-420 420-440 440-460 460-480 480-500

No. of sherds in the 16 sq m counting unit 20 17 37 37 19 35 35 28 23 43 62 130 109 22 16 26

0-20 20-40 40-60 60-80 80-100 100-120 120-140 140-160 160-180 180-200 200-220 220-240 240-260 260-280 280-300 300-320 320-340 340-360 360-380 380-400 400-420 420-440 440-460 460-480 480-500 500-520 520-540

70 29 6 2 15 67 37 20 7 47 222 24 153 131 30 85 80 47 95 91 53 11 21 56 74 15 38

Farm-house Farm-house edge Sandy area Sandy area Sandy area (low sand) Takyr Takyr Sandy area (low sand) Sandy area Farm-house edge Farm-house Takyr Farm-house Farm-house Sandy area Farm-house Farm-house Farm-house Farm-house Farm-house Farm-house Sandy area Sandy area Sandy area Sandy area Sandy area Sandy area

Takyr Sandy area (low sand) Takyr Takyr Takyr Sandy area (low sand) Takyr Takyr with sand Takyr with sand Takyr with sand Farm-house Farm-house Farm-house Sandy area Sandy area (low sand) Sandy area

On-the-line Transect (2002) Site No. 203 (Yaz II-III) The southern portion of this area has been completely modified by agricultural exploitation activities. To the north, several farm‑house units were still preserved at the moment of our visit in October 2002. UTM Coordinates

Distance from origin

411396, 4211124

10 m 30 m 50 m

No. of sherds in the 16 sq. m. counting unit 60 15 391

26

Context Takyr Takyr Farmhouse

Transects and Other Techniques for Systematic Sampling UTM Coordinates

Distance from origin

411388, 4211171

70 m 90 m 110 m 130 m 150 m 170 m 190 m 210 m 230 m 250 m 270 m 290 m 310 m 330 m 350 m 370 m End point

411445, 4211194

411419, 4211277 411405, 4211256 411407, 4211276 411422, 4211294

411380, 4211426

No. of sherds in the 16 sq. m. counting unit 200 70 31 11 132 199 303 79 140 216 325 179 151 200 294 186

27

Context Farmhouse Farmhouse sloping edge Takyr Takyr Farmhouse Farmhouse Farmhouse Takyr Sandy area Farmhouse Farmhouse Farmhouse Takyr Takyr Farmhouse Farmhouse (edge) Sandy area

Chapter 2

A GIS for the Archaeology of the Murgab Delta

1

Barbara Cerasetti Introduction

The GIS application, using the ArcMap 8.2 software, resolved various levels of difficulty such as the large size and high diversification of the data to be organised and the systematisation of archaeological data with analytical procedures. The goal of the work determined the type and amount of data to be processed, and the map was simplified into a documentation work whose purpose was mainly to record the location, size, type and time span of each site. In order to obtain the precision and consistency required in the GIS organisation, all types of data were codified on the basis of a definite scale of values. In the case of the Murgab Delta, characterised by several millennia of settlement remains distributed over a large territory, all information had to be classified into codified categories developed ad hoc in different databases. The GIS was used as an instrument for assembling and managing the vast arrays of diversified archives to be merged into a historical analysis of settlement and population dynamics. Digitised satellite images, and aerial photographs in general, are providing another tool for the definition of ancient landscapes, before and beyond the direct fields of observation (Figure 2.2).

The application of a Geographic Information System (GIS) (Wheatley and Gillings 2002; Rondelli and Tosi 2005) to manage a large amount of data, like that coming from an extensive area such as the Murgab Delta (approximately 20,000 sq km) (Figure 2.1), is now becoming the standard method adopted by a growing number of archaeological projects. Different GIS programmes are actually used for the same purpose, in order to organise numerous and different data into a homogeneous structure that allows us to undertake several types of research. In building up the Murgab Delta GIS, we had to tackle both the development and modification of archaeological research techniques, on the one hand, and hardware and software technological evolution on the other. For this reason we were forced to expend a large part of our energies in rendering an enormous quantity of data suitable for an Informative System. A First Setting of the Informative Archaeological Map: the GIS and the Remote Sensing Data

Considering the available information, at the beginning we prepared a vector cartographic digitisation of historical maps (Abbott 1843; Stewart 1881; Lumsden 1885) on different layers, in order to reconstruct the ancient irrigation system previous to the building of the Karakum Canal, which began in 1954 (Grinberg 1963: 8), and in order to obtain cartographic information for measuring the hydrodynamic and environmental evolution of the delta. A careful analysis of topographical, geographical, historical, environmental and geological maps allowed us to achieve an acceptable reconstruction of the hydraulic history of the Murgab River Delta from the Middle Bronze Age (MBA) to the Achaemenid period [Iron Age (IA) 3-4] (Figure 2.3). This was fundamental for a comprehension of the settlement-pattern distribution of the archaeological sites in the ancient landscape during the different chronological periods. The management of the data concerning the archaeological map of the Murgab Delta by means of a GIS system presented various levels of difficulty (Cerasetti 2000-2001). Considering the large size and high diversification of the data on hand, a set of graphs was produced as well, in order to provide a first level of representation of the changing population distribution

The choice of five 1:200,000 topographic military maps as reference cartography for the “Archaeological Map of the Murgab Delta” (AMMD) (Gubaev et al. 1998) was agreed upon by the coordinating agencies of the project because of the limited access to smaller-scale maps for security reasons. We recently integrated the AMMD cartography by including six 1:100,000 military maps that covered the central and southern regions of the Murgab Delta. The vector version of the 1:200,000 map mosaic is a product of the formar Cartographic Centre of the State Security Committee of the USSR in Moscow. All the strategic information was included, since the map editions used for digitising were largely outdated for military purposes. After the independence of Turkmenistan (1991), the relevant national security agencies refused access to any map on a scale smaller than 1:200,000. The 1:10,000 military maps are available only for a better ground definition of critical areas such as the Merv region. The final results of the remote sensing analyses were made possible by the Short-Term Mobility Grant (2005) of the Consiglio Nazionale delle Ricerche, Rome.  See the relative bibliography.  We thank G.A. Koshelenko and V. Gaibov. Institute of Archaeology of Russian Academy of Sciences of Moscow, for the availability of 1:100,000 topographical maps.  We thank G. Herrmann and T. D. Wiliams, Institute of Archaeology, University College London (UCL), for the availability of 1:10,000 topographical maps. 1

The collaboration with P. Mozzi and A. Ninfo, Department of Geography of University of Padua, and M. Cremaschi and A. Perego, Department of Earth Science of University of Milan, has been fundamental for the success of the project on the irrigation system reconstruction of the Murgab Delta. 

29

B. Cerasetti

Figure 2.1 View of the Murgab Delta from the south.

Figure 2.2 Some examples of satellite images used for the reconstruction of the ancient landscape: fluvial ridges (light grey) by P. Mozzi, A. Ninlo, A. Perego) with the distribution of the archaeological sites (round simbols).

30

A GIS for the Archaeology of the Murgab Delta

Figure 2.3 Ancient irrigation system reconstruction from the Middle Bronze Age to the Achaemenid period on the basis of SRTM images (2000).

in relation to the transformation of the palaeo‑channel network.

activities drastically had transformed the landscape and the settlement pattern all over the territory. A multimedia GIS makes it possible to manage many different kinds of data in thematic maps and to perform complex spatial analyses, integrating information from theoretically infinite levels of recorded data (Cattani et al. 2003).

In addition to the historical and archaeological data, aerial photos and a panchromatic and multispectral satellite imagery of the region enabled the use of new analytical methods (Bewley et al. 1999; Donoghue et al. 2006). This type of support has been used on a large scale to define ancient landscapes, in Near and Middle East archaeological projects (Adams 1981; Forte et al. 1998; Francfort and Lecomte 2002). We therefore directed our research towards the interpretation of the digital and cartographic data on a geo‑archaeological basis. Here the possibility offered by the system to compare the geo‑morphological landscape (Cremaschi 1998) with the distribution of ancient sites using the Digital Elevation Model (DEM) imagery as a background turned out to be particularly useful. The digital representation of continuous changes of relief within this space (Burrough 1986: 39; Gaffney and Stančić 1996: 23) and of the patterned variations in elevation in the territory of the delta evidently suggested the presence of palaeo‑channel segments (Cerasetti 2002) (Figure 2.4). In the past, the increase in the aridity of the climate, in spring dryness and in human economic

Main Research Applications One of the main aims of the AMMD project was to reconstruct the palaeo‑channel network of the Murgab Delta and to understand the fluctuations in the location of the major archaeological sites and of the relative water supplies. After the first reconstruction of the hydrographic palaeo‑system on the basis of historical cartography (Figure 2.5), the integration of raster and vector data made it possible to reconstruct palaeo-delta dynamics, testifying the regression of the Murgab Delta, mainly due to an ongoing desertification process that reached its climax around the middle of the 2nd millennium BC. An important wave of technical improvements came from the field of satellite imagery: on one hand, higher-resolution liberalised images allowed a direct visibility of large-tomedium archaeological sites [multispectral Landsat TM (1990) and ETM7 (2001); panchromatic 1-m resolution

See the relative bibliography. See the relative bibliography.

 

31

B. Cerasetti

Figure 2.4 Archaeological site distribution on the basis of SRTM images (2000) (by P. Mozzi, A. Ninfo, A. Perego).

IKONOS (2001)] (Galiatsatos et al. submitted); on the other hand, the US Government released the panchromatic CORONA KH-4 images (1972), dating well before the major transformations brought about by the modern irrigation works (MacDonald 1995; Day et al. 1998)10 (Figure 2.2).

Shuttle Radar Topography Mission (SRTM)11, employed to solve this specific problem, obtained elevation data on a near-global scale, in order to generate the most complete high-resolution digital topographic database of the Earth. The SRTM consisted of a specially modified radar system that flew onboard the Space Shuttle Endeavour during an 11-day mission in February 2000. It allowed us to penetrate the thick blanket of vegetation and alluvium deposits that has deeply transformed the morphology of the delta with time, and to perform a detailed analysis of the original river system originating from the Iolotan’ hub (Galiatsatos et al. submitted). Obviously the results obtained on the basis of the SRTM mosaic were integrated with the other vertical platforms such as aerial photos and CORONA and IKONOS images, as well as with topographical military maps of 1941 and 1990 (Cerasetti forthcoming) (Figure 2.6).

The main obstacle for the reconstruction of the palaeo‑system in the core of the modern delta was the thick vegetation that followed the conversion of the desert areas into agricultural fields. The satellite imagery obtained by the satellite sensor system operating in the visible ranges of the electromagnetic spectrum was not suitable for the reconstruction of the irrigation system of the Merv region, which is deeply irrigated in order to support the nearby modern town of Bajram-Ali. The We wish to thank T. D. Williams and G. Herrmann and D. Powlesland, The Landscape Research Centre (UK), for having helped us to acquire high-resolution satellite imagery of the Merv region (Ziebart et al. 2002).  See the relative bibliography. 10 All technical information concerning the remote sensing data employed for this research is consultable in Cerasetti forthcoming. 

The SRTM is an international project spearheaded by the National Geo‑spatial-Intelligence Agency (NGA) and the National Aeronautics and Space Administration (NASA) (http://glcf.umiacs.umd.edu/data/ srtm/index.shtml). 11

32

A GIS for the Archaeology of the Murgab Delta

Figure 2.5 Chronological sequence of Murgab Delta irrigation system based on historical cartography.

which are moving southwards, and by the expansion of modern irrigation works. In this case a detailed analysis of the SRTM images, integrated by positive-film CORONA ones, were fundamental for carrying out the systematic recording of sites and ancient riverbeds across the Murgab Delta core region. The agriculture-induced vegetation, covering large areas of the Murgab Delta, dramatically reduces the effectiveness of traditional land-survey techniques. The use and interpretation of CORONA and SRTM image combinations made it possible to locate and understand the pattern of the main supply systems such as natural channels and artificial canals, ancient jazireh12 (continuous, cultivable fine alluvial spaces between two water‑courses, characterised by a high agricultural yield) (Rodríguez-Iturbe and Rinaldo 1997; Buchanan 2002; Rondelli and Mantellini 2003), and wells (Cerasetti and Mauri 2002: 2-3). The Landsat images from the Thematic Mapper (TM) and Enhanced Thematic Mapper (ETM) sensors, with 30 m of spatial resolution, turned out to be essential for integrating the CORONA and SRTM imagery in the desert area, in and around the present delta fan (Figure 2.7).

Using this combination of remote sensing data, it became possible to investigate the environmental, geo‑morphological and archaeological history of the Murgab Delta with a higher degree of confidence. The integration into the GIS of the data from different archives, collected during ten years of intensive work by several specialists in different scientific disciplines, was used to reconstruct the settlement distribution pattern in a virtual archaeological landscape (Cerasetti and Mauri 2002). By using the GIS as a tool for the systematic testing of alternative explanations and simulation models, it became possible for us to utilise the archaeological data for enhancing historical interpretations of the formative stages of Central Asian Civilisations. The Murgab River flooding has apparently buried hundreds of sites and related irrigation systems under metres of silt. The exposed sections along modern irrigation canals suggest that, in the southern section of the delta fan, the early levels are buried more than five metres below the present surface (Cremaschi 1998: 22, Fig. 3). Between and beyond the present channels of the open fan, the visibility of the ancient irrigated farmlands is almost totally hindered by the sands of the Karakum,

The Arabian word jazirah which means “island” is roughly equivalent to the Greek word Mesopotamía, the land “between the rivers”. 12

33

B. Cerasetti

Figure 2.6 Archaeological site distribution on the basis of different platforms.

Conclusions13

a reference sequence in relation to the distribution of the settlements (Figure 2.5).

A more specific analysis of the physical structure of the delta, using all available satellite platforms, aerial photos and small-scale maps from different sources, was made difficult by the expansion of cultivations watered by the Karakum Canal, and the progress in remote sensing was counteracted by the spread of the cultivations that encroached on the area. These factors crippled our ability to distinguish landscape details in the braided channels of the delta and to recognise and evaluate the original ancient jazireh, which are now practically invisible and have been altered by modern works. However, by combining the different sources, we have developed a good degree of understanding, in spite of the fact that the situation becomes rather complicated at least starting from the Iron Age, when larger irrigation canals were made by diverting the water flow into dry channels of an earlier delta. The analyses were subsequently focused on the structure of the inter-nodes, in order to produce a basic model of the irrigation network, active for each period, and to establish

The modern Iolotanskaya Dam, built in 1907, follows the original natural water system. The dam carries the water coming from the Sultan’yab Canal, which originates 15 km to the south, from the Sultanbent Dam (today Kolkhozbent), to numerous sovkhoz located to the southeast of the large town of Bajram-Ali. The Sultan’yab Canal, as testified in the wonderful map made by Major-General Sir Peter Lumsden (1885), was considered until now the main water system that carried river water towards the Merv region. On the basis of the recent results coming from the SRTM image analyses, it is essential to revise the interpretation of the oriental water system of the delta, because Lumsden’s map is the basis of a great number of subsequent topographical maps. Since the MBA, the Iolotan’ sub-delta, which originates from the main course of the Murgab River close to the modern Dzhanybek sovkhoz and carries river water to the northward canals, was the main water system of the oriental region of the Murgab Delta. The SRTM images unmistakably show that the Merv region was irrigated by the channels of the Iolotan’ system; only after IA 3-

Particular thanks are due to Maurizio Tosi for his invaluable contributions, included in this article, for the final interpretation of the Murgab River system. 13

34

A GIS for the Archaeology of the Murgab Delta

Figure 2.7 3D views with contour lines obtained by means of ArcScene (by A. Perego).

4, when the Achaemenid central power became stable, main hydraulic structures such as the Sultanbent Dam were built. As shown by Lumsden’s map, the territory to the south of Merv became suitable for farming thanks to the Sultan’yab; the modern system follows the previous one, and so do the main modern canals that now carry the river water. Today the stretch of the natural branch of the Iolotan’, between the junction-point with the main course of the Murgab River and the modern lock of Turkmenkala, has been abandoned and replaced by a new canal situated 2 km to the east, probably in order to facilitate the subsequent exploitation of the Sultan’yab Canal.

Margiana economy for the Achaemenid policy was one of the main reasons of the radical repression of Darius I of the Margiana secession. The Water lack must be a problem for the subsistence of an increasing population, and the control of the water source of the Murgab River must correspond to a ‘territorial control’ of the Margiana region” (Cerasetti and Tosi 2004: 102-103) (Figure 1.8). The Parthian period was characterised by the expansion of the city of Merv and its suburban surroundings, and the final configuration of the Sultan’yab irrigation system was established. The expansion of the ancient capital, as a nodal centre of the Murgab Delta region, was fundamental for the development of the hydraulic engineering works. A vast restructuring of the Murgab water system concerned the oriental part of the delta, and immediately after 1000 B.C. the growth of the work‑force, due to the strengthening of the central power, enabled the implementation of a completely new strategy. The construction of the new parallel water system of Sultan’yab, to the east of Iolotan’, ensured the capture of water and a more direct and efficient irrigation of the whole south-eastern area, south of the Bronze Age fan.

During the Middle and Late Bronze Age the Murgab Delta was still a continuous alluvial plain where irrigated farm‑lands filled in all the spaces between the water‑courses. The settlements were spread along the main channels of the Iolotan’ system, and the irrigation was divided into parallel channels opening from northward flowing main‑streams. The political transformations of the Achaemenid Empire led to the expansion of the settlements and to the eastward extension of the cultivated lands, with new canals; a marked retreat and contraction of the southern stretch of the river began. Water was brought from main-flow channels downstream of the Iolotan’ hub; the foundation of Erk-kala probably started the first partial deviation of the Sultan’yab and the creation of the vast expanse of the south-eastern delta alluvium. The Achaemenid province of Margush established and consolidated fortress lines as shifting frontiers, “probably defending the cultivated area and the main waterworks. The oasis fertility and the importance of

The retreat of the Murgab water was not caused directly by the environmental and geo‑morphological changes, but was probably the effect of human encroachments on the natural water system during the first period of the Iron Age. The SRTM images suggest that the citadel of Erkkala was probably irrigated by the easternmost channels of the Iolotan’ sub-delta; the foundation of Gyuar-kala 35

B. Cerasetti involved a considerable increase in population and a remarkable expansion of the cultivated areas, which could hardly be supported by the old irrigation system. During IA 1 the area north of Merv was still populated, but during the Achaemenid period and the first part of the Hellenistic period the delta extensively retreated towards the south. During the Late-Hellenistic and Parthian period the northern fortresses controlled an area that by then had become a desert. The site of Yaz‑depe clearly indicates the way the change of control took place. In IA 1, Yaz was the largest site, and during the first stage of IA 3, when the fortifications of Erk-kala had already been established, it still controlled the northern provinces. Probably at the end of the 4th century B.C. Yaz-depe was abandoned and the area was occupied by the Parthian fortress of Geobekly.

Cattani, M., B. Cerasetti, S. Salvatori and M. Tosi 2003. “The Murgab Delta in Central Asia 1990-2001: GIS from a Research Resource to a Reasoning Tool for the Study of Settlement Change in Long-Term Fluctuations”, in Doerr and Sarris (eds.) 2003���������� : 125-131. Cerasetti, B. 2000-2001. “A Geographic Information System for Ancient Margiana”, Annali Istituto Universitario Orientale di Napoli 60-61: 197-214. Cerasetti, B. 2002. “A 5000-Years History of Settlement and Irrigation in the Murgab Delta (Turkmenistan). An Attempt of Reconstruction of Ancient Deltaic System”, in Burenhult and Arvidsson (eds.) 2002: 21-27. Cerasetti, B. (forthcoming). Remote Sensing in Archaeology: River Fluctuations in the Murgab Delta from the Middle Bronze Age to the Achaemenid Empire. (BAR. ������������������������������� International Series). Oxford. Cerasetti, B. and M. Mauri 2002. “The ������������������ Murgab Delta Palaeochannel Reconstruction on the Basis of Remote Sensing from Space”, in Space Applications for Heritage Conservation, Strasbourg (France), November 2002. Strasbourg. Cerasetti, B. and M. Tosi 2004. �������������������������� “Development of the “Open Frontier” between Iran and Central Asia: the Murgab Defensive Systems in the Antiquity and the Variants of the Silk Road across the Karakum”, Parthica 6: 1-6. Cremaschi, M. 1998. “Palaeohydrography and Middle Holocene Desertification in the Northern Fringe of the Murgab Delta”, in Gubaev, Koshelenko and Tosi (eds.) 1998: 15-25. Day, D. A., J. M. Logsdon, and B. Latell (eds.) 1998. Eye in the Sky: the Story of the Corona Spy Satellites. Washington-London. Doerr, M. .and .A. .Sarris .(eds.) 2003. CAA 2002, “The Digital Heritage of Archaeology”. Computer Applications and Quantitative Methods in Archaeology. Proceedings of the 30th Conference, Heraklion, Crete, April 2002, Heraklion. Donoghue, D. N. M., A. R. Beck, N. Galiatsatos, K. B. McManus and G. Philip 2006. “The Use of Remote Sensing Data for Visualising and Interpreting Archaeological Landscapes”, in Baltsavias, Gruen, van Gool and Pateraki (eds.) 2006���������� : 317-326. Forte, M., P. Mozzi and M. Zocchi 1998. “Immagini satellitari e modelli virtuali: interpretazioni geoarcheologiche della regione del Sistan meridionale”, Archeologia e Calcolatori 9: 271-290. Francfort, H.-P. and O. Lecomte 2002. ������������������������ “Irrigation et sociétés en Asie centrale des origines à l’époque achéménide”, Annales 3: 625-663. Gaffney, V. and Z. Stančić 1996. GIS Approaches to Regional Analysis: a Case Study of the Island of Hvar. Ljubljana. Galiatsatos, N., D. N. M. Donoghue and G. Philip (submitted) High Resolution Elevation Data Derived from Stereoscopic CORONA Imagery with Minimal Ground Control: an Approach using IKONOS and SRTM Data, Photogrammetric Engineering and Remote Sensing.

At present the GIS of the Murgab Delta system is being enhanced by the increasing availability of different computer application programmes that allow us to manage an enormous quantity of data coming from the long-term fieldwork project. The simultaneous advancement of archaeological techniques and technological analyses in southern Turkmenistan will greatly help, and also define, an integrated interpretative approach to complex multi-vector processes that are hidden below the changing pattern of the archaeological evidence in the cultural history of the Murgab Delta. References Cited Abbott, J. 1843. Narrative of a Journey from Heraut to Khiva, Moscow, and St. Petersburg, during the Late Russian Invasion of Khiva; with Some Account of the Court of Khiva and the Kingdom of Khaurism, I. London. Adams, R. McC. 1981. Heartland of Cities. Surveys of Ancient Settlement and Land Use on the Central Floodplain of the Euphrates. Chicago. Andrianov, B. V. 1969. Drevnie orositel’nye sistemy Priaral’ya. Moskva. Baltsavias, E., A. Gruen, L. van Gool and M. Pateraki (eds.) 2006. Recording, modeling and visualization of cultural landscape. London. Bewley, R., D. Donaghue, V. Gaffney, M. van Leusen and A. Wise (eds) 1999. Archiving Aerial Photography and Remote Sensing Data. Oxford. Buchanan, M. 2002. Nexus. Small ��������������������� Worlds and the Groundbreaking Science of Networks. New York. Burenhult, G & Arvidsson, J (eds) Archaeological Informatics - Pushing the Envelope - CAA 2001 - Computer Applications and Quantitative Methods in Archaeology, Proceedings of the 29th Conference, Gotland, April 2001. BAR International Series 1016. Oxford Burrough, P. A. 1986. Principles of Geographical Information Systems for Land Resources Assessment. Oxford. 36

A GIS for the Archaeology of the Murgab Delta Middle Zeravshan Valley”, The Silk Road Foundation. News Letter 1: 1. Rondelli, B. and M. Tosi 2005. “GIS and Silk Road Studies: Monitoring Landscape and Population Changes at Samarkand and in the Middle Zeravshan Valley”, in Uno (ed.) 2005: 459-489 Stewart, C. E. 1881. „������������������������������������ The Country of the Tekke Turkomans, and the Tejen and Murgab Rivers”. Proceedings of the Royal Geographical Society and Monthly Record of Geography. London. Uno, T. (ed.) (forthcoming) �������������� Reading Historical Spatial Information from around the World. Studies of Culture and Civilization based on Geographical Information Systems Data. International Research Center for Japanese Studies, 24th International Research Symposium (Kyoto, 7-11 February 2005). Kyoto. Wheatley, D. and M. Gillings 2002. Spatial Technology and Archaeology: the Archaeological Application of GIS. New York. Ziebart M., P. Dare, T. Williams and G. Herrmann 2002. “Acquisition, registration and application of IKONOS space imagery for the cultural World Heritage Site at Merv, Turkmenistan.” Space Applications for Heritage Conservation, Strasbourg (France), November 2002.

Gardin, J.-C. and P. Gentelle 1976. “Irrigation et peuplement dans la plaine d’Aï Khanoum de l’époque achéménide à l’époque musulmane”, Bulletin de L’École Française d’Extrême-Orient LXIII: 5-99. Grinberg, L. M. 1963. Karakumskij kanal. Ashgabat. Gubaev, A., G. A. Koshelenko and M. Tosi (eds.) 1998. The Archaeological Map of the Murgab Delta. Preliminary Reports 1990-95. Reports and Memoirs. Series Minor, III. 1st AD. Rome. Lisizina, G. N. 1965. Oroshayemoye zemledeliye epokhi eneolita na yuge Turkmenii. Moskva. Lisitsina, G. N. 1969. “The ���������������������������� Earliest Irrigation in Turkmenia”, Antiquity XLIII: 279-288. Lock, G. and Z. Stančić (eds.) 1995. Archaeology and Geographical Information Systems: a European Perspective. London. Lumsden, P. 1885. Countries and Tribes bordering on the Koh-i-Baba Range. Proceedings of the Royal Geographical Society and Monthly Record of Geography. London. MacDonald, R. 1995. “Si aprono gli archivi della Guerra Fredda”, Sistema Terra. Rivista Internazionale di telerilevamento 3: 4-15. Rodríguez-Iturbe, I. and A. Rinaldo 1997. Fractal River Basins. Cambridge��. Rondelli, B. and S. Mantellini 2003. ������������� “Methods and Perspectives for Ancient Settlement Studies in the

37

Chapter 3 The Murgab Delta in Central Asia 1990-2001: the GIS from Research Resource to a Reasoning Tool for the Study of Settlement Change in Long-Term Fluctuations Maurizio Cattani, Barbara Cerasetti, Sandro Salvatori and Maurizio Tosi Introduction

Oasis to the east (Figure 3.1). Similar in its layout to an open hand, the system is divided into a lower or “palmshaped” section to the south, where continuous cultivations form a true “Mesopotamia”, and an upper section with the lower channels radiating like “fingers” towards the north and flowing across dry sediments and desert sands. At present these branches form sequels of oases that extend the cultivated lands by approximately other 30-40 km. The fluctuations of the delta landscape during the Holocene can be analysed on the basis of the dimensions and relative proportions of these four divisions.

The Murgab River drains the northernmost corrugations of the Hindu Kush along the southern borders of Central Asia, forming a medium-sized land-locked fertile corridor between the highlands of Afganistan and the Karakum Desert in Turkmenistan. Once out of the narrow mountain valleys, the Murgab receives the last tributary waters, and its course runs northwards for other 150 km, encased in the lime‑stone basement and in a series of Pleistocene gravel conoids. As the gradients descend towards the shallow lowlands, the river fans out into a delta that irrigates a flood plain of some 35,000 sq km, divided by the elongated shallow waters of the Dzhar swamp into two distinctive sub‑systems: the Aravali Delta to the west and the Merv

Water and silt have turned the Murgab Delta into one of the largest farm‑lands of Central Asia, and into a propulsive area in the formation of the early Iranian Civilisation.

Figure 3.1 The Murgab Delta: simplified general map of natural channels and ancient irrigation networks. Note the irrigation sub‑system around Merv fed by the Sultan’yab water collector, probably built in Iron Age times.

39

M. Cattani, B. Cerasetti, S. Salvatori and M. Tosi Known to the Greeks as Margianaē or Margiana in Latin, the country is first mentioned as Margush in the lists of the provinces ruled by the Achaemenid king of kings since the end of the 6th century BC. After Alexander’s conquest in 332 BC and the opening of direct trade relations with China, Margiana developed as a nodal point along the Silk Road. Trade and industry made its capital Merv, founded around 500 BC a legendary place: after the Arab conquest, the city gave its name to the whole region. Today its ruins form an impressive compound of different cities, covering an area of several hundred hectares: in 1997 UNESCO added it to the World Heritage List (Herrmann et al. 2000).

towards north-north-west, from a maximum height of 250 m asl at the breaking point in Iolotan’ to a minimum height of 170 m asl at the end of the water flows near the caravanserai of Sheikh Mansur, 200 km downstream. The Murgab buried hundreds of sites under several metres of silt. The only reliable window of observation over some considerable sections of fossil landscapes were left behind by a southward retreat of the delta, presumably after 1000 BC, and probably in connection with the creation of the Sultan’yab water collector and the vast irrigation schemes around Merv (Figure 3.1). This area of higher visibility extends for some 140 km east-west and 200 km northsouth across the eastern delta, between latitudes 37°20’ - 38°40 N and 61°20’ - 62°20’ E, about 200 m asl. The exposed surfaces are fine alluvial sediments that were farmed during most of the 2nd millennium BC, occupied by mounds and shallow sites of the Late Bronze Age, while earlier ones of the Middle Bronze Age are mostly buried under silt.

Archaeological surveys and excavation works began in the nineteen-fifties and were focused on the upper and lower sections of the eastern delta, in connection to Merv and its close surroundings (Masson 1959; Masimov 1979; Sarianidi 1990; Gubaev et al. 1998). Not surprisingly, explorations revealed that the history of agriculture and settlement in the Murgab Delta had begun in later prehistoric times, long before the foundation of Merv. The earliest evidence dates to the beginning of the 3rd millennium BC, in the Early Bronze Age (Sarianidi 1990). Like all the other alluvial flood plains across the arid lands between the Aral Sea and the Indian Ocean, the Murgab Delta was an area of convergence in the complex political and cultural developments that accompanied the emergence of early states and urban societies east of Mesopotamia and of the Iranian Plateau (Tosi et al. 1992; Sarianidi 1993; Hiebert and Lamberg-Karlovsky 1992).

Immediately to the south of this area, only the Iron Age sites become visible, while Bronze Age ones lie 2 to 4 metres below the present ground level. Further southwards, Parthian and Sasanian mounds dominate the skyline, and neither Iron Age nor Bronze Age sites can be detected, even in the exposed sections of irrigation canals; this suggests that the early levels are buried more than 5 metres below. Northwards, the situation does not improve. Between and beyond the present channels of the open fan, the visibility of the ancient irrigated farm‑lands is almost totally hindered by the southward-moving sands of the Karakum, which cover all signs of prehistoric occupation with a continuous blanket.

It is a well-established assumption among archaeologists and historians that in dry alluvial lowlands, settlement fluctuations are almost exclusively related to irrigation and agricultural productivity. Since irrigation requires, besides water, the organisation of human labour, the scale of the irrigation works is expected to be related to the level of political complexity. Settlement hierarchies and the projected sizes of the related farm‑lands have long been considered direct signatures for scaling the magnitude of political systems. The limited extension and relative isolation of the Murgab Delta provide ideal conditions for testing these theoretical assumptions. Large-scale irrigation works altered the layout of the delta several times during the Late Holocene, from the emergence of early states around 3000 BC to the most radical transformations brought about by the Soviet Power after 1960 with the construction of the Karakum Canal, the largest water‑works scheme ever carried through.

The Archaeological Map of the Murgab Delta In its original formulation, The Archaeological Map of the Murgab Delta was designed to carry out the systematic recording of sites and palaeo‑channels across the Merv Oasis, before most of them disappeared because of the expansion of the irrigation works for the Karakum Canal. Its straightforward aim was to reconstruct landscape and settlement variations along the traditional research lines established by R. McC. Adams in the Mesopotamian lowlands (Adams 1965) and by J.-C. Gardin in Northern Afganistan (Gardin 1980). After the very first seasons it became evident that in the north a very large number of earlier sites was buried under sand or silt, while among the later Parthian-Sasanian and Medieval ones in the south only the higher mounds had survived the intensification

What distinguished the Murgab Delta from other landlocked alluvial areas in Middle Asia was its relative stability during the Holocene. Its branching water‑courses were far less erratic than those of Helmand or Tarim, and as a consequence of this there are no vast fossil landscapes dotted with surviving ghost towns, as in Seistan or Xingjian. The result was that over the past 5000 years the sediments built up a stair-like sequence of descending platforms

The Archaeological Map of the Murgab Delta was originally designed as a joint research project by the Institute of Archaeology of the Russian Academy of Sciences (IARAS) in Moscow, the State University of Turkmenistan (TSU) in Ashgabat, the Italian Institute for Africa and the Orient (IsIAO) in Rome and the University of Naples “L’Orientale” in Naples. 

40

The Murgab Delta in Central Asia 1990-2001 of agriculture. There was no possibility of developing any reliable reconstruction of the agricultural landscape or of drawing a population estimate from the settlement data. However, though on the one hand visibility was critically limited, on the other hand the technical means of survey work had been radically improved by the “geonomics revolution” that hit archaeology in the early nineteennineties. GPS were available to us from the third field‑work season, together with Total Stations directly linked to computers powered by a variety of softwares for recording and handling very large amounts of data. A second wave of technical improvements came from new developments in the satellite imagery made available to the public: on the one hand higher-resolution images ensured the direct visibility of medium-to-large sites, on the other hand the US Government released the CORONA images dating to the nineteen-sixties, before the transformations brought about by the construction of the Karakum Canal.

archaeological aspects that emerged as the survey revealed the high level of preservation in the best-exposed sections of the delta country. Rules and parameters had to be redefined. First of all, the fact that – with various degrees of density – artefacts are ubiquitous across the flood‑plain whenever even a small patch of alluvial soil is exposed among the sands, made the distinction between site and non-site became a matter of discriminatory thresholds, to be arbitrarily determined after an evaluation of the context. The fact is that the stability of the delta has allowed the preservation of the least conspicuous settlement remains: most of them, 70-80% of the surfaces identified as sites, were shallow scatters of artefacts with deposits less than 1 metre high. Mounds still reveal the remains of central or important settlements, marking their rank, because their mass derived from brick platforms or other monumental remains. They can be used as nodal points in Thiessen Polygons or any other Central Place Theory (CPT) representation, but no population estimates could be made beyond the narrow limits of the micro-regional dimension, left exposed between the sand and the silt. We wonder to what extent we can rely on earlier survey works in other alluvial lowlands in south-west Asia, if they were based on population estimates based on an evaluation of the number of mounds.

Quite obviously the project had to be radically re-organised: from a systematic documentation work and a quite linear historical reconstruction, it was turned into a complex web of methodological issues. This was not the only motivation for changing the first framework of the project. The complexities faced by the working teams were exponentially increased by the

Figure 3.2 The Archaeological Map of the Murgab Delta. Transects carried in Togolok and Takhirbai areas. Legend: graduated size circles according number of sherds; symbols and polygons = Late Bronze Age sites; dotted area = massive aeolian sands; meander lines = ancient river‑beds recognized from aerial photographs; enclosed green areas = takyr playas; straight lines = modern canals.

41

M. Cattani, B. Cerasetti, S. Salvatori and M. Tosi Systematic walking transects were carried out across the exposed section of the Merv Delta between couples of main sites, counting sherds per standard units of surface (Figure 3.2). The resulting intersection lattice allowed the definition of settlement spaces and their functional repartition into different classes of indicators, along the lines developed for the Middle East by T. J. Wilkinson (1982; 1989). Figure 3.2 indicates some of several transects carried out in the Murgab Delta survey. In the best-detailed area, between Togolok and south of Takhirbai 3, five different transects connect central points in the settlement lattice. From Togolok 1 to Sites No. 126-148, sherd counting allowed us to disarticulate the complex aggregations forming the MBA-LBA centre of Togolok (Site No. 190) and to identify some secondary small sites, for a better reconstruction of the rural landscape. It also made it possible to view the edge of alluvial deposits that hide Bronze Age settlements towards the south. Between Sites No. 148 and 126, an increasing density of potsherds corresponds to the first settlement complexes of the Achaemenid period. Similar evidence, with the exclusive presence of Iron Age pottery, is attested in the transect carried out between Sites No. 148 and 64 (Takhirbai 1). More transects were carried out in order to evaluate the impact of aeolian sands in the LBA, focusing on the lack of settlements (transect between Sites No. 638 and 172), or on the presence of Andronovo sites in desert zones (transect between Sites No. 64 and 972). Furthermore, the transect between Sites No. 172 and 215 indicates a continuous but scanty presence of Late Iron Age pottery, suggesting manuring rather than actual settlement remains.

works, would destroy all sites of Category 3, over 90% of those of Category 2, and only a fraction of the higher mounds of Category 1 (Kirkby and Kirkby 1976; Miller Rosen 1986). The exceptional preservation of few camp‑sites in Margiana, as well as that of a very large number of shallow settlement areas, was made possible by the retreat of irrigation and the advance of the desert. We should point out that once camp‑sites are targeted in walking transects, a higher proportion of them can be recovered on top of settled sand dunes. This is the result of two different conditions: the fact that the nomads had camped in the desert or on the sand patches advancing over the farm‑lands, or the fact that the potsherds scattered on the surface are the result of eruptions from topsoil underlying the alluvial silt. The two situations can be easily distinguished by carrying out a closer scrutiny and small test excavations. Since both these situations often occur on the same site, the detection and study of campsite remains requires targeted research procedures. As a rule, a statistical analysis of quantitative data spread sheets would rarely be meaningful if based on a direct incorporation of the surviving evidence. It is necessary to project the restricted patches of surviving evidence onto reconstructing models. Ethno‑graphic data on patterns of mobility, herd composition and camp architecture become the essential tool for any future theoretical construction, if we are to overcome the limitations of a record formed of mounds (Nechaeva et al. 1943; Pletneva 1981). Intrusive Camp‑site Scatters of Alien Pastoralists

In general, our survey work across Margiana indicates three main categories of sites: 1.

Mounds, formed of the massive remains of superimposed architectural masses, which correspond to long-duration central sites;

2.

Low Elevations, from artefact concentration and shallow architectural remains, which correspond to medium-to-short-duration dwellings and industrial areas of permanent settlements;

3.

Scatters of Artefacts, with almost no surviving sediment, which indicate seasonal occupation.

In the course of every archaeological project there is a degree of unexpected discovery, and the AMMD has been no exception to this rule. The closer scrutiny of the ground surface from a multiplication of walking transects produced the unexpected identification of dozens of seasonal camp‑sites scattered across the alluvial plain (Figure 5.8). They indicate that, during a developed stage of the Late Bronze Age, around 1700 BC, a significant space across the delta was allocated to seasonal animal breeders. The important aspect is that the ceramic assemblage associated to these camp‑site remains does not have any relation with the local tradition of fine-tempered wheel-made pottery. The majority of the sherds are a coarse ware with incised or impressed decoration (ICW), considered the most characteristic signature of the Andronovo Culture Complex, which spread across the Eurasian steppes during the Bronze Age for most of the 2nd millennium BC (Teploukhov 1927; Gryaznov 1966). With minor typological variations, its material culture covered a large part of the steppe grasslands east of the Ural Mountains, up to the borders of China. For many of the specialists, Andronovo marks the beginning of pastoral nomadism and represents the formative stages of the civilisation of Scythian and Saka in the 1st millennium BC (Kuz’mina 1994; for a recent critical review of the whole problem,

There are intermediate situations between these three categories, but the lack of natural elevations from terraces or rocky outcrops creates a single uniformitarian situation across the whole delta landscape. Many more sites are buried under the silt of later alluviation, and are only exposed by occasional and archaeological excavations; they cannot be included in a classification strictly related to survey work. We assume that the expected rate of preservation is directly related to building volume of and artefact density per square unit of surface. Natural erosion, and to a greater extent ancient and modern agricultural 42

The Murgab Delta in Central Asia 1990-2001 see Lamberg-Karlovsky 2002). Vast barren deserts cross the continent landmass, dividing the Eurasian steppes from the farm‑lands along the mountains that form the northern borders of Iran and India. The rivers that drain into Central Asia from the northern watershed of the Hindu Kush, including the Murgab, form fertile corridors that deeply cut the desert waste lands and connect those outposts of agricultural civilisation to the grasslands of the northern steppe, home of different evolutionary pathways. The discovery of Andronovo ceramics intruding into specialised camp‑sites among the established agricultural communities in the Murgab Delta gave direct indications for the beginning of this process of interchange.

LBA towns and villages. These occurrences increase in density and frequency particularly during the final phases of the Late Bronze Age. Vessels or potsherds are found sometimes above the floors of main buildings of the farmers’ central towns (e.g. Togolok 1) or abundantly spread on the surface of several sites (e.g. Takhirbai 3). Andronovo sites are particularly abundant along radial axes starting from Site No. 67 (Takhirbai 4), where a more intensive and systematic survey has been carried out. In this case we can assume that there were complementary pastoral activities that integrated the farming production of a sedentary population. In particular it is evident that pastoral camp‑sites are prevalently distributed across desertified zones, transversally to irrigation canals flowing from the north. The Togolok area is characterised by a presence of steppe-like pottery that is more important than that of LBA sites, and this suggests an increase in population density during their final phases. According to the evidence from Takhirbai 3 (Masson 1959), during the Final Bronze Age a higher level of integration between nomadic and sedentary peoples was attained. However the presence of an Andronovo graveyard still indicates a degree of ethno-cultural diversity.

This exceptional preservation of nomadic camp‑sites, as well as of a very large number of shallow settlement areas, was made possible by the retreat of irrigation and the advance of the desert. Walking transects, when they were focused on camp‑sites, allowed us to recover a higher proportion of them on top of settled sand dunes. Sites with materials of steppe-like tradition, including ICW potsherds, fired clay or stones, are frequently located in playas (takyr) free from the sand cover. Several sites are situated on top of stabilised sand dunes, confirming the co-occurrence of sand invasion and nomadic camp‑sites. This stems from two different causes: the fact that the pastoralists had camped in the desert or on the patches of sand that were already invading the farm‑lands, or the fact that the scattering of the potsherds on the surface is the result of an eruption from the silt top‑soil buried underneath. A distinction between these two situations can be easily made by means of a closer scrutiny and small test excavations; in many cases, while often both these situations occur in the same site.

The intrusive phase of camp‑sites, dated between 1700 and 1400 BC, is concurrent with significant changes in the landscape. A detailed geo‑morphologic survey of the eastern delta carried out by M. Cremaschi (1998) indicates that a direct correspondence may have developed between the establishment of nomadic camp‑sites and the advance of aeolian sands over the alluvial farm‑lands. Undoubtedly, synergies rather than conflicts can better explain the fine network of intersecting farmers and pastoralists that resulted from a compensating strategy devised by the local farming communities in order to meet the diminishing returns from irrigation farming across a territory invaded by sand and salt. The small seasonal camp‑sites presumably represent not an “invasion” event, but a convergence process that went on perhaps for two hundred years at the middle of the 2nd millennium BC around 1300 BC, with the emergence of the new Yaz I culture, characterised by painted pottery, there is no more evidence of a material culture related to the steppe. The ceramic types of this period both from settlements and from camp‑sites are consistently the same.

In a few other instances, camp‑sites with ICW are located within areas of cultivated fields and canals, around

In order to evaluate, also in quantitative terms, the impact of the northern pastoralists on the local population of established farmers, we need to analyse the evidence, comparing it with the environmental settlements and cultural changes over a long period, that is more than a thousand years between the Middle Bronze Age and the Early Iron Age, before and after the Andronovo intrusion. At present we can divide this period into five or six phases (Figure 3.3), though this division is probably still too coarse to allow us to obtain an appropriate profile of the variability across the supra-regional dimension of the delta.

Figure 3.3 Chronological scheme of Bronze Age in Margiana and relationships with surrounding areas.

43

M. Cattani, B. Cerasetti, S. Salvatori and M. Tosi The widespread distribution of camp‑sites across the exposed surfaces indicates that the original scale of the phenomenon can only be established through a detailed re-examination of all the exposed surfaces. More walking transects will be necessary to detail the extent and duration of the camp‑sites on the basis of the density of the small potsherd scatters, since the scale is too small to allow air photographs or other means to support the evidence. Although very few of the sites found so far have any archaeological deposits left, excavations may provide interesting clues for a reconstruction of the particular societal and economic conditions.

Andronovo. Moreover, if we plot the Andronovo sites on the basis of size classes, the results are deceptive, since it is highly possible that many of them are not only seasonal camp‑sites. Figure 3.4 indicates the frequency per sizes, distributed into four main site classes. Highest values are around 1,000 sq m, equivalent to an area of approximately 30 x 30 m. This may correspond to small camp‑sites with one or only a few dwellings and other small features relative to pastoral or related processing activities. In calculating the size, we must consider that the remains of herding sheds are hardly recognizable in field research. The other dimensional classes may indicate either larger camp‑sites with an industrial activity or settlements where there is an integration between herding and farming peoples. Some of the largest camp‑sites found so far, for instance Site No. 1211, are large enough for permanent habitations and multiple industrial activities, and contain indications of them.

The GIS as a Tool for Model Building and Research Planning The presence of encampments and other seasonal installations of animal breeders was a common feature across the rural landscape of the Ancient Orient (Hole 1974; Cribb 1991). The exceptional conditions of preservation offered by the Murgab Delta open a new perspective for the modelization of past economies on the basis of survey data. Any future quantitative elaboration will have to begin from the consideration that no matter how many camp‑sites are found, they will always remain a fraction of the total number.

Thus, in order to develop an appropriate analytical tool within a GIS frame (ArcView), we need to establish an intense interplay between the field and the computer, almost on a day-by-day basis, in order to incorporate the different options for classification categories, developing the scale measure. The optional definition must be tested in parallel procedures, at best by means of simulations that compare the cultural and functional data from the archaeological record with the environmental variables relative both to alluvial regression and to sand infiltration. The fact is that we have to establish to what extent the success of the nomads was determined by political decisions or environmental adaptations. One is a function of the other.

In order to address these further explorations to the historical dimensions of the problem, a GIS might provide us with the most appropriate analytical tools for visualising the concurrent variability across the archaeological record. Standard methods of site ranking, mostly based on classifications per type and size, turned out to be totally inadequate for organising the complexity and discontinuity of the sites identified across the Murgab low‑lands. In pastoral camp‑sites there are elements of functionality not directly related to the size and number of households (Khazanov 1994). The animal herds are a measure of wealth and power, and their size is only loosely connected with the number of people. This is particularly evident when we have no clue for the techniques used to control the animals and the pastures, as is still the case for

Considering the cultural integration reached around 1300 by the Yaz I culture, the Andronovo infiltration turns out to be an episode of convergence that may have lasted from 100 to 300 years, and we need to define its occurrence in the greatest detail, in order to transfer the archaeological data to the level of historical interpretation. For this reason, the study of the Andronovo in Murgab is a great opportunity to test the value of the GIS for future investments in archaeology. Finally, a most significant aspect where the GIS can be developed into a research tool is its application in reconstructing the ancient landscape. Present-time geomorphological analyses are often off the right path for scientific research and can be used only if it is demonstrated that there have been no changes with respect to the ancient landscape. As a reasoning tool, the GIS appears to be highly useful, because it combines archaeological and geo-morphological records. Areas lacking data do not correspond to an absence of ancient settlements. Some geo-morphological evidences such as buried Bronze Age remains suggest that alluvial deposits created by some of the ancient river branches may hide the location of several settlements in the area. A presumable area of alluvial sediments deposited progressively during the first part of

Figure 3.4 Frequency of ICW site surface.

44

The Murgab Delta in Central Asia 1990-2001 the Iron Age was located by combining points of depth of the Bronze Age soils with the absence of contemporary sites on the surface. By reasoning with selected data in the GIS, we can direct future researches so as to exclude the presence of ancient remains or to determine their position. For this purpose, test trenches and observations on the most exposed sections, together with bore-drillings, will make it possible to estimate the thickness of alluvial deposits in the same area.

Hiebert, F. and C. C. Lamberg-Karlovsky 1992. “Central ��������� Asia and the Indo-Iranian Borderlands”, Iran XXX: 115. Hole, F. 1974. “Tepe Tul’ai, an Early Campsite in Khuzistan, Iran”, Paléorient 2/2: 219-42. Khazanov, A. M. 1994. Nomads and the Outside World. Cambridge. Kirkby, A. and M. J. Kirkby 1976. “Geomorphic Processes and the Surface Survey of Archaeological Sites in Semiarid Areas”, in Davidson and Shackley (eds.)����������� : 229-253. Kohl, Ph. L. 1984. Central Asia. Palaeolithic Beginnings to Early Iron Age. ������ Paris. Kuz’mina, E. E. 1994. Otkuda prishli Indoarii? Material’naya kul’tura plemen andronovskoj obshchnosti i proiskhozhdenie indoiranzev. Moskva. Lamberg-Karlovsky, C. C. 2002. ����������������� “Archaeology and Language: The Indo-Iranians”, Current Anthropology 43: 63-88. Masson, V. M. 1959. Drevnezemledel’cheskaya kul’tura Margiany. (Materialy i issledovaniya po arkheologii SSSR, no. 73). Moskva. Masimov, I. S. 1979. “Izuchenie pamyatnikov epokhi bronzy nizov’ev Murgaba”, Sovetskaya Arkheologiya 1: 111-131. Miller Rosen, A. 1986. Cities of Clay. The Geoarchaeology of Tells. Chicago. Nechaeva, N. A., I. A. Mordvinov and I. A Mosolov. 1943. Past’ishcha Kara-Kymov i ik ispol’zovanie. (Izdatel’stvo TurkmenFAN). Ashkhabad. Pletneva, S. A. (ed.) 1981. Stepi Evrazii v epoky srednevekov’ya. ����������������� Arkheologiya SSSR. Moskva. Sarianidi, V. I. 1990. Drevnosti strany Margush. Ashkhabad. Sarianidi, V. I. 1993. ���������������������������������� “Margiana in the Ancient Orient”, IASSCCA Information Bulletin 19: 5-28. Teploukhov, S. A. 1927. ����������������������� “Drevnie pogrebeniya v Minusinskom krae”, Materialy po etnografii III (2): 57-112. Tosi, M., S. Malek Shahmirzadi and B. K. Thapar 1992. “The Bronze Age in Iran and Afghanistan”, in Dani and Masson (eds.): 191-224. Wilkinson, T. J. 1982. “The Definition of Ancient Manured Zones by Means of Extensive Sherd-Sampling Techniques”, Journal of Field Archaeology 9: 323-33. Wilkinson, T. J. 1989. “Extensive Sherd Scatters and Land-Use Intensity: Some Recent Results”, Journal of Field Archaeology 16: 31-46.

References Cited Adams, R. McC. 1965. Land behind Baghdad: a History of Settlement on the Diyala Plains. Chicago. Adams, R. McC. and H. J. Nissen 1972. The Uruk Countryside. The Natural Setting of Urban Societies. Chicago. Burenhult G. and J. Arvidsson (eds.), Archaeological Informatics: Pushing the Envelope CAA 2001. Computer Applications and Quantitative Methods in Archaeology. Proceedings of the 29th Conference, Gotland, April 2001. BAR International Series 1016. Oxford. ������� Cerasetti, B. 2002. “A 5000-Years History of Settlement and Irrigation in the Murgab Delta (Turkmenistan). An Attempt of Reconstruction of Ancient Deltaic System”, in Burenhult and Arvidsson (eds.)2002�������� : 21-27. Cremaschi, M. 1998. “Palaeohydrography ������������������������������ and Middle Holocene Desertification in the Northern Fringe of the Murgab Delta”, in Gubaev, Koshelenko and Tosi (eds.) 1998: 15-25. Cribb, R. 1991. Nomads in Archaeology. Cambridge University Press. Dani, A. H. and V. M. Masson (eds.) 1992. History of Civilisations of Central Asia, vol. I, Paris. Davidson, D. A. and M. K. Shackley (eds.) 1976. Geaoarchaeology: Earth Science and the Past. London Gardin, J.-C. 1980. “L’archéologie du paysage bactrien”, Compte-Rendus de l’Académie des Inscriptions et Belles-lettres: 480-501. Gryaznov, M. P. 1966. Andronovskaya kult’tura. Moskva. Gubaev, A., G. A. Koshelenko and M. Tosi (eds.) 1998. The Archaeological Map of the Murgab Delta. Preliminary Reports 1990-95. Reports and Memoirs. Series Minot, III. IsIAO, Rome. Herrmann, G. et al. (eds.) 2000. ������������������������ “The International Merv Project. Preliminary Report on the Ninth Season (2000)”, Iran XXXVIII: 1-150.

45

Chapter 4 Non-graphic Information Systems and Diachronic Transformations in Margiana Joshua Wright

Introduction

between settlements in Margiana. I choose this because a portion of the basic work of settlement hierarchy has already been done with this survey data for the major periods of interest (Salvatori 1998b; Genito 1998b). This paper offers a look at the Margiana data from a direction that is free of the divisions created by the conventional chronological framework. Though this chronology (Table 4.1) does form an essential part of this project, it is in the position of a set of data being testing rather than a framework in which testing is to take place.

Settlement pattern analysis is the primary archaeological method of studying socio-political organisation in complex societies. Over the past 50 years, a regional survey has been carried out in a wide variety of areas and conditions (Willey 1953; Parson 1971; Puleston 1983; Smith and Parsons 1989; Fish and Kowalewski 1990; Shelach 1999; Holdaway et al. 2000; Phillips and Watson 2000; Honeychurch 2003; Wilkinson 2003) and using many different tactics (Parson 2004). Survey has proved to be a highly effective method of increasing our understanding of, and altering previous interpretations of, past political and spatial organisation. Throughout this time the methods used to collect and analyse survey data have changed with the times and made use of new technologies that have become available. These methods and technologies, however, have always been tempered by the need for good, solid, artefact sequences and field walking.

Five different analytical investigations will be undertaken in the paper that follows: first, the nature of changes in settlement regions in relation to the movement of desert and river alluviation. Second, the dynamics of site size, population and the nature of change of settled areas. Third, an examination of how the sites are interrelated across their given chronological boundaries and what this could mean for social continuity in the region. Fourth, a brief examination of changes in water control and access. And finally, a study of self-organised criticality in settlement growth and decline.

Margiana ‑ geographically the delta of Murgab River, between the Kopetdag Mountains and the Karakum Desert ‑ has yielded a palimpsest of archaeological landscapes spanning many millennia and including the advent of complex societies, cities and empires in the Near East and Central Asia. These data have been recovered by recent regional surveys and through a long period of excavations by Soviet, Turkmen and foreign scholars. The results of this work provide data that can be used to study the sociopolitical and spatial dynamics of complex societies in the long term, the nature of particular historical periods and the delineation of those historical periods.

The data for this project is drawn from a series of surveys performed in Margiana during the nineteen-nineties (Gubaev et al. 1998; Cattani, personal communication). A large amount of survey, archaeology and historical data was collected, and was analysed separately by chronological period. Gubaev’s (1998) volume includes some basic analysis of Bronze Age settlement patterns (Salvatori 1998a, 1998b) and historical discussion of the Iron Age (Genito 1998a), Achaemenid (Genito 1998b) Parthian (Bader et al. 1998), Seleucid and Medieval periods among others. Other work on the region includes that of Sarianidi and Masson (1972 for a general survey), Hiebert (1992, 1994), and Hiebert and Lamberg-Karlovsky (1992), to name only a few.

When I first studied the theory and practice of geographic information systems it was impressed upon me that a GIS is only a visual database and that the organisational power of a GIS is built on the organisational, and analytical, power of the database that backs it up. To that end, we were trained to look first at the database of a GIS and understand its structure and what sort of analysis could be done with the non-graphical data alone. This is what I have done here. Many of the analyses and interpretations here could be carried over to the geographic and graphic medium of the GIS, but that is beyond the scale of the current project, which is only an exploratory and interpretation exercise.

This paper does not set out to argue against any of these works, to test their conclusions, or offer a more in-depth look at regional history. Instead it will make an independent analysis of long-term regional survey data (Cattani, personal communication) in the hope of offering some parallel suggestions about cultural historical process in Margiana and making a simple contribution to the study of this region.

This is a diachronic analysis of changes in the large-scale patterns of land use and in the networks of interrelations 47

J. Wright Period BA MBA LBA

Bronze Age Middle Bronze Age Late Bronze Age

Dates 2400-1500 BC 2400-2100 BC 2100-1500 BC

Span 900 300 600

FBA

Final Bronze Age

1800-1500 BC

300

IA IA 1 IA 2 IA 3 IA 4 HEL SEL? PAR SAS EISL ISL/MED

Iron Age

1500 BC 1500-1000 BC 1000-600 BC 600-300 BC 450-300 BC 4th cen. BC 330-130 BC c. 130 BC-213 AD 216-650 AD 650-850 AD 850-1250 AD

1200 500 400 300 150

Hellenistic Seleucid Parthian Sassanian Early Islamic Islamic/Medieval

Notes Namazga V, Kelleli Gonur Phase Namazga VI, Gonur Dashly Namazga VIa Namazga VIb Takhirbai, Dashly Terminal LBA Togolok Phase Yaz I Yaz II Yaz III Achaemenid

470 430 200 400

Table 4.1 Chronological periods used in this paper. (Drawn from Kohl 1984, 1981); Gubaev et al. 1998; Boyle and Fisher 1999)

Environmental Change: Shifts in the Alluvium and Desert

northern end of the delta. On the other hand, the shift could be due to a shift in political and economic control towards the southern end of the delta, perhaps with the founding of the city of Merv (Hiebert 1992; Herrmann et al. 1994).

As the environment of Margiana changes through time, the fertile region of the oasis shifts. At the southern end of the Oasis the lateral movement of the river is more pronounced, with land (and archaeological sites) being destroyed or buried. This is not a substantially higherelevated region that the rest of the oasis. At the northern desert, end of the oasis, the sand is constantly encroaching especially in periods when the climate is arid. This pattern of change affects the location of archaeological sites in two ways: first it makes the discovery of older sites both at the northern and southern ends of the oasis more difficult. And second, probably more important that the first point, the shifting fertility of the oasis affects where people will choose to live.

There are a few other details of note between the other ranges. One of them is the Late Bronze Age expansion into the southern reaches of the oasis, though it could be suggested that the core region of the Bronze Age remained within the range shown by the Final Bronze Age range. Other factors that could affect this distribution include a shift in site types. A change away from depe or large open sites and agricultural regions which will leave sherd pavements in the desert, both easy to detect on survey, could make the site ranges contract towards their densest areas. A further interesting suggestion is that the Late Bronze Age and the Parthian period could be periods of an influx of pastoralists into the oasis and have seen a resultant change in land use patterns.

In order to understand this, the sites were grouped on a south-north line for each period. The ranges for each period are shown in Figure 4.1. When this data was prepared, a single extreme outlier was removed from both the Parthian and Middle Bronze Age.

Site Size Dynamics Site size is often considered to be a proxy for regional population. In this example we will do the same, with the caveat that in this region site size reflects for the most part the non-nomadic population. Figure 4.2 shows the total area of sites in the survey area in terms of sites per-century during any given period. The curve shows a falling area of sedentary population in the Early Iron Age and between the Late Iron Age and the Parthian period.

The most distinctive thing about this chart is the drastic shift between the Parthian and Late Iron Age (IA 4) periods. With the available data is only possible to speculate about the cause of this change. On the one hand, it could be due to a major shift in environmental conditions. Either an increased period of fluvial activity that wiped out record of all earlier sites at the southern end of the delta or an increased period of aridity that forced an abandonment of 48

Non-graphic Information Systems and Diachronic Transformations in Margiana

Figure 4.4 Percentage of sites that are also sites of the previous period. This table shows sharp discontinuities at material culture boundaries and connected stability within periods.

Figure 4.1 Relative south to north distribution of sites in each period.

Figure 4.3 presents a more in-depth look at variations in site size over time. It depicts a count of sites per century in any given period, compared with a curve of average site size. Combined, these two values illustrate the relative agglomeration of population — between many small sites and few large sites. As in the previous example (Figure 4.2), falling lines are a decrease in settled area or a spreading out of population if the number of sites rises while the average size falls. A rising curve displays the reverse — if the site count falls while the size rises, sedentary population is coalescing. Figure 4.3 can be interpreted as an illustration of population and settlement history. The Bronze Age is a period of relative stability in site size with a large increase in the number of sites in the Late Bronze Age, an expansion of settled areas and settled population in an obvious interpretation. The transition to the Iron Age brings population together, decreasing the number of sites and increasing their average size. In IA 1 and IA 2 the site size is dropping while the numbers remain the same. This suggests fewer settled people and perhaps a spread-out settlement pattern. IA 3 shows a major increase in the number of sites and total settled area but size is still relatively low. This suggests an overall stability and consistency of settlement during the middle of the Iron Age. During the final period of the Iron Age (IA 4), sites are growing in size and settled population is decreasing. Overall, the Iron Age is a period of fluctuating settled population. The Iron Age-Parthian transition drops the site count to the lowest level for millennia, after which population and site size begin to climb towards the Medieval period.

Figure 4.2 Total site size area as a proxy for settled population.

Site Continuity across Periods More than one third of the sites found on the survey continue from one period to the next. As a single figure, this suggests a strong continuous habitation in which the foundation

Figure 4.3 The count of sites and average site size per century in any period.

49

J. Wright and abandonment of sites was a very local process — not concerned with larger cultural and technological changes, but only with the immediate situation of a site.

distributed in relation to water (natural or canalised). This simple model was created by plotting out the sites along a flat east-west continuum (using their UTM coordinates). A dispersed or evenly arranged pattern of sites should suggest a broad use of the landscape, while a strongly centred pattern suggests sites gathered along a north-south axis — such as a major channel.

To examine this, the percentage of sites that continued from the previous period and/or continued into the next was determined for each period where sufficient data was available. The most informative results are displayed in Figure 4.4.

Figure 4.5 depicts the east-west distribution of sites across the survey area during the various periods. The periods with the broadest distribution of sites are the Bronze Age (MBA and LBA) and the Parthian period (PAR). This would suggest that at the transition to the Iron Age and by Sasanian times, forces were at work that drew settlements in, around the north-south axis, and that this was more obvious in the later periods. In the Iron Age, centralisation is less clear; considering the interpretation above (growth of existing sites rather than founding of new ones), here these data suggest that individual sites waxed and waned ‑ perhaps due to shifts in river courses ‑ but there was net growth in the settled area. In both cases, the difference between the two types of settlement distributions is clear.

The curve of continuity percentages (Figure 4.4) shows that there is not a continuous rate of continuity, but instead several periods of sharp discontinuity in which the settlement system was largely re-organised. These are shown in the curve by the steep slopes facing right. There are sharp discontinuities with the advent of the Iron Age and the beginning of the Parthian period. The Iron Age itself seems to be a period of strong continuity and change in place, not due to shifts in site location or land use. The fact that the total site size area (Figure 4.2) climbs in the Iron Age, combined with the presence of a large number of continuous sites, suggests that existing sites were growing in size, rather than that new sites were being founded. Again, this is an indication of settlement continuity.

Self-Organised Criticality: the Nature of Intersite Connections in Margiana

Considering this data from a meta-analysis point of view, what it really represents is the amount of change in the ceramic assemblages that are used to date the sites. There must be sharp and obvious discontinuity in every trough while a rising line traces sherd assemblages that are commonly grouped together.

The study of self-organised critical systems (see Bak 1996 for a complete introduction) is an outgrowth of broader studies of complexity. Self-organised critical systems organise a systematic continuum between larger, infrequent events or components and small, frequent parts. The most notable manifestation of this is cascades of change in a system followed by periods of relative stability. There always are patterns of interaction between the components of the self-organised system. These interactions also have an effect upon the action, and success, of other interacting components. The archetypal example is the organisation of an avalanche in a sand pile, but the mathematical signature of self-organised criticality can be found in many places in nature and culture (Bentley and Maschner 2001: 36). Selforganised criticality is a phenomenon of natural systems that may ‑ or may not ‑ exist in archaeology. If we accept that it does, it provides a way of describing change in large interacting systems, and discovery road signs indicating points of interest at which further research ‑ that will recover culture, ideology and agency ‑ can be directed.

Water Control and Access Water management is a very important question when considering settlement in marginal and desert areas. The organisation of water management has implications for community and labour organisation, government and power, as well as economic production (Salvatori 1998b). For this study of Margiana the nature of change in water management through time will be examined by taking advantage of the distinctive geography of the desert oasis. Geographically Margiana is a triangular delta opening out into the desert. The rivers flow from south to north with (when the landscape is unmodified) areas of thick tamarisk, swamp and moving sand in between them. Rivers form braided channels with water and nourish the land immediately around them; to extend fertile land beyond the immediate area of the rivers, people need to build and maintain irrigation works. At any given time there is a band of land between the mountains and the desert that can be used for agriculture; over the long term this band migrates north or south with the rise and fall of rivers, but otherwise it remains roughly the same.

What are the general implications of self-organised criticality for the understanding of interacting systems? A system with a large number of connections makes it more difficult for abstract inequalities to build up between agents, but results in many small changes and fewer larger ones (Bentley and Maschner 2001: 46). Without almost any thought at all, the implications of this statement for hierarchy and heterarchy are obvious — a hierarchical system, in which levels of hierarchy are disconnected from one another, is more likely to undergo massive change in its

What is of interest to us here is a simple model of how sites are distributed across the deltaic landscape, and therefore 50

Non-graphic Information Systems and Diachronic Transformations in Margiana

Figure 4.5 Relative east-west distribution of sites. Showing the dispersed organization of the Late Bronze Age and Parthian periods and the central orientation of the other periods.

51

J. Wright more closely interconnected sections (levels of hierarchy) than a heterarchical system where an entire society will vent pressures in small changes before a large, societywide change takes place. A further implication comes from the dichotomy of small-world (local) connections and dispersed, wide-world, connections. In a self-organised model, change spreads more rapidly through long-range connections than through short-range ones. A smallworld system has many close connections, which serve to maintain an abstract conservatism among agents, while agents which interact through a few — long — channels are more likely to be able to change. An immediately obvious use of this model is the study of trade systems; it suggests that locally conservative economic patterns would be much more enduring and conservative than a long-range, entrepot-based system — at least initially; but once there were enough long-range connections, the long-range system would become more enduring than the local system. Finally, a system with many interconnected agents can be seen to have a greater chance of preserving those agents for a longer period. Fewer interconnections are more likely to be swept by avalanches of change. This is, of course, all math and theory. There are no applications of these ideas to history and archaeology. Like so many similar ideas, they could provide a useful heuristic device, but not a work of dogma about processes in the past.

2.

A self-organised critical system: in this case the distribution would follow a power law distribution with proportionally more examples of small life-spans. This would present a strongly skewed graph, which, when presented as a log-log distribution, would be a straight line.

4.

The final hypothesis suggests the sort of lifespan distribution you might find when the life‑span of a style is determined by its position in an intuitive or divisive typology (Whallon 1972; Sinopoli 1991). In short, a system created by archaeologists. Bentley and Maschner suggest that, due to the way the typology is constructed (by dividing things into smaller and smaller piles), the plot of the number of different life spans would be a log-normal curve centred on an average life‑span, the most common result of the division process.

Figures 4.6-8 show the results of displaying the lifespan of every site recovered by the survey for which the data was available (n=883). Figure 4.6 shows the result of a test of hypothesis 1 and 2. The data does not follow any sort of linear pattern and therefore suggests that the appearance and disappearance of sites is not related in a self-organised critical system. Figure 4.7 is testing hypothesis 3. Again

Mathematical worlds can be constructed in which selforganised criticality flows easily through systems and its interaction and results can be tracked. It is somewhat more difficult to do this with observed data, and observed archaeological data is still more troublesome. This search for self-organised criticality in archaeological data, in this case the settlement system of Margiana, will follow the example of Bentley and Maschner’s (2001) study of change in ceramic typologies. They examined the ”life-spans” of various ceramic styles from archaeological regional surveys and historically documented productions, in order to examine the nature of interactions between ceramic producers. For this study, they made four hypotheses about the nature of the interactions and what graphical representations of them one might expect to see (Bentley and Maschner 2001: 52-54): A random relation: the number of different life spans will be a level linear distribution—there is no relation between duration and number of cases.

A non-interactive system: in this case one would expect a similar time span for everything and no interactions determining the end or beginning of a life-span. This would appear as an exponential distribution favouring short lifespans (the most likely to end). Using a log distribution again, this would be a straight line on a single log graph.

This portion of this paper will test the same four hypotheses using the life‑spans of archaeological sites, as determined by their ceramic chronologies. The aim is to discover something about the nature of interaction and change in the past.

Method

1.

3.

Figure 4.6 Log-plot of the total number of sites of various year spans. In a self-organized system of long term settlement change a linear cluster would be expected. The high regression line shows that here are many sites of intermediate durations.

52

Non-graphic Information Systems and Diachronic Transformations in Margiana

Figure 4.9 The data of figure 7, a log-normal plot, displayed as a histogram with an example normal curve. This emphasizes the nature of the distribution.

Figure 4.7 Plot of the log of the number of sites of a given span. If site foundation and abandonment was an unorganized process one would expect an exponential distribution which would appear on this plot as a linear cluster of points.

Figure 4.10 The log-normal plot of Figures 4.7 and 4.8 presented over a wider range in order to emphasize the number long duration sites on the right tail of the curve.

observations that can be made from it. A cluster of a few sites that lasted longer than the rest, visible on the righthand tail of the histogram (Figure 4.9) — expanded in Figure 4.10 to make it more obvious — suggests that some places were continuously settled for several millennia.

Figure 4.8 Plot of site span counts. This creates a nearly lognormal curve which suggests that the spans of sites could be due to the nature of the chronology used to date the sites.

The immediate impression is that these sites are the largest depe sites in the oasis. This is not always the case. The site sizes and types include the whole gamut from small scatters to named depe.

the points are scattered and do not follow any linear pattern suggesting that site life spans are not non-interactive. Finally, Figure 4.8 shows a count of site life‑spans. Figure 4.9 displays this same distribution along with a normal curve, to illustrate that the distribution of site life spans could well be the result of the typology that is used by the archaeologists to anchor them in time.

The chaotic nature of the data plots (Figures 4.6 and 4.7) and the counter-intuitive position of the regression lines demonstrate that there are a large number of sites with similar ages. This is brought out further in the test of hypothesis 4. In conclusion, these brief tests show that with the data available it is not possible to suggest much about the nature of sites or long-term inter-site interaction

Despite the complete failure of this data to organise in a self-organised critical way, there are a few interesting 53

J. Wright in Margiana, but that this is due to the nature of the chronology in use more than anything else.

to the passage of time. A similar possibility holds for the analysis of site continuity as well. As mentioned above, the apparent continuity between sites could be due to the ease of recognising the boundaries between some pottery types, and not between others. In fact, all the results above could be explained in terms of flawed data and that would probably be a possible explanation for many of the patterns. But this is no way to proceed: as archaeologists we must balance the quality of our data with our ability to move forward. One way to move forward, when further extensive field‑work and analysis is not possible, is to refine the data.

Discussion These five brief analyses provide a historical narrative of settlement and land use change through time. The Bronze Age is a period of relative stability, with a growing number of sites spreading out over the landscape as time passes. This landscape could well have been one of stable irrigation. The shift to the Iron Age was a drastic transition. The population settlement pattern changes, the settlements are drawn together and the population becomes less scattered; the material culture of the sites shows a marked discontinuity with the preceding Bronze Age. Throughout the Iron Age there is a period of continuity and stability; site numbers grow but their form and locations do not change. Individual sites appear to wax and wane, perhaps following the fortunes of the river channels they were built close to.

This analysis has been hampered by the fact that it could not get far enough away from the pottery-based chronology — which infused all analyses: size, periodicity etc. To make really pottery-free suggestions about chronology and socio-economic change it is necessary to build a strong radiocarbon-based chronology for the region, ideally by dating sites but more likely by dating the ceramics in many sequences. Of course, more spatial data would help to identify spatial patterns — this means continuing survey work. But even the simple measure of turning the analytical suggestions made here into propositions to be tested with a GIS would add more to the work. Any expansion of work must include more environmental data; suggestions about climate and environmental change are central to our understanding of cultural and socioeconomic change. Finally, the definition of social change and its archaeological evidence could be refined along with the data. Is population change, political re-organisation or ideology the absolute marker of change? Or should we be weaving together complex bundles of threads? In fact, all the results of the analysis above could be explained in terms of insufficient data, but this is no way to proceed, as archaeologists we must balance the quality of our data with our ability to move archaeological interpretation forward. One way to move forward, when refining the data and definitions is not possible or not enough, is to offer interpretations that can be fitted, tested and modified as data improves.

The end of the Iron Age is a dynamic time. The Achaemenid and Hellenistic conquests of the region and the transition to Parthian rule took place; the settled population decreased. Another sharp material culture discontinuity is seen during this period. There is a notable shift in the regions of the oasis that are settled, and the settlements once again spread out across the oasis rather than gather around channels, though this could be the result of the general southward shift of settlements. Comparing this analysis with those related to the survey data and regional history itself, the fit of the data is relatively close. The Bronze Age is the only period in which the fit does not seem close. Salvatori (1998a, 1998b) describes a shift in settlement integration and fragmented settlement system. This does not fit well with the picture of steady growth and widely distributed sites produced by this analysis. The description of the landscape as one of “orchards and fields” (Gubaev et al. 1998: XVII) seems to fit, however. Other than that, the Iron Age is a homogenous period with a sharp boundary between it and the preceding Bronze Age. At its end the population drops and the population distribution is re-ordered as the central pay of the oasis moves southward towards the area of Merv (Genito 1998b). The Parthian period (Bader et al. 1998) is one of expanding rural settlement centred in the south of the oasis.

After having said all this, from the information at hand we can assert that there are three historical periods in Margiana, each with its own distinctive character. The Bronze Age was a period of stability and integration. I would argue that the major walled centres are not indicative of a period of endemic conflict (Salvatori 1998a: 51), but of either momentary conflicts or longer-term ideological strategies. The Iron Age was a less organised period than the preceding ones. Sites rose and fell, there were fewer larger centres, and so on. Even so, it was a period of stability and continuity. Both the Bronze and Iron Age are characterised by pulsating polities (Demarest 1992) that grow and expand in influence and then recede back into obscurity. The difference is that the Bronze Age polities maintained themselves through periods of obscurity better. This suggests that in there was a fundamental difference

All the interpretations offered here must be tempered with the very real possibility that the data is problematic in many dimensions. The results of the fifth analysis above suggest that the site typology and chronology may be due mainly to the nature of the ceramic chronology that is used to construct the typology, rather than directly due 54

Non-graphic Information Systems and Diachronic Transformations in Margiana Boyle, J. A. and W. B. Fisher (eds.) 1991. The Cambridge History of Iran. Cambridge. Chilton, E. S. 1999a. “One Size Fits All, Typology and Alternatives for Ceramic Research”, in Chilton (ed.) 1999b: 44-60. Chilton, E. S. (ed.) 1999b. Matrial Meanings, Critical Approaches to the Interpretations of Material Culture. Salt Lake City. Demarest, A. 1992. “Ideology in Ancient Maya Cultural Evolution, the Dynamics of Galactic Polities”, in Demarest and Conrad (eds.) 1992: 135-158. Demarest, A. and G. W. Conrad (eds.) 1992. Ideology and Pre-Columbian Civilisations. Dole, G. E. and L. Camero (eds.) 1960. Essays in the Science of Culture. New York. Durrenburger, P. and G. Palsson (eds.) 1989. Anthropology in Iceland. Iowa City. Fish, S. K. and S. Kowalewski (eds.) 1990. The Archaeology of Regions, A Case for Full-coverage Survey. Ford, J. A. 1954a. “Comment on A.C. Spaulding “Statistical Techniques for the Discovery Of Artifact Types”, American Antiquity XIX (4): 390-391. Ford, J. A. 1954b. “The Type Concept revisited”, American Anthropologist 56: 42-54. Genito, B. 1998a. “The Achaemenids in the History of Central Asia”, in Gubaev, Koshelenko and Tosi (eds.). 1998: 149-158. Genito, B. 1998b. “The Iron Age in the Merv Oasis”, in Gubaev, Koshelenko and Tosi (eds.). 1998: 89-95. Gubaev, A., G. Koshelenko and M. Tosi (eds.) 1998. The Archaeological Map of the Murgab Delta Preliminary Reports 1990-1995. ���������������������������������� Reports and Memoirs. Series Minor III, Rome. Herrmann, G., K. Kurbansakhatov 1994. “The International Merv Project. Preliminary Report on the Second Season (1993)”, Iran XXXII: 53-75. Hiebert, F. T 1992. “The Oasis and City of Merv (Turkmenistan)”, Archéologie Islamique 3: 111-127. Hiebert, F. T. 1994. Origins of the Bronze Age Oasis Civilization in Central Asia. American School of Prehistoric Research Bulletin 42. Peabody Museum of Archaeology and Ethnography, Cambridge, MA. Hiebert, F. T. and C. C. Lamberg-Karlovsky 1992. “Central Asia and the Indo-Iranian Borderlands”, Iran XXX: 1015. Hodder, I. (ed.) 1987. The Archaeology of Contextual Meanings. Cambridge. Holdaway, S., P. Fanning and D. Witter 2000. Report of the Western New South Wales Archaeological Project. Auckland. Honeychurch, W. 2003. Inner Asian Warriors and Khans: a Regional Spatial Analysis of Nomadic Political Organization and Interaction. PhD Dissertation, Dept. of Anthropology, University of Michigan, Ann Arbor. Kohl, P. L. 1984. Central Asia. Palaeolithic Beginnings to the Iron Age. Paris. Kohl, P. L. (ed.) 1981. The Bronze Age Civilizations of Central Asia, Recent Soviet Discoveries. New York.

in social organisation between the two periods. The long continuity of socio-economic organisation from the Parthian to the Medieval period mirrors the thousand-year stable trajectory of the Iron Age and Bronze Age before that. Between each of these three periods, there are sharp unconformities, in which settlement and land use patterns change. At the periphery of this archaeological vision there are the nomadic pastoralists of Asia, from the 2nd millennium BC and on; nomadic pastoralists were a constant presence on the cultural and economic stage of Asia. Unfortunately they are hard to detect; the Margiana survey offers glimpses — possible Andronovo pottery contemporary with the Late Bronze Age, historical mentions of Cimmerians or Parthians. It is obvious that the nomads were always there, playing an integral part in long-term and short-term cultural changes in Margiana. As a final thought, this paper shows that though there are patterns in data, that data — and those patterns — are only as strong as the typology on which they are based. Typology in general is a complex and long-debated issue (Kreiger 1944; Spaulding 1953, 1960; Ford 1954a, 1954b; Steward 1954; Rouse 1968; Read 1982, 1989; Sørensen 1987; Adams and Adams 1991; Bestley 1993; Chilton 1999), but it is not an outdated issue. The typology, and chronology, of sites that we see here could be just as etic as any system designed to study ceramics or stone tools. Any study of landscape archaeology includes elements of typology in the categorisation of sites (Barrett 1994; Lansing 1991; Tilley 1994). The next step in the analysis of the landscape of Margiana is to go beyond the distribution of sites through time and begin to think about sites and settlement systems in their own time and their own space. References Cited Adams, W. Y. and E. W. Adams 1991. Archaeological Typology and Practical Reality. Cambridge. Bader, A., V. Gaibov, A. Gubaev and G. Koshelenko 1998. “The Parthian Period”, in Gubaev, Koshelenko and Tosi (eds.): 189-190. Bak, P. 1996. How Nature Works: the Science of Selforganized Criticality. New York. Barrett, J. C. 1994. Fragments from Antiquity, an Archaeology of Social Life in Britain, 2900-1200 BC. Oxford. Bentley, R. A. and H. D. G. Maschner 2001. “Stylistic Change as a Self-organized Critical Phenomenon: an Archaeology Study in Complexity”, Journal of Archaeological Method and Theory 8 (1): 35-66. Bestley, N. J. 1993. “Type and Typology and Writing the Past: the Case of Cotswold-Severn Monuments”, Archaeological Review from Cambridge 12 (1): 91102.

55

J. Wright Kohl P. L. 1984, Central Asia, Paleolithic beginnings to the Iron Age. (Editions Recherche sur les Civilisations). Paris. Kreiger, A. D. 1944. “The Typological Concept”, American Antiquity IX (3): 271-288. Lansing, J. S. 1991. Priests and Programmers. Princeton. Masson, V. M. and V. I. Sarianidi 1972. Central Asia, Turkmenia before the Achaemenids. New York. Parson, J. R. 1971. Prehistoric Settlement Patterns in the Texcoco Region, Mexico. Memoir No. 3, Museum of Anthropology, University of Michigan. Ann Arbor. Parson, J. R. 2004. “Critical Reflections on Forty Years of ‘Systematic Regional Survey’”. Montreal. Phillips, T. and A. Watson 2000. “The Living and the Dead in Northern Scotland 3500-2500 BC”, Antiquity 74: 786-792. Puleston, D. 1983. The Settlement Survey of Tikal. Tikal Reports No. 13, University Museum Monograph No. 48, University of Pennsylvania. Philadelphia. Read, D. W. 1982. “Toward a Theory of Archaeological Classification”, in Whallon and Ford (eds.) 1982: 5692. Read, D. W. 1989. “Intuitive Typology and Automatic Classification: Divergence or Full Circle?”, Journal of Anthropological Archaeology 8: 158-188. Rouse, I. 1968. “Prehistory, Typology, and the Study of Society”, in K. C. Chang (ed.) Settlement Archaeology, Palo Alto: 10-30. Salvatori, S. 1998a. “The Bronze Age in Margiana”, in Gubaev, Koshelenko and Tosi (eds.) 1998: 47-55. Salvatori, S. 1998b. “Margiana Archaeological Map: The bronze age settlement pattern”, in Gubaev, Koshelenko and Tosi (eds.) 1998: 57-65.

Shelach, G. 1999. Leadership Strategies, Economic Activity, and Interregional Interaction. Social Complexity in Northeast China. Fundamental Issues in Archaeology. Kluwer Academic/Plenum. Sinopoli, C. M. 1991. Approaches to Archaeological Ceramics. New York. Smith, K. and J. Parsons 1989. “Regional Archaeology in Iceland: Prospects and Problems”, in Durrenburger and Palsson (eds.) 1989: 179-202. Sørensen, M. L. S. 1987. “Material Order and Cultural Classification: the Role of Bronze Objects in the Transition From Bronze Age to Iron Age in Scandinavia”, in Hodder (ed.) 1987: 90-101. Spaulding, A. C. 1953. “Statistical Techniques for the Discovery of Artifact Types”, American Antiquity XVIII (4): 305-313. Spaulding, A. C. 1960. “The Dimensions of Archaeology”, in G. E. Dole and L. Caniero (eds.) 1960: 437-456. Steward, J. H. 1954. “Types of Types”, American Anthropologist 56: 54-57. Tilley, C. 1994. A Phenomenology of Landscape, Paths Places and Monuments. Berg. Whallon, R. 1972. “A New Approach to Pottery Typology”, American Antiquity 37: 13-33. Whallon, R. and J. A. Ford (eds.) 1982. Essays on Archaeological Typology. Evanston Wilkinson, T. 2003. Archaeological Landscapes of the Near East, University of Arizona Press. Tucson. Willey, G. 1953. Prehistoric Settlement Patterns in the Virú Valley, Peru. Bulletin 155, Bureau of American Ethnology, Smithsonian Institution, Washington, D.C.

56

Chapter 5

The Margiana Settlement Pattern from the Middle Bronze Age to the Parthian-Sasanian: a Contribution to the Study of Complexity Sandro Salvatori Introduction

many fossil remains of the endoreic outflow system that, per se, distinguishes Central Asia (Miroshnikov 1987, 1992). In regions that are chiefly characterised by dry conditions from the Neogene on (Gerasimov 1978: 323), the variations in human peopling and farming activities are often thought to be strongly correlated with variations in the flow rate of the rivers. It is common opinion that the human populations of those regions, in order to survive, were compelled to adjust themselves to the changes in the environment resulting, over a period of several thousand years, from the constant reduction in the flow rate of these rivers. This reduction, moreover, was combined with the

The present-day geo‑morphological configuration of the desert plains of Central Asia (Figure 5.1) has been determined not only by the ceaseless shaping action of erosion and aeolian transportation, but also by the action of rivers, which, with flow rates and channels that were often different from the present ones, have left visible traces of their ancient courses. The Murgab Delta, i.e. the Margiana of the classical period, in south-eastern Turkmenistan, is only one of

Figure 5.1 Satellite photo of Central Asia (from NASA World Wind).

57

S. Salvatori effects of other variables, such as continental tectonic movements, the barrier created by alluvial deposits in the distal portions of the internal deltas, and the influence of the phenomenon known as “Ber’s law”, which can lead to significant deviations in the courses of rivers (Gerasimov 1978: 332).

to be complied with by each of the aggregation groups that had been found. However, there was an exception: the settlement of Gonur 1, whose area was more than 40 hectares, according to an estimation made by V. Sarianidi and reported by Ph. Kohl (1984: 146). Because of its unusual size, for a long time this site was regarded as a regional capital (Sarianidi 1993).

If we consider its action over several millennia, this more or less gradual reduction of the flow rate of the endoreic rivers of Central Asia undoubtedly turns out to have played an important role in the environmental transformations. The dynamics of the peopling, and the cultural developments bound to them through adaptation processes that could change the structure of human social organisations even in a short space of time, are often regarded as dependent variables of regimes of instability and environmental crisis. Regardless of their possible degree of success, these mechanisms, in any case, seem to proceed towards an intensification of production and a hierarchical organisation of the forms of power.

Nine aggregations of sites were then defined: they were interpreted as “micro-oases” and were usually given the name of the nearest phreatic well (Kelleli, Egri Bogaz, Taip, Adzhi Kui, Gonur, Auchin, Adam Basan, Togolok, Takhirbai). On the basis of a rough sequence developed intuitively from the results of the first trial excavations performed on a certain number of sites in the delta area, a preliminary outline of the history of the settlement of the region was proposed (Sarianidi 1981: 188; Masimov 1981a: 218; Hiebert 1994a: 72,174-175); it is still accepted by many scholars. According to this reconstruction, there was an initial settlement period that was limited to the north-western part of the delta (Kelleli area) and dates back to the final phase of the period called Namazga V (end of the Middle Bronze Age), in the known sequence relative to the piedmont of southern Turkmenistan. This initial period was followed by demographic expansion, which corresponded with a considerable increase in the number of settlements and also with a southward shift in the barycentre of the settlements of the region (Gonur phase, from the name of the most extensive archaeological site identified in the central area of the delta). This phase was ascribed to the early stages of the Late Bronze Age; it was presumably followed by a further southward shift in the barycentre of the settlements during the subsequent Togolok phase. In this reconstruction, the greater density of the Iron Age settlements in the southernmost part of the delta presumably represents a further stage in the drying up of the delta, which had been going on for several centuries. As the years went by, some aspects of this reconstruction (which gave rise to a linear model of the drying up of the Murgab Delta and at the same time was based on it, through a clearly tautological process) were adapted to the results of the most recent research activities in this area.

A reconstruction of the historical geo‑morphology of these environments and of the conditions that caused the variation in the flow rate and/or course of the rivers is essential for locating the traces of human settlements and for understanding the development and socio-political transformations, as well as the territorial dynamics. From this point of view, the Murgab Delta, thanks to the excellent preservation of the archaeological data, still constitutes a privileged laboratory. The surveys and excavations performed in this area during the first stages of the archaeological investigation of the Murgab Delta (1949-1989) gave rise to several, often conflicting, theories, which attempted to explain both the role played by this region in the broader sphere of the cultural development of Central Asia, and the internal dynamics of its peopling during the Bronze and Iron Ages. In those years almost all the Murgab Delta was explored. Though the surveys were not adequately supported yet by the help of the necessary aerial‑photography coverage, they led to the discovery of more than a hundred proto-historical settlements. The Russian and Turkmen archaeologists immediately realised that the maps they were gradually developing revealed significant aggregations of settlements. The first attempt at arranging the data in an organic system was made by I. S. Masimov (1981b), who identified three categories of sites, on the basis of their area: 1) less than 5 hectares; 2) between 5 and 10 hectares; 3) more than 10 hectares. This three-level hierarchy, at that time, seemed

In fact, as we have demonstrated elsewhere (Salvatori 1998a; Cattani and Salvatori, Chapter 1), the microoases into which the delta was divided are not discrete This interpretation is actually the result of the chronological levelling of the settlement system, brought about by the first explorers in this area. The settlement of Gonur 1, for instance, consists of two mound formations, a larger one to the north and a smaller one to the south. The northern mound can be entirely ascribed to the Namazga V period (Middle Bronze Age), while the southern one, except for its basal level, can be ascribed to the subsequent Namazga VI period (Late Bronze Age). At present a picture is emerging that is much more complex and diversified than was previously supposed, and the spatial organisation pattern proposed by Masimov in 1981 is now decidedly outdated.  Takhirbai 3: Masson 1959; Sarianidi 1990. Auchin-depe: Masson 1959; Sarianidi 1957. Kelleli, Taip, Adam Basan, Adzhi Kui: Masimov 1984, 1986; Masimov and Lyapin 1977. Togolok-1, Togolok-21, Gonur-1 North and South: Sarianidi 1990, 1993, 1998. 

This situation is undergoing a process of rapid worsening and transformation due to the massive operations for putting the territory of the ancient delta to crops. A thick network of irrigation canals has been dug during the last few years, bringing into the delta the water of the Amudar’ya through the great Karakum Canal, which acts as the main collector. Hundreds of archaeological sites are now threatened with destruction or have already been destroyed by farming operations. 

58

The Margiana Settlement Pattern from the Middle Bronze Age to the Parthian-Sasanian chronological reference units, because each of them contains settlements that belong to different phases of the Margiana sequence.

of its history, this system had been connected to a river collector, an ancient branch of the Amudar’ya (Marcolongo and Mozzi 1998: 2), into which there flowed the water of the active branches of the Murgab River.

The problem obviously does not consist only in the interpretation of the settlement system, but also in the need to achieve reconstructive hypotheses that agree with the entire geo‑political system of Central Asia in the Bronze Age, where strong material features that reveal areas of cultural homogeneity can be clearly recognised. In other words, the fact that the earliest settlements in the Murgab area were ascribed to the final part of the Namazga V period led the scholars to regard them as the consequence of a migratory flow from the piedmont of southern Turkmenistan caused by a crisis of the urban system in that area (Masson 1959; Biscione 1977; Biscione and Tosi 1979; Masimov 1979), and perhaps also by the negative effects of a peak in dryness (Masson 1986a: 81). The farming population of those regions presumably shifted towards the Murgab Delta and occupied an area that was more favourable to the preservation and development of a production structure that was based on irrigation and had been strongly organised as such in the original regions. The favourable conditions met in the delta presumably promoted a considerable increase in population during the subsequent Late Bronze Age. It is supposed that this increase in population resulted not only in an intensification in the settlement of the delta, but also in a further migratory wave towards the regions crossed by the Amudar’ya, that is towards Bactria and Sogdiana. An alternative reconstruction, of which at least two variants are known, has been proposed by other researchers. It postulates the settlement of Margiana during the Namazga VI period (Late Bronze Age) by a population coming from Bactria, presumably followed by a penetration, albeit quite limited, into the piedmont of southern Turkmenistan, where the crisis of the urban system had left some room for a small-scale colonisation from outside (Alekshin 1983; Francfort 1984: 174; 1989: 375; Jarrige 1986).

This ancient river collector north of the sub-fossil delta of the Murgab is now almost completely full of sand coming from the Karakum Desert. Sand, in the form of long northsouth dunes, at present also covers a large part of the delta. Together with the alluvial deposits carried by the river in ancient times, this sand has probably hidden the earliest traces of the human peopling of the region (Cremaschi 1998). The systematic survey carried out for the project of the “Archaeological Map of the Murgab Delta” considerably increased our knowledge of the peopling of the area over the entire period from the Middle Bronze Age to the Islamic period. As a result, we were able to correct and redefine the picture that had been obtained on the basis of the early pioneering explorations of the nineteen-fifties and seventies. Middle Bronze Age What now emerges with an increasing clearness is a generalised distribution, over the area of the delta, of settlements that turn out to be contemporary to those ascribed to the Namazga V period in the piedmont of southern Turkmenistan, though some peculiarities in the material culture suggest that this is a specific regional variant (Salvatori 1998b; Salvatori, Chapter 6). In terms of an absolute chronology based on a series of C14 determinations, this period covers the span between 2400 and 2100 BC, and was characterised, in Margiana, by the development of phenomena of urban concentration very similar to the well-known ones of the piedmont of southern Turkmenistan, from the region of Abiverd (Namazgadepe) to that of the Meana-Chaacha system (Altyn-depe) (Masson 1956, 1981; Kohl 1984).

The results of the most recent researches, in particular of the “Archaeological Map of the Murgab Delta” project (Gubaev et al. 1998), and the critical reassessment of the data that had long been available, now make it possible to propose a model of the peopling of the Murgab Delta that is quite different from the one we have just described, and above all is incompatible with the migratory hypotheses stemming from it.

The large settlement of Gonur 1 North, which covers an area of about 35 hectares, and whose sequence is entirely within the Middle Bronze Age, is a clear example of a protourban centre that is quite comparable to the urban centres of the piedmont. The area where there is the greatest concentration of sites datable to the Middle Bronze Age is, to this day,

The extensive survey activities of the first research stages were replaced, from 1989 on, by intensive methods combined with broad-range geo‑morphological investigations based on remote-sensing techniques and integrated by ground-level investigations facilitated by the possibility of using a complete aerial- and satellitephotograph coverage (Cremaschi 1998; Marcolongo and Mozzi 1998). The ancient delta was thus found to be a part of a system that was much more extensive than had been previously supposed: in a comparatively ancient phase

The international project, in its intensive stage, was carried out, in the years between 1990 and 1994, under the joint leadership of A. G. Gubaev (State University of Turkmenistan), G. A. Koshelenko (Russian Academy of Sciences of Moscow) and M. Tosi (Istituto Italiano per il Medio ed Estremo Oriente, now Istituto Italiano per l’Africa e l’Oriente). From 1995 to 2002 only verification and study campaigns were organised in a discontinuous way.  Except for the Late Bronze Age graveyard excavated on the top of the depe, after it had been abandoned. 

59

S. Salvatori exposed takyr yielded evidence of pottery concentrations on mounds or pottery scatters on flatlands, attesting that during the Bronze Age the delta area was practically free of sand formations and entirely available for agricultural exploitation. This evidence is the best proof of the fact that during the Middle Bronze Age the Murgab Delta was open to an agricultural exploitation that was limited only by the available irrigation technology.

the central-northern part of the delta. However, several settlements of the same period have been identified in the southern part. In our opinion, the lower density observed in the south may be the result of the fact that the most ancient sites are hidden by the progressive accumulation of alluvial sediments and by the superimposition of more recent anthropic levels. This double possibility seems quite plausible, and is corroborated, among other things, by the findings of a deep test trench carried out in the nineteen-eighties at Togolok 1 by V. I. Sarianidi (1990) and by the evidence supplied by a site identified by M. Cremaschi in the section of a modern canal (Cremaschi 1998). The deep levels reached by the trial excavation at Togolok-1 yielded material datable to the Middle Bronze Age, of which no trace had been detected among the surface material. Extending this evidence, it does not seem unreasonable to assert that the southern part of the delta, too, was inhabited during the Middle Bronze Age, no less than the central and northern parts. The settlement density in the central-northern part of the delta during the Middle Bronze Age and the proven presence of sites from that period also south of the area of Togolok lead us to reconsider the mode of the peopling of the Murgab Delta from an angle quite different to that of the surmises accepted up to now. At the same time, the geomorphologic study and the new archaeological data indicate the direction of a new and different interpretation of the drying up of the delta, which had previously been explained in terms of a simplistic model of gradual regression. In order to throw further light on the new dimension of the problem, in the years 1995-2002, we organised and completed a specific sub-project, within the project for the “Archaeological Map of the Murgab Delta”. Our goal was to demonstrate, on an experimental basis – by means of a series of survey transects chiefly oriented in the east-west direction, i.e. transversally to the direction of the flow of the water of the ancient Murgab, and interconnecting some of the largest archaeological sites of the area – that there was no interruption in the settlement system of Margiana during the Bronze Age. The following transects were carried out (Figure 1.1): Taip - Adzhi Kui, Adzhi Kui - Gonur, Gonur - Auchin (not completed because extensive areas in the planned route turned out to be flooded in relation to the project of putting the area to crops), Kelleli - Egri Bogaz 4, Egri Bogaz 4 - Gonur, Egri Bogaz 2 - Adzhi Kui, Adzhi Kui-1 – Togolok-1, Egri Bogaz 2 - Taip 1, Site No. 725 - Taip 1, Site No. 212 - Site No. 357, Site No. 638 - Site No. 172, Site No. 1412 - Site No. 1411, Site No. 156 - Site No. 232, Site No. 232 - Site No. 144, from the road to Auchin - to Site No. 340, and Site No. 234 - Site No. 280. This strategy made it possible to considerably enrich the inventory of the Bronze Age sites known in the region and to identify two large settlements (Sites No. 723 and 746) that fit significantly in the Thiessen polygonal structure we can draw on the basis of the distribution of the larger sites. The most important result, however, was by far the documentation of a practically uninterrupted presence of Bronze Age ceramics along the transect routes: most of the

The new data thus collected demonstrate that the settlement system of Margiana during the Bronze Age had not been generated or conditioned by an environmental situation dominated by presumptive oases, but by spatial rules connected to human group dynamics and by the structural complexity of the political and administrative forms of the territory. The application of Thiessen polygons to the settlement system of Margiana, despite the fact that we are still far from having detected all the Bronze Age settlement traces in the region, supplies an picture of a settlement system organised not on the basis of the fragmentary, discontinuous availability of usable soil (oases), but on the basis of an organic pattern of total occupation of the delta flatlands. The data acquired during recent years indicate, on the one hand, a generalised distribution of Middle Bronze Age sites over the entire delta area, and, on the other hand, an equally generalised presence of sites datable to the Late Bronze Age. Unfortunately there remain two mostly inaccessible areas: the northernmost part of the delta, where surveys are hindered or prevented by the presence of a continuous coverage of sand dunes, and the southern delta area, where modern farming activities are more intense and have spared only the largest and latest settlements. Though we realise that the present state of the research does not offer an adequately accurate corpus of chronological data, and that therefore it is not possible to achieve the necessary division of the Late Bronze Age sites into subphases, we can make an initial, approximate attempt at identifying the rules that governed the settlement system and the degree of integration of the settlement hierarchies (Johnson 1977, 1980). As regards the Middle Bronze Age, we utilised only the sites detected by the surveys of the Murgab Archaeological Map Project and those that, though identified in previous stages of the research, can be ascribed with certainty to this period on the basis of the material that has been published. The Gonur North site is estimated in the literature to have an approximate extension of 24 hectares. On the basis of On the northern side, north of the line that theoretically joins Kelleli to Egri Bogaz, where among the sands we found some takyr that were still bare, we identified further settlements and farms of the Middle Bronze Age. In the areas north and east of these settlements the alluvium was completely covered with sand. 

60

The Margiana Settlement Pattern from the Middle Bronze Age to the Parthian-Sasanian

Figure 5.2 MBA site size distribution. Thiessen polygons.

the topographical survey carried out by M. Mascellani for our project and of the excavation data that demonstrate its extension also in the basal levels of Gonur South (Hiebert 1994: 113), we can estimate that its extension is about 35 ha.

on this problem, did not yield any evidence that could be dated to the Middle Bronze Age. Given these limitations, which have been partly lessened by the execution of the survey transects mentioned above, the evidence concerning the Middle Bronze Age settlement system is undoubtedly incomplete; but the elements we now possess demonstrate quite convincingly that there developed a hierarchically organised settlement system having the large site of Gonur North at its centre. Around this site, the smaller ones were arranged in a radial pattern, forming a consistent lattice and with modular distances. The settlement system, therefore, seems to have organised itself on at least three levels of site size and two levels of administrative control (Figure 5.2). The trend towards the realisation of a strong administrative integration is shown by the rank-size index, which attains the value 0.016 with Gonur North evaluated at approximately 35 ha (Figure 5.3). It is evident that the site was central in terms of concentration of administrative functions, and this is confirmed by the presence of the large palace that towers

Some areas of the delta seem to have little or no evidence relative to this period, because of various types of limitations: 1) Northern sector: apart from the Kelleli area, the entire northern sector to the east of the latter area is intensely covered with dunes, both hardened and in course of formation. The few exposed takyr areas, however, show evidence of the presence of protohistoric settlements. 2) Central-eastern sector: many sites detected in this sector (Auchin area) by the Soviet surveys of the nineteen-fifties and seventies were recently destroyed when the area was put to crops. On the other hand, there are no systematic publications about the material collected on the surface of those sites, and this made it impossible to discriminate between Middle Bronze Age sites and Late Bronze Age ones. Auchin 1, the largest settlement in this area, which we checked in the autumn of 2000, focusing particularly

A device suggested by Johnson (1980) was adopted: the index was calculated on the basis of a fault point, usually identified at the extension of 1 ha, and called “RSI cut point”. 

61

S. Salvatori

Figure 5.3 Rank-size curve of MBA sites.

Figure 5.4 Rank-size curve of LBA sites.

62

The Margiana Settlement Pattern from the Middle Bronze Age to the Parthian-Sasanian

Figure 5.5 Thiessen Polygons LBA.

over the Middle Bronze Age settlement (Sarianidi 1998); but it is equally evident that the administrative activity was distributed over the territory through a consistent series of second-rank sites that controlled essentially modular territories.

hierarchy. In other terms, the centrality of a site was no longer recognised, and, as regards rank-size, the decided convexity of the curve (RSI = + 0.351) denotes a strong independence of the groups of sites (Figure 5.4). During this period, all the largest settlements (Figure 5.5), at the centres of the polygons, appeared to be fortified villages, internally organised also with storage structures that almost reproduced the model of the Middle Bronze Age palace of Gonur North. If we observe the polygonal lattice, keeping in mind that during this period (Late Bronze Age) the total area of Gonur 1 South did not exceed 5 hectares, the system appears as a series of juxtaposed polygons without a predominant centre. In this case there may have been a situation of territorial fragmentation that seemed to proceed following self-regulation rules that were of a proxemic type and were not integrated in a central political and administrative system. The fragmentation of the system in terms of a parcelled territorial control led to effects that can be ascertained from the angle of the produced or imported goods (Salvatori, Chapter 6).

The classical projection of the Thiessen Polygons onto the Middle Bronze Age sites of Margiana produces a rather regular grid that highlights the centrality of the Gonur North site. Such a configuration excludes, at least for this period, the presence of oasis-type concentrations, and points rather to a widespread, ubiquitous distribution: in other words, the forces that determined it were apparently generated not by specific environmental situations, but by rules relative to group organisation dynamics. Late Bronze Age Precisely in the course of this process of integration of the administrative control system with the settlement system, the region apparently underwent a deep crisis, probably of institutional origin; this crisis, during the subsequent Late Bronze Age, led to a fragmentation of the system, with the dethroning of the highest ranks of the administrative

In our present state of knowledge, this long period cannot be dealt with in a more analytic manner, because of the unsatisfactory details in the publication both of the stratigraphic sequences and of the pottery associated to 63

S. Salvatori

Figure 5.6 ICW site distribution.

the various sub-phases into which it could be divided. It is certain that the period can be roughly subdivided into at least three horizons: at present they are called Taip phase, Gonur South phase and Togolok phase, and are clearly separate from the subsequent Takhirbai 3 Phase, which, because of its characteristics, should perhaps be defined as Final Bronze Age. These subdivisions can be identified with some certainty as regards the excavated settlements; however, since the stratigraphic control at those sites was very poor, it has not been possible yet to develop a pottery sequence reliable enough to allow a precise division into phases (except for that of Takhirbai 3) of the materials coming from the surface collections. This situation therefore makes it impossible to study the territorial developments in a detailed manner, phase by phase, and compels us to supply an undoubtedly flattened image of this period.

notice the arrival of groups bearing a material culture unknown to the indigenous agricultural communities. This intrusion is indicated by the appearance of hand-made pottery decorated with incisions, typical of the Bronze Age cultures of the steppes of Central Asia. In a great number of places throughout the Murgab Delta (Figure 5.6) there has been reported the presence of this type of pottery, unanimously ascribed by the specialists to the Tazabag’yab phase of the Andronovo horizon: tempered ware with abundant lithic inclusions and an incised decoration with comparatively varied geometric motifs. These intrusive groups, presumably nomadic pastoralists and breeders, appeared in the Murgab Delta during the final phases of the Late Bronze Age and settled on the sands that by then had encroached on extensive areas of the region. Only during the Takhirbai 3 stage, right at the end of the Bronze Age, they seemed to integrate with the indigenous population, as confirmed now by the excavations conducted by M. Cattani in Sites No. 1211 and 1219 (Cattani, Chapter 9). As far as we know at present, only in the settlements of this last phase the steppe pottery was associated with the local one, attesting to the integration of the new‑comers in the socioeconomic system of the region. At the same time, we must point out that during the Final Bronze Age

At the end of the Bronze Age, the advance of sand had become an extensive problem in the central-northern part of the delta, and, as we shall see, eventually determined not only a decisive shift in the barycentre of the settlement areas, but also a considerable demographic imbalance. Together with the advance of sand, however, it is also possible to 64

The Margiana Settlement Pattern from the Middle Bronze Age to the Parthian-Sasanian

Figure 5.7 FBA site (full circle) distribution and the associated ICW sites (triangles).

Figure 5.8 FBA-ICW rank-size.

65

S. Salvatori

Figure 5.9 Yaz I site distribution.

Figure 5.10 Rank-size of Yaz I sites.

66

The Margiana Settlement Pattern from the Middle Bronze Age to the Parthian-Sasanian

Figure 5.11 Yaz II site distribution.

the settlement system underwent a significant reduction and southward shift (Figure 5.7), partly, in our opinion, because of the dramatic advance of sand, and partly, perhaps, because of a political and organisational crisis in the native farmers’ groups. The rank-size analysis carried out on the sites of this period produces a markedly concave curve (RSI = -0.939) (Figure 5.8), probably distorted by the overall area of Takhirbai 3. For the sake of clarity, we must point out that the Takhirbai 3 site has yielded materials from the Late Bronze Age, the Final Bronze Age (including the materials ascribable to steppe populations) and the Iron Age. Unfortunately, so far no breakdown of the respective distributions has been attempted, so we have no idea of the actual extension of the settlement during each of the chronological and cultural phases whose presence has been recognised on the basis of the pottery collected on the surface.

The southward shift of the settlement system was still evident. The distribution of the sites seemed to cover the entire central-southern part of the delta from east to west, demonstrating that the advance of sand had by then made it impossible to cultivate a rather extensive part of the ancient delta. The number of delta branches that were still active had dramatically decreased, and the settlement episodes were clustered along their courses. The rank-size index was now +0.319 (Figure 5.10), indicating that the territorial fragmentation was still high. During the following phase, Yaz II (unless the picture is distorted by the fact that an upper-rank site has not been found), the system seemed to undergo a collapse in its political and administrative integration similar to that detected in the Late Bronze Age (Figure 5.11), but apparently having a less severe impact. The rank-size index was +0.284, which means a curve that is less convex than that of the previous period (Figure 5.12). During this period, the hydrographical system apparently became stable, as indicated presumably by the increase in the number of settlements along the western branch of the delta.

Iron Age At the beginning of the Iron Age (Yaz I) there seemed to be a resumption of the process of political integration of the territory (Figure 5.9), revealed by a new, progressive expansion of the settlement system in the delta, with the occupation of areas bordering on the main active branches.

During the subsequent Yaz III phase (RSI = 0.276) (Figure 5.13), the situation that emerged seemed to tend to forms 67

S. Salvatori

Figure 5.12 Rank-size of Yaz II sites.

Figure 5.13 Rank-size of Yaz III sites.

68

The Margiana Settlement Pattern from the Middle Bronze Age to the Parthian-Sasanian

Figure 5.14 Yaz III and IV site distribution.

Figure 5.15 Rank-size of Yaz IV sites.

69

S. Salvatori

Figure 5.16 Parthian period rank-size.

Figure 5.17 Sasanian period rank-size.

70

The Margiana Settlement Pattern from the Middle Bronze Age to the Parthian-Sasanian of higher integration. The system had presumably become stable, in terms of a more complex hierarchisation of the sites. A very large central site (Site No. 174) reappeared, while the location of the sites, apart from their numerical oscillation, remained stable along the lines determined by the delta branches that were still active. However, during the late Yaz IV phase (Figure 5.14), the western branch, where, during the previous periods, Sites No. 379, 384 and 388 had flourished, together with their smaller satellites, was apparently affected by further water-flow problems, because these settlements disappeared. Another important fact is that the vast perifluvial areas of diffused settlement in the eastern part of the delta were still occupied (Sites No. 211, 212 and 213), though we do not know whether this occupation was only partial. The rank-size curve (RSI = -0.234) (Figure 5.15) had now become concave, anticipating the situations of the subsequent periods. It seems therefore that a state-like organisation, probably integrated in an extraregional system, had then appeared in Margiana.

and administrative integration, by then had been almost completed, and the subsequent phase was clearly indicated by the territorial systems of the Parthian period (RSI = 0.830) (Figure 5.16) and of the Sasanian one (RSI = -0.536) (Figure 5.17), when the rank-size curve appeared, for the first time, to be decidedly concave, in connection with the large centre of Erk-kala. This configuration clearly shows that the role of the dominant centre exceeded the regional limits and that Margiana therefore was now included in a broader political and administrative system. Erk-kala. was the main political and administrative centre of the region, and its rank should be considered in a context that was much broader than that of the Parthian-Sasanian imperial system. Conclusions From the point of view of the peopling of the area, despite the various environmental crises that, from the Late Bronze Age on, promoted the progressive southward advance of the front of the Karakum sand, determining an undeniable shift in the settlement barycentre of the region, and despite the evident crises of the political and social establishment (which we have seen, for instance, during the transition from the Middle Bronze Age to the Late Bronze Age), the overall trend was a constant growth, revealed by the parameter of the dwelling surface expressed in hectares for each period (Figure 5.18). Moreover, if we analyse the sizes of the settlements of the various periods in terms of percentages, we can see that the settlement structure did not undergo any variations, except in the main settlements, i.e. in those settlements that, because of their size, can be classified as urban, or rather protourban (Figure 5.19).

During the Yaz III phase, the local system, apart from the rank-size values, seemed to be strongly determined by a considerable organisational capability whose political centre was probably outside of the area under examination. The occupation of the territory was remarkably concentrated along the axis of the delta branches, which were probably kept active through an upstream control: we believe that only a centralised control could have achieved the guarantees required for the realisation of a settlement and production system concentrated along predetermined, stable axes (Cattani and Salvatori, Chapter 1). Moreover, the system seemed to be equipped with defensive lines that ensured its protection both to the north-east and to the northwest. These defensive lines consisted of a considerable number of fortresses that bordered the settlement area and protected its northern boundary (Figure 1.8).

This is the most surprising result of the research under way: the crises of the system were apparently not always connected with climatic and environmental determinants, as previously surmised. The demographic arrangement of the region (which we have a chance to observe over a long period in this privileged geographic laboratory), though sometimes it undergoes negative impact due to environmental deterioration, at other times seems to conform to rules of a completely different nature. Undoubtedly, starting from the Final Bronze Age, the southward shift of the system depended directly on the climatic and environmental events that transformed the irrigated flat‑land of the delta landscape into an environment that was more and more extensively attacked by the advance of the desert sand.

The transition towards a state organisation, with a strong control of the territory in terms of productive

It seems, however, that these progressive transformations of the Murgab Delta landscape did not significantly affect the demographic dynamics of the region, except at the end of the Late Bronze Age and during the Final Bronze Age; so the different organisational arrangements of the settlement system must be due to other causes. In order to understand the peculiar, but somehow regular oscillations,

Figure 5.18 Distribution of population growth according to site size.

71

S. Salvatori

Figure 5.19 Graph of distribution of sites in the different periods.

of the system, we must, on the one hand, give up linear evolutionary patterns, and on the other hand regard the complex political and social organisation as an in fieri structure that moved toward constantly different and instable forms of balance, by trial and error and through gradual or abrupt, forward or backward transitions. It was therefore a far-from-linear movement, and its course is not easy to represent in a two-dimensional space. The case of Margiana, in this sense, is enlightening. More than once the establishment of central place (in which the administrative control functions seemed to coalesce and the early ramifications of a state political power seemed to develop, affecting the territorial integration process) underwent a selective collapse that involved only the dominant centre without having severe demographic consequences. Only after several trials and errors, the system became stable and a state appeared, with forms that were quite distinctive and decidedly different, for instance, from the incipient ones of the Middle Bronze Age with its three hierarchical levels. The full concavity of the rank-size curves reveals that during the Parthian-Sasanian period Margiana was only a part of a system that was geographically much broader and politically otherwise integrated. It is not a coincidence that the new organisation, which cannot be easily enclosed in a polygonal pattern, began to emerge during a period that was historically marked by the appearance of the Median-Achaemenid kingdom, and became stable, remaining practically unchanged throughout the span of the Hellenistic, Parthian and Sasanian periods.

To return briefly to the main assumption of this paper, we would like to emphasise the fact that the crisis of the organisational system did not always coincide with the environmental crises. Between the Middle Bronze Age and the Late Bronze Age, when the large centre of Gonur North was abandoned, and the slopes and top of the mound were used as a burial ground by the inhabitants of the smaller fortified village of Gonur South, the delta flood‑plain was still practically free of sand. During the early stages of the Late Bronze Age there actually took place an irradiation towards the region to the east of Gonur, with a dense settlement system organised around Auchin. Then, when the encroachment of sand from the north and the attendant reduction in the flow rate of the Murgab River determined a southward shift in the settlement barycentre, there took place, surprisingly, a revival of forms of centralised integration of the system (Yaz I e II periods). This process was resumed during the subsequent phases of the Iron Age, despite the fact that the environmental situation had started progressively worsening again, leading to the rapid depopulation of the northern part of the western delta branch. The integration process did not stop even later on; on the contrary, it achieved its most complete and stable form during the Parthian-Sasanian period, despite the further worsening of the environmental conditions which had led to a considerable southward shift of the population.

72

The Margiana Settlement Pattern from the Middle Bronze Age to the Parthian-Sasanian Now that we have downsized the environmental variable as a direct cause of the recurring system-wide crises in the organisational configuration of the territory, we will have to look into the process of the formation of political power or into the dialectics among groups of power, in order to find the basic causes of this far-from-linear advance towards a complex society; this advance seemed to bring about a crisis precisely in the urban element, which, as a rule “… tends to preserve its structure; that is, the hierarchical organization, the position of cities in the hierarchy of sizes, and also the main socio-economic differences” (Pumain 1997: 115).

Johnson, G. A. 1980. “Rank-Size Convexity and System Integration: a View from Archaeology”, Economic Geography 56 (3), 234-247. Kohl, Ph. L. (ed.) 1981. The Bronze Age Civilization of Central Asia. New York. Kohl, Ph. L. 1984. Central Asia. Palaeolithic Beginnings to Early Iron Age. Paris. van der Leew, S. E. and J. McGlade (eds.) 1997. Time, Process and Structured Transformations in Archaeology. London. Marcolongo, B. and P. Mozzi 1998. “Outline of Recent Geological History of the Kopet-Dagh Mountains and the Southern Kara-Kum”, in Gubaev, Koshelenko and Tosi (eds.) 1998: 1-13. Masimov, I. S. 1979. “Izuchenie pomatnikov epokhi bronzi nizovii Murgaba”, Sovetskaya Arkheologiya 1: 111-131. Masimov, I. S. 1981a. ���������������������������������� “The Study of Bronze Age Sites in the Lower Murgab”, in Kohl (ed.) 1981: 194-220. Masimov, I. S. 1981b. “Raskopki v Kellinskom oazise”, Arkheologicheskie Otkritiya - 1980 Goda: 466-467. Masimov, I. S. 1984. “Zhiloi dom na Kelleli”, Pamiyatniki Turkmenistana 2/38: 16-18. Masimov, I. S. 1986. “Novie issledovaniya pamyatnikov epokhi bronzy na Murgabe”, in Masson (ed.) 1986b: 171-181. Masimov, I. S. and A. A. Lyapin 1977. “Arkheologotopograficheskie raboti ve nizov’yokh drevnego Murgaba”, Arkheologicheskie Otkritiya, 1976 Goda: 552. Masson, V. M. 1956. “Raspisnaya keramika Iuzhnoi Turkmenii po raskopkam B. A. Kuftin”, in Trudy YuzhnoTurkmenistanskoj Arkheologicheskoj Kompleksnoj Ekspedizii 7, Leningrad: 291-373. Masson, V. M. 1959. Drevnezemledel’cheskaya kultura Margiani. (Materiali i issledovaniya po arkheologii SSSR, no. 73). ������� Moskva. Masson, V. M. 1981. Altyn-depe. (Trudy YuzhnoTurkmenistanskoj Arkheologicheskoj Kompleksnoj Ekspedizii 18). Leningrad. Masson, V. M. 1986a. “Ancient Central Asian Civilization: Trends of Development, Assimilation of Ecological Niches, Cultural Links (Antiquity to Early Iron Period)”, Information Bulletin IASCCA 11: 76-84. Masson, V. M. (ed.) 1986b. Drevnie Civilizacii Vostoka. Tashkent. Miroshnikov, L. I. 1987. “Central Asia: Towards the Issue of Geographic Boundaries of a Historical-Cultural Region”, Information Bulletin IASCCA. Moscow: 5357. Miroshnikov, L. I. 1992. “A Note on the Meaning of the Term ‘Central Asia’ as used in this Book”, in Dani and Masson (eds.) 1992: 477-480. Pumain, D. 1997. “City-size Dynamics in Urban Systems”, in van der Leew and McGlade (eds.) 1997: 97-117. Salvatori, S. 1998a. “Margiana ���������������������������������� Archaeological Map: the Bronze Age Settlement Pattern”, in Gubaev, Koshelenko and Tosi (eds.) 1998: 57-65.

References Cited Allchin, B. (ed.) 1984. South Asian Archaeology 1981. Cambridge. Alekshin, V. A. 1983. “Burial Customs as an Archaeological Source”, Current Anthropology 24 (2): 137-145. Biscione, R. 1977. “The Crisis of Central Asian Urbanization in the II Millennium B.C. and Villages as an Alternative System”, in Deshayes (ed.) 1977: 113127. Biscione, R. and M. Tosi 1979. Protostoria degli stati Turanici. Napoli. Brice, W. C. (ed.) 1978. The Environmental History of the Near and Middle East since the Last Ice Age. ������� London. Cremaschi, M. 1998. “Palaeohydrography ������������������������������ and Middle Holocene Desertification in the Northern Fringe of the Murgab Delta”, in Gubaev, Koshelenko and Tosi (eds.) 1998: 15-26. Dani, A. H. and V. M. Masson (eds.) 1992. History of Civilizations of Central Asia. Volume I. The Dawn of Civilisation: Earliest Times to 700 B.C. ������ Paris. Deshayes, J. (ed.) 1977. Le Plateau Iranien et l’Asie centrale des Origines a la Conquête Islamique. Paris. Francfort, H.-P. 1984. ��������������������������������� “The Early Periods of Shortughai (Harappan) and the Western Bactrian Culture of Dashly”, in Allchin (ed.) 1984: 170-175. Francfort, H.-P. 1989. Fouilles de Shortughai. Recherches sur l’Asie centrale protohistorique, 2 volls. Paris. ������ Gerasimov, I. P. 1978. “Ancient Rivers in the Deserts of Soviet Central Asia”, in Brice (ed.) 1978: 319-334. Gubaev, A., G. A. Koshelenko and M. Tosi (eds.) 1998. The Archaeological Map of the Murgab Delta. Preliminary Reports 1990-95. Resorts and Memoirs. Saries Minor III. Rome. Hiebert, F. T. 1994. Origins of the Bronze Age Oasis Civilization in Central Asia. (American School of Prehistoric Research Bulletin 42). Cambridge. ���������� Jarrige, J.-F. 1986. “Mundigak ������������������������������������ and Mehrgarh: la question des relations entre le Baluchistan et la Turkménie méridionale au chalcolithique et à l’àge du bronze”, Information Bulletin IASCCA 11: 56-68. Johnson, G. A. 1977. “Aspects of Regional Analysis in Archaeology”, Annual Review of Anthropology, 6: 479508. 73

S. Salvatori Sarianidi, V. I. 1990. Drevnosti Strany Margush. Ashkhabad. Sarianidi, V. I. 1993. “Excavations ��������������������������������� at Southern Gonur”, Iran 31: 25-33. Sarianidi, V. I. 1998. Margiana and Protozoroastrism. Athens.

Salvatori, S. 1998b. “The ��������������������������������� Bronze Age in Margiana”, in Gubaev, Koshelenko and Tosi (eds.) 1998: 47-55. Sarianidi, V. I. 1957. “Keramichicheskie pechi drevnej Margiany”, Kratkie soobcheniya, Institute istorii material’noj kul’tury 69: 72-78. Sarianidi, V. I. 1981. ��������������������������������� “Margiana in the Bronze Age”, in Kohl (ed.) 1981: 165-193.

74

Chapter 6 Cultural Variability in the Bronze Age Oxus Civilisation and its Relations with the Surrounding Regions of Central Asia and Iran Sandro Salvatori

Introduction

the Middle Chalcolithic and the early phases of the Late Chalcolithic in the area of the Tedzhen River and of the piedmont of southern Turkmenistan (Lyonnet 1996). The formal relations between the materials of this early period of the Sarazm sequence are chiefly anchored to the southern Turkmenistan sequence, and straddle the final phase of the Yalangach period and the beginning of the subsequent Geoksyur period, because, though the pottery shapes are typical of the former horizon, their manufacture (which is present almost everywhere: red polished pottery with black spots) evokes the typical production of the beginning of the subsequent period. To this we must add the absence of painted decorations, an indication that seems to tally with evidence coming from the basal level of Geoksyur and from contemporary levels in other sites of the piedmont, such as Altyn‑depe and Chong‑depe. Period II of Sarazm is connected directly and with absolute certainty with Level 2 of Geoksyur 1, not only because of the presence of hemispherical bowls with polychrom decoration, but also because of the overall pottery finds. Period III of Sarazm still presents a strong Turkmen component connected with the late horizon of the site of Geoksyur 1 (Level 1), but the relations with Baluchistan become more marked, because of the presence of elements that find close parallels at Mundigak III, Sur Jangal II and Amri ID.

The remarkably vast spread of the Oxus Civilisation, which encompassed, starting from the Middle Bronze Age, the piedmont of southern Turkmenistan (Altyn-depe, Namazgadepe, etc.), the delta and lower reaches of the Murgab, the Margiana of the classical period, the regions of Uzbekistan and southern Tadzhikistan on the Amudar’ya, the righthand tributaries of this river, and northern Afganistan (oases of Daulatabad, Dashly, Nichkin and Farukhabad), needs to be studied in a carefully-focused, specific way. The purpose of this essay is to analyse – on the basis of the currently available archaeological documentation, with all its countless collection and publication limits – the convergences and differences within this cultural system during the Middle Bronze and Late Bronze Ages, in other words – in terms of an absolute chronology based on a still-too-limited number of radiocarbon determinations – between the middle of the third millennium BC and the middle of the second. At the same time, we will also endeavour to outline a picture of the relations that this Asian civilisation had with the Iranian plateau (including the alluvial plains of south-western Iran and the overall Elamite world), the Indus Valley, Baluchistan and the regions of the Arabian Peninsula that face the Gulf. The exploration and partial settlement of other areas seem to have involved, much before the period examined by us, the Chalcolithic groups of the southern Turkmen piedmont and Tedzhen Delta. As early as the Middle Chalcolithic Age, they had started “colonising” new areas to the northeast, advancing up to the Zarafshan River. A. I. Isakov’s excavations (Askarov 1977.) at Sarazm, on the middle reaches of the river, unearthed – for the beginning of the cultural series of the site (Period I) – ceramic material that was quite similar to that known for the final phases of

Period IV seems to be a period of great development for the settlement of Sarazm, and though Turkmenian pottery ascribable to the final phase of the Late Chalcolithic (Namazga III) (upper level of Chong‑depe and final period of Kara‑depe) is still present, the presence of pottery from the Baluchistan becomes considerably more substantial. The various styles known in Baluchistan are present here (Quetta, Amri, Nal and Sur Jangal), as had been observed, for instance, at Mundigak (Lyonnet 1996). This expansion of the communities of the Middle-Late Chalcolithic of southern Turkmenistan, which was also responsible for the presence of Late Chalcolithic material both at Shahr-i Sokhta in the Iranian Sistan (Bonora et al. 2000; Sarianidi 1983) and at Mundigak in Afganistan (Casal 1961), unaccountably left some gaps in areas that were quite desirable for farming communities. I refer in particular to a region situated precisely along the theoretic axis that joins the Tedzhen River to the Amudar’ya River and then to the Zarafshan River: the Murgab Delta. Here, so far, the only find datable to the Chalcolithic Age is a terracotta figurine with the typical Geoksyurian features,

We are accepting the use of this term here, though with some doubts, because it tends to shift the barycentre of the civilisation it defines, centring it on the two banks of the great Central Asian river and somehow marginalising the piedmont of southern Turkmenistan, in spite of the fact that this area contains large and important urban centres of the Early and Middle Bronze Ages such as Altyn‑depe, Namazga‑depe and Ulug‑depe. Another region that is likewise marginalised is that of the delta and the lower reaches of the Murgab, where the peopling during the Middle and Late Bronze Ages was particularly substantial. On the other hand, a reason that supports the use of this term in place of the widespread BMAC (Bactria Margiana Archeological Complex) is that the latter, starting from regional names, marginalises the piedmont of southern Turkestan even more; it has been used, moreover, in a way that tends to confine cultural integration exclusively to the Late Bronze Age (Namazga VI). 

75

S. Salvatori

Figure 6.1 Comparative C14 chronology between Margiana and Bactria.

collected on the surface in the area of Kelleli by I. S. Masimov (Kohl 1984: Pl. 14, centre, bottom). The lack of data in this area may probably be ascribed to the fact that the northern parts of the Murgab Delta were covered by sand of the Karakum from the second millennium BC on, while the central-southern parts received a constant input of alluvial deposits from the river. Another reason for the absence of data may be the complete lack of archaeological exploration activities in vast areas to the west of the Aravali basin and in the overall area west of the Merv Oasis, a crucial region that separates the Tedzhen Delta from the Murgab one.

No matter what drive gave rise to this sort of exploratory movement (other episodes of this expansion also have some implications of colonisation/occupation of territory: the Uruk drive towards Syria and Kabur or the Egyptian proto-dynastic one towards the Palestinian corridor), it seems certain that during this period (between the fifth millennium BC and the fourth) the foundations were established of a considerably extensive geographic knowledge and of a comprehensive “opening” of relationships between the cultures and human groups that occupied an extremely vast area, from the eastern shores of the Mediterranean Sea to the Indus Valley and from the Arabian Peninsula to inner Middle Asia. The effects of all this became measurable above all during the third millennium BC, with the intensification of all sorts of exchanges, not excluding the élite ones and the ideological ones, and the appearance, among other things, of the first forms of market exchanges.

The Chalcolithic expansion of the population of the southern Turkmen piedmont and the Tedzhen area may resemble, on the one hand, the earlier (fifth millennium BC) explorative movement of the Obeid peoples along the northern coast of the Arabian Peninsula, facing the Persian/Arabian Gulf (Burkholder 1972; Masry 1997; Oates 1986), and, on the other hand, the mutually more or less simultaneous exploratory movements of the Late Uruk/Jemdet Nasr/ED I communities of Mesopotamia towards the Iranian plateau and of the Proto-Elamite population (Jemdet Nasr/ED I) towards eastern Iran and Sistan (Lamberg-Karlovsky 2001b; Amiet and Tosi 1978; Amiet 1979, 1983).

As regards Margiana, however, the lack of data datable to the Chalcolithic Age and Early Bronze Age opens a series of interpretation problems with reference to the settlement of this important area, which had a high agricultural potential. A similar and perhaps more serious problem arises for Bactria, where the only known materials dating to the Middle Bronze Age come from the graveyards in the 76

Cultural Variability in the Bronze Age Oxus Civilisation oases of northern Afganistan that had been looted in the nineteen-seventies. In this case, however, various types of evidence allow us to ascribe the gaps in the data (at least, but not only, for the Middle Bronze Age) to inadequate research strategies in the survey operations carried out by the Soviet archaeologists. Though we are not yet in possession of new data that we presume may come (we hope in the near future) from northern Afganistan, we are able to see that a comparative analysis of the C14 dating available to this day in the two areas (Figure 6.1) makes it possible to suspect the presence of an Early Bronze Age phase in Bactria and to date the two palace-like buildings of Sapalli‑depe and Dashly 3 to the Middle Bronze Age. Other aspects, apart from the evidence supplied by the radiometric dating, will be discussed below, particularly in the paragraph about the Middle Bronze Age architecture in the area.

by these three sites does not give a satisfactorily broad picture of the defensive architecture of this area during the Middle Bronze Age, it does reveal an essential unity in the architectural culture of the period. Another contemporary site, of which only a small part has been excavated, Adzhi Kui 8 (Sarianidi 1990: Fig. 2) possesses a defensive structure that appears to be less grand and severe. Here too, however, though the proportions of the building are more limited, there is a tower next to the entrance. This series of defensive buildings characterised by quadrangular towers at the corners and along the sides apparently includes also the structure of Taip 1 North, excavated by I. S. Masimov and unfortunately unpublished as yet. In Bactria, rectangular towers associated, in this case too, with double curtain walls, are present in the circular building of Dashly 3. The chronology of the settlement is very uncertain. When the structure had lost its function, some graves were dug in it: they obviously belong to a later period, and can be dated to the Namazga VI period and in any case to the Late Bronze Age. It is not unlikely, therefore (and it would tally with some instances of C14 dating, as we have mentioned above) that the circular structure may date from the previous period of the Middle Bronze Age.

The picture of this Central Asian area that appeared during the Middle Bronze Age was that of a comparatively homogeneous culture, distributed along the piedmont of southern Turkmenistan, the Murgab Delta (the Achaemenid Margush or classical Margiana) and the oases or microdeltas of northern Afganistan. We have just stressed the fact that this extremely vast region, during this period, seemed to belong to the same culture; but this culture, together with strong convergences, also showed ongoing regionalisation phenomena, therefore elements of discontinuity in the material production of the various regions.

Another chronologically uncertain datum is the position of other two “palatial” buildings, quite similar to each other in their layout: the complex of Sapalli‑depe (Askarov 1977: Fig. 4), in northern Bactria, and that of Dashly 3 (Sarianidi 1977: Fig. 15), in southern Bactria. In both cases the architectural complex is characterised by a discontinuous double-curtain wall produced by the combination of Tshaped elements. In the building of Dashly 3, the entrance door is formed of a rectangular avant-corp divided into three parts, with the central access characterised by three successive constrictions that were probably closed by wooden doors. The “palatial” complex of Sapalli‑depe is more regular in its perimeter than that of Dashly 3. After it had lost its function, it was used as a graveyard. Here the funerary outfits of the graves belong to an early stage of the Late Bronze Age, so, in this case too, it is not unlikely that the building may date to the Middle Bronze Age.

Middle Bronze Age Architecture The Middle Bronze Age architecture of Margiana is known as yet only in an unsatisfactory way, though it is possible to point out some interesting features, particularly in the boundary walls or fortifications of the largest centres excavated so far. On this subject, in addition to what has already been known for years (Sarianidi 1990, 1993b; Hiebert 1994a; Salvatori 1998a), we must mention the publication, albeit partial, of the excavations performed by V. Sarianidi at Gonur 1 North (Sarianidi 1998a, 2002). The structure of the boundary wall of the palace-like complex, with its rectangular towers placed along the sides and at the corners of the double curtain walls, closely resembles the fortification structures of the Middle Bronze Age sites of the north-western part of the delta, Kelleli 3 and 4 (Masimov 1984, 1986). Though the evidence yielded

Finally, we must mention the great defensive structure of Altyn‑depe, formed of the city walls, of which, however, we know only the layout of the southern gate. Besides its monumental character, we should point out the presence, as at Dashly 3, Gonur 1 North and Kelleli 4 and -3, of two rectangular buttresses on its sides (Masson 1988: Fig. 11).

As Ph. Kohl has pointed out, the Soviet surveys in the oases of northern Afganistan were not systematic; moreover, vast areas of open takyr, clearly visible in the satellite photographs, were not examined, despite the fact that their characteristics were quite similar to those of the areas that were being explored. These severe limits in the documentation relative to the development of the peopling of that region made it impossible to use that documentation “…to prove the thesis that emigration from southern Turkmenistan alone was responsible for the emergence and development of Bronze Age communities in northwestern Afganistan “ (Kohl 1984: 160). 

The impression given by the totality of the Middle Bronze Age defensive structures in the area of the Oxus Civilisation is that of a great homogeneity and strong convergence, despite the fact that they are divided into regional variants,

77

S. Salvatori reflecting a phenomenology that, as we shall see, also played a part in other forms of material production.

coming from the graves mentioned above has strong formal analogies with the series of Middle Bronze Age Margiana, with production differences that concern almost exclusively the clay and firing, and a predominance of reddish or pinkish surfaces, often unslipped. When slip is present, it is whitish, as in Margiana.

The analysis that follows will deal, in the order stated, with locally-produced and imported classes of objects that make it possible to evaluate the degree of regionalisation inherently expressed by the prevailing variants (we believe that, in some cases, these classes of objects, because of their widespread distribution also outside of the Oxus Civilisation, belong to the symbolic dimension of exchanges and are real elements of a transcultural system). Then we will deal with those classes of objects that, because of their distribution, seem to belong to exchange subsystems and therefore yield information about specific, sometimes exclusive (bidirectional) routes of segments of a broader international trade organisation.

At present, the southernmost attestation of this ceramic culture seems to consist in the finds of Takhta Bazar, on the middle reaches of the Murgab, where there appear, in tomb contexts, all the significant elements, including pottery, of the Middle Bronze Age of Kelleli/Gonur 1 North. These finds allow us to identify one of the communication routes between Margiana and southern Bactria; this route was certainly active also later, during the Late Bronze Age. In general terms, we can state that one of the most marked differences between the production of Altyn (i.e. of the piedmont) and that of Margiana consists in some strongly carinated beakers, which are practically absent in Margiana, and in the lower frequency, in Margiana, of large bowls with long, narrow cylindrical spout.

Pottery The problem of the relationship between the Middle Bronze Age pottery production of Margiana and that of the piedmont has been dealt with, we believe in a comprehensive manner, in relation to the currently available data, by B. Udeumuradov in a study of 1993. The formal convergence of the Middle Bronze Age (Namazga V) pottery types of the large urban settlement of Altyn-depe, in the piedmont, with those in the area of Kelleli and (we can now add) with the materials of the graveyard of this period at Gonur (Salvatori 1993, 1994a, 1994b, 1995a) and of the small contemporary settlement of Adzhi Kui 9 (Salvatori 2002) is extraordinarily strong, though from the viewpoint of production it is more than evident that each of these complexes was manufactured with materials peculiar to each of these two areas. Unfortunately we know little or nothing – or rather, little or nothing has been published – about the pottery production associated with the looted Bactrian graves, because private collecting was mainly focused on artefacts in bronze, other metals and stone, including semi‑precious stones. The scanty ceramic items present in private collections have not been published yet, except for a very small number of specimens. As far as I have been able to see in private collections, the pottery

The differences, which obviously are not limited to these few elements, are not very important, however, and appear to be normal within the process of regionalisation of such a widely distributed culture. Anthropomorphic Terracotta Figurines The production of female terracotta figurines, in our opinion, allows us to appreciate the degree of complexity of the regional development. Though the overall structure of the figures is quite similar in Margiana and in the piedmont, the details indicate two different traditions, i.e. a difference in the lexicon that allows us to identify the two types of production with a remarkable certainty. As V. M. Masson had correctly pointed out (Masson 1988: 9293; 1986: 40-41), this production, though present in both the regions, has precise features in each of them, and this makes it possible to tell them apart from a stylistic point of view. The production of Altyn‑depe is characterised by the portrayal of hair, which flows down the back in a single braid and in several long braids on the chest (Masson 1988: Pls. VIII, IX). This element is completely absent in the Margiana production. Here the figurines are characterised by a triangular head, often with two holes on the top, and without any trace of applied elements to represent hair (Sarianidi 1990: Pls. XVI-XXII). The fact that the two types of production are contemporary is attested by the presence of a few specimens of the Margianan-type figurines at Altyn‑depe (Masson 1988: Pl. IX.3, 7; Udeumuradov 1993: Fig. 21).

Little can be added to Udeumuradov’s study, at least until the publication (if ever it will be possible) of the pottery sequence of the largest Middle Bronze Age site in Margiana, i.e. Gonur 1 North. About this, however, we are very doubtful, because V. Sarianidi, in his latest publication about the excavations in this large settlement (Sarianidi 1998a), shows a strong tendency to underestimate stratigraphical facts. This Russian archaeologist asserts that the rooms of the palace of Gonur 1 North were filled exclusively with sand (Ibidem: Fig. 40): this allegation, completely devoid of imagination, is simply meant to hide the fact that, during the excavation of the palace, he did not make any attempt to identify and document the sections within the rooms, though this, as we all know, was of the utmost importance. This essential deficiency in the excavation methodology and documentation becomes even more serious when we consider that (as we can witness, having seen it directly) during the excavations the pottery that was collected was accumulated in long rows out of the excavation trenches: the rows roughly corresponded to the depth at which the objects had been found. This operation, too, was carried out without any stratigraphic checks, i.e.without any reference to recognised stratigraphic units, but only with respect to the macrocuts performed for removing the soil. 

Female anthropomorphic ceramic figurines, on the contrary, are absent in the burial complex of southern Bactria dating to the Middle Bronze Age. An explanation of this may be the fact that, as far as we know from the 78

Cultural Variability in the Bronze Age Oxus Civilisation finds of both Altyn-depe and Margiana, this type of object practically never appears in funerary contexts.

assert, on the basis of the documentation published so far, that in Margiana there prevail the types with geometric motifs, whereas in Bactria there frequently appear portrayals of winged deities, anthropomorphic figures, animals such as monkeys, and very complex designs of snakes or eagles.

On the other hand, we believe that we should not overlook the fact that, though the Bactrian materials of the Middle Bronze Age come exclusively from the looted graveyards of northern Afganistan, to this day no settlement of this period has been excavated in that region. This fact greatly limits our possibility of achieving a satisfactory assessment of the variability of this production over the entire area under examination, and leaves many historically and culturally important problems unsolved.

The distribution of this class of seals, with specific reference to the Central Asian types, is very interesting, because it makes possible to understand the vastness and complexity of the trading network that involved the world of the Oxus Civilisation during the second half of the third millennium BC. The current documentation consists of some seals with a strong Bactrian characterisation found in the Indus Valley (During-Caspers 1994a, 1994b); seals and seal impressions from the Bactria-Margianan area found in graves at Shahdad (Hakemi 1997: Graves 31, 36, 55, 74, 115, 136, 163, 166); seals belonging to the production of the piedmont of southern Turkmenistan imprinted on locally-produced pottery and found in the graveyard of Shahdad (Ibidem: Graves 62, 122, 308, 377); and seals in the Bactria-Margianan tradition found at Susa and Tepe Yahya (Amiet 1986: Fig. 105; Pittman 2001: Figs. 10.5859).

Compartmented Bronze Seals This type of seal is widespread in the entire area of the Oxus Civilisation and it is an essential key to the interpretation of the regionalisation phenomena. First of all we must mention that it was the discovery, in archaeologically controlled sites in Margiana, of a substantial number of compartmented stamp seals in bronze and sometimes silver that made it possible to date to the Middle Bronze Age the enormous number of seals of this type that were being sold in the antique market of Kabul in the nineteenseventies and that came from the looted burial grounds of the oases of north-western Afganistan (Salvatori 2000). There are three larger groups of compartmented seals in metal (chiefly bronze), and they come respectively from the excavations of Altyn‑depe (Kircho 2001: Figs. 1213); from Margiana (Middle Bronze Age graveyard of Gonur-1 North: Salvatori 1994b: Figs. 11.4, 13.6; 1995a: Fig. 10:G.91/9; Gonur-1 North and other Middle Bronze Age sites in Margiana: Sarianidi 1990, 1998b; Masimov et al. 1998; Masimov and Salvatori, Chapter 7); and from the graveyards (but also from other contexts) of southern Bactria (Sarianidi 1977, 1986, 1993a; Salvatori 1988a; Pittman 1984; Baghestani 1997; Tanabe 1983).

A counterpart of the Bactria-Margianan bronze compartmented seals found in the Indus Valley (DuringCaspers 1994a) is represented by some seals from Harappa and other objects coming undoubtedly from the Indus Valley and found at Altyn‑depe (Masson 1988: Pl. XXII.1ab; Kircho 2000: Figs. 2.I.5, 2.IV.10-11); by some pottery shapes of Indian origin coming from the graveyard and the settlement of Gonur 1 North (Salvatori 1994b: Fig. 5.1; 1995a: Fig. 18, G. 431/3; Sarianidi 1998a: Fig. 13.8-9); by an etched carnelian bead collected by Sarianidi in one of the sites in Margiana surveyed by him; and finally by a Indus Valley inscribed stamp seal with the representation of an elephant (Sarianidi 2005: Fig. 114). The designs of monkeys on compartmented seals coming from southern Bactria are probably of Indian origin as well. Though the evidence of contacts between the two areas of ancient urbanism consists of a small number of objects, in any case these objects demonstrate the existence of bidirectional exchanges. The seals, objects with a strong personal characterisation, indicate the presence of “merchants” of the Indus Valley in Central Asia; and the converse is also true.

The peculiar characteristics of each of these three groups of seals are easy to recognise. At Altyn‑depe, and as a rule in the piedmont (for instance at Tekkem-depe: Kohl 1984: Pl. 20b), closed-back compartmented stamp seals predominate decidedly, mostly with cross and stepped pyramid motifs (Kircho 2001: Fig. 12.1-18). Zoomorphic seals are frequent as well (Ibidem: Fig. 13.1-12), while openwork seals are more rare (Ibidem: Fig. 13.23-24b); the latter are much more frequently documented in the Bactrian-Margianan series. Both in Bactria and in Margiana the types known at Altyn are present but are an absolute minority.

Another, even more interesting, aspect is the evidence yielded by the graveyard of Shahdad, where the seals and seal impressions on locally manufactured vessels are flanked by other objects that are typical of the Oxus Civilisation. The number of these objects is high enough to make it possible to surmise that at Shahdad, during the Middle Bronze Age, there was a substantial trading

The two complexes – that of Bactria and that of Margiana – are apparently much more similar to each other; the Bactrian one, however, differs because of the presence, in considerable numbers and variants, of apparently distinctive themes such as monkeys and male or female figures sitting or controlling animals. As a rule, we can Sarianidi 2001, 2002 and unpublished material from the Middle Bronze Age graveyard of Gonur 1.

The objects published in Kircho 2000 are ivory dice that find an exact parallel in the palace of Gonur 1 North (Sarianidi 1998a: Fig. 21.16-7).





79

S. Salvatori community coming from Central Asia, and their typology makes it possible to recognise the presence of individuals coming both from the piedmont of southern Turkmenistan and from the Bactria-Margianan regions. The diagnostic elements that can be referred to Altyn‑depe bronze stamp seals production consist of impressions on red ceramic jars; this is true for the glyptic production of the BactriaMargianan area as well.

2) Site No. 1220, cf. Salvatori, Chapter 8) a seal fragment from Togolok 1: Sarianidi 1990: Pl. LXXXV.12; 3; a seal, again from Togolok 1: Sarianidi 1990: Pl. LXXXV.7). One of the two seals from the graveyard of Gonur has a cuneiform inscription; its stylistic characteristics indicate that it was imported from an Iranian environment and belonged to the post-Sargonid Akkadian period. The other two seals belong to a well-known series of which there are specimens at Yahya IVB (Pittman 2001: Figs. 10.46, 48-49, 51; Lamberg-Karlovsky 2001b: 273), at Shahdad (Hakemi 1997: Graves 163, 166, 193), and, more generally, in south-eastern Iran (Amiet 1986: Figs. 133, 134, 137). The Bactrian specimen published by Winkelmann (1997) also belongs to this group. The specimens from Gonur, since they come from the Middle Bronze Age graveyard, may be dated to the period between 2400 and 2100 (but if the grave context were published, this might help narrow down the time range); this dating, incidentally, agrees with the analysis of the problem recently carried out by LambergKarlovsky (2001b: 271-273) for the specimens from Yahya. The importance of this group of seals consists above all in its distribution. Since a seal is an object that expresses the identity of its user/owner (in Mesopotamia, as a matter of fact, the personal property of the object is often stressed by an inscription that indicates the owner’s name), we must conclude that the group of seals described above denoted individuals belonging to the same cultural sphere and field of activity. The fact that these seals reappear at Susa, Tepe Yahya, Shahdad, in Margiana (Gonur 1) and in Bactria indicates a precise line along which, as we have seen in our analysis of small chlorite containers for cosmetics, specific trading transactions took place within the boundaries of a well-established merchant organisation.

The seals of Middle Asian origin found at Susa, finally, are typologically related to the Bactria-Margianan type also. We must point out the presence of compartmented stamp seals belonging to the sphere of the piedmont of southern Turkmenistan at Tepe Hissar (Schmidt 1937: Fig. 118) and at Anau (Pumpelly 1908: Fig. 257). This is not surprising, if we consider that a possible route for the connection between the Meana-Chaacha area and Shahdad passes precisely through Tepe Hissar-Khinaman, as demonstrated by a series of finds and of very specialised techniques that connect Tepe Hissar and Khinaman with the eastern Iranian settlement: the bronze objects of Khinaman (Curtis 1988) converge stylistically with the contemporary production of Hissar IIIB on the one hand (Schmidt 1937: 191, Fig. 112) and with that of Shahdad on the other hand (Hakemi 1997: axes: Graves 36, 47; dishes with embossed decoration: Graves 115, 122, 232, 297), while both at Shahdad and at Hissar there reappears the unusual technical convergence of the systematic use of the chisel shoulder as a support for manufacturing flintstone drills for obtaining necklace beads (Bulgarelli 1974: 22-24). Towards Shahdad, a centre that in another occasion we defined as a boundary and hinge between the Elamite world and Outer Iran (Salvatori 1978: 13-14), there converge several connection lines coming from the various regions of the Oxus Civilisation; during the second half of the third millennium BC, these regions show a great dynamism based on trade, but also on cultural relationships, as is demonstrated not only by the extremely wide diffusion of elements peculiar to this culture, but also by its capability of absorbing and assimilating, on the level of artistic expression, some elements that are typical of other great civilisations of the Middle-East.

Metal Phials, Parade Axes, Palettes, Applicators for Cosmetics As regards the metal production, other than that of compartmented seals, it is possible to propose some conjectures of regionalisation or strong convergence. For instance, let us examine how some toilet objects such as spatulas with bent tang, mirrors, phials and applicators for cosmetics, though they appear frequently in all the area of the Oxus Civilisation and have very strong formal analogies, can help us, at least partly, to interpret the convergence/divergence system.

Cylinder Seals (Margiana, Bactria, Indus Valley, Iran) The presence of cylinder seals is rarely attested east of Mesopotamia and in Iran, out of the Elamite area, during the second half of the third millennium BC. The few specimens known in Central Asia and dated to this period come from southern Bactria (Winkelmann 1997: Fig. 1a-c) and, above all, from Margiana; but here too the number is almost negligible (graveyard of Gonur, two specimens: Sarianidi 2001: 197, Pl. 10.7; 2002: Fig. on p. 278. Surface finds: 1) graveyard of Gonur: Sarianidi 1998b: Fig. 27.3;

The sector of metal phials is the one that apparently shows the regionalisation incidence more clearly: the Bactrian production differs for its more elaborate plastic decoration, which sometimes achieves extremely sophisticated An initial reading of the inscription was given to Sarianidi by T. Sharlach (Sarianidi 2002: 334).  Similar conclusions, though in a chronological framework that partly differs from the one proposed here, had been reached also by Pierre Amiet, who, among other things, proposed an interpretation of the buildings of Sapalli‑depe and Dashly 3 as trading compounds comparable to the Assyrian karum in Cappadocia (Amiet 1986: 194, 206). 

See Salvatori, Chapter 8.



80

Cultural Variability in the Bronze Age Oxus Civilisation surfaces decorated with bas-relief and high-relief scenes (Pottier 1984: Pl. XVIII). Obviously we must not overlook the fact that the number of illegally excavated graves is rather great in the oases of north-western Afganistan, while in Margiana we have only the graves of the burial ground of Gonur 1 North, i.e. a very limited sample, considering that about 90% of those graves were looted in antiquity (Salvatori 1995a: 5). In the piedmont, so far, metal phial finds have been rather rare. We can mention the specimen from Sumbar (Kohl 1984: Pl. 21b), typologically identical to the specimens from the graveyard of Sapalli (Askarov 1977: Pl. XXVI), and probably ascribable to the beginning of the Late Bronze Age. At Altyn-depe, three specimens of the type with a sub-globular body are known (Kircho 2000: Fig. 2; 2001: Figs. 7.7 and 18, 10.3;); in this case they all come from Namazga V contexts. In Margiana, too, very few metal phials are known. Three specimens, of which one is a ceramic imitation, come from Middle Bronze Age grave contexts (Salvatori 1993: Figs. 7, 12: G.C. 2/8; 1994b: Figs. 11.5, 14.1); a specimen from Togolok-24 is typologically unique, and almost certainly belongs to an advanced period of the Late Bronze Age (Sarianidi 1990: Pl. XXXVI.27), if it is not a culturally decontextualised object.

Though it is extremely unbalanced, the currently available documentation seems to indicate that in Bactria this object attracted a particular attention in the production sphere or in the local demand, and that this led to the development of a distinctive style that might be defined, so to speak, as Baroque. A similar phenomenon reappears in the production of ceremonial axes (Pittman 1984: Fig. 36; Salvatori 1988: Figs. 102-103), while the production of axes for everyday use seems to have a common denominator on a typological level in the entire area. (This is true also for metal flameshaped mace heads: Sarianidi 1998b: Fig. 24.13; Pottier 1984: 48, Pl. XLIII.317, 320; Askarov 1977: Pl. LVII.23). Out of Bactria, this type of ceremonial axe is rather rare; but we must mention the specimen from Khurab (During‑Caspers 1972), which is closely related to a specimen from Susa, one from Shahdad and several from southern Bactria. As regards bronze mirrors, the simple disk type is known throughout the area and also elsewhere in Iran. Their simplicity, however, prevents us from regarding them as elements having a strong chronological value, though they can serve to establish a cut-off point between the Middle and Late Bronze Age. The disk-shaped mirror with smooth edges is present throughout the third millennium BC. It appears in the graveyard of Gonur 1 North in at least two specimens (Grave 37: Salvatori 1994b: 19 No. 25, Fig. 11.8; Grave Channel 4: Idem 1993: 7 note 9, Fig. 8). Again in Margiana it appears in the graveyard of Takhta Bazar (Udeumuradov 1993: Fig. 32.3-4, 7) and is frequently attested in collections coming from southern Bactria (Pottier 1984: Pl. XXXII; ��������������������������������������� Amiet 1988: Fig. 10a; Sarianidi 1977: Fig. 40.1-3). Out ����������������������������������������� of the area of Bactria and Margiana, this type of mirror has been found at Altyn-depe in Graves 843-845 (Kircho 1988: Fig. 4.10), in Grave 721 and in the so-called “Hoard 2” of the “Tower” excavations (Kircho 2000: Fig. 2.III.4 and IV.1), at Mundigak IV.1 (Casal 1961: vol. II, �������������������������������������������������� Fig. 139.17; vol. I, ���������������������������� 249), at Shahr-i Sokhta; in graves from the third quarter of the third millennium BC (Grave 712, inv. 7641 and Grave 725, upper, inv. 8015: Salvatori and Piperno n.d.), at Mehrgarh VIII (Santoni 1988: Fig. 2.6); among the materials of the Vakhsh culture (P’yankova 1981: Fig. 16.10-12) and at Sarazm IV, in Tadzhikistan (Askarov 1977: Fig. 64.2). During the subsequent period, the Late Bronze Age (Namazga VI), a variant of this type of mirror, in laminar bronze with raised edge, became widespread both in Bactria (Sapalli-depe: Askarov 1973: Fig. 25.14-16; Idem 1981: Fig. 7.13-14) and in Margiana (Hiebert 1994a: Fig. 12.2 from Grave 40 of the Namazea VI graveyard of Gonur 1). This technical and formal variable turns out to be a clear criterion of chronological and cultural discrimination.

The production of southern Bactria is undoubtedly the most varied, both because of the richness of the decorations on the external surface of these small containers (the only decorated specimen from Margiana is one that comes from the Middle Bronze Age graveyard of Gonur 1 North and has a snake that seems to serve as an ear for the small container: Salvatori 1994b: Fig. 7), and because of the variety of the phial forms, which in many cases have a zoomorphic or anthropomorphic shape (Pottier 1984: Pls. XVIII-XIX; Pittman 1984: Figs. 12, 14-16; Sarianidi 1979: Fig. 3.3; Tanabe 1983: Figs. on pp. 40-42).

This is not surprising, since, as we have already stated above, the graveyard near the Gonur 1 North site is known to have been looted in approximately 90% of its graves during the Late Bronze Age, as we have been able to demonstrate by excavating some of these graves that had subsequently been covered by a small kiln for the production of faience objects (Salvatori 1995a: 11). A certain number of objects coming from the Late Bronze Age Margiana are undoubtedly older than the context in which they have been collected. Among these, we should mention a little white marble head which undoubtedly belonged to a Bactrian-type composite figurine and was found in a recess in the wall near the northern gate of the fortified complex of Gonur 1 South (Sarianidi 1998a: Fig. 17.1). The fact that these figurines date to the Middle Bronze Age has been demonstrated, on the other hand, by the discovery of a complete specimen in the Middle Bronze Age graveyard of Gonur (this object can be seen at the National Archaeological Museum of Ashgabat). Other objects that are quite likely to come from the looting of that graveyard are the numerous little stone columns found in the central building of Togolok 21, where they were being reused as raw material for producing several types of objects (Sarianidi 1990: Pl. LXXXIX; Hiebert 1994a: Fig. 9.19). This category of items also includes some fragments of chlorite objects decorated in the “intercultural” style, i.e. belonging to the “série ancienne“, in the terminology of P. de Miroschedji (1983). In other words, the recycling of older objects retrieved by grave looters was probably quite intense in Margiana, so it is not surprising to find some older objects in more recent contexts. 

81

S. Salvatori Mirrors with handles can be divided into two main types: 1) with built-in cast tang; 2) with handle that is applied and usually has an elaborate shape. The former is probably widespread over the entire area of the Oxus Civilisation, though as yet we do not know any specimens of its coming from Margiana. Without attempting to make a complete list, we should mention that this type of mirror is present in a Namazga V grave at Altyn-depe (Grave 252: Kircho 2000: Fig. 2.I.1), and that many specimens, in two variants, with smooth-edged or wavy-edged handle, are known from the looted graves of southern Bactria (Pottier 1984: Fig. 38, No. 265, 266). Among these Bactrian specimens, a further variant consists in the presence both of circular and dropshaped disks. However, the fact that it is impossible to reconstruct the grave contexts in which these objects were found prevents us from determining whether these formal variations are connected with chronological differences.

are known from Namazga VI contexts in the piedmont or in Margiana. This type of object is frequently present among the materials of southern Bactria (Pottier 1984: Pl. XVI.104-108; Amiet 1977: Fig. 15.11-12), but is absent in the deposits of the Namazga VI excavated settlements and in the burial grounds of that region. Out of Bactria and Margiana, it appears in the deposit of material, all from the Namazga V12 period, of the Hotel Serena of Quetta in Pakistan (Jarrige and Hassan 1989: Fig. 2). Also worthy of attention is a series of applicators with flat, complex-cross-shaped head that is variously distributed over the area of the Oxus Civilisation. In southern Bactria we know a certain number of specimens, now scattered in museums and private collections (Pottier 1984: Fig. 23; 172-175). This type of object is also present at Altyn‑depe, in graves datable to the Namazga IV period (Masson 1988: Pl. XXXVIII.4, Grave 845; XXXVIII.5, Grave 828; other two specimens from Grave 845 are illustrated in Kircho 2001: Fig. 6: 4b-5). In Margiana, no occurrences of objects like these are known, but there is a type with flat lanceolated head in the Middle Bronze Age graveyard of Gonur 1 (Salvatori 1994b: Fig. 11.6). This type appears also at Altyn-depe, in Namazga IV tomb contexts (Grave 627: Kircho 2001: Fig. 6.13) and in Namazga V ones (Grave 721: Masson 1988: Pl. XL.6; Grave 252: Kircho 2000: Fig. 2.I.3). Out of the Oxus Civilisation, the type with cross-shaped head reappears at Mundigak IV.1 and IV.3 (Casal 1961: vol. I, 249; vol. II, Fig. 140.19) and at Shahr-i Sokhta, in tomb contexts from phase 5-4 (Graves 712 and 725 Inferior: Piperno and Salvatori, n.d.), therefore ascribable to the period 2600-2400 BC (Salvatori and Tosi 2005).

This type of mirror, generally speaking, reappears also in northern Bactria at Sapalli (Askarov 1977: Pl. XXXVII.1, 3) and at Dzharkutan 4c in a grave (G. 24) of the first phase, i.e. the Sapalli phase of the local sequence (Shirinov and Baratov 1997: Fig. 4.1). The specimen from Dzharkutan, which dates to an early period of the Late Bronze Age, attests to the continuity of the life of this type of mirror between the Middle and Late Bronze Age. Another type of mirror, with applied “anthropomorphic” handle, is known in a great number of specimens from southern Bactria (Pottier 1984: 39-40) and, at least in two specimens, from Sapalli, in southern Bactria (Askarov 1977: Pl. XXXVII.6, 8). Out of Bactria only one specimen from Mehi, in Baluchistan, is known (Stein 1931: Pl. XXXII). As Pottier has pointed out (1984: 40), the specimen from Mehi must be regarded as imported, so this type of mirror with “anthropomorphic” handle10 may now be regarded as a typically Bactrian product11. The specimens from Sapalli, however, suggest that this type of mirror may be ascribed to a transition stage between the Middle and Late Bronze Age. The type may represent a typically Bactrian variant, but the lack of documentation in Margiana, where only circular mirrors with raised edge are known for the Late Bronze Age, does not allow further speculations on the cultural meaning of this apparently specific feature.

Judging from its chronological position in datable contexts, it seems that this type of applicator for cosmetics cannot be dated to later than 2400 BC, so the Bactrian specimens may perhaps be one of the few elements that afford a clue on an Early Bronze Age settlement phase unidentified as yet in this area. The other type, on the contrary, seems to cover the entire span of the third millennium BC, i.e. the Namazga IV and V periods, considering its chronological position at Altyn-depe and Gonur 1. Chlorite Containers

The same conclusions are reached through the examination of the subrectangular bronze palettes documented in the Middle Bronze Age graveyard at Gonur (Salvatori 1993: Fig. 12, G.C. 2/10; 1994b: Fig. 11.7), in the contemporary one at Takhta Bazar (Udeumuradov 1993: Fig. 32.5) and at Altyn‑depe (Masson 1981: Pl. XV, 8; Kircho 2000: Fig. 1.4); no specimens of these palettes, on the contrary,

The chlorite containers coming from Middle Bronze Age levels or graves are decorated with incised or excised motifs, whose graphic complexity and manufacturing techniques are variable (Sarianidi 1990: Pl. XXVIII.1, 6, 10, 11; 1998a: Fig.17.9, 19.9; 2001: Pl. 9.4, 7). These types of production are comparable, in their form and decoration, to several Bactrian pieces from the antique market (Pottier 1984: Pl. �������������������������� XXVIII.225, 226, 231; Pl. XLII; Salvatori 1988: Figs. 84, 85). �������������������� Generally speaking,

This particular type of applied handle was the object of a formal analysis by P. Amiet (1977: 116), who compared it to the Egyptian Hathor’s-head-shaped mirrors. Pittman (1984: 39), on the other hand, believes that they rather resemble the trunk of a palm tree. 11 In Deposit II of Mehi, however, there are other Bactrian-type items datable to the Late Bronze Age (During-Caspers 1992). 10

For a discussion on the dating of this complex, see Salvatori 1995b.

12

82

Cultural Variability in the Bronze Age Oxus Civilisation

Figure 6.2 Chlorite cosmetic phiales typology.

Margiana

C1-2

Bactria

Susa

Shahdad

Jiroft

Yahya

A1-2

A1-2

A1-2

A1

B1-2

B1-2

B1 (?)

D

D

Shahr-i Sokhta

Hili Tomb A

Abraq

Tarut A2 (?)

B1 (?)

B2

C1-2 D

D

E F

F Table 6.1 Distribution of the main cosmetic phial types (cf. Figure 6.2).

83

S. Salvatori this class of items, because of some of its forms and its decorative system, finds close parallels in items coming from out of the area of the Oxus Civilisation, particularly (but not exclusively) from the Iranian plateau, for instance from Susa, Tepe Yahya and Shahdad, always in contexts from the second half of the third millennium BC, and can be traced back to Miroschedji’s (1983) série antique. This production was widespread ìn the Mesopotamian, Iranian and Harappan spheres, over a span of time that, in terms of Mesopotamian chronology, covers the era from the Early Dynastic III to the post-Sargonid Akkadian period.

from the excavations of Susa (Miroschedji 1983: Pl. VIII and Fig. 11). They belong to Type A, described above; Type B (concave undecorated neck and more or less everted rim: B1: undecorated body, B2 decorated body); and to Type D, which includes phials on architectural pedestal. Another much larger group was yielded by the excavations in the graveyard of Shahdad (Hakemi 1997). Here all three of the groups found at Susa (Types A, B and D) are present. At Tepe Yahya, one of the centres of chlorite mining and processing, there probably exists only one fragmentary specimen, belonging to Type A (Lamberg-Karlovsky 1970: Fig. 21.R), and another specimen, not in chlorite but in calcite, belonging to Type B1 (Potts 2001: Fig. 4.38). Both of these specimens come from Yahya IVB; the second one, more exactly, from IVB515. A specimen belonging to Type B2 is present in Grave A of Hili North (Vogt 1985) and can be dated to the third quarter of the 3rd millennium BC (Cleuziou 1989: 70).

Chlorite Phials In our present state of knowledge, the category of chlorite containers does not seem to be able to offer clues that allow us to detect clear signs of regionalisation within the area of the Oxus Civilisation. The most widespread typology in this area is that of the chlorite phial with quadrangular-section body, cylindrical neck and narrow, slightly projecting rim. In Margiana so far only a few specimens are known: three from the palace of Gonur North (Sarianidi 1998a: Fig. 19.1; 2002: Fig. on p. 126); other three from Togolok 21 (Sarianidi 1990: Pl. XXIX.2-4); a fragmentary one made of white stone, probably coming from the graveyard of Gonur (Sarianidi 2001: Pl. 25.3); and another one from the Site No. 392 Late Bronze Age site (Masimov et al. 1998: Fig. 4.1).

We should also point out the presence, in the wide range of chlorite vessels retrieved or excavated in the Island of Tarut, in the Persian/Arabian Gulf, of a fragment belonging perhaps to Type A2 (Zarins 1978: Pl. 68.143), and of another one belonging to Type F (Ibidem: Pl. 70.584). Type F is represented by other two chlorite specimens only in Bactria (Pottier 1984: Fig. 20.152, 153), and by a pottery one from Dashly 3 (Sarianidi 1976: Fig. 51 No. 24). The pattern outlined above clearly shows an absolute coincidence between the complexes of Susa, Shahdad and, probably, Tepe Yahya and Jiroft (Majidzadeh 2003). The diffusion of Type A also in Bactria is an element of contact between the Asian series and the Iranian ones (the absence of this type in Margiana may be fortuitous).

From a typological point of view, the specimens from Gonur and Togolok may be included in a single type, characterised by its cylindrical neck (Type C), usually low and undecorated (Figure 6.2). The specimen from Site No. 392, on the other hand, is unique, because it has a dice-eyes decoration on a probably truncated-cone-shaped neck, and an incised design of seated figures (Type E)13.

The almost total absence of this category of containers at Yahya, which was one of the chlorite processing centres, and in the entrepôts of Tarut and Failaka, may mean that these containers were produced by another or several other atelier, and that their commerce or distribution was specifically oriented to routes other than the Mesopotamian ones, to which the production of Yahya was apparently not directed either (Kohl 2001). All this indicates once more that we are far from being able to outline a satisfactory picture of these forms of production, which show very strong similarities with each other, both in their shapes and in their ideological and technological substratum.

In southern Bactria, the number of known specimens is higher (Pottier 1984: Figs. 18-21; Pittman 1984: Fig. 17), and they can be divided into three main groups. The first one (Type A) is characterised by an incised decoration on its neck, formed of parallel lines or of a combination of parallel lines with an excised raised zigzag line, or, sometimes, of incised motifs usually consisting of a zigzag line whose triangular spaces are filled with parallel lines14. The second group (Type C) has a cylindrical neck that is more (C1) or less (C2) high and has no decoration. The third group (Type F) is a container with a parallelepiped body, base with rounded, projecting band, and short, outlined neck (Table 6.1; Figure 6.2).

The specimen from Yahya, which, we repeat, is alabaster and not chlorite, has been discussed by Lamberg-Karlovsky (2001b: 273), whose description of it diverges in several details from the drawing supplied by Potts. If the drawing is correct, the object belongs to Type B, present, as shown in the summary table, only at Susa and Shahdad. However the material of which the phial is made may turn out to be chronologically revealing: two specimens similar in shape and material have been found at Shahr-i Sokhta, respectively in Rooms LVIII and LIX of the “House of Stairs”. They belong to Period II, Phase 5 of the local sequence, and, in terms of absolute chronology, date to approximately 2600-2500 BC (Salvatori and Tosi 2005). 15

No chlorite objects of this type are known in the piedmont of southern Turkmenistan, while they are widespread in Iran. It is known that a certain number of specimens come For a formal analysis of this object, see Masimov et al. 1998: 36-37). This type may be divided into two variants: A1, with high, narrow neck, and A2, with wider, lower neck. 13 14

84

Cultural Variability in the Bronze Age Oxus Civilisation As regards Central Asia, the low number of finds in Margiana perhaps is due to the fact that the only known Middle Bronze Age graveyard examined so far in the region had been looted in antiquity; but we cannot exclude other possibilities. The specimen published by Sarianidi as coming from the palace of Gonur North16 seems to confirm that this type may probably be dated to the second half of the third millennium BC. Here, too, we cannot positively exclude that this production persisted also during the Late Bronze Age, as the three samples from Togolok and that from Site No. 932 seem to indicate.

graveyard of Shahdad) and probably a bronze beaker with pelican-beak spout, a type that is widespread in Susiana and in the region of Gilvran (Calmeyer 1969: 14-16) during a period presumably between the Early Dynastic III and the Early Akkadian periods (Ibidem: 16)17. At Susa this type is present, according to Le Breton (1957: 119), in Susa Db-c: this confirms the dating indicated above. Finally, we must mention the presence of a small subcylindrical “alabaster” column. In this grave, too, on the whole the dating elements seem to suggest a period between 2400 and 2300 BC. In Grave 61, a phial is associated with a small alabaster bowl from Series I and a chlorite beaker from Series IVa (the reference, obviously, is to the shape only, and not to the material). All the pottery that is present is produced locally; among the bronze objects we should mention a flat-bottomed dish with flaring rim, fairly frequent in the graves of Shahdad. Though the dating elements are decidedly limited, the dating suggested by this grave is presumably clustered between 2400 and 2300 BC.

However, a few further remarks are necessary here, in the light of what we can obtain from the analysis of the Shahdad grave outfits. The graveyard of the large Iranian pre-protohistorical centre, which extended, during the third millennium BC, over an area of more than 100 hectares at the margin of the Lut Desert (Meder 1979; Salvatori 1978; Salvatori and Vidale 1982), is the only archaeological situation in which it is possible to check the context of the finding of this type of container. As we pointed out above, at Shahdad there are three of the five types into which this production can be roughly divided. Only one of these types, in its two variants (Types A1 and A2), is present with a great number of specimens also in southern Bactria. Before we evaluate the diversified distribution of this type of container and attempt to interpret its meaning, we should point out some facts that help us to classify it chronologically, by analysing its associations with the graves of the burial ground of Shahdad.

In Grave 74, a phial is present together with locallymanufactured pottery, but we must point out the presence, on a red ware jar, of the impression of a Bactrian-type compartmented seal. In Grave 116, two specimens of phials for cosmetics are associated with two chlorite boxes divided into four compartments. One of these boxes has a lid topped with a high-relief snake. The grave also contained a circular mirror with flat margin (see above), and, besides the usual locally-produced pottery, a little black-on-buff jar.

Here we find Type A, in its two variants, A1 and A2, in at least sixteen graves. In Grave 39 a chlorite beaker with a decorative system belonging to Kohl’s “intercultural” style or to Miroschedji’s “série ancienne” is associated with a small alabastrine calcite bowl of Series I of Casanova’s typology (1991), a black-on-red painted ceramic vessel, and a locally produced red ware jar that bears the impression of a bronze compartmented seal with star motif. The type of the alabaster bowl is widespread both in Mesopotamia (Ibidem: 32 and note 34) and in Susiana (Susa: “Vase à la Cachette”: Amiet 1966: Fig. 152.A), in contexts datable to the late pre-Sargonid period; this dating matches also with the chronological position of the objects manufactured in the “intercultural style” (Kohl 2001). On this basis, it would be reasonable to surmise a date around 2400-2300 BC.

In Grave 119, besides a second phial (Type B1, however), there is a little black-on-buff jar belonging to a series related to the production types of Susa IVa (Carter 1980) and Godin III.6 (Cuyler Young 1969; Cuyler Young and Levine 1974; Henrickson 1984). This grave also contained a bronze vessel comparable to specimens from Susa Dde (Le Breton 1957: Figs. 40.32-33; 41.20d). Here too the close relationship with objects from Susa IVa and Godin III.6 indicates a chronological span between 2400 and 2300 BC. In Graves 121 and 140, the phials are associated with black-on-buff and black-on-orange pottery18 and, in Grave 121, also with a chlorite bowl belonging to a type widespread at Yahya (Lamberg-Karlovsky 1970: Fig. 23, Pl. 24.B). In Grave 140 there is an anthropomorphic clay statue as well.

In Grave 57, the small alabaster bowl is associated with an alabastrine calcite beaker of Casanova’s Series IVb-c (1991: Figs. 3-4), a type of object that is widespread throughout the second half of the 3rd millennium BC (see further on). In the grave furniture, besides the locally-produced red pottery vessels, there are a garment pin with hourglassshaped head (a type that appears very frequently in the

Grave 166 is particularly interesting, because here the phial is associated with a Bactrian-type bronze compartmented This type of bronze vessel is present also in southern Bactria (Pottier 1984: 243, Fig. 33). 18 We must emphasise the convergence of these painted products with the items from tepe Yahya, rather than with the later Omanite and Emirate products (Méry 2000) 17

We must, however, remind the reader that we do not know where the object was actually collected or what its stratigraphical context was. 16

85

S. Salvatori seal and a cylinder seal belonging to a series that is well known because it appears at Yahya, Susa and in Central Asia (see below, Figure 6.3). The grave assemblage also contains a calcite bowl, probably with a S-shaped profile; a chlorite box divided into four compartments; a bronze pin with hourglass-shaped head; and a little clay “model of a house”. All the associated pottery is locally-produced red ware. The chronological position of this grave is more problematic than that of other graves, though is seems likely that the range defined by the associations may be from 2300 to 2100 BC.

III.6 and an alabaster beaker from Series IVa. A beaker from Series IVa is associated with a phial in Grave 310. The two specimens present in Grave 328 are associated with an alabaster vessel, unfortunately not shown in the figures, a millstone and a bronze dish with flaring rim. Finally, we do not obtain any help from the reappearance of this type of phial in Grave 347, because the other objects of the outfit have not been shown in the figures. Our report makes it fairly evident that the chronological position of this type of chlorite phial is between 2400 and 2300 BC, with a possible extension approximately up to 2100 BC.

No help is given by Graves 179 and 201. In the former, two specimens of chlorite phials are associated only with locally-produced pottery. In the latter, the phial is associated with local pottery, bronze objects that have not been shown in the figures, and a little subcylindrical “alabaster” column.

A similar chronological indication is given by an analysis of the associations of Type B, which is present only at Susa and Shahdad (while calcite phials appear, in contexts of the middle of the 3rd millennium BC, at Shahr-i Sokhta and Yahya IVB5): subtype B1 has a range from 2400 to 2300, while subtype B2 probably covers a more extensive range, between 2400 and 2100 BC.

Grave 212 is more interesting from our point of view: here the association between the small alabaster bowl from Series I and a calcite beaker from Series IVa reappears. This grave also contained a bronze dish with flaring rim, a garment pin with hourglass-shaped head, a bronze vessel, and a Bactrian-type bronze compartmented seal. The possible dating of this grave is between 2400 and 2300 BC.

The chronological evidence supplied by the graves of Shahdad makes it possible to date the Bactrian specimens as well to the second half of the 3rd millennium BC. In our opinion, this perspective should be evaluated very carefully. The distribution of Type A, which is a link between the Central Asian series and the Iranian ones, seems to suggest

Proceeding with our list, in Grave 305 the phial is associated with a pot belonging to type Susa IVa - Godin

Figure 6.3 Trade routes as defined by chlorite phiales types polarization.

86

Cultural Variability in the Bronze Age Oxus Civilisation Alabaster Vessels

that there was a direct, privileged route between the Iranian geo‑political sector and at least one of the subregions of the Oxus Civilisation19. The greater typological complexity found in the chlorite phials of the Iranian area leads us to suppose that they were probably produced in Iran, though it is quite likely that the production of this type of object was polycentric, as demonstrated by Kohl for the chlorite products both of the “intercultural style” and of the “série récente”. It seems reasonable to surmise that there was at least another production centre, that it produced Type C, which at present seems to be widespread only in Margiana and Bactria; and that it was hypothetically active between the end of the 3rd millennium BC and the beginning of the 2nd, if we assume that the Type C1 objects found at Togolok were not decontextualised. If, on the contrary, we consider this possibility true, the only object of the series that has been dated on the basis of the context where it was found would be that of Gonur 1 North, so this type too would come under the limits of the second half of the 3rd millennium BC. Following up this hypothesis, we would conclude that all the chlorite phials from Margiana and Bactria, as well as the Iranian ones, must be dated only to the Middle Bronze Age. This hypothesis is independently corroborated by the absence of this type of object in the dwelling and graveyard contexts of the Late Bronze Age unearthed by regular excavations both in southern and northern Bactria.

Alabastrine calcite vessels are widespread over the whole Middle Eastern area, throughout the third millennium BC. The typologies are quite diversified, and some of them appear to characterise the 1st half of the millennium: this is true, for instance, of the truncated-cone-shaped bowls of Shahr-i Sokhta (Ciarla 1979; 1981). Here we will examine only the types coming from contexts that can be dated to the 2nd half of the millennium. The excavations carried out under my leadership in the graveyard of Gonur between 1992 and 1994 yielded only two specimens: a twin vase of exquisite workmanship (Salvatori 1993: Fig. 9)21 and a subcylindrical one with horizontal rim (Idem 1995a: Fig. 10, G.91/6). Other specimens of alabastrine calcite vessels were found in the same graveyard by V. Sarianidi in the subsequent years; among these we should mention a great number of specimens with subcylindrical cup and tall, truncated-cone-shaped foot. The containers from Gonur have a very interesting technical feature: an umbilicate internal bottom22, which implies the use of a distinctive drilling technique. This feature, as far as I know, is present, within Central Asia, only on a limited number of specimens. In Margiana it reappears on the tall-footed subcylindrical cups found in the graveyard of Takhta Bazar (Udeumuradov 1993: Fig. 34.3); in Bactria on a certain number of subcylindrical bowls and tall-footed subcylindrical cups (Pottier 1984: Figs. 26.197, 201, 204; 27.208; 28.214; 29.216, 218, 220); in the piedmont of southern Turkmenistan on a subcylindrical bowl from Grave 843-845 (Kircho 1988: Fig. 4.7), on a specimen from Grave 636 of Altyn-depe and on another one from Grave 677, both of them datable to the end of the Namazga IV period23. The distinctive umbilicate internal bottom, on the other hand, presumably does not appear on specimens of a similar type found in

Though we do not believe that these chlorite phials for cosmetics were produced in Central Asian centres, their typological distribution (Figure 6.3) seems to show clearly that some basically trading mechanisms were active, that they worked along preferential lines, and that this helped to intensify regional diversification within culturally homogeneous areas. Even if, as we have endeavoured to illustrate, during the Middle Bronze Age the area involved in the Oxus Civilisation was characterised by basic bonds strong enough to induce us to define it (highlighting its similarities) as a consistent, homogeneous cultural system, there emerge equally clear regional differences that affect both the symbolic system (anthropomorphic figurines, figurative motifs on bronze compartmented seals) and the economical system (differences in the typology of circulating goods)20.

At least another twin vase was found by Sarianidi in the Middle Bronze Age graveyard of Gonur North; it is similar to the one found in Grave 252 at Altyn-depe (Kircho 2000: Fig. 2.I.8). We must point out that in this Namazga V grave there were some stick-shaped bone dice (Ibidem: Fig. 2.I.5) identical to those found by Sarianidi (1998a: Fig. 21.16-17) in the palace of Gonur; a bronze phial with an applicator with flat, lanceolated head; a laminar bronze palette; and a bronze mirror with flat tang. Other two specimens of alabastrine calcite twin vases are known in southern Bactria (Pottier 1984: Fig. 29.219-220); one of them certainly had an umbilicate bottom. 22 This technical and morphological detail was noticed by T. F. Potts (1993: 129) in the Mesopotamian and Bactrian specimens, without making any further remarks on them. 23 In other containers from Altyn-depe, similar in their shape and material and all of them datable to the Namazga IV period, the internal surface of the bottom is not umbilicate. This particular feature, on the other hand, does not appear in any calcite container present in graves ascribable to the Namazga V period. A specimen from Ulug-depe (Masson and Sarianidi 1972: Pl. 35, bottom) is known only in photograph, so we do not know whether its internal bottom surface is umbilicate. 21

We must also point out the presence, in southern Bactria, of chlorite boxes divided into four compartments, with sides decorated with incisions (Pottier 1984: Fig. 31.226; cf. Shahdad: Hakemi 1997: Graves 35, 116, 120, 165, 166), and of chlorite dishes with central subcylindrical receptacle (Pottier 1984: Fig. 30.222; cf. Shahdad: Hakemi 1997: Grave 193). 20 This last aspect is worthy of being emphasised by quoting the hypothesis proposed by Raffaele Biscione (1985), who postulated a relationship between the series of the so-called Nestorian seals of the Ordos region and the Bactrian compartmented seals. The possible relationship between Bactria and Ordos is – now it seems clear – nothing else than the indication of another segment of an extremely vast system of exchanges, probably arranged in a chain of tangent or partly overlapping rings. 19

87

S. Salvatori contemporary Middle Eastern contexts24; it is present, however, at Shahdad, at least on a tall-footed goblet from Grave 139 (Hakemi 1997: 297, No. 1438), and on most of the specimens of alabastrine calcite from Susa, including a cylindrical specimen with flaring rim from the “Vase à la Cachette” deposit, datable approximately to 2400 BC (Amiet 1986: 125, Fig. 96.7; Casanova 1991: 33 note 37). It is interesting, perhaps, to point out that at Susa the umbilicate internal bottom reappears also on some types of vessels that can be compared only with Mesopotamian specimens and never appear in contexts further east than Susa.

During the first half of the 3rd millennium BC, one of the manufacturing places was undoubtedly the Iranian Sistan (Shahr-i Sokhta: Ciarla: 1979, 1981, 1985; Tepe Graziani: Tosi 1989: 24); but here, though some types reappearing in Margiana, Bactria, the piedmont of southern Turkmenistan, Baluchistan, Shahdad, Susa and Mesopotamia are present, they do not have the technically important feature we have been examining in this essay. Other production centres are not known so far or cannot be dated to the first or second half of the 3rd millennium BC; but this is undoubtedly due to a shortcoming of research27. At any rate, the argumentation carried on up to now, though it has not solved the problem of the identification of the centre or centres of production of vessels with umbilicate bottom (which reveals the use of a drill head whose lower face has a concave centre) decidedly places this production within a time span fully consistent with the dating relevant to the Namazga V period in Turkmenistan (2400-2200/2100 BC). This supplies a further chronological reference for some of the decontextualised items of southern Bactria28.

As far as I know, the only other vessels that show this technical and formal feature are a number of specimens found in Mesopotamia that fortunately bear Rimush and Naram-Sin inscriptions (Potts 1986; Potts 1989): this qualifies them as part of the loot brought back from the campaigns carried out by these Akkadian sovereigns against Elam and Barahshi/Marhashi25 in the case of Rimush, and against Magan26 in the case of Naram-Sin. Obviously the fact that these specimens date, at the latest, to the years of the campaigns of these two sovereigns is certain and fixed by these inscriptions, and so is their origin from regions to the east of Mesopotamia.

Small Stone Columns and Rods A further link between the Oxus Civilisation, the extra-Elamite Iranian plateau, the area of the “Elamite confederation”, southern Afganistan and Baluchistan consists in the diffusion of two types of objects such as small stone columns and rods.

Finally, the umbilicate internal bottom seems to be present on a specimen from Ur datable to the Protodynastic III period (Heinz 1989: Fig. 11); this dating matches with the one estimated for the final Namazga IV specimen from Altyndepe. It seems clear to us that the production of alabastrine calcite or alabaster vessels during the 3rd millennium BC throughout the Middle East was polycentric, like the chlorite or steatite production (Kohl 1975; 1979; 2001; Lamberg-Karlovsky 2001b), and the feature highlighted by us here demonstrates the presence of several artisan traditions connected with this production.

The stone columns may presumably be one of the distinctive elements of a phenomenon of connection between the political or trading élité of the various urban centres present in that immense territory. We believe that it is expedient, here, to move on to analyse in detail not only the contexts where these objects were found, but also the different weight that these contexts must take on in the definition of their chronological framework and sphere of meaning.

As regards the series of alabastrine calcite vases from the deposit of Quetta, at present it is not possible to verify whether the umbilicate internal bottom is present or absent, because we do not have any drawings showing their section (Jarrige 1987: Fig. 3; Jarrige and Hassan 1989: Fig. 2). This problem exists also for the two alabaster vases of the temple of Barbar IIa, in the island of Bahrein (Glob 1958: Figs. 2, 4) and for the other whole or fragmentary specimens from the Arabian Peninsula (Potts 1993: 133 note 41). The few specimens of subcylindrical bowls from Shahr-i Sokhta come chiefly from surface contexts and do not have an umbilicate internal bottom (Tosi 1968: Figs. 19-20; 1969: Fig. 40; 1983: Fig. 17; Salvatori and Vidale 1997: Fig. 248); neither do the fragments collected by Fairservis (1961: Figs. 29: i-k; 37: 45-6) in Site No. 109 of Gardan Reg. 25 This region may presumably be identified as Kerman, where Shahdad is placed, according to Steinkeller (1982) or as Makran, according to Vallat (1985: 52). 26 The location of Magan/Makkan on the opposite shores of the Persian/ Arabian Gulf (Makran on the Iranian shore and Oman on the Arabian one) is by now a widely accepted opinion. The relevant literature is extremely vast, so we will only mention D. T. Potts (1986), because, in his work, besides concisely summarising the problem of identification, he also comprehensively deals with the diffusion of alabastrine calcite vessels. In any case, a text-oriented discussion can be found in Heimpel 1988. 24

Small Columns Finds in primary contexts This group includes the specimens found in the Middle Bronze Age graveyard of Gonur 1 in Margiana29, and those For an examination of the problem of the distribution of possible raw material sources, see Casanova 1991: 53-61. 28 The lack of data about the ateliers where these alabastrine calcite, aragonite or alabaster objects were manufactured is a strong limitation to the attempt to acquire a realistic view of the routes for the exchange or purchase of these items (Potts 1989). An overall picture has been tentatively outlined by T. F. Potts (1993), but we are still quite far from grasping the complex system of the interchange through the Iranian plateau. 29 Some of these little columns were shown to me by V. Sarianidi during a visit to Gonur in the autumn of 1998. 27

88

Cultural Variability in the Bronze Age Oxus Civilisation coming from the graveyard of Shahdad, which are twentythree specimens30. The associations of the Shahdad graves fit with those of the graveyard of Gonur 1 in indicating a dating between 2400 and 2100 BC.

building remains (Ibidem: 216-219, Fig. 132). Here the stratigraphical context was dated to the Hissar IIIC period, but since it was formed of “hoards”, it is possible to surmise that the material belonged to an earlier period, probably IIIB of the Hissar sequence.

Other four specimens must be added to this group: three of them come from Excavation 9 of Altyn-depe and can be dated, because of their stratigraphical position and cultural context, to the Namazga V period (Masson 1988: Pl. XXXV.1-3); the fourth one comes from Grave 362 and can be dated to the Middle Bronze Age (Alekshin and Kircho 2005: Pl. 70).

Another find we must report is that of four small columns in the deposit or “cenotaph” of Quetta31 (Jarrige and Hassan 1989: 153, Figs. 4-5). The context where they were found and the presence, among other things, of gold leaf scraps (Jarrige 1987: Fig. 2) seems to suggest a hoard. The pottery present in the deposit belongs to the BactriaMargianan type, but from a typological point of view it can be dated to the Late Bronze Age. The rest of the material is evidently a sort of hoard, and this explains the presence of earlier items such as the tall-footed subcylindrical bowls and the body (only the body) of a Bactrian-type composite figurine.

There is uncertainty about the position to be given in this list to the specimen from Excavation 7 of Room 7 at Altyn-depe (Masson 1988: 65-68, Figs. 21-22). It may be a primary context, but it also may be a situation similar to that of the Treasure Hill deposit of Hissar described further on. Finally, the presence of a small column was reported at Kara-depe in a possibly Namazga IV context (Hiebert 1994a: 154).

Archaeological finds from contexts that cannot be reconstructed

A great number of small columns were found at Tureng Tepe, “..tant en place au sommet de la terrasse que dans la couche d’écroulement du monument…” (Deshayes ���������������� 1977: 99, Fig. 4), and dated to the beginning of Period IIIC1 of the Iranian site.

Susa: two specimens (Amiet 1986: Figs. 97.4, 101, 102). Piedmont of southern Turkmenistan: little column found at Tekkem-depe, whose discovery context has not been published yet (Kohl 1984: 141, Pl. 20a).

About possible finds of little columns in graves excavated at Dashly 3 we only have an uncertain report (Sarianidi 1977: 101; Deshayes 1977: 101 note 6).

Surface finds in archaeological contexts Iranian Sistan: communication).

Finds in secondary contexts

Shahr-i

Sokhta

(Tosi,

personal

This group includes the specimens from Togolok 21. About twenty-nine whole or fragmentary small columns were unearthed by V. Sarianidi (1990) in the central building of Togolok 21, and some of them were undoubtedly being reduced in order to manufacture other types of objects (Hiebert 1994a: 154). In other words, in this context the little columns are nothing other than raw material for a secondary production.

Tepe Abdali: two specimens collected on the surface (Tosi, personal communication).

A little stone column comes from Togolok 24, where it was apparently found in Grave 30 (Sarianidi 1990: Pl. LXXXIX.19); about this grave only a concise report (Ibidem: Pl. LV) and only one of the vessels of the outfit (Ibidem: Pl. LVII.16) have been published. Presumably this is the only Late Bronze Age grave in which a small column has been found.

Southern Afganistan: Chah Ismail (Balsan 1972: Fig. on p. 207).

Surface finds in non-archaeological contexts Southern Afganistan: Godar-i Shah, Shela Rud (Besenval and Francfort 1994: Fig. 1.9);

Objects retrieved in the antiquarian market Southern Bactria (Pottier 1984; Sarianidi 1977; 1993a: Fig. 6; Amiet 1986: Figs. 157-158).

A very different situation is that of the discovery of at least six specimens at Tepe Hissar in Hoard I of Treasure Hill (Schmidt 1937: Figs. 96-97), at least one in Hoard II (Ibidem: Figs. 98-99), and other three in the CF 37

The marble disk found in the deposit is undoubtedly connected with similar objects present in grave contexts at Shahdad (Grave 20: Hakemi 1997); with the disks found by Schmidt in Hoards I and II, in CF 37 and also in a room ascribed to Period IIIB at Tepe Hissar (1937); and with the disk present among the decontextualised materials of Shela Rud, in southern Afganistan (Besenval and Francfort 1994: Fig. 1.9). 31

Graves 57, 60, 75, 80 (two specimens), 104, 134, 139, 176, 188a, 192, 201, 206, 211, 214, 277 (two specimens), 280, 286, 294 and 371. 30

89

S. Salvatori This list shows that the discovery contexts of the stone columns with central groove on the upper face, and sometimes also on the lower one, are rather diversified; but before we comment this class of materials it will be convenient to perform a similar classification of the stone rods that appear in association with the small columns in some primary contexts.

MARGIANA Taip Gonur Togolok Takhirbai 3

BACTRIA Sapalli Dzharkutan Kuzali Molali Bustan

Table 6.2 Suggested connection between Margiana and Bactria Late and Final Bronze phases.

Subcylindrical Stone Rods

were destined to be recycled as raw materials seems to be demonstrated by the finds of Togolok21.

Finds in primary contexts Middle Bronze Age graveyard of Gonur 1 (in association with small columns: e.g. Grave 1500: Sarianidi 2001: 198).

Late Bronze Age

Graveyards of Shahdad (in some cases in association with small columns)32.

Before we concisely review the same area during the Late Bronze Age, we must first point out that during this period the population appeared to decrease drastically in the piedmont of southern Turkmenistan. Altyn-depe disappeared, and Namazga-depe was reduced to a small village of about 2 hectares. The area of other settlements of this period, e.g. Tekkem-depe, El’ken, Grisha and, further east, Ulug-depe, was less than 2 hectares (Kohl 1984: 137).

Altyn-depe: Grave 362, datable to the Middle Bronze Age and associated with a stone column (Alekshin and Kircho 2005: Pl. 70). Finds in secondary contexts Tepe Hissar: at least one pair of long cylindrical stone rods comes from Hoard I of Treasure Hill (Schmidt 1937: Figs. 96-97), together with samll columns and stone disks, partly reproducing the series of associations of the “cenotaph” of Quetta.

In contrast with this noticeable depopulation of the southern piedmont, there was a sharp increase in the population of Margiana and Bactria (both northern and southern). As far as we can observe in Margiana, most of the cities and villages of the Middle Bronze Age were abandoned (for instance Adzhi Kui 8 and 9, Gonur 1 North and many other small farming settlements of the region). Other cities and villages were still inhabited in the subsequent period, but we do not know whether this situation arose after a period of abandonment (Togolok 1, for instance, is a site that presents an apparently continuous sequence between the Middle and Late Bronze Age, but it is unknown whether there were interruptions between the two periods). Many new settlements were established both in the areas occupied by settlements in the previous period (areas of Gonur, Togolok, Kelleli and Egri Bogaz) and in areas that so far have not yielded archaeological evidence datable to the Middle Bronze Age (area of Auchin).

Quetta: a great number of fragmentary specimens (Jarrige 1987: Fig. 4; Jarrige and Hassan 1989: Fig. 4) were found, among other things, together with four small columns, a marble disk and a stone “sceptre” similar to a specimen from Room 7, Excavation 7 of Altyn-depe (Masson 1988: Pl. XXXV.5). Objects retrieved in the antiquarian market Southern Bactria (Pottier 1984: Pl. 5). To sum up, we can say that these objects, when they are in primary contexts, can easily be dated to the second half of the 3rd millennium BC, and are often associated with each other. The frequency with which they are found in grave contexts seems to suggest that these objects were not distinctive elements of a political élité, but belonged to a category of people associated with a commercial or mercantile circuit. The intrinsic value of these objects, in any case, was high, even at the time of their decontextualisation, as shown by the care with which they had been recovered and preserved in hoard deposits, together with whole and fragmentary golden objects, as in Hoard I of the Treasure Hill of Hissar. The fact that many of these little columns

At Gonur, some time after the abandonment of the huge site, a smaller fortified settlement, Gonur 1 South, was established and its inhabitants used the northern depe as a burial ground. In Bactria many fortified sites were established both north of the Amudar’ya River (Sapalli, Dzharkutan, Kuzali, Molali, Bustan, to name only a few) and south of it, where the peopling datable to the Late Bronze Age was quite substantial (Sarianidi 1977; Kohl 1984)33. The large centres of Dashly 3 and 1, excavated in the nineteenseventies by V. Sarianidi, are only a few of the numerous Late Bronze Age sites identified in northern Afganistan. 33

Graves 75 (with a small column), 142, 193, 206 (with a small column), 211 (with a small column), 267 and 274. 32

90

Cultural Variability in the Bronze Age Oxus Civilisation Architecture

settlement of northern Bactria, there is practically no documentation about materials ascribable to the built-up area. On the contrary, there is an extensive documentation about the numerous graves from the burial ground that was established after the site had been abandoned. The fact that the graves are not coeval with the architectural structure is evident when one carefully examines the survey data of the burial ground and the excavation photographs: some of them are unpublished, but are exhibited in the corridor of the Institute of Archaeology of the Academy of Sciences of Samarkand.

Apart from Sapalli and the “palace” of Dashly 3, which had an absolutely distinctive layout (Askarov 1981: Fig. 2; Sarianidi 1977: Fig. 15; for an older chronology see above), the other centres developed a defence system characterised by an outer wall with circular towers at its corners and semicircular towers along its sides (Margiana: Auchin 1, Togolok 21, Togolok 1, Gonur 1 South etc.; southern Bactria: Dashly 1, Tikar 4)34. Once again, we must regret a severe lack of stratigraphical documentation for the Late Bronze Age settlements excavated by Sarianidi; because of this lack, we are not able to establish a series of the above-mentioned architectural system variants or of the sequential variants for each of the sites mentioned, as we would have wished to do. It is certain, however, that during the Late Bronze Age there developed throughout the area of Bactria and Margiana a consistent system of defensive architecture, characterised, as we have already stated, by fortifications with circular towers at the corners.

The same phenomenon was observed at Gonur, where the largest depe, the Middle Bronze Age one, was used as a burial ground by the inhabitants of the Late Bronze Age settlement of Gonur 1 South. In this case, too, the graves, from a later period (and still unpublished) often cut into the wall remains of the earlier period. We cannot but point out that situations of this sort reappeared frequently in the sites of Bactria and Margiana, and that the way the data from these sites were published, even in a recent past, has given rise to “naive” cultural interpretations. In relation to these facts, some people have mentioned intra muros graveyards (Francfort 1989: 343), and now we know how incorrect this is.

Pottery The pottery coming from the excavations of many of these settlements has been studied by P’yankova (1989, 1993, 1994). She has been able, chiefly thanks to the better-known sequences of northern Bactria, to establish a connection between the pottery production of Margiana and that of Bactria, proposing a correlation system with several sequential phases (Table 6.2).

Apart from these (at present largely insurmountable) difficulties presented by the archaeology of the Oxus Civilisation, it is possible to attempt to reassemble some of the pieces of this mosaic, at least for some production sectors of the craft system, starting from the area that from some points of view is the best-known, i.e. Margiana.

The system proposed by P’yankova could not include the pottery from the important site of Taip, in Margiana, which was excavated by I. S. Masimov and unpublished as yet35. We must also point out that P’yankova was inevitably compelled to define the cultural phases sometimes exclusively on the basis of graveyard pottery assemblages, and sometimes on the basis of assemblages coming from a settlement. In the latter case there are several problems due to the method with which the materials were collected during the excavations and their correlation with specific architectural phases.

The disappearance of the large settlement of Gonur North at the end of the Middle Bronze Age seemed to give rise to a general change in the settlement system, which broke up into a series of territorial segments gravitating around some centres whose size was limited when compared with that of the central site of the earlier system, and in any case was between 10 and 14 hectares. As I have already stressed elsewhere (Salvatori 1998b), this apparent crisis of the settlement system of Margiana seemed to be due more to political causes than to demographic ones. As a matter of fact, if we consider the total settled area, the Late Bronze Age greatly exceeded the previous period. The same conclusions were apparently reached also by Lamberg-Karlovsky, who, in an essay of 1994, connected the layout of Late Bronze Age Margiana settlements with the later castle systems widespread in Central Asia: the qala of Khorezm, up to the ethnographic records from present-day Afganistan. The American scholar brilliantly connected this architectural type with the organisational system of the qanat (Lamberg-Karlovsky 1994: 400). Lamberg-Karlovsky’s hypothesis is fascinating, and implies a strong independence of the individual fortified centres and probably a fragmentation of the political power or a decrease in its integration.

A single example, among the many possible ones, can help us to explain our personal frustration vis-à-vis the problem of data reliability. At Sapalli-depe, as we have pointed out above, there was a quite distinctive architecture with a quadrangular layout characterised by T-shaped rooms. In the publications of the excavation of this important a: 51-52. Cf. Francfort 1989: 340-343. The currently available information about the fortified settlement of Taip 1 North refers almost exclusively to the discovery of an important series of cylinder-stamp seals. As far as we have been able to see by repeatedly visiting the site, also together with the archaeologist who led the excavation, the defensive structure of the site followed the typical model of Middle Bronze Age settlements: the boundary wall has quadrangular towers at its corners, and not circular ones as in all the Late Bronze Age sites of Margiana. 34 35

91

S. Salvatori A model like this may possibly be a correct description of Margiana during the Late Bronze Age, when, after the disappearance of the central site, i.e. of the regional-level centre (probably in terms of classical chiefdom) of Gonur 1 North, a polycentric system emerged: at its vertexes there were the second-level sites of the previous system (or what replaced them) and in some cases new sites that occupied still-available portions of agricultural land. The situation of Bactria is less clear: here we are still unacquainted with the Middle Bronze Age sites; it is possible, however, that many of them were under the Late Bronze Age levels, as in the case, presumably, of Dashly 3 (see above). In northern Bactria the situation appears to be similar to that of Margiana, but here too we cannot make a comparison with the Middle Bronze Age sites.

Arabian Gulf, where possible contacts during this period are confirmed by D. T. Potts’s recent finds at Tell Abraq (Potts 1993). This picture agrees with the great development of the sea route of the Gulf at the beginning of the second millennium BC (Zarins 1989: 81-83). Bronze products dating to the Late Bronze Age, such as garment pins with compartmented head, are homogeneously scattered over the entire area of Bactria and Margiana, as well as in the piedmont of southern Turkmenistan (Masimov et al. 1998: 36). In this perspective, the overall picture indicates a stronger convergence of the material production and symbolic system (Francfort 1994) than in the more centralised system of the Middle Bronze Age. This strong cultural homogeneity that characterises the cultural development of the Late Bronze Age in all its phases, however, seems, at least apparently, to contradict other pieces of evidence having a strongly social characterisation.

Considering the fragmentation of political power, it would be reasonable to expect a lower convergence of material production. The actual situation, on the contrary, was quite different. As we have seen above, P’yankova was able to detect very strong correlations between the pottery assemblages of the two regions, both synchronically and diachronically, i.e. over the entire span of the Late Bronze Age until its final phase, which in Margiana has been called Takhirbai 3 phase by Masson and in Bactria has been called Bustan phase (Avanesova 1995).

A paradigmatic case that is worthy of being stressed is that of Dzharkutan, in southern Uzbekistan, where at least three contiguous graveyards (Askarov 1977: Fig. 21) coexist, all three apparently in use throughout the Late Bronze Age. A situation of this type presumably indicates the coexistence, in that area, of at least three lineage segments, each of which has been using its own burial ground. In this case, the difference is apparently expressed not by the specific peculiarity of material culture elements, but, much less visibly, by the fact of belonging to ideologically distinct groups. Social discontinuity, or rather its expression, does not seem to impinge on material culture, but only on the dialectics within local groups, which is difficult to detect in the archaeological record.

Glyptic Glyptic production, too, underwent a univocal change in all the area of Bactria and Margiana. It seems possible to detect a transition stage that can be called Taip phase in Margiana and Sapalli phase in northern Bactria: it is characterised by the inclusion of cylinder-stamp seals, bronze non-compartmented stamp seals and garment pins with compartmented head, and above all by the appearance of stone stamp-seals (very often in chlorite or serpentinite), amulet seals, as V. Sarianidi has called them (1981), with double face and rectangular shape, and pseudoseals that were often circular (cylinder section), had a biface decoration and in some grave contexts appeared to have been used as necklace beads (Sapalli: Askarov 1977: Pl. XLIII.1). Out of Bactria, seals of this period are rather rare. Some whole or fragmented specimens have been found in the largest centres of the Indus Civilisation (During-Caspers 1994b), at Sibri (Santoni 1988: Fig. 2.25) in Baluchistan, and at Bahrein (Crawford and Al Sindi 1995), in the Persian/Arabian Gulf. The cylinder stamp‑seal too seemed to shift towards the Indus Valley (Sibri: Santoni 1984: Fig. 8.4A; Jarrige 1985: Fig. 6), where it was assimilated or somehow integrated into the local tradition (Knox 1994).

What is evident from the archaeological point of view, however, is a process of strong contraction, throughout the area of the Oxus Civilisation, of categories of goods recognisable as elements of an “international” exchange circuit, in marked contrast with the situation that can be documented for the Middle Bronze Age. All this, obviously, may perhaps be fortuitous36, and due to a limit of the archaeological research. To this day, however, the available documentation suggests a reduction in the volume of Central Asian trade activities during the Late Bronze Age, and its contemporary re-orientation excluding the Iranian plateau. The only Bactrian-Margianan stone seals of this period out of the area of origin have been found, as we have already mentioned, at Mohenjo‑daro As a matter of fact, we must point out that it was barely possible to imagine the existence of this Middle Asian Bronze Age world before the beginning of the nineteen-seventies (see the admirable report/memorial in Lamberg-Karlovsky 2001a), before the arrival, on the antiquarian market, of the mass of materials from the looted graveyards of northern Afganistan; before Sarianidi’s excavations in this region and shortly afterwards in Margiana; before the Soviet scholars’ excavations in southern Uzbekistan and Tadzhikistan; and also before M. Tosi’s excavations at Shahr-i Sokhta, C. C. Lamberg-Karlovsky’s at Tepe Yahya and A. Hakemi’s at Shahdad. 36

To sum up, seals tell us that during the first centuries of the second millennium BC the exchange circuit was not interrupted, but was certainly directed towards the Indus Valley, and from there along the sea route of the Persian/ 92

Cultural Variability in the Bronze Age Oxus Civilisation and Harappa, in the Indus Valley, and at Bahrein, in the Persian/Arabian Gulf. A wedge of penetration of BactrianMargianan groups can be detected in southern Afganistan (Nad-i Ali: Besenval and Francfort 1994) and perhaps in Iranian Sistan (Shahr-i Sokhta phase 0), and above all to the east, precisely towards the Indus Valley, at Mehrgarh VIII, Sibri and Quetta, in Baluchistan.

This strong political and cultural integration of the system implies a great capability to mobilise new organisational efforts in the spheres of primary production (presence of artificial canals near Gonur and elsewhere in Margiana), of the production of luxury goods (bronze, gold and silver metallurgy; lapicide craft) and of mercantile trade. This situation led to the activation of interrelation flows through which the Oxus Civilisation centres were incorporated: an “international” network that included the Indus Valley, the Iranian plateau, the alluvial plain of Khuzistan and perhaps Sumerian-Akkadian Mesopotamia38. Merchants from these Central Asian proto-state centres were almost certainly present at Shahdad: they were firmly established in the heart of this large urban centre, towards which there converged Elamite people from Susiana and the Zagros (Susa IVa and Godin III.6 pottery type), and probably also people coming from the southern coast of the Persian/ Arabian Gulf.

Conclusions The currently available evidence, though still distributed in a patchy fashion, makes it possible, as we have seen, to distinguish, within the Oxus Civilisation, at least three cultural phases, each of them characterised by typologically and stylistically specific material assemblages, and above all by distinctive organisational, socio-economic and political lines of development. Disregarding, here, the problems relative to the Early Bronze Age about which we have evidence only in the westernmost part of the system, i.e. along the piedmont of southern Turkmenistan (Kohl 1984), we can state that during the Middle Bronze Age it seems certain that there developed some proto-state forms of political organisation that contained incipient elements of political and administrative control centralised on a regional level. The characteristics of the territorial organisation observable in Margiana during this period37 – with its centre in the large site of Gonur 1 North, and secondary-magnitude, mostly fortified, centres arranged around it with a modularity that complies with the principles of Cristaller’s theory – clearly reveal the type of trend under way in the region. The distribution and fortified configuration of the minor Middle Bronze Age sites allow us to catch a glimpse of an organisation based on chiefdoms coordinated by a major centre where the dominant group resided.

Undoubtedly, as indicated by the presence of typical “Iranian” seals in Margiana and Bactria, merchants from Central Iran were present in the Oxus cities and made the bidirectional trading flow even more steady and intense. A similar phenomenon linked the Oxus centres to the cities of the Indus Valley. Bronze compartmented seals of the Bactria-Margianan type are present at Mohenjo‑daro, while seals and other ivory objects typical of the Harappan civilisation are present in Middle Bronze contexts at Altyndepe and Gonur. In addition, the relation with the Indian area is indirectly confirmed by a great number of images of monkeys on the compartmented seals of southern Bactria. We should not overlook a third route on which the BactrianMargianan trade flow may have hinged during the Middle Bronze Age: the one that heads for China, through the (as yet weak) relationship of which we can catch a glimpse thanks to the formal and technical connections between the bronze seals production of this region and the group of seals from the Ordos.

The emergence of a territorial pattern, controlled by a central site characterised by a large palace-like building that occupied a large part of the Gonur 1 North area, is clearly a symptom of a process that tended to the establishment of a proto-state configuration. To confirm the type of process that was under way, we can mention some evidence, for instance the presence, at Gonur 1 North, of clay tokens (Sarianidi 1998a: 41), which are commonly interpreted as embryonic elements of administrativeaccounting systems (Schmandt-Besserat 1992), probably the appearance of a first embryo of writing (Klotchkov, personal communication), and the presence of a corpus of signs that may suggest a writing or proto-writing system, in the piedmont of southern Turkmenistan, at Altyn-depe, Khapuz-depe and Taichanak-depe (Masson and Sarianidi 1972: Fig. 31).

As we saw when we dealt analytically with regional variability, during the Middle Bronze Age the Oxus Civilisation was still a variegated world with more or less definite regional differentiation within the sphere of craft and artistic production, and probably the same commercial openness towards Elamite Iran; this world, as we have seen, involved all the regions that we have grouped under the name Oxus, and developed with a some autonomy in its compartments, probably in an unhomogeneous manner, so as to give rise to selective, discrete flows of material39. The collapse of the system at the end of the Middle Bronze Age, i.e. around or shortly after 2000 BC, appears to be The cylinder seal described in this book undoubtedly is indicative of a direct contact with Mesopotamian prototypes, though we cannot be sure that this contact took place in Mesopotamia, rather than in western Elam or Susiana. 39 See what emerges from the analysis of chlorite phials, above. 38

See Salvatori, Chapter 5.

37

93

S. Salvatori the effect, on the one hand, of the failure of the protostate model and, on the other hand, of the resumption of political control and power by the individual subregional chiefdoms, each of them pivoting on a fortified centre that was able to control a limited portion of the regional territory. The Late Bronze Age phase began, therefore, with a conservative-type political upheaval, characterised by a decisive repossession of the command levers by “tribal” segments that were strongly embedded in specific portions of the territory. During the same period, some evident metamorphoses took place in the military and defensive technique, with the appearance of a new type of fortification characterised by circular towers enclosing a palace, dwellings and production structures40.

one consists of a strong, progressive deterioration of the environment, with a considerable reduction of the flow of the water coming from the Pamir regions, and, particularly in Margiana, the concomitant advance of the sand of the Karakum Desert, carried by the prevailing northern winds. The second phenomenon consists of the progressive entrance into the Murgab Delta and northern Bactria of groups of nomadic pastoralists who were bearers of the distinctive pottery of the Andronovo groups of the Eurasian steppes (Cerasetti 1998). This second phenomenon crossed the entire territory of the Oxus Civilisation, and apparently went through two stages. A first stage of separate coexistence with the native, sedentary farmer groups was characterised by the presence of an extremely great number of sites with steppe pottery on areas which had already been encroached on by sand, around the Late Bronze Age built-up areas. A second stage, that had already been identified by V. M. Masson (1959) in the nineteen-fifties and called Takhirbai3 phase, but that we prefer to call the Final Bronze Age phase, saw the integration of these groups of nomadic pastoralists within the settlements of the local sedentary farmers. This second phase was a period of great reduction in the population of the Murgab Delta and perhaps also of the Bactrian region, probably in connection with the two phenomena mentioned above: the deterioration of the environment and the arrival of the nomadic population. We are certain, however, that right afterwards there began a new recovery stage, with the appearance and consolidation of a new archaeological culture that opened up the horizons of the Iron Age civilisation in Central Asia.

In our opinion, these transformations were not unrelated to some political circumstances that seemed to produce, during the same period, political changes or territorial re‑arrangements both in Iran and in the Indus Valley. Undoubtedly the material symbols of the trading relation network that characterised the archaeological complexes of the previous period lost their meaning, and were deprived of their function partly by hoarding and partly by recycling processes, while the residual trading flow was apparently diverted towards the Indus Valley, and, through that route, along the Gulf, to the detriment of the previous trading flows that crossed the Iranian plateau. In any case there remained, at least in Margiana, traces of administrative control techniques such as bullae and seals, though the latter appeared in shapes and materials that were quite different from, and unrelated to, the previous tradition of the metal compartmented seal. However we cannot but notice the drastic reduction or disappearance, in the archaeological inventory of exotic materials, of metal products with a high degree of symbolic content and technical elaboration, and of objects made out of precious materials such as gold or stones such as lapis lazuli41. Though we believe that the data are still inadequate, we may guess that in the new social and political organization resulting from the failure of the protostate model, wealth no longer followed the route of a more or less widespread redistribution, but became the object of concentration in the hands of the head of the community. A mechanism of this type may perhaps justify the impression of a great homogeneity in the grave furnitures present in the burial grounds of the Late Bronze Age both in Margiana and in Bactria.

We are aware of the fact that the picture outlined here will probably have to undergo substantial adjustments with the progress of researches, but we believe that we can already submit to historical and archaeological consideration an image of the discontinuity of this historical process that is dynamic, diversified and definitely not levelled down by simplistic schematisations. References Cited Alekshin, V. A. and Kircho L. B. 2005. Khronologya epokhi posdnego Eneolita - srednej bronzy sdrednej Azii (pogrebeniya Altyn-depe). Sankt-Peterburg. Al Khalifa, S. H. A. and M. Rice (eds.) 1986. Bahrain through the Ages. The Archaeology. London. Allchin, B. (ed.) 1984. South Asian Archaeology 1981. Cambridge. Amiet, P. 1966. Elam. Paris. Amiet, P. 1977. “Bactriane proto-historique”, Syria 54 (12): 89-121. Amiet, P. 1979. “Les sceaux de Shahr-i Sokhta”, in Van Lohuizen-De Leeuw (ed.) 1979: 3-6. Amiet, P. 1983. “The Archaic Glyptic at Shahr-i Sokhta (Period I)”, in Tosi (ed.) 1983: 199-210 Amiet, P. 1986. L’âge des échanges inter-iraniens 35001700 avant J.-C.. Paris.

During the Late Bronze Age, moreover, two phenomena appeared to severely affect cultural and political developments within the Oxus Civilisation. The first See, for instance, the settlement of Gonur 1 South (Sarianidi 1993b). Only a single grave datable to the Late Bronze Age, excavated by Sarianidi in the nineteen-eighties, at the top of the hill of Gonur 1 North, where the burial ground of the smaller settlement of the southern depe is, and, as far as I know, unpublished as yet, can be compared, for the richness of its outfit , with the Middle Bronze Age graves. 40 41

94

Cultural Variability in the Bronze Age Oxus Civilisation Amiet, P. 1988. “Antiquities ������������������������������������������ of Bactria and Outer Iran in the Louvre Collection”, in Ligabue and Salvatori (eds.) 1988: 159-180. Amiet, P. and M. Tosi 1978. “Phase 10 at Shahr-i Sokhta: Excavations in Square XDV and the Late 4th Millennium B.C. Assemblage of Sistan”, East and West 28: 9-31. Askarov, A. A. 1973. Sapallitepa. Tashkent. Askarov, A. A. 1977. Drevnezemledel’cheskaya kultura epokhi bronzi yuga Uzbekistana. Tashkent. Askarov, A. A. 1981. “Southern Uzbekistan in the Second Millennium B.C.”, in Kohl (ed.) 1981: 256-271. Askarov, A. A. and B. N. Abdullaev 1983. Dzharkutan. Tashkent. Avanesova, N. 1995. “Bustan VI, une nécropole de l’âge du Bronze dans l’ancienne Bactriane (Ouzbékistan méridional): témoignages de cultes du feu”, Arts Asiatiques L: 31-46. Baghestani, S. 1997. Metallene Compartimentsiegel aus Ost-Iran, Zentralasien und Nord-China. Rahden. Balsan, F. 1972. An Registan Inexploré (sud-afghan). Paris) Besenval, R. and H.-P. Francfort 1994. ��������������� “The Nad-i Ali „Surkh Dagh“: a Bronze Age monumental platform in Central Asia?”, in Kenoyer (ed.) 1994: 3-14. Biscione, R. 1985. “The So-called‚ Nestorian Seals‘: Connections between Ordos and Middle Asia in MiddleLate Bronze Age”, in Gnoli and Lanciotti (eds.) 1985: 95-109. Bonora, G. L., C. Domanin, S. Salvatori and A. Soldini 2000. ����������������������������������������� “The Oldest Graves of the Shahr-i Sokhta Graveyard”, in Taddei and De Marco (eds.) 2000: 495520. Bulgarelli, G. M. 1974. ��������������������������������� “Tepe Hissar. Preliminary Report on a Surface Survey, August 1972”, East and West 24 (1-2): 15-27. Burkholder, G. 1972. “Ubaid Sites and Pottery in Saudi Arabia”, Archaeology 25: 264-269. Calmeyer, P. 1969. Datierbare Bronzen aus Luristan und Kirmanshah. ������� Berlin. Carter, E. 1980. “Excavations in Ville Royale I at Susa: the Third Millennium B.C. Occupation”, Cahiers de la Délégation Archéologique Française en Iran 11: 11134. Casal, J.-M. 1961. Fouilles de Mundigak. ������������ I-II. Paris. Casanova, M. 1991. La vaisselle d’álbâtre de Mésopotamie, d’Iran et d’Asie centrale aux IIIe et Ie millénaires av. J.-C.. Paris. Cerasetti, B. 1988. ���������������������������������� “Preliminary Report on Ornamental Elements of «Incised Coarse Ware»”, in Gubaev, Koshelenko and Tosi (eds.) 1998: 67-74. Ciarla, R. 1979. “The Manufacture of Alabaster Vessels at Shahr-i Sokhta and Mundigak in the 3rd Millennium BC: a Problem of Cultural Identity”, in Gnoli and Rossi (eds.) 1979: 319-35. Ciarla, R. 1981. “A Preliminary Analysis of the Manufacture of Alabaster Vessels at Shahr-i Sokhta and Mundigak in the 3rd Millennium B.C.”, in Härtel (ed.) 1981: 45-63.

Cleuziou, S. 1989. “The Chronology of Protohistoric Oman as seen from Hili”, in Costa and Tosi (eds.) 1989: 47-78. Costa, P. M. and M. Tosi (eds.). Oman Studies. Papers on the Archaeology and History of Oman. ������ Rome. Crawford, H. and K. Al Sindi 1995. ��������������� “A Seal in the Collections of the National Museum, Bahrain”, Arabian Archaeology and Epigraphy 6: 1-4. Curtis, J. 1988. “A Reconsideration of the Cemetery at Khinaman, South-East Iran”, Iranica Antiqua XXIII: 97-128. Cuyler Young, T. Jr. 1969. Excavations at Godin Tepe: First Progress Report. Toronto. Cuyler Young, T. Jr. and L. D. Levine 1974. Excavations of the Godin Project: Second Progress Report. Toronto. ��������� Deshayes, J. 1977. “À ��� propos �������������������������������������� des terrasses hautes de la fin du IIIe millénaire en Iran et en Asie Centrale”, in Gardin (ed.) 1977: 95-111. During-Caspers, E. C. L. 1972. “La �������������������������� Hachette trouée de la sépulture E de Khuraab, dans le Balouchistan Persan, Examen retrospectif”, Iranica Antiqua IX: 60-64. During-Caspers, E. C. L. 1992 “Intercultural/Mercantile Contacts between the Arabian Gulf and South Asia at the Close of the Third Millennium B.C.”, Proceedings of the Seminar for Arabian Studies 22: 3-28. During-Caspers, E. C. L. 1994a. “Widening Horizons; Contacts between Central Asia (the Murgabo-bactrian Culture) and the Indus Valley Civilization towards the Close of the Third and the Early Centuries of the Second Millennium B.C.”, Annali 54 (2): 171-198. During-Caspers, E. C. L. 1994b. “Non Indus Glyptic in a Harappan Context”, Iranica Antiqua XXIX: 83-105. Dyson, R. H. Jr. and S. M. Howard (eds.) 1989. Tappeh Hesar. Reports of the Restudy Project, 1976. Firenze. Fairservis, W. A. Jr. 1961. Archaeological Studies in the Seistan Basin of South-Western Afganistan and Eastern Iran. New York. Francfort, H.-P. 1989. Fouilles de Shortugaï. Recherches sur l’Asie Centrale protohistorique. Paris. Francfort, H.-P. 1994. ������������������������������������ “The Central Asian dimension of the Symbolic system in Bactria and Margiana”, Antiquity 68 (259): 406-418. Frifelt, K. and P. Sørensen (eds.) 1989. South Asian Archaeology 1985. ������������ Copenhagen. Gardin, J.-C. (ed.) 1977. Le Plateau Iranien et l’Asie centrale des origines a la conquéte islamique. Leurs relations à la lumiére des documents archéologuques. Paris Gardin, J.-C. (ed.) 1985. L’archéologie de la Bactriane ancienne. Paris. Gardin, J.-C. (ed.) 1988. L’Asie Centrale et ses rapports avec les civilisations orientales des origines à l’Âge du Fer. Paris. Glob, P. V. 1958. “Alabaster ������������������������������������������ Vases from the Barbar Temples. The Danish Archaeological Bahrein-Expedition’s Fifth Excavation Campaign”, KUML: 138-145. Gnoli, G. and L. Lanciotti (eds.) 1985. Orientalia Iosephi Tucci Memoriae Dicata. Rome. 95

S. Salvatori Gnoli, G. and A. V. Rossi (eds.) 1979. Iranica. Napoli. Gubaev, A., G. A. Koshelenko and M. Tosi (eds.) 1998. The Archaeological Map of the Murgab Delta. Preliminary Reports 1990-95. Reports and Memoirs, Series Minor , III. Rome. Härtel, H. (ed.) 1981. South Asian Archaeology 1979. Berlin. Hakemi, A. 1997. Shahdad. Archaeological Excavations of a Bronze Age Centre in Iran. Rome. Heimpel, W. 1988. ��������� “Magan”, Reallexikon der Assyriologie 7: 195-199. Heinz, M. 1989. “Die Steingefässe aus Süd- und Mittelmesopotamischen Zeit”, Baghdader Mitteilungen 20: 197-224. Henrickson, R. C. 1984. “Šimaški ����������������������������� and Central Western Iran: the Archaeological Evidence”, Zeitschrift für Assyriologie 74 (1), 98-122. Hiebert, F. T. 1994a. Origins of the Bronze Age Oasis Civilization in Central Asia. ���������� Cambridge. Hiebert, F. T. 1994b. ������������������������������������� “Production Evidence for the Origins of the Oxus Civilization”, Antiquity 68 (259): 372-387. Isakov, A. I. 1977 Isakov, A. I. 1991. Sarazm. Dushanbe. Jarrige, C. and V. Lefèvre (ed.) 2005. South Asian Archaeology 2001. ������ Paris. Jarrige, J.-F. 1985. “Les relations entre l’Asie Centrale méridionale, le Baluchistan et la vallée de l’Indus à la fin du 3e et au début du 2e millénaire”, in Gardin (ed.) 1985: 105-118. Jarrige, J.-F. 1987. “A ���������������������������������������� Prehistoric Elite Burial in Quetta”, Newsletter of Baluchistan Studies 4: 3-9. Jarrige, J.-F. 2005 Jarrige, J.-F. and M. U. Hassan 1989. “Funerary Complexes in Baluchistan at the End of the Third Millennium in the Light of Recent Discoveries at Mehrgarh and Quetta”, in Frifelt and Sørensen (eds.) 1989: 150-166. Kenoyer, J. M. (ed.) 1994. From Sumer to Meluhha. Contributions to the Archaeology of South and West Asia in Memory of George F. Dales, Jr.. �������� Madison. Kircho, L. B. 1988. “The ����������������������������������� Beginning of the Early Bronze Age in Southern Turkmenia on the Basis of Altyn-depe Materials”, East and West 38 (1-4): 33-64. Kircho, L. B. 2000. “Bogatoe ����������������������������������� pogrebenie epokhi srednej bronzy na Altyn-depe”, Arkheologicheskie Vesti 7: 7076. Kircho, L. B. 2001. �������������������������������������� “Metallicheskie izdeliya Altyn-depe”, in Kircho (ed.) 2001: 60-84. Kircho, L. B. (ed.) 2001. Osobennosti proizvodstva poseleniya Altyn - depe ve epokhu paleometalla. (Materialy Yuzhno - Turkmenistanskoj arkheologicheskoj compleksnoj ekspedizii, vy. 5). Sankt-Peterburg. Knox, A. A. 1994. “New Indus Valley Cylinder Seal”, in Parpola and Koskikallio (eds.) 1994: 375-378. Kohl, Ph. L. 1975. “Carved ������������������������������������� Chlorite Vessels: a Trade in Finished Commodities in the Mid Third Millennium B.C.”, Expedition 18 (1): 18-31.

Kohl, Ph. L. 1979. ���������������������������������� “The ‚World Economy‘ of West Asia in the Third Millennium B.C.”, in Taddei (ed.) 1979: 55-85. Kohl, Ph. L. (ed.) 1981. The Bronze Age Civilization of Central Asia. Recent Soviet Discoveries. New York. Kohl, Ph. L. 1984. Central Asia. Palaeolithic Beginnings to the Iron Age. ������ Paris. Kohl, Ph. L. 2001. ���������������������������������� “Reflections on the Production of Chlorite at Tepe Yahya: 25 Years Later”, in LambergKarlovsky and Potts (eds.) 2001: 209-230. Kohl, Ph. L., G. Harbottle and E. V. Sayre 1979. “Physical and Chemical Analyses of Soft Stone Vessels from Southwest Asia”, Archaeometry 21: 131-159. Lamberg-Karlovsky, C. C. 1970. Excavations at Tepe Yahya, 1967-1969. Cambridge. Lamberg-Karlovsky, C. C. 1994. ���������������� “The Bronze Age Khanates of Central Asia”, Antiquity 68 (259): 398405. Lamberg-Karlovsky, C. C. 2001a. “Excavations ��������������������� at Tepe Yahya. The Biography of a Project”, in LambergKarlovsky and Potts (eds.) 2001: XIX-XLI. Lamberg-Karlovsky, C. C. 2001b. “Excavations ��������������������� at Tepe Yahya: Reconstructing the Past”, in Lamberg-Karlovsky and Potts (eds.) 2001: 269-280. Lamberg-Karlovsky, C. C. and D.T. Potts (eds.) 2001. Excavations at Tepe Yahya, Iran 1967-1975. The Third Millennium. Cambridge. Le Breton, L. 1957. “The Early Periods at Susa, Mesopotamian Relations”, Iraq XIX (2): 79-124. Ligabue, G. and S. Salvatori (eds.) 1988. Bactria. An ancient oasis civilization from the sands of Afganistan. Venezia�. Lyonnet, B. 1996. Sarazm (Tadjikistan). Céramiques (Chalcolithique et Bronze Ancien). Paris. Majidzadeh, Y. 2003. Jiroft: the Earliest Oriental Civilization. Tehran. Masimov, I. S. 1984. “Zhiloj dom na Kelleli”, Pamyatniki Turkmenistana 2/38: 16-18. Masimov, I. S. 1986. “Novye issledovaniya pamyatnikov epokhi bronzy na Murgabe”, in Masson (ed.) 1986: 171-181. Masimov, I. S., S. Salvatori and B. Udeumuradov 1998. “Preliminary Analysis of the Bronze Age Material Collected by the Margiana Archaeological Project and a First Chronological Assessment”, in Gubaev, Koshelenko and Tosi (eds.) 1998: 35-46. Masry, A. H. 1997. Prehistory in Northeastern Arabia. London. Masson, V. M. 1959. Drevne zemle del’ cheskaya kul’tura Margiany. (Materialy i issledovaniya po arkheologii SSSR, vol.73), Moskva - Leningrad. Masson, V. M. 1957. Masson, V. M. 1981, Altyn-depe, (Trudy yuzhnoturkmenistaskoj arkheologicheskoj kompleksnoj ekspedici 18). Leningrad. Masson, V. M. (ed.) 1986. Drevnie zivilizazii Vostoka. Tashkent. Masson, V. M. 1988. Altyn-depe. Philadelphia. 96

Cultural Variability in the Bronze Age Oxus Civilisation Masson, V. M. and V. I. Sarianidi 1972. Central Asia: Turkmenia Before the Achaemenids. ������� London. Meder, O. G. 1979. Klimaoekologie und Siedlungsgang auf dem Hochland von Iran in vor- und frühgeschichtlicher Zeit. (Marburger Geographische Schriften 80). Marburg/Lahn. Méry, S. 2000. Les céramiques d‘Oman et l‘Asie moyenne. Une archéologie des échanges à l‘Âge du Bronze. Paris. Miroschedji, P. de 1983. “Vases et objets en stéatite susiens du Musée du Louvre”, Cahiers de la Délégation Archéologique Française en Iran� 3: 9-80. Oates, J. 1986. “The Gulf in Prehistory”, in Al Khalifa and Rice (eds.) 1986: 79-86. Parpola, A. and P. Koskikallio (eds.) 1994. South Asian Archaeology 1993. (Annales Academiae Scientiarum Fennicae, Series B, Vol. 271). Helsinki. P’iankova, L. 1981. “Bronze Age Settlements of Southern Tadjikistan”, in Kohl (ed.) 1981: 287-310. Pyankova, L. 1989. “Pottery Complexes of BronzeAge Margiana (Gonur and Togolok 21)”, IASCCA Information Bulletin 16: 27-54. P’iankova, L. 1993. “Pottery of Margiana and Bactria in the Bronze Age”, IASCCA Information Bulletin 19: 109-127. P’yankova, L. 1994. “Central Asia in the Bronze Age: Sedentary and Nomadic Cultures”, Antiquity 68 (259): 372-387. Piperno, M. and S. Salvatori n.d. ������������������������� Shahr-i Sokhta Graveyard Excavations 1972-1978. ����� Rome. Pittman, H. 1984. Art of the Bronze Age. Southern Iran, Western Central Asia, and the Indus Valley. New York. Pittman, H. 2001. “Glyptic Art of Period IV”, in LambergKarlovsky and Potts 2001: 231-268. Pottier, M.-H. 1984. Matériel funéraire de la Bactriane méridionale de l’Âge du Bronze. ������ Paris. Potts, D. T. 1986. “The Booty of Magan”, Oriens Antiquus 25: 271-285. Potts, D. T. 1993. “A New Bactrian Find from Southeastern Arabia”, Antiquity 67 (256): 591-595. Potts, D. T. 2001. “Phase IVB5”, in Lamberg-Karlovsky and Potts (eds.) 2001: 105-143. Potts, T. F. 1989. “Foreign Stone Vessels of the Late Third Millennium B.C. from Southern Mesopotamia: Their Origins and Mechanisms of Exchange”, Iraq 51: 123164. Potts, T. F. 1993. “Patterns of trade in third-millennium BC Mesopotamia and Iran”, World Archaeology 24 (3): 380-402. Pumpelly, R. (ed.) 1908. Explorations in Turkestan, Expedition 1904. Prehistoric Civilizations of Anau. 2 vols. Washington. ����������� Salvatori, S. 1978. “Problemi di protostoria iranica: note ulteriori su di una ricognizione di superficie a Shahdad (Kerman, Iran)”, Rivista di Archeologia II: 5-13. Salvatori, S. 1988. “Early Bactrian Objects in Private Collections”, in Ligabue and Salvatori (eds.) 1988: 181-187.

Salvatori, S. 1993. “The Discovery of the Graveyard of Gonur-depe 1 (Murgab Delta, Turkmenistan): 1992 Campaign Preliminary Report”, Rivista di Archeologia XVII: 5-13. Salvatori, S. 1994a. “A ����������������������������������� Late Third Millennium Graveyard at Gonur Tepe 1 (Murgab Delta, Turkmenistan)”, in Parpola and Koskikallio (eds.) 1994: 657-666. Salvatori, S. 1994b. “Excavations ����������������������������������� at the Namazga V Late Graveyard of Gonur 1 (Murgab Delta, Turkmenistan). Preliminary Report on the 1993 field season”, Rivista di Archeologia XVIII: 15-39. Salvatori, S. 1995a. “Gonur-depe ������������������������� 1 (Margiana, Turkmenistan): the Middle Bronze Age Graveyard. Preliminary Report on the 1994 Excavation Campaign”, Rivista di Archeologia XIX: 5-37. Salvatori, S. 1995b. “Protohistoric Margiana: on a Recent Contribution (Review of “Information Bulletin IASCCA” 19, 1993)”, Rivista di Archeologia XIX: 138-55. Salvatori, S. 1998a. “The Bronze Age in Margiana”, in Gubaev, Koshelenko and Tosi (eds.) 1998: 47-55. Salvatori, S. 1998b. ���������������������������������� “Margiana Archaeological Map: the Bronze Age Settlement Pattern”, in Gubaev, Koshelenko and Tosi (eds.) 1998: 57-65. Salvatori, S. 2000. “Bactria and Margiana Seals. A New Assessment of their Chronological Position and a Typological Survey”, East and West 50: 97-135. Salvatori, S. 2002. “Project “Archaeological Map of the Murgab Delta” (Turkmenistan): Test Trenches at the Bronze Age Sites of Adzhi Kui 1 and 9”, Ancient Civilizations from Scythia to Siberia 8 (1-2): 107-178. Salvatori, S. and M. Tosi 2005. “Shahr-i Sokhta Revised Sequence”, in Jarrige (ed.) 2005: 281-292. Salvatori, S. and M. Vidale 1982. “A Brief Surface Survey of the Protohistoric Site of Shahdad (Kerman, Iran)”, Rivista di Archeologia VI: 5-10. Salvatori, S. and M. Vidale 1997. Shahr-i Sokhta 19751978: Central Quarters Excavations. Preliminary Report. ����� Rome. Santoni, M. 1984. ��������������������������������� “Sibri and the South Cemetery of Mehrgarh: Third Millennium Connections between the Northern Kachi Plain (Pakistan) and Central Asia”, in Alchin (ed.) 1984: 52-60. Santoni, M. 1988. “Aspects máterels des cultures de Sibri et de Mehrgarh VIII (plaine de Kachi, Baluchistan, Pakistan) à la fin du troisième et au début du deuxième millénaires”, in Gardin (ed.) 1988: 135-141. Sarianidi, V. I. 1976. “Issledovanie pamyatnikov Dashlinskogo oazisa”, Drevnyaya Baktriya 1: 21-86. Sarianidi, V. I. 1977. Drevnie zemledel’zy Afganistana. Moskva. Sarianidi, V. I. 1979. “New ������������������������������� Finds in Bactria and IndoIranian Connections”, in Taddei (ed.) 1979: 643-659. Sarianidi, V. I. 1981. ������������������������������������� “Seals-Amulets of the Murgab Style”, in Kohl (ed.) 1981: 221-255. Sarianidi, V. I. 1983. ������������������������������������� “The Pottery of Shahr-i Sokhta I and its Southern Turkmenian Connections”, in Tosi (ed.) 1983: 183-198. 97

Sarianidi, V. I. 1986. Die Kunst des alten Afganistan: Architektur, Keramik, Siegel, Kunstwerke aus Stein und Metall. �������� Leipzig. Sarianidi, V. I. 1990. Drevnosti strany Margush. Ashkhabad. Sarianidi, V. I. 1993a. “Reperti inediti da tombe Battriane depredate”, Mesopotamia XXVIII: 5-20. Sarianidi, V. I. 1993b. ��������������������������������� “Excavations at Southern Gonur”, Iran XXXI: 25-37. Sarianidi, V. I. 1998a. Margiana and Protozoroastrism. Athens. Sarianidi, V. I. 1998b. Myths of Ancient Bactria and Margiana on its Seals and Amulets. ������� Moscow. Sarianidi, V. I. 2001. Necropolis of Gonur and Iranian Paganism. ������� Moscow. Sarianidi, V. I. 2002. Margush. Ancient Oriental Kingdom in the Old Delta of the Murgab River. Ashgabat. ��������� Sarianidi, V. I. 2005. Gonurdepe. City of Kings and Gods. Ashgabat. Schmandt-Besserat, D. 1992. Before Writing. From Counting to Cuneiform. ��������������� 2 vols. Austin. Schmidt, E. F. 1937. Excavations at Tepe Hissar (Damghan). Philadelphia. Shirinov, T. and S. Baratov 1997. “Bronzezeitlische ������������������ Grabstätten aus der Nekropole Džarkutan 4c (SüdUzbekistan)”, Archäologische Mitteilungen aus Iran und Turan 29: 65-120. Stein, A. 1931. An Archaeological Tour in Gedrosi. (Memoirs of the Archaeological Survey of India 43). Calcutta. Steinkeller, P. 1982. ����������������������������� “The Question of Marhashi: A Contribution to the Historical Geography of Iran in the Third Millennium B.C.”, Zeitschrift für Assyriologie 72: 237-265. Taddei, M. (ed.) 1979. South Asian Archaeology 1977. Naples. Taddei, M. and G. De Marco (eds.) 2000. South Asian Archaeology 1997. ����� Rome.

Tanabe, K. 1983. Animals in the Arts of the Ancient Orient. Tokyo. Tosi, M. 1968. “Excavations at Shahr-i Sokhta, a Chalcolithic Settlement in the Iranian Sistan. Preliminary Report on the First Campaign, OctoberDecember 1967”, East and West 18 (1-2): 9-66. Tosi, M. 1969. “Excavations at Shahr-i Sokhta. Preliminary Report on the Second Campaign, September-December 1968”, East and West 19 (3-4): 283-386. Tosi, M. 1983. “Development, Continuity and Cultural Change in the Stratigraphical Sequence of Shahr-i Sokhta”, in Tosi (ed.) 1983: 127-179. Tosi, M. (ed.) 1983. Prehistoric Sistan. 1. Rome. Tosi, M. 1989. “The Distribution of Industrial Debris on the Surface of Tappeh Hesār as an Indication of Activity Areas”, in Dyson and Howard (eds.) 1989: 13-24. Udeumuradov, B. N. 1993. Altyn-depe i Margiana: svyazi, khronologiya, proiskhozhdenie. Ashgabat. ��������� Van Lohuizen-De Leeuw, J. E. (ed.) 1979. South Asian Archaeology 1975. �������� Leiden. Vallat, F. 1985. “Élements de géographie élamite (résumé)”, Paléorient 11 (2): 49-54. Vogt, B. 1985. “The Umm an-Nar Tomb A at Hili North: A Preliminary Report on Three Seasons of Excavation, 1982-1984”, Archaeology in the United Arab Emirates IV: 20-37. Winkelmann, S. 1997. “Ein neues trans-elamisches Siegel”, Archäologische Mitteilungen aus Iran und Turan 2 (9): 135-146. Zarins, J. 1978. “Typological Studies in Saudi Arabian Archaeology. Steatite Vessels in the Riyad Museum”, Atlal 2: 65-93. Zarins, J. 1989. “Eastern Saudi Arabia and External Relations: Selected Ceramic, Steatite, and Textual Evidence: 3500-1900 BC”, in Frifelt and Sørensen (eds.) 1989: 74-103.

Chapter 7 Unpublished Stamp-seals from the North-western Murgab Delta Iminjan S. Masimov † and Sandro Salvatori

In 1998 V. I. Sarianidi published a catalogue of the seals of Bactria and Margiana, in which, besides illustrating a very great number of seals, he presented a concise analysis of their iconographical system and an interpretation of it within the historical and cultural context of the ancient eastern world (Sarianidi 1998).

Tadzhikistan), and constantly bring to light new, and sometimes unexpected, specimens of this important class of objects. In this essay, we will examine a group of seals found during the researches that were carried out between the nineteenseventies and the nineteen-nineties, and that for various reasons are still unpublished or inadequately published. Most of these seals come from the western area of the ancient Murgab Delta, where archaeological settlements dating back to a period ranging from the Middle Bronze Age (Namazga V) to the Achaemenid period have been found.

The seals presented in the book, though numerous (1802, 254 of them coming from Margiana), are not the total of those known at present, not only because some important European collections are unpublished as yet, but also because archaeological researches are still under way in Margiana and northern Bactria (Uzbekistan e

We will begin this review from the northern sector of the delta, where up to now a great number of protohistorical settlements, distributed on the alluvial plain (takyr) of Kelleli, have been identified. All the settlements in this area, excluding two, date back to the Middle Bronze Age. Out of the eight settlements that were located here, four were investigated archaeologically: Kelleli 1, 3, 4 and 6. One of the first seals was found in the Kelleli 1 site. It is a bronze circular seal with three birds of prey arranged radially around a circular centre (Figure 7.1.3). This seal is an example of the work of handcraftsmen who had thoroughly mastered the use of their technical instruments, and fits into a typologically and iconographically specific production. The motif of radially arranged animals is wide-spread: examples of it have been found at Altyndepe (Masson 1988: Pls. XVI.9; XXIX.6; XXXVII.4), in Margiana (in this case the animals are capridae: Sarianidi 1994: Fig. 57.3.b), and in southern Bactria (Pittman 1984: Fig. 25c). In the settlement of Kelleli 3, there was a part of a fortress whose layout was rectangular and whose perimetric walls had quadrangular towers at the corners. Here some bronze seals, most of them with a cross motif, were found. Along the eastern front of a perimetric corridor, a stepped-lozenge seal with rounded corners was found. (Figure 7.2.3). At its centre, there are two diagonally intersecting lines forming an X, one of the variants of the cross motif. Another seal, with a Maltese cross shape, has slightly flaring ends Masimov 1981: Fig. 2.11. For the problem of broader-range correlations between seals with radially arranged animals, see Salvatori 2000: 124. Other specimens coming from southern Bactria, both with eagles and with goats, mouflons or ibexes, are known to be present in unpublished private collections.  

Figure 7.1 Metal stamp-seals from: 1) Taip 1; 2) Kelleli 4 ; 3) Kelleli 1; 4) Kelleli 4.

99

I. S. Masimov and S. Salvatori

of them, however, undoubtedly belong to the class of cross-shaped seals. In the nearby settlement of Kelleli 4, a monumental building that had undoubtedly belonged to an important family, was brought to light. Its layout was square, and, like the boundary wall of Kelleli 3, it had an internal perimetric passageway. Instead of defensive towers, at its corners there were rectangular rooms used for dwelling or for economical activities. In this settlement, V. I. Sarianidi, at the beginning of his excavation, collected a zoomorphic seal depicting a caprid with stubby legs and short horns (Figure 7.2.5). The seal, like others found in the area of Kelleli (Figure 7.2.6), is compartmented, has a closed back and bears an image of a caprid that is comparatively widespread both in the piedmont of southern Turkmenistan (Altyn-depe), in Margiana (Figure 7.2.6) and in Bactria. This type of zoomorphic seal with open margins is not very frequent in Margiana and Bactria, perhaps for purely accidental reasons; it is frequent, on the contrary, at Altyndepe (Kircho 2001: Figs. 13.1-9, 12). It has been found also in Sistan, but unfortunately almost all the specimens known up to now come from surface collections (Ferioli et al. 1979: Figs. 5a, 4c, 4h), except for an unpublished specimen from Grave No. 731, datable to Period III, i.e. to the first part of the second half of the 3rd millennium BC. From Kelleli 4, particular attention is deserved by a triangular seal with open margins and triangular centre (Figure 7.1.2); this, as far as we know, is an unusual shape in the bronze glyptic of Central Asia. Another seal from Kelleli 4 belongs to a type that is widespread not only in Central Asia, but also in Iranian Sistan. Its shape is that of a stepped lozenge with an Xshaped motif at its centre (Figure 7.2.4). This type of seal has a rather long life span: its first appearance, as far as we know at present, dates back to the 27th century BC. From the same site there also comes an interesting seal that belongs to a type quite frequent in Margiana, southern Bactria and the piedmont of southern Turkmenistan. It is a circular compartmented openwork seal, with a single, simple collar and a cross-shaped motif (Figure 7.1.4). A similar, though not identical, specimen is present among the finds of Altyn-depe (Kircho 2001: Fig. 13.23), while several specimens of this type have been collected in Margiana (Sarianidi 1981: Figs. 7.27; 8.6) and Bactria (Sarianidi 1977: Fig. 48.3; Salvatori 1988: Fig. 56).

Figure 7.2 Metal stamp-seals from: 1) Kelleli 3 ; 2) Taip 1 ; 3) Kelleli 3; 4) Kelleli 4; 5) Kelleli 4; 6) Taip 1 (surface).

(Figure 7.3.8). Unlike other seals of the same class known at Altyn-depe (Masson 1988: Pls. XXIX.7; XVI.6), this seal does not have a circular centre, but a cross-shaped element that makes it comparable to another seal from Altyn-depe (Ibidem: Pl. XVI.1). A third seal, which was found at Kelleli 3, on the surface of the hill, belongs to the openwork type, and has an internal cross motif (Figure 7.3.7). Most of the bronze seal fragments found on the surface of this site or unearthed during its excavation do not make it possible to identify the decorative motif; some

      

100

Masimov 1981: Fig. 2.14. Salvatori 2000: Type 4.b, 110. Masson 1988: Pl. XVII.11, 14; XXIX.9; XXXVII.5. From Taip 1: Masimov 1981: Fig. 2.9. Amiet 1986: Fig. 179. Salvatori and Tosi 2005. Salvatori 2000: Type 5, 113-116.

Unpublished Stamp-seals from the North-western Murgab Deltae

Figure 7.3 Metal stamp-seals from: 1) Taip 8; 2) Adzhi Kui 7 ; 3) Taip 1; 4) Kelleli 3; 5) Taip 1; 6) Kelleli 4; 7) Kelleli 3 (surface); 8) Kelleli 3; 9) Adzhi Kui 7; 10) Taip 1; 11) Adam Basan 12; 12) Kelleli 3; 13) Taip 1.

101

I. S. Masimov and S. Salvatori

Figure 7.5 Stone stamp-seals from: 1) Adam Basan 10; 2) Adzhi Kui 6; 3) Adzhi Kui 10; 4) Taip 1; 5) unrecorded; 6) Adam Basan 16; 7) Adzhi Kui 3. Figure 7.4 Metal and stone stamp-seals from: 1-2) Adam Basan 20; 3-7) Taip 1.

cut in ancient times, perhaps as a result of a regularisation of the object after it had been chipped. It is well known that the swastika motif is widespread in the glyptic of the Indus Valley (Joshi and Parpola 1987). A very few examples of this decoration are present in Central Asia and Iran. A specimen that is practically identical was found at Altyn-depe, in the piedmont of southern Turkmenistan (Masson 1988: 68, Tab. XXII.1b). This specimen comes from Room 7 of Excavation 7, one of the rooms, or rather the sanctuary room, of a building that has been classified as a funerary complex. According to V. M. Masson, the complex probably dates to the beginning of the Namazga V period (Masson 1988: 75). The excavator pointed out that this seal, as well as another one with an inscription formed of two signs (Ibidem: Tab. XXII.1a), had been imported from the Indus Valley (Ibidem: 118).

Excavations were performed also in the settlement of Kelleli 6, the smallest of the group. Here a series of dwelling structures were found that are different, in their building technique, from those of the area of Kelleli: they are not built with regular bricks, but with irregular blocks of clay, and their layout is quite irregular. Nonetheless, the pottery unearthed here is not different from that found in the Middle Bronze Age settlements, so it is possible to surmise that the archaic character of the building technique depends only on the nature of the settlement, which perhaps was only a small farming village. From the deposits of Kelleli 6 there comes an alabaster compartmented stamp‑seal. It is square, and on one face it has a swastika motif made with deep incisions and surrounded by a thin frame (Figure 7.8.8). On the verso there is a low, oval suspension ring with through hole. One of the margins of the seal has been

The swastika motif, though stylistically quite distant from the specimens of Altyn and Kelleli 6, is also present on 102

Unpublished Stamp-seals from the North-western Murgab Deltae

Figure 7.7 Stone stamp-seals from: 1-3) Taip 1; 4) Kelleli 4; 5) Adam Basan 4; 6) Egri Bogaz 1.

some differences between the handles of the three seals. In the specimen from Kelleli 6, the handle is on the back, in a central position, is oval and has a through hole. In the specimen from Altyn-depe, the handle is sub-cylindrical and has a through hole. The seal from Bactria, on the contrary, has a ring handle in an eccentric position.

Figure 7.6 Stone stamp-seals from: 1) Adam Basan 4; 2) Taip 1; 3) Adam Basan 5; 4) unrecorded; 5-7) Taip 1; 8) Adzhi Kui 1.

a lenticular button-shaped seal from Margiana (Masimov 1981: Fig. 4.10) datable to the Late Bronze Age; on a bronze closed-back compartmented seal (Sarianidi 1990: Tab. XXXVI.2) that was found on the surface at Gonur 1 and can be dated to the Middle Bronze Age; and on a bronze closed-back compartmented specimen that was collected on the surface in one of the sites of the area of Togolok (Ibidem: Tabs. XXXVI.11; LXXXVIII.2)10.

South-east of the settlements of the Kelleli area, at a distance of about 20 km, separated by a vast expanse of hardened sand, there are the sites that were found near the well of Egri Bogaz. One of the largest sites of this area was called Egri Bogaz 1, and here a biface, dark-green serpentinite seal was found (Figure 7.7.6). A corner of this seal is missing, probably as a result of a breakage in antiquity, because the breakage surface has been smoothed. On one face there is a horse on whose back a saddle-cloth with rectangular motifs has probably been depicted. The front legs of the horse are straight, while the back ones are slightly angulated; its muzzle is lowered, as if it were grazing. The other face of the seal has a wide breach in it. As far as it is possible to gather, it has been incised with

Finally, we must point out a specimen from Bactria that is stylistically homogeneous with the two seals from Altyndepe and Kelleli 6 (Sarianidi 1998: No. 1353). There are The seal comes from Togolok 21, according to the caption of Pl. XXXVI, or from Togolok 24, according to the caption of Pl. LXXXVIII. 10

103

I. S. Masimov and S. Salvatori

Figure 7.9 Stone stamp-seals from: 1) unrecorded; 2) Taip 8; 3) Adam Basan 10; 4) unrecorded.

oblique lines that converge towards the centre, and in the triangular area there is the twisting body of a snake. Further south of Egri Bogaz, at a distance of about 10 km, there are the sites of the Adzhi Kui area. It consists of about ten low mounds where about ten seals were found during the surface surveys of the nineteen-seventies (Masimov 1981: Figs. 2.1-10; 4.9, 13-14). Subsequently, at Adzhi Kui 7, a small, round bronze stamp-seal with a very prominent loop handle on the back was collected. The functional face is decorated with a petal motif and round pistil (Figure 7.3.2). From Adzhi Kui 10 there comes an oval, biface serpentinite seal with plano-convex section, bearing the image of a stylised spider or otherwise a millipede on one face and of a snake on the other (Figure 7.5.3). The snake’s body has been made with a series of drill strokes. The seal has a longitudinal through hole. At Adzhi Kui 1, a fragment of a round serpentinite stamp-seal was collected. On its back, around the prominent, perforated handle, a zigzag motif has been incised. On the functional face, only a few grooves can be detected (Figure 7.6.8)11.

Figure 7.8 Stone stamp-seals from: 1) Adam Basan 20; 2) unrecorded; 3) Taip 1; 4) Adzi Kui 9; 5-7) Taip 1; 8) Kelleli 6.

Further finds made in 1977 at Adzhi Kui 1 and 9 are shown in Salvatori 2000: Figs. 5 and 6. 11

104

Unpublished Stamp-seals from the North-western Murgab Deltae

The group of sites of Taip is about 80 km north-west of Bajram-Ali, along the road that leads to Kelleli. Here, besides a few Bronze Age settlements, some Early Iron Age sites and a camp-site of the steppe cultures were found. The largest site, Taip 1, is about 12 hectares and is formed of two distinct hills, the civil settlement area and a sort of fortified citadel. Here the first four cylinder stamp-seals known in Central Asia were found. Some more were found later on in the sites of Togolok 1 and Gonur 1 South. At Taip 1, in all, more than thirteen seals were collected, and part of them have already been published. Among the unpublished ones, we would like to mention some stone stamp-seals with an incised cross motif; other stone seals have zoomorphic figurations. In this site there was also found a bronze closed-back stamp-seal. It has the shape of a corolla with a central pistil formed of two concentric circles (Figure 7.1.1)12.

Figure 7.10 Stone stamp-seals from: 1) Adzhi Kui 6; 2) Adam Basan 5.

We should also mention a fragment of an amulet seal decorated on both faces with incised snakes (Figure 7.6.2); a rectangular amulet seal, almost complete, that has on one face a star motif and four rectangular depressions, and on the other a bird with its wings open in heraldic position (Figure 7.6.6); another ovoidal amulet seal that has on one face a cross motif with incised lines in its interstices and on the other a series of radially-arranged lines that start from a central cup hole. The margins of this seal are outlined (Figure 7.7.3). A square serpentinite amulet seal has on one face a fantastic winged animal and on the other a plant with two broad divergent leaves at its base, smaller leaves along its stalk and a flower at the top13. Two through holes have been made at the end of this seal (Figure 7.7.2). We should also mention a fragment of a stone stamp‑seal, probably rectangular, with outlined margins. The functional face has a motif with parallel lines that are made of dots obtained with drill strokes surrounding an unidentified central element. On the back, a perforated band-like handle has been preserved (Figure 7.6.5). Another serpentinite stamp-seal, dark green and round, has a motif of seven petals around a circular pistil. A series of rope-shaped motifs is arranged on the outside. On the back, a perforatedring handle has been preserved (Figure 7.4.6). Finally, we should mention a further circular stone seal that has, on its functional side, a motif formed of three crescents and a vertical line that intersects the upper crescent at its centre and divides the space into two parts, each of which contains a triangle with an oblique line above it. On the

Figure 7.11 Cylinder stamp-seals from Taip 1.

Published in Masimov 1981: Fig. 2.8. A similar seal from Dzharkutan is known (Askarov and Abdullaev 1983: Tab. XXXII.1). 13 This motif has been frequently reproduced, with numerous variants, on biface seals both in Margiana and in Bactria (Sarianidi 1990: Fig. 32). We will mention here only a few specimens, e.g. a stepped- lozenge seal from Sapalli-depe (Askarov 1977: Pl. XLIV.2) and a similarly-shaped seal from southern Bactria (Sarianidi 1981: Fig. 17). As regards Margiana, we can mention as an example the specimen from Gonur 1 South published by Sarianidi (1981: Fig. 11). 12

Figure 7.12 Sherd impressed with a cylinder stamp-seal from Taip 1.

105

I. S. Masimov and S. Salvatori

Figure 7.13 Sherd impressed with stone and metal stamp-seals from Kelleli 3.

back, around the rectangular perforated handle, there is a crown of radial lines (Figure 7.4.4).

vessel walls, as V. I. Sarianidi (1998: 317-318) has aptly documented in his recent catalogue.

At Taip 8, on the surface of the remains of a kiln, a terracotta stamp-seal was collected. It has a high, sub-cylindrical perforated handle, and its functional face has a series of six cup holes arranged in a circle around a central one (Figure 7.9.2). This specimen, which has been already published (Masimov 1981: Fig. 1.3), is worthy of being pointed out because it has been made from figurine ware and because of the high quality of its firing.

Further examples of these impressions were found at Kelleli 3. On a fragment of a jar there are the repeated impressions of a metal closed-back compartmented seal and the motif of the corolla of a flower with circular pistil (Figure 7.13.1). The same impression reappears on a pottery fragment that probably belongs to the same jar (Figure 7.14.4). Another pottery fragment has the impression of a motif made of concentric circles, which do not allow us, however, to determine the shape of the seal that produced them (Figure 7.13.2). The double impression present on another ceramic fragment may be attributed to a round compartmented seal with cross motif and oblique lines starting from the segments of the cross and forming a sort of spider-web (Figure 7.13.3); this type of seal was well-known in Margiana (Gonur 1 North: Sarianidi 1990: Tab. XXXVI.1), in southern Bactria (Salvatori 1988: Fig. 54), and, in a simplified form, also in northern Bactria (Sapalli-depe: Askarov 1977: Tab. XLVI.9).

To conclude this presentation of unpublished seals, we should mention those coming from a last group of sites scattered in the area near the well of Adam Basan. Here, in addition to Bronze Age sites, there are several settlements that can be dated to the Early Iron Age and to the Achaemenid period. More than half a dozen seals were found in these sites: they are bronze and stone seals, and amulet seals, chiefly in serpentinite, and on the whole are not different from the ones we have just described (Figures 7.3.11, 7.5.1, 7.6.1), including the cross motif (Figure 7.4.2).

At Taip 1, three pottery fragments were found: the first one has the impression of a metal stepped lozenge (Figure 7.14.1); the second one has the impression of a corollashaped seal on the outside of the rim of a large vessel

In the settlements of the Murgab Delta, beside the seals, seal impressions were also found, both on cretulae and on 106

Unpublished Stamp-seals from the North-western Murgab Deltae

Figure 7.14 Sherds impressed with stone and metal stamp-seals from Taip 1.

(Figure 7.14.2); and the third one has a circular impression containing a motif formed of a dovetail element and a linear mark with a circular sign above it (Figure 7.14.3). Because of its size, it may be the impression of the base of a cylinder stamp-seal.

confirm some conclusions that have been drawn during the last few years by archaeological research and historical thought. The bronze compartmented seals coming from the Middle Bronze Age sites of the Kelleli area and other areas of the delta are closely connected with the production that is known in the piedmont of southern Turkmenistan, above all in the multilayered site of Altyn-depe. On the whole, however, the bronze production of Margiana in the Middle Bronze Age (Namazga V) differs from that of the piedmont because of the numerical weight of open work types in Margiana: at Altyn-depe, metal closed-back seals prevail16. On the other hand, the production of Margiana differs from that of southern Bactria because of the lower incidence, or even total absence, of anthropomorphic images or of mixed ones (man/woman and animal: depictions of human characters with the head of an animal, particularly of a bird of prey), depictions of male or female “deities” accompanied by wild or fantastic animals, or depictions of monkeys. The overall formal richness of the metal seal

Seal impressions are a category of finds that is comparatively widespread in Margiana. They are produced both with stamp-seals like the ones described above and with cylinder seals. Impressions of cylinder seals (or rather of cylinder stamp-seals) have been found at Taip 1 (Figure 7.12)14 and Gonur 1 South15. In the latter site, seal impressions were found not only on the walls of pottery vessels, but also on cretulae and bullae (Sarianidi 1998: 317-18; 1998b: Fig. 28). At the end of this presentation of mostly unpublished seals coming from the western part of the Murgab Delta, we can Masimov 1981: Fig. 10; Hiebert 1994b: Fig. 7. Hiebert 1994b: Fig. 4.32.3.

14

For the distribution of these two classes of seals, cf. Salvatori 2000.

15

16

107

I. S. Masimov and S. Salvatori

production of southern Bactria is quite evident, on the basis of the data available up to now. However, if we are to tackle the problem in an organic manner, we must wait for the systematic, complete publication of the excavations carried out by V. I. Sarianidi, from 1992 at Gonur 1 North, the largest of the Middle Bronze Age sites of Margiana, and from 1995 in the vast contemporary graveyard to the west of that settlement. Here Sarianidi also found some silver compartmented seals with complex images17. Only when the finds of these recent excavations have been published it will be possible to evaluate, in comparative, statistical and formal terms, the regional variability that at present seems to be clearly emerging.

The stone stamp-seals seem to have a similar evolution in Margiana and Bactria (northern and southern), both in their shapes and in their decorative motifs, perhaps indicating a greater integration of material culture during the Late and Final Bronze Age than during the Middle Bronze Age, period in which the regional variants within this vast territory of Central Asia were quite marked.

It is possible, in any case, to fix some certainties about the reality of the regionalisation processes in the production of the three great areas of diffusion of the Namazga V facies of the Middle Bronze Age. An element that clearly reveals the development of regionalisation phenomena is the production of female anthropomorphic figurines, which is shared by the piedmont of southern Turkmenistan and Margiana during the Middle Bronze Age (see Salvatori, Chapter 6).

One of the seals published in the catalogue, with the image of a horse, is probably even more recent, because it contains a line of Aramaic writing on the horse’s back (Sarianidi 1998: No. 1496).

In the catalogue published by V. I. Sarianidi, besides the Bronze Age seals, there appear, in a smaller, but significant, number, some specimens from a more recent period, i.e. the Iron Age, and in particular some seals from the Achaemenid period and the classical one.

As a rule, the hunting scenes with men on horseback may depend on Assyrian models, though in the Achaemenid sphere (Ibidem: Nos. 1482, 1486-1488). Side by side with the Iron Age seals published in Sarianidi’s book, we may mention a specimen of a cylindrical seal found during the American expedition led by R. Pumpelly (1908), who carried out an exploratory visit in Margiana at the beginning of the twentieth century. This seal was published much later by E. E. Kuz’mina (1971); it contains the image of a horseman with a long spear, who is piercing a wild animal that is attacking his horse.

These regions, particularly Margiana and Bactria, preserved, also during a later period (Late and Final Bronze Age) a very strong integration of material culture (P’yankova 1994), which includes not only some peculiar aspects of defensive architecture, but also the pottery production, and (more relevant to the topic of this essay) the glyptic production. The Late Bronze Age saw a radical change, both in Margiana and in Bactria, in the general situation of seals production. Metal seals practically disappeared, and the few specimens that were produced during this period reveal a radical technical change, with solid casting and chiselling of the decorative motifs. But what characterises the glyptic production is the generalised transition to stone seals, both for typical cylinder stamp-seals (Figures 7.11.1-4, 7.12) and for more widespread stamp-seals, both with single face and with two faces (Figures 7.4.1-7, 7.5.1-7, 7.6.1-8, 7.7.1-6, 7.8.1-7).

References Cited Amiet, P. 1986. L’âge des échanges inter-iraniens 35001700 avant J.-C.. ������ Paris. Askarov, A. A. 1977. Drevnezemledel’cheskaya kul’tura epokhi bronzy yuga Uzbekistana. Tashkent. Askarov, A. A. and K. Abdullaev 1983. Dzharkutan. Taskhent. Ferioli, P., E. Fiandra and S. Tusa 1979. “Stamp-seals and the Functional Analysis of their Sealings at Shahr-i Sokhta II-III (2700-2200 B.C.)”, in Van Lohouizen-De Leeuw (ed.) 1979: 7-28. Gafurov, B. G. (ed.) 1971. Istoriya iranskogo gosudarstva i kul’tury. Moskva. Hiebert, F. T. 1994a Origins of the Bronze Age Civilization in Central Asia. Cambridge. Hiebert, F. T. 1994b. ������������������������������������� “Production evidence for the origins of the Oxus Civilization”, Antiquity 68: 372-387. Jarrige, C. (ed.) 2005: South Asian Archaeology 2001. Paris. Joshi, J. P. and A. Parpola 1987. Corpus of Indus Seals and Inscriptions: I. Collections in India. ��������� Helsinki. Kircho, L. B. 2001. �������������������������������������� “Metallicheskie izdeliya Altyn-depe”, in Kircho (ed.) 2001: 60-84. Kircho, L. B. (ed.) 2001. Osobennosti proizvodstva poseleniya Altyn-depe ve epokhu paleometalla.

As far as we know, the cylinder stamp-seal18 is equally widespread in Margiana and in Bactria19, and strongly confirms the previously mentioned process of cultural convergence between the two regions. In this sphere, unfortunately, there are no data about a key region such as the piedmont of southern Turkmenistan, where the evidence relative to the Namazga VI period is too limited. Sarianidi, personal communication. With this name, we classify a particular type of cylinder seal that does not have a longitudinal through hole, as in the classical production, but a perforated handle on one of its ends, while on the other end there is an incised motif that was used for impressing, as well as the motif or scenes depicted on the cylindrical side of the object. 19 We must point out that many specimens of this type of seal are present among the material coming from northern Afganistan and preserved in hitherto unpublished collections. 17 18

108

Unpublished Stamp-seals from the North-western Murgab Deltae

Pumpelly, R. 1908. Exploration in Turkestan. Washington. Salvatori, S. 1988. “Early ����������������������������������� Bactrian objects in private collections”, in Ligabue and Salvatori (eds.) 1988: 181187. Salvatori, S. 2000. “Bactria and Margiana Seals. A New Assessment of their Chronological Position and a Typological Survey”, East and West 50: 97-145. Salvatori, S. and M. Tosi 2005. “Shahr-i ������������������������ Sokhta Revised Sequence”, in Jarrige (ed.) 2005: 281-292. Sarianidi, V. I. 1977. Drevne zemledel’zy Afganistana. Moskva. Sarianidi, V. I. 1981. ��������������������������������� “Margiana in the Bronze Age”, in Kohl (ed.) 1981: 221-255. Sarianidi, V. I. 1990. Drevnosti strany Margush. Ashkhabad. Sarianidi, V. I. 1994. “Margiana and the Ino-Iranian world“, in Parpola, ���������������������������������������������������� A. and P. Koskikallio (eds.) 1994: 667-680. Sarianidi, V. I. 1998. Myths of Ancient Bactria and Margiana on its seals and amulets. Moscow. Van Lohouizen-De Leeuw, J. E. (ed.) 1979. South Asian Archaeology 1975. Leiden.

(Materialy Yuzhno - Turkmenistanskoj arkheologicheskoj compleksnoj ekspedizii, vy. 5). Sankt-Peterburg. Kohl, Ph. L. (ed.) 1981. The Bronze Age Civilization of Central Asia. Recent Soviet Discoveries. New York. Kuz’mina, E. E. 1971. “Zilindricheskaya pechat’ iz Mervskogo oazisa so szenoj edinoborstva”, in Gafurov (ed.) 1971. Ligabue, G. and S. Salvatori (eds.) 1988. Bactria: an ancient oasis civilization from the sands of Afganistan. Venice. Masimov, I. S. 1981. “Novye nakhodki pechatej epokhi bronzy s nizovij Murgaba”, Sovetskaya Arkheologiya 2: 132-150. Masson, V. M. 1977. “Pechati protoindiskogo tipa iz Altyn-Depe”, Vestnik Drevnej Istorii 4: 147-155. Masson, V. M. 1988. Altyn-Depe. Philadelphia. Masson, V. M. 1996. Istoricheskie rekonstrukzii v arkheologii. ������� Samara. Parpola, A. and P. Koskikallio (eds.) 1994. South Asian Archaeology 1993. Helsinki. P’yankova, L. 1994. “Central Asia in the Bronze Age: Sedentary and Nomadic cultures”, Antiquity 68: 355372. Pittman, H. 1984. Art of the Bronze Age. Southeastern Iran, Western Central Asia, and the Indus Valley. New York.

109

Chapter 8 A New Cylinder Seal from Ancient Margiana: Cultural Exchange and Syncretism in a “World Wide Trade System” at the End of the 3rd Millennium BC Sandro Salvatori In October 2000, during the surveys performed within the Italian-Turkmen joint project “Archaeological Map of the Murgab Delta” (Figure 8.1), a cylinder seal in white stone with rust-coloured streaks, of exquisite workmanship (Figures 8.2-3), was found on the surface of Site No. 1220 (41N 413728, 4211999).

production of Mesopotamian glyptic, from the Early Dynastic II period through the period of the post-Sargonid Akkadian dynasty. When considered individually, the figures or pairs of animals or hybrids are known to have an even longer life in the Near East glyptic, but the scene depicted on the seal from Margiana can be most significantly compared with the archaic or archaising production of the area of Mesopotamia proper. The hero at the centre of the seal, who is protecting two bulls that are being attacked from the back by two lions appears on pre-Sargonid seals from Lagash (Figure 8.6) (Amiet 1961: Pls. 82 Nos. 1085, 1092, 1094; 83 No. 1102; 101 No. 1350.) and from the first dynasty of Ur (Amiet 1961: Pls. 80 No. 1064; 81 Nos. 1072, 1074.). On two seals of the first dynasty of Ur (Figure 8.7), there appears the image of the hero, who is protectively embracing two human headed bulls (Amiet 1961: Pl. 81 Nos. 1078, 1080). In one of these two seals, the scene that is depicted also includes the lions that are attacking, from the back, the human headed bulls, which, as in the specimen examined in this essay, turn their heads towards their aggressors. It is precisely this structural variant of the scene that is reproduced in the seal from Margiana, with the only difference that the lions, because of lack of space, are depicted in a chiastic position. However, this, too, is a solution that has been borrowed from the Mesopotamian glyptic (Figure 8.8), where it appears quite frequently in the seals of the Early Dynastic period and of the subsequent age, up to a late period, and where it is adopted in all the surrounding regions that have somehow assimilated the thematic and stylistic features of the carvers of the land between the two rivers (Amiet 1961: Pls. 75 Nos. 989, 990, 994, 995, 1000, 1001; 80 Nos. 1064, 1065 1066; 82 Nos. 1084, 1086 1088, 1090; 83 Nos. 1100, 1103, 1104; Frankfort 1939: Pls. XII. c; XIV.g; XVI.b, e; XXV.g.).

Site No. 1220 is located along the track that leads to Togolok 1, approximately 7 km south of this centre. It is a small settlement of about 0.3 hectares, with a low hill that implies the presence of a small farm. The ceramic materials that was recorded and partly collected on the surface of the site can be dated to the Late Bronze Age (Figure 8.4). Among the other items that were collected there, we should also point out a circular chlorite stamp seal with subtriangular rounded-out loop handle on the back (Figure 8.5.1), and a fragment of a biface circular chlorite seal, with outlined margin and, on one of its faces, the clearly recognisable motif of an eagle with spread wings (Figure 8.5.2). The cylinder seal examined in this essay has a motif that is frequently reproduced in the repertory of Mesopotamian glyptic, with which the carver was probably acquainted. It is the well-known motif in which at the centre there is a human figure, and by its sides there are two bull-men whose heads are turned back towards two chiastically arranged lions, who are attacking them from behind. The style of the carving is very accurate and shows a great technical skill both in the use of the drill (some of whose strokes have been deliberately not eliminated in the subsequent carving operation, in order to mark out the anatomical joints of the human figure and animals) and in the masterly depiction of the essential, connotative traits, e.g. the muscles and the animals’ fur.

Unfortunately, in the seal from Margiana the head of the hero or deity is much worn by use, more than other areas of the cylinder, and it is difficult to identify his physiognomy with certainty; in any case he wears a short beard and perhaps a braid that comes down behind his ear. It is not possible to determine whether he is wearing a hat, or, if he does, what type of hat it is.

It seems indubitable that the scene depicted here is directly inspired and consciously borrowed from the typical The project began in 1990, coordinated by M. Tosi (1990-1998) then by S. Salvatori (1999-2005), for the Istituto Italiano per l’Africa e l’Oriente, by A. G. Gubaev (1990-2002), for the State University of Turkmenistan, and O. Gundogdiyev (2005) for The State Institute of Cultural Heritage of the Peoples of Turkmenistan, Central Asia and the Orient under the President of Turkmenistan.  H. 3.15 cm; Ø 1.6 cm; Ø hole (bipolar) 0.5/0.45 cm.  For a review of the researches on the cultures of Margiana in the Bronze Age see: Hiebert 1994; Gubaev et al. (eds.) 1998. 

An important aspect is the fact that, unlike the figure on whom he is clearly modelled in the context of the composition, and who is always portrayed heroically in the nude in the Mesopotamian production (not only of 111

S. Salvatorii

Figure 8.1 Map of southern Central Asia.

the Early Dynastic period), in the seal from Margiana the hero is wearing a long flounced skirt with a high studded belt that emphasises the slimness of his waistline and the trapezoidal shape of his bust. The flounced skirt appeared in the Akkadian period, and, as far as we can deduce from the glyptic of that period, was worn, often with a belt, by bare-chested male deities. The flounced dress, in several forms, but always tending to cover at least one of the wearer’s shoulders, went on being used for a long time after this period, with stylistic variants that do not resemble our specimen.

almost as if the carver wished somehow to emphasise his artistic independence from the reference prototypes. And this independence, considering the indubitable qualities of the carver, can only be meant to declare his identification with a specific artistic and cultural tradition. In Mesopotamia, the hero and bull-man were always depicted not only in front view, but also with a long, curly beard. On the seal from Margiana, on the contrary, the human headed bulls have a short beard and a braid of hair that comes down their face, behind their ear. Precisely this particular representation of the bull-man’s head leads us to the ambience of Bactria-Margiana, where we find this stylistic and formal feature in the rows of sitting figures within the agricultural and convivial scenes depicted on embossed and chiselled silver vessels coming from the Middle Bronze Age graveyards looted in the nineteenseventies in northernAfganistan (Figures 8.9-10). However, we must also mention an iconographic relationship with the depiction of a bull-man or of a lion with human head, as Pittman suggests (Pittman 2001: 239 and Fig. 10.57.), on a well-known stone stamp seal from Tepe Yahya: here we find the braid behind the ear, the lozenge-shaped eye with rounded corners and the muscles of the animal’s thigh represented with a similar characterisation.

Another significant difference lies in the fact that all the heads, both of human figures and of animals, are shown in profile, in decided contrast with the style of the Mesopotamian production of the Early Dynastic period (Frankfort 1939: Pls. XII-XIV.): here the faces of the bearded hero and of the human headed bull were always shown from the front. In similar scenes of attack, the heads of the lions were mostly shown in three-quarter view, so as to ensure that both eyes were clearly visible. During the Sargonid period, lions attacking their prey were more and more frequently depicted in profile, though, at the same time, the mode that was in vogue during the Early Dynastic III period also persisted. During the period of the Akkadian dynasty, the representation in profile apparently became, once and for all, the favourite one. In our case, as we have mentioned above, the heads are all in profile,

The seal from Margiana is therefore a further testimony of the strong interrelationship between the Bronze Age civilisation of Bactria-Margiana and the adjacent Near Eastern area, perhaps via the Iranian plateau, as many other clues lead us to believe is quite probable (Amiet 1986; Winkelmann 2000: 43-95). It reveals to us, once again, how the ancient world, with a highly adaptative attitude,

The motif of the lion attacking the bull-man appears also on a famous seal impression from Susa (Amiet 1972: No. 1363), which, from an iconographical point of view, follows the Mesopotamian model of the Early Dynastic III period rather closely, also in some apparently secondary stylistic features (cf. Frankfort 1939: Pl. XII a, b). 

112

A New Cylinder Seal from Ancient Margiana

Figure 8.2a Cylinder seal from Site No. 1220.

Figure 8.2b Cylinder seal from Site No. 1220.

used the systems of signs, including the figurative ones, to achieve syncretic formulations that somehow acted as a bond and as an exchange language in long-distance political interrelationships.

a carver has made, on a figurative design clearly derived from Mesopotamia, some significant changes that reflect the political and religious ideological system of the Syrian centre (Matthiae 1979: 26-29; 1984: Pls. 25-26; Matthiae et al. 1995: Figs. 222-231).

The prolonged work of adapting motifs mostly originating in the Mesopotamian sphere has been thoroughly investigated in some important essays by Sylvia Winkelmann (Winkelmann 1997: 135-146; 2000: 4395), so it is not worth-while to return to this subject here, except to state that the seal under examination is also a part of this complex system of “image mirroring” in which the shapes, symbols, themes and deities are “distorted” in order to adapt them to the cultural sphere in which they are functionally accepted and often absorbed through the powerful filter of art and religion (Amiet 1976: 15-32). High-level expressions of this phenomenon are known to us, for instance, through the Palatine cretulae of Ebla (Ancient Bronze Age IVA: about 2350-2300 BC), where

We must point out that in Site No 1220, on whose surface the cylinder seal under examination was collected, we were not able to identify any trace of diagnostic ceramic material datable to the Middle Bronze Age. The primary context of this small settlement, with reference to the surface material, undoubtedly dates from the Late Bronze Age; but the cylinder seal presented here does not fall into the Late Bronze Age of Bactria-Margiana, because of the scene that is depicted, the style and quality of the carving, and the care with which the anatomical details of the figures are rendered.

113

S. Salvatorii

Figure 8.2c Cylinder seal from Site No 1220.

Figure 8.2d Cylinder seal from Site No 1220.

Cylinder seals were found in the region of BactriaMargiana, though in a very limited number, only in contexts dating from the Middle Bronze Age; they should be separated, typologically and chronologically, from the cylinder stamp‑seals, which seemed to have a rather short life span in an early period of the Late Bronze Age (Salvatori 2000: 133). In Margiana, in particular, as regards cylinder seals proper, one complete specimen and another fragmentary one were collected on the surface of Togolok 1 (Sarianidi 1990: Tab. LXXXV.7, 12). As far as I know, only very unsatisfactory photographs have been published of these specimens, both of the originals and of their impressions, and no drawing of the impression has been published, so it is problematic to attempt a stylistic and

iconographical interpretation of them. Other three cylinder seals were found, in this case too by V. Sarianidi, in the graveyard of Gonur 1 North: two from graves and one on the surface. The first one (Sarianidi 2001: 197 and Pl. 10.7; 2002: 333-335 and Fig. on p. 326. ) is quite interesting, because it is a seal with a cuneiform inscription, in postSargonid, Akkadian provincial style, presumably Iranian. The presence of an object like this in a grave in the Middle Bronze Age graveyard of the large protourban centre of Gonur 1 North irrefutably confirms the strong bond between the civilisation of Bactria-Margiana and the Iranian world during the second half of the third millennium BC. The other two seals (Sarianidi 1998: Fig. 27.3; 2002: Fig. on p. 278.) may be easily included in a series of cylinder

A list of these seals, though incomplete, can be found in Salvatori 2000: 134 (with the warning that in the References “Amiet 1989” must be corrected to “ Sarianidi 1989: Fig. 12.5”).





A preliminary reading of the inscription was given to Sarianidi by T. Sharlach (Sarianidi 2001: 334).

114

A New Cylinder Seal from Ancient Margiana

Figure 8.3 Artist rendering of the cylinder seal from Site No 1220.

Figure 8.5 Stamp-seals from Site No 1220.

Figure 8.4 Late Bronze Age pottery from Site No 1220.

Figure 8.6 Cylinder seal from Lagash (from Amiet 1961: No. 1350).

115

S. Salvatorii merchants; this, by itself, may justify the presence of such a broad range of very specific objects of undoubtedly allochthonous origin (Salvatori 1995: 38-55; 1998: 4755.). The system of staples or trading agencies of the third millennium BC, probably similar to the Assyrian karums in Anatolia or to the village of Meluhha at Ur (Parpola et al. 1977: 129-165.), so far has not been documented archaeologically as the slightly later Anatolian system, but merchant trading is well attested over the entire span of that millennium, by the cuneiform sources both from Syria (Pettinato 1986: 177.) and from Mesopotamia (Pettinato 1972: 43-166.).

Figure 8.7 Cylinder seal of the first Ur dynasty (from Amiet 1961: no. 1378).

So what seems to bring our seal (which certainly cannot be placed in a period too remote from the Early Dynastic III prototypes on which the overall structure of the composition depends) into an Akkadian chronological sphere is not only the presence of specific iconographic elements, such as the abandon of the three-quarter position of the lions’ heads and the adoption of flounced skirts, but also the broader context in which there took place the great diffusion of Mesopotamian and south-western Iranian motifs over the entire Iranian plateau and in the regions of Bactria and Margiana.

seals with very similar designs, formed, at present, of two specimens from the graveyard of Shahdad (Hakemi 1997: 320 obj No. 1792, 324 obj. No. 1824.), five specimens from Tepe Yahya (Potts 1981: Figs. 1-2; Amiet 1986: Fig. 132.1, 3, 5, 6, 9.), one from Susa (Winkelmann 1997: Fig. 1.g; Amiet 1994: Fig. 3.), one from the Treasure of Tôd (Amiet 1986: Fig. 132.11.), in Egypt, and at least four in private collections (Winkelmann 1997: Fig. 1.c; Amiet 1986: Fig. 132.2, 10; Porada 1993: Pl. 31). The specimens of this series that come from stratigraphically verified contexts can be ascribed to the Akkadian, probably postSargonid, period (Potts 1981: 135-142; Amiet 1986: 165 and following; Winkelmann 1997: 137-138.). As demonstrated by the distribution of this series of seals and by numerous other pieces of evidence of different nature, including the rich documentation of Central Asian elements present in the graveyard of Shahdad (Hakemi 1997; Salvatori and Tosi 1997: 121-138.), during the second half of the 3rd millennium BC a strong integration was taking place, probably in connection with the presence of an intense and complex “international” system of longdistance exchanges, that saw an intensification of contacts and cultural exchanges between the Iranian world (Hissar, Khinaman, Shahdad, Tepe Yahya and Susa), Central Asia (piedmont of southern Turkmenistan, Bactria and Margiana), and the Indus Valley (Amiet 1986; Salvatori 1995: 38-55; 2000; Winkelmann 2000.).

References Cited Amiet, P. 1961. La glyptique mésopotamienne archaique. Paris. Amiet, P. 1972. Glyptique susienne. Paris. Amiet, P. 1976. “Introduction à l’étude archéologique du Panthéon systématique et des Panthéons locaux dans l’Ancien orient”, Orientalia 45:15-32. Amiet, P. 1986. L’âge des échanges inter-iraniens: 35001700 avant J.-C. Paris. Amiet, P. 1994. “Un sceau trans-élamite a Suse”, Revue d’Assyriologie 88 (1): 1-4. Charpin, D. and F. Joannes (eds.) 1992. La circulation des biens, de personnes et des idèes dans le proche- Orient ancien, (XXXVIIIe R.A.I.). Paris. Crawford, H. 1992. “En Early Dynastic Trading Network in North Mesopotamia“ in Charpin and Jeannes (eds.) 1992: 77-82. Curtis, J. (ed.) 1993. Early Mesopotamia and Iran. London. Frankfort, H. 1939. Cylinder Seals. A Documentary Essay on the Art and Religion of the Ancient Near East. London. Gubaev, A., G. A. Koshelenko and M. Tosi (eds.) 1998. The Archaeological Map of the Murgab Delta. Preliminary Reports 1990-95. Reports and Memoirs. Sewries Minor III. Rome. Hakemi, A. 1997. Shahdad. Archaeological Excavations of a Bronze Age Center in Iran. Rome. Hiebert, F. 1994. Origins of the Bronze Age Oasis Civilization in Central Asia. Cambridge.

In our opinion, the cylinder seal presented here fits in this context: its regional re-elaboration of themes and personages undoubtedly of Mesopotamian origin reveals a strong tendency of the culture that has produced it to assimilate, through skilful synthetic manipulations, the themes and personages of a ideological and religious system that was allochthonous, but undoubtedly wellknown, probably through a direct contact, like the one promoted by the trading settlements that, we know, were active at least since the middle of the 3rd millennium BC (Parpola et al. 1977: 129-165:. Crawford 1992: 77-82.). In other works, I have already put forward the hypothesis that at Shahdad there was an entrepôt of Central Asian 116

A New Cylinder Seal from Ancient Margiana

Figure 8.8 Cylinder seal from Lagash (from Amiet 1961: No. 1100).

Figure 8.9 Artist rendering of the decoration on a Bactrian silver vessel (private collection).

Figure 8.10 Artist rendering of the decoration on a Bactrian silver vessel at Louvre Museum (AO 28518).

117

S. Salvatorii Lamberg-Karlovsky, C. C. and D. T. Potts (eds.) 2001. Excavations at Tepe Yahya, Iran 1967-1975. The Third Millennium. Cambridge. Matthiae, P. 1979. “Appunti di iconografia eblaita, I”, Studi Eblaiti I/2; 17-31. Matthiae, P. 1984. I tesori di Ebla. Bari. Matthiae, P., F. Pinnock and G. Scandone Matthiae (eds.) 1995. Ebla. Alle origini della civiltà urbana. Trent’anni di scavi in Siria dell’Università di Roma “La Sapienza”. Milano. Parpola, S., A. Parpola and R. H. Brunswig, Jr. 1977. “The Meluhha Village. Evidence of Acculturation of Harappan Traders in the Late Third Millennium Mesopotamia?”, Journal of the Economic and Social History of the Orient XX (II): 129-165. Pettinato, G. 1972. “Il commercio con l’estero della Mesopotamia meridionale nel 3. Millennio av. Cr. alla luce delle fonti letterarie e lessicali sumeriche”, Mesopotamia VII: 43-166. Pettinato, G. 1986. Ebla. Nuovi orizzonti della storia. Milano. Pittman, H. 2001. “Glyptic Art of Period IV”, in LambergKarlovsky and Potts (eds.) 2001: 230-268. Porada, E. 1993. “Seals and Related Objects from Early Mesopotamia and Iran”, in Curtis (ed.) 1993: 44-53. Potts, D. T. 1981. “Echoes of Mesopotamian Divinity on a Cylinder Seal from South-eastern Iran”, Revue d’Assyriologie 75: 135-142.

Salvatori, S. 1995. “Protohistoric Margiana on a Recent Contribution. Review of IASCCA (International Association for the Study of the Cultures of Central Asia) Information Bulletin, 19, Moscow”, Rivista di Archeologia XIX: 38-55. Salvatori, S. 1998. “The Bronze Age in Margiana”, in Gubaev, Koshelenko and Tosi (eds.) 1998: 47-55. Salvatori, S. 2000. “Bactria and Margiana Seals. A New Assessment of their Chronological Position and a Typological Survey”, East and West 50: 97-145. Salvatori, S. and M. Tosi 1997. “Postscriptum. Some Reflections on Shahdad and its Place in the Bronze Age of Middle Asia”, in Hakemi 1997: 121-138. Sarianidi, V. I. 1990. Drevnosti strany Margush. Ashkhabad. Sarianidi, V. I. 1998. Margiana and Protozoroastrism. Athens. Sarianidi, V. I. 2001. Necropolis of Gonur and Iranian Paganism. Moscow. Sarianidi, V. I. 2002. Margush. Ancient Oriental Kingdom in the Old Delta of the Murgab River. Ashgabat. Winkelmann, S. 1997. “Ein neues trans-elamisches Siegel”, Archäologische Mitteilungen aus Iran und Turan 29: 135-146. Winkelmann, S. 2000. “Intercultural Relations between Iran, the Murgabo-Bactrian Archaeological Complex (BMAC), Northwest India and Failaka in the Field of Seals”, East and West 50: 43-95.

118

Chapter 9 Excavations at Sites No. 1211 and No. 1219 (Final Bronze Age) Maurizio Cattani

Palaeo-environmental Context of the Murgab Delta in the Bronze Age and the Definition of its Cultural Aspects

The surface evidence reveals the co-presence of wheelmade pottery that can generally be dated to the Namazga VI period (Late and Final Bronze Age), and of coarse, hand-made pottery, typical of the regions of the Central Asian steppes, in many cases decorated with lines and other geometric motifs, incised or impressed (the last-named pottery is classified here as ICW, Incised Coarse Ware). This site is outstandingly important for understanding the cultural dynamics during the final stages of the Bronze Age, when the nomadic populations that were bearers of ICW ceramics were apparently interacting with the sedentary populations of the Murgab Delta. The site actually displays all the characteristics of a nomadic camp with permanent installations, and contains a great quantity of nomadic pottery, located at the margins of the areas of ancient cultivation. The nearest settlements that can be attributed to sedentary farmers are within a radius of 4 to 6 km, while the largest site in the area, Takhirbai 3, is 8 km to the north-east..

The analysis of the geomorphological and palaeoenvironmental context of the eastern area of the Murgab Delta led to further, more in-depth investigations into the area around Takhirbai; in particular, following the traces of the palaeo-channels, these endeavoured to reconstruct the landscape and settlement of this area during the Bronze Age. During the surface surveys carried out in 2000, and in particular during the execution of a walking transect between the vast Site No. 212, Iron Age settlement area, and the new Takhirbai Canal (see Chapter 1), two great areas were identified that were characterised by an abundant presence of incised coarse ware pottery and of wheel-made pottery belonging to the final stages of the Bronze Age; these two areas were called respectively Site No. 1211 and Site No. 1219. The sites (Site No. 1211: GPS coordinates UTM 41N 427029, 4205612; Site No. 1219: GPS coordinates UTM 41N 427029, 4205612) are located a few kilometres to the north-west of the large Parthian-Sasanian fortress of Garry Kishman, in a predominantly desert area that has recently undergone agricultural exploitation. It is worth pointing out that these are the southernmost Bronze Age sites that have been located as yet in the Murgab Delta. Moreover, the combined fact that steppe ceramics were present and that the archaeological stratigraphy was well preserved reinforced the high interest in the site and resulted in an intensification of the researches.

During the campaign of 2000, a preliminary surface investigation was carried out in Site No. 1211, with an instrumental topographical survey, a collection of scattered pottery items, and a systematic collection on 1 x 1 m squares in the areas of greater density of finds. A small trial trench (4 x 3 m) was also made, in order to verify the conservation of the archaeological deposit and to structure the stratigraphic exploration of the site. The trial trench was made in an area that was particularly interesting because of the presence of surface evidence consisting of fired-clay structures and great quantities of ICW pottery. The wide-range surface collection enabled us to locate a second area, apparently separated from the first one: it was given an independent number, Site No. 1219. However, a more accurate analysis revealed that both the areas probably formed a single settlement. Farmed fields have now divided it, modifying the visibility of the surface and interrupting the continuity of the materials. More specifically, after having noticed that the farmed fields are located in a low-lying area along the south-north axis, i.e. in the direction of the slope, we surmised that they corresponded to a natural palaeo-channel, datable, perhaps, to the Bronze Age, and that the two settled areas, Site No. 1211 (7100 sq m to the east) and Site No. 1219 (5800 sq m to the west) were located on its banks (Figure 9.1). The archaeological areas are both situated on the top of ancient fluvial hummocks, where they were safe from possible floods. An artificial canal, which is connected to

The purpose of the surface survey transect was to explore an area that had not been examined previously, located between the traces of two palaeo-channels, of which the eastern one probably dates from the Achaemenid era and the western one is evidently connected to the Bronze Age sites. The transect started from the eastern margin of the Murgab Delta, at the latitude of the Achaemenid Age settlements (Yaz III or Late Iron Age), Sites No. 211, 212 and 213, and of depe No. 173 and 174, and crossed a expansive area characterised by takyr and by low sand dunes that interrupt the visibility of the ancient ground. Beyond this area, which revealed a low presence of sparse pottery fragments in the remaining takyr, proceeding westwards, the territory appears to be less encroached upon by aeolian sand: in this interface are Sites No. 12111219, which probably date from the Bronze Age. 119

M. Cattani by the nomadic peoples whose archaeological traces had been found chiefly in the sand-covered areas around the fields tilled by the sedentary farmers. In this case, we can point out that there probably were other situations in which nomadic populations could live both in farmed areas and in peripheral areas involved in the incipient desertification process. During the campaigns of 2001 and 2002, the area of the archaeological test was extended, and the stratigraphic sequence was explored down to virgin soil. The main results were the discovery of a large area with structures for food storage, and the collection of a great quantity of ICW pottery further attesting to the occupancy of the site by the nomadic community. The excavation made it possible to unearth a great number of vessels, some of them in wheel-made ceramics, others in rough ceramics with incised decorations, and others in baked clay. Three of the wheel-made pottery vessels contained at least three different types of cereals (triticum, hordeum and an unidentified one) that will have to be subjected to laboratory analyses. These vessels appear to have undergone burning, probably in a fire. It seems likely that the area belonged to a structure for grain storage and processing. The presence of a millstone confirms the grain processing activity.

Figure 9.1 Layout of Sites No. 1211 - 1219. Topographical survey by G. L. Bonora with some integrations by M. Cattani.

Particularly interesting is the abundant decorated pottery found in the site. There are strong resemblances, both in the shapes and in the decorative motifs, with other finds in the Murgab Delta and with the pottery collected in similar settlements in the area of Gudzhaili, which corresponds to the southern stretch of the Makhandar’ya River, north of the Zaman-Baba Lake (Bukhara, Uzbekistan): here the remains of the Zaman-Baba Culture were found together with Late Andronovo settlements (Gulyamov et al. 1966), which are identified as the Tazabag’yab variant, regarded at present as an autonomous cultural aspect (Avanesova, personal communication).

this palaeo-channel but whose use has not been ascertained yet, was found in the southern area of Site No. 1211, during the stratigraphic checks carried out along the section of some small modern canals. It is likely that this canal had a water-controlling function, presumably in connection with the need to irrigate the cultivated fields. The plan for the future continuation of the research includes the digging of a transversal trench in order to verify the chronology and characteristics of the palaeo-channel. In Site No. 1211 a systematic collection was done on 1 x 1 m squares within a central area of 10 x 10 m, and in a north-south transect in which an effort was made to locate the margins of the settlement.

Surface Analysis and Topographical Notes within the Site

The presence of recent irrigation canals, moreover, made it possible to locate some Bronze Age structures along the sections of the deposit. The stratigraphic section that was thus cleaned revealed, starting from its base, a deposit of sand, covered by an anthropogenic deposit, 30 to 40 cm thick, containing Bronze Age pottery and some evidence of clay fireplace structures; all this was covered by an aeolian deposit approximately 10 or 20 cm thick. This sequence confirms the reconstruction proposed by M. Cremaschi (1998: 19), in which an environment that was suitable for farming during the Bronze Age was gradually covered with sand during the subsequent stages of the Iron Age.

After a preliminary surface inspection in the area where pottery and other archaeological remains (fired-clay fragments and millstones) were scattered, the site was marked off and described. Since recent farming works had considerably changed the original situation, we endeavoured to identify the area that was still intact, in order to be able to carry out a systematic surface collection. We thus located an area that was slightly sloping to the south-west, and was bounded to the east by a sand dune and to the west by tilled fields. We divided this area into a grid of 1 x 1 m squares and a square of 10 x 10 m. The elements of the grid were indicated with an alphanumerical code (letters on the north-south axis and numbers on the

The stratigraphic sequence at Site No. 1211 is particularly significant for the identification of the ecozones occupied 120

Excavations at Sites No. 1211 and No. 1219 (Final Bronze Age)

Figure 9.2 Site No. 1211. Surface collection of pottery fragments. Central area with percentage of wheel-made and hand-made pottery.

Figure 9.3 Site No. 1211. Surface collection of pottery fragments. Symbol graduated distribution of wheel made pottery.

east-west one), over a total area of 70 x 20 m (Figures 9.2-3).

The examination of the percentages of hand-made and wheel-made pottery allowed us to identify two areas where there is a higher density of the former type, though superimposed on a substratum of wheel-made pottery. This fact may be interpreted as a sign of the presence of nomadic-type structures, and led us to carry out a trial excavation.

The surface collection was carried out by picking up, square by square, all pottery fragments whose size exceeded 1 x 1 cm, identifying their type (WM = wheel-made, HM = hand-made) and counting the fragments of each type. The diagnostic fragments (rims, bases, decorated walls) were recovered and marked with the year, site, square code and provenance (surface). The same criteria were followed along the two lines, to the north and to the south of the central square.

The Archaeological Excavation On the basis of the results of the surface collection, an area was selected for a trial excavation whose purpose was to detect the possible presence of structures and to verify the substantiality of the archaeological deposit. In 2000 we had limited ourselves to simply removing the most superficial level in squares A-C 4-6. During the subsequent campaign, the excavation proceeded until the sterile layer was reached, and in 2002 the excavation was extended to an area of 55 sq m. A further trench was opened in squares A8 – A16 in order to examine the stratigraphy present under the surface level of aeolian sand (Figure 9.4). In the eastern part, the anthropogenic deposit turned out to be covered with a layer of clayey silt: in other words there was a takyr formation with scanty anthropogenic elements, quite similar to SU 1.

The systematic surface collection was quite useful for defining the limits of the settlement and for identifying the areas where trial excavations, if any, were to be done. Though the surface of the site is partly covered with sand, we were able to locate the southern limit of the site (Squares AD2 to AC8 and A14). Along this axis, we found only some traces concerning a small oven and some hearths. The northern limit, on the contrary, is not clear. Several areas with a high density of pottery concentration were found up to the margin of Site No. 1219. The western and eastern limits turned out to be indefinite because of the presence, respectively, of farmed fields and sand dunes. 121

M. Cattani

Figure 9.4 Site No. 1211. Test trench excavation. Plan with main finds.

122

Excavations at Sites No. 1211 and No. 1219 (Final Bronze Age) SU 13. Lenticular level formed of light-grey powdery sand with sparse potsherds and animal bones. Maximum thickness 5 cm. ������������ Ash deposit.

The organic-sediment level also yielded a small number of pottery fragments, leading us to believe that the trench was in a marginal area with respect to the central structures.

SU 14. ���������������������������������� Clay level. Covers SU 4 and SU 13.

Stratigraphic Units Detected:

SU 15. Clay level, hardened by fire, under Vessel 4.

SU 1. Horizontal level of sandy clay, compact and light brown. Maximum thickness: 10 cm. It contains a few minute potsherds. It can be interpreted as an alluvial level or takyr formation that covers the Bronze Age structures. Covers SU 2 and 3.

SU 16. Filling level formed of crumbly clay. Present in square AD5. Fills SU 17 post-hole.

SU 2. Horizontal layer of silt, rich in organic material, dark grey, with thickness ranging from 2-3 cm to 15 cm. The soil is crumbly and breaks up into small clods. This ground contains a great number of vessels and clay containers placed in small pits. SU 2 can be identified as a stage of occupation with repeated human activities. This level is partly interrupted in squares AA4 to AA6. It covers SU 3, 4, 6, 13 and 14. SU 3. Silt level, light brown. It corresponds to virgin soil. The upper surface of the level is partly sloping in a westto-east direction. Not excavated. SU 4. Carbonaceous sand level, rich in organic material. Soft, powdery texture. Found in squares AB4-5, AC3-5 and AD4-5. It fills the bottom of the pits where the vessels have been placed, covers SU 7, 9, 11 and 13, and is very similar to SU 6. SU 5. Silt level, brown. Present in squares A7-8. SU 6. Carbonaceous sand level, rich in organic material. Soft, powdery texture. Present in squares A-B 4-5. Covers SU 3. SU 7. Filling level formed of crumbly silt with a few fragments of ICW pottery. Present in squares AD4-5. Fills the SU 8 post-hole and covers SU 23. SU 8. Post-hole, cylindrical and with flat bottom. Diameter 33 cm, depth 40 cm. Filled by SU 7, 23 and 24. Cuts through SU 3. SU 9. Crumbly silt filling level. Present in square AC4. Fills SU 10 post-hole. SU 10. Post-hole, conical. Diameter 17 cm, depth 20 cm. Filled by SU 9. Cuts through SU 3. SU 11. Filling level formed of crumbly silt. Present in square AB3. Fills the SU 12 post-hole. SU 12. Post-hole, conical. Diameter 14 cm, depth 20 cm. Filled by SU 10. Cuts though SU 3. Figure 9.5 Site No. 1211. Stratigraphic matrix of test trench.

123

M. Cattani

Figure 9.6 Site No. 1211. Hand-made pottery with incised ornaments and wheel-made pottery.

SU 17. Post-hole, conical. Diameter 14 cm, depth 20 cm. Filled by SU 16. Cuts through SU 3.

The stratigraphic matrix (Figure 9.5) can be summarised in four archaeologically relevant phases: above the virgin soil (SU 3), the first evidence of occupation consists of some post-holes (SU 8, 10, 12, 17, 19 and 21) belonging to an architectural structure. Connected with these structural traces, there are some layers characterised by fire activity (SU 4, 6) and followed by the deposition of a great number of clay containers and ceramic vessels, covered with anthropogenic soil that is more rich in organic material (SU 2) or less rich (SU 5, 22). At the end of the period of use there follows a stage of abandonment with alluvial deposition (SU 1) and erosion processes on the surface.

SU 18. Filling level formed of crumbly silt. Present in square AD5. Fills SU 19 post-hole. SU 19. Post-hole, conical. Diameter 15 cm, depth 22 cm. Filled by SU 18. Cuts through SU 3. SU 20. Filling level formed of crumbly silt. Present in square AD5. Fills the SU 21 post-hole. SU 21. Post-hole, conical. Diameter 15 cm. depth 22 cm. Filled by SU 20. Cuts through SU 3.

In a delimited area, corresponding to squares A4, A5, B4, B5 and C4/5, partly covered or inserted in SU 2, we located and unearthed a great number of unfired and fired clay containers, and ceramic vessels (Figure 9.6). The former had been fired on the premises, because we noticed the typical progressive redness from the bottom to the outer surface. At least ten of these containers were situated in a narrow area and had partly collapsed on one another. A single phase was recognised, with the presence, along the northern margin, of a wheel-made vessel resting on other two containers. The latter type of container is easily

SU 22. Horizontal level of clayey silt, light brown with few pottery fragments. Present in squares A-B7-8. Interpreted as residual palaeosoil. SU 23. Filling level formed of coal and carbonaceous sand, under SU 7. Fills the SU 8 post-hole and covers SU 24. SU 24. Filling level formed of crumbly silt, with a low percentage of coal, under SU 23. Fills the SU 8 post-hole. 124

Excavations at Sites No. 1211 and No. 1219 (Final Bronze Age)

Figure 9.7 Site No. 1211. Stratigraphic section of artificial ditch located at the southern part of the site.

recognised by its carinated shape with truncated-coneshaped walls and restricted opening.

SU 104. Sand level, dark grey (10 YR 3/1), with abundance of charcoal.

Once SU 4 and SU 6 had been completely excavated, virgin soil appeared. It was characterised by very compact clayey silt with a shallow depression, where the SU 2 containers described above were previously situated.

SU 105. Sand level, light brown with sparse carbonaceous grains. SU 106. Sandy clay lens, yellow-red (5 YR 5/8). Clear-cut limit.

The stratigraphic sequence in the surrounding areas was explored by cleaning some sections that had been exposed by the digging of modern canals for irrigating the fields. The stratigraphy shows a sequence with virgin soil, chiefly composed of sandy sediment, covered by 30-40 cm of anthropogenic soil containing archaeological material, which in turn is covered by a layer of aeolian sand. At the southern limit of Site No. 1211, the presence of an ancient canal, south-east to north-west was detected. The canal is not visible on the surface, because it is covered by alluvial clayey sediment, very similar to those that form takyr. Only at a depth of 80 cm the thin alluvial level makes it possible to recognise the flow of water. The bottom was reached during the campaign of 2002, and the profile decidedly shows the artificial origin of the canal (Figure 9.7).

SU 107. Level of clay mixed with charcoal, grey-brown (10 YR 5/2). The lateral limits overlap with SU 106. Undefined lower limit. SU 108. Clay level, brown (10 YR 5/3). Clear-cut lower margin. SU 109. Sand level, yellow with light-grey spots (5 Y 7/2). Clear-cut lower margin. SU 110. Sand level, whitish (10 YR 8/1), formed by the decay of aquatic vegetation. Clear-cut lower margin. SU 111. Sand level, light brown.

The stratigraphic sequence of the deposits within the bed of the canal is described below:

SU 112. Sand level, light yellow-brown (10 YR 6/1). Clay lamination frequent in the central band.

SU 100. Aeolian sand level, yellowish brown (10 YR 5/4), compact, with sparse traces of vegetation at the top. The lower limit is jagged.

SU 113. Clay level, light grey (10 YR 7/1). SU 114. Sandy silt level, with numerous clayey clasts, carbonaceous lenses and grey-and-yellow mottled sediment crossed by clay lenses in the middle. Pale brown (10 YR 6/3).

SU 101. Sand level, light brown, compact, with rare carbonaceous grains. Partly pedogenized, with vertical cracks. The lower limit is clear-cut.

SU 115. Clay level, divided into two parts by a thin sand lens. Upper part light yellowish brown (10 YR 6/4), lower part brown (10 YR 5/3).

SU 102. Level of fired clay in little blocks (5 to 10 cm thick), reddish, with grey carbonaceous soil.

SU 116. Surface of canal cutting.

SU 103. Sand level, silty, grey (2.5 Y 5/0), rich in organic material. Preserves traces of residual humidity. 125

M. Cattani

Figure 9.8 Site No. 1219. Stratigraphic section of test trench 1.

SU 117. Ancient soil formed of a clayey deposit, olive grey (5 Y 4/2).

sandy soils that are now being tilled intensely with deep furrows for the cultivation of watermelons.

SU 118. Sand level, light yellow brown (2.5 Y 6/4). Presumably of alluvial origin.

This site was investigated during the campaigns of 2001 and 2002 with stratigraphic sondages, in order to ascertain the texture and conservation of the archaeological stratigraphy. In 2001 a sondage carried out on the northern limit of the clayey area had exposed a stratigraphy formed of a layer of silty sand, about 10 cm thick and rich in organic substances (chiefly ash), above a very compact clayey layer, which in turn was above several sterile sand layers. The anthropic layer was rather homogeneous and sparsely characterised by cultural remains, in terms both of structures and of finds. The results of the sondage were interpreted as evidence that the site was a peripheral area of the settlement.

SU 119. Silty sand level, hard, olive grey (2.5 Y 6/6). Excavation in the Late Bronze Age Built-up Area. Site No. 1219 The archaeological area of Site No. 1219 is situated a few tens of metres north-west from Site No. 1211, and together with it forms a single built-up area that covers a space of slightly more than one hectare. Farming activity, which has increased during the last few years, has partly disturbed the archaeological stratification and has unearthed, in the northeastern sector, some traces of dwelling structures, with abundant material culture remains that can be associated to the populations which borrowed the characteristic ICW pottery and to a predominantly pastoral economy that coexisted with the sedentary groups during the late and final stages of the Bronze Age in the Murgab Delta.

During the campaign of 2002, other two sondages (3-4) were performed within the clayey surface. Despite the fact that the surface was occupied by a great number of pottery fragments, the sondages uncovered a shallow layer with traces of fired-clay artefacts and a few ceramics, or the evidence of an apparently sterile clayey layer overlapping the sterile sand layers that had already been exposed during the sondage of 2001. This clayey layer may be interpreted as a natural platform created by the floods of a stream. The extensive clayey surface corresponds to an alluvial

Site No. 1219 is characterised by an extensive clayey surface, slightly elevated above the surrounding area, not occupied by cultivated areas, and surrounded by silty126

Excavations at Sites No. 1211 and No. 1219 (Final Bronze Age)

Figure 9.9 Site No. 1219. Plan of sunken dwelling.

Figure 9.10 Site No. 1219. Stratigraphic section of test trench 2 with sunken dwelling unit.

127

M. Cattani

Figure 9.11 Site No. 1219. View of clay steps during the excavation.

Figure 9.12 Site No. 1219. Fireplace with cicular shape.

128

Excavations at Sites No. 1211 and No. 1219 (Final Bronze Age)

Figure 9.13 Site No. 1219. Eastern oven, dated to the 2nd phase.

Sondage 2 unearthed the remains of a sunken dwelling unit, sub-rectangular with an area of 6 x 5.3 m and a depth of approximately 50 cm (Figures 9.9-10). The structure, which has been partly damaged by agricultural works, has vertical walls and a flat bottom with some small pits and post-holes. On its margin, an entrance point was identified: it consisted of a step formed of a clay block (Figure 9.11). The area was divided into squares oriented to 35° west, following the orientation of modern fields.

hummock with silty sediment covering more sandy layers. Presumably this is the evidence of the aggradation activity of a palaeo-channel existing before the Bronze Age, whose reduced fluvial activity was perhaps contemporaneous with the sites under investigation. Very different results were yielded by three sondages that were carried out over areas ranging from 1 x 2 m to 2 x 6 m in the currently farmed area to the east of the clayey platform, in zones characterised by a concentration of organic material formed of ash and carbonaceous material, and by ceramic finds, both wheel-made and hand-made. Sondages 1, 3 and 4 revealed a stratigraphy characterised by a sandy layer with areas of concentration of ash and ceramics that can be ascribed to combustion or dumping activities. In sondage 1 (Figure 9.8) there appeared a sequence of tabular, horizontal layers having a constant thickness of about 10-15 cm, formed of sand, silt and clayey silt, and alternating with each other. Only at the top of this sequence there was a circular depression (1.60 m in diameter and 30 cm thick) filled with soft sediment, rich in ash and an organic substance, with sparse potsherds distributed without any concentration. This layer is the evidence of human occupation.

The dwelling area is also characterised by a circular hearth formed of a rectangular-section clay curb that encloses a space with traces of fire-reddening (Figure 9.12) and two ovens, presumably relative to two distinct stages, located at the corners of the structure (Figure 9.13). The ovens consist of fired-clay walls with an abundant filling of silt and charcoal. The excavation stages were made difficult by the agricultural work that had disrupted the surface by ploughing it and had removed the deposit up to the sterile ground, where two deep furrows had been made in order to provide drainage channels. Since the agricultural work was bound to periodically worsen the state of conservation of the site, it had a forcing effect on the course of the excavation, which had to be completed within the end of the campaign.

Sondage 4 (3 x 2 m) revealed stratigraphy with a 30-cmthick layer, very rich in finds (pottery, bones, fired clay) that extended horizontally beyond the limits of the sondage. 129

M. Cattani

Figure 9.14 Site No. 1219. Coarse ware found above the floor.

After the removal of the superficial layer of soft sandy silt that had been ploughed and already damaged by the agricultural work, and was very rich in ash and organic substances, there appeared a layer of aeolian sand (SU 1), limited to the central area of the structure. The layer is almost sterile, because it contains only small fragments of wheel-made pottery, presumably in a secondary position.

SU 3 Grey level (5 Y 6/0) and is chiefly formed of dusty sand, rich in ash, and containing a small amount of charcoal bits and potsherds. Beneath SU 3 there appears a layer that is very rich in an organic substances, with abundant charcoal bits and a great number of potsherds (SU 4), of a very dark shade of grey (2,5 Y 3/0). This is the archaeological evidence that corresponds to the degradation of the last stage of use of the structure; it is 20 cm thick, has a horizontal configuration near the margin of the structure, and starts sloping towards the centre, thinning down to a thickness of 8 cm; its colour fades and merges with that of SU 3 in the northern area.

In the northern part, the sand layer contains sandy-lime lenses (SU 2) whose formation is similar to that of takyr, i.e. resulting from the standing water on a horizontal surface. Beneath SU 1, in the southern part, there is SU 3, which outcrops from under the sowable ground near the margin of the structure. This stratigraphic differentiation, which was revealed after the removal of the sowable layer, indicates that at the centre of the structure there was a depression filled with aeolian sand and characterised by water formed phenomena. This behaviour is typical of sunken structures: after they have been abandoned they fill up, around their external margins, with wall-collapse materials and organic remains, while at the centre a depression is created by the compression of the layers and the fact that there are less deposits.

Interposed between SU 3 and SU 4, there is a clay lens (SU 19), slanting from the margin of the structure to the inside and limited only to the western part. This SU, too, is a result of standing rain-water after the abandonment of the structure. The stratigraphic sequence under SU 4 is more diversified, and contains units that are not uniformly distributed on the entire surface of the structure, but that correspond to structural elements and remains of human activities or accumulation.

130

Excavations at Sites No. 1211 and No. 1219 (Final Bronze Age) SU 13. Silt layer with scattered charcoal, limited to the southern part of the structure: its thickness is 20 cm and its colour grey (5 Y 6/0);

The sequence outside of the structure shows a superimposition of alternating silt layers with a higher or lower content of sand or clay.

SU 14. Thin layer of clayey silt, presumably corresponding to a renovation of the floor; it is gradually scattered and broken up into small fragments in the northern part of the structure.

SU 9. Clayey silt, approximately 20 cm thick. SU 8. Sandy silt, approximately 20 cm thick. SU 10. Clayey soil, dark greyish brown (10 YR 4/2).

SU 15. It is very rich in an organic substance and contains a great number of potsherds and dark-grey charcoal (2,5 Y 4/0). This layer, too, is approximately 25 cm thick at its southern margin and becomes thinner towards the centre of the structure, while the intensity of its colour and its anthropic characteristics diminish as well. It is evident that the compression and wash away effect due to the percolation of rainwater at the centre of the structure has geometrically and chemically transformed the characteristics of the layer, which presumably corresponds to the first stage of occupation of the structure.

SU 11. Silty sand, light yellowish brown (2,5 Y 6/4), 20 cm thick. SU 12. Olive-coloured sand (5 Y 5/4), 20 cm thick, but not excavated entirely. The materials found in the stratigraphic units include a great number of fired-clay blocks, some small potsherds and many bones. On the bottom of the structure there were the fragments of a single cooking pan (Figure 9.14) and other fragments of hand-made coarse ware. Among these there are numerous fragments with incised decoration.

SU 16. It was detected under SU 15, only in the central part of the structure. It contains abundant charcoal and potsherds in a primary position; together with the previous stratigraphic unit, it forms the most direct evidence of the first stage of the life of the structure.

The sunken dwellings were used since the Chalcolithic Age, as the discovery of Botai shows (Levine et al. 1999), with surfaces of 30 to 70 sq m. It is worth noting that at Botai, too, there is the mechanism of formation of layers with a central depression, visible to this day on the surface.

SU 17. It is a layer of filling of hearth SU 20. It contains ash and carbonized sand.

What determines the use of this type of structure is usually the geological and sedimentologic condition of soils formed through aeolian deposition (loess) or fluvial deposition (alluvium), and almost always characterised a very limited presence of stones: in the Chalcolithic Age of Kazakhstan there are also other structural situations consisting of stone masonry, when the surrounding environment is rocky, and this is often associated with the choice of hills or slopes for settlement.

SU 18. Filling of contemporary pit SU 19. SU 19. Present days pit. SU 20. It is a hearth formed of a curb with a rectangular section and with a circular layout. SU 21. Oven situated in the south-western corner. It refers to the second stage of occupation. Largely destroyed in the upper part, it has a plan with oval shape and walls thick from 5 to 10 cm.

In steppe areas, but above all in alluvial plains, sunken dwellings are the structure most frequently adopted by nomadic or semi-nomadic populations. In Khorezmy, the researches carried out by Tolstov and Itina unearthed dozens of sunken dwellings, rectangular and with an entrance corridor (Itina 1977).

SU 22. Oven situated in the north-eastern corner. It refers to the first stage of occupation. Similar to the previous, with oval shape. SU 23. Clay steps. On the southern side of the dwelling along the vertical bank there are two steps cut in the virgin soil to be interpreted as a small stair to reach the floor. They have rectangular plan, thick from 20 to 30 cm.

In the Murgab Delta, this type of structure coexists with the building method most frequently adopted by the sedentary farmers’ communities, i.e. raised unfired-clay constructions. This architecture, based on the use of bricks or more generally adobe, attains a degree of complexity that creates real cities; it reappears also in lesser sites with small raised constructions.

SU 24. Fragment of a clay floor between SU 4 and SU 26. SU 25. Sandy layer

The exploration carried out so far refers mainly to the large built-up areas; the smaller sites have not been explored yet, but, thanks to the observations afforded by the surface investigations, it is possible to state that the unfired-clay

SU 26 = SU 15.

131

M. Cattani References Cited

building technique was reproduced everywhere in the farming communities. The small clay mounds detected during the surface survey are actually the remains of brick constructions that have collapsed and have been subsequently uniformed by post-depositional events.

Cremaschi, M. 1998. “Palaeohydrography ������������������������������ and Middle Holocene Desertification in the Northern Fringe of the Murgab Delta”, in Gubaev, Koshelenko and Tosi (eds.) 1998: 15-25. Gubaev A., G.A. Koshelenko and M. Tosi (eds.) 1998. The Archaeological Map of the Murgab Delta. Preliminary Reports 1990-95. Reports and Memoirs. Series Minor, III. IsIAO, Rome Gulyamov, Ya. G., U. Islamov and A. Askarov 1966. Pervobytnaya kul’tura i vozniknovenie oroshaemogo zemledeliya v nizov’yakh Zarafshana. Tashkent. Itina, M. A. 1977. Istoriya stepnykh plemen Iuzhnogo Priaral’ya. Moskva. Levine, M., Y. Rassamakin, A. Kislenko and N. Tatarintseva (eds.) 1999. Late Prehistoric Exploitation of the Eurasia Steppe. Cambridge.

It appears, therefore, that sunken constructions are one of the cultural characteristics pertaining to pastoral communities. The findings of Sites No. 1211 and 1219 confirm this characteristic. The presence of hearths within the structures indicates a use that may have covered a long time span, exceeding a merely seasonal occupation. This characteristic confirms that the interaction between nomadic pastoral communities and sedentary agricultural ones was firmly established, with a particular concentration in the final phases of the Bronze Age.

132

Chapter 10 The Final Phase of the Bronze Age and the “Andronovo Question” in Margiana Maurizio Cattani Introduction

of nomadic populations from the north (Kuz’mina and Lyapin 1984), whereas other authors claimed that there had been a peaceful contact between the two groups (Sarianidi 1975).

During the surface examinations carried out between 1990 and 1996 for the compilation of the “Archaeological Map of the Murgab Delta” (Gubaev et al. 1998), a great number of the located sites were characterised by the presence of the hand-made tempered ware with incised and impressed decoration, classified as Incised Coarse Ware (ICW) (Cattani and Genito 1998; Cerasetti 1998). This type of pottery had been found occasionally during previous researches in the Murgab Delta and in the piedmont of Kopetdag, and had been called “steppe ceramics” by Sarianidi (1975) on the basis of the similarity with the ceramic production found in the steppe between the Ural Mountains and the T’jan Shan (Kuz’mina and Lyapin 1984). The ceramics in this region are much more frequently included in the production of the Andronovo culture (Tolstov 1948; Gryaznov 1966; Kuz’mina 1971, 1994).

The study and researches carried out during the last few years require a deeper investigation into the steppe civilisations of the Bronze Age and above all an examination of the “source” context of the cultural expressions included in the denomination “Andronovo culture”. (NB: since 2001 the Italian Institute for Africa and the Orient, in collaboration with the University of Bologna, has activated new researches in Central Asia; in particular, with the new Archaeological Mission in Kazakstan, it has been able to tackle directly one of the contexts that are most rich and suitable for dealing with the problems of the development of nomadic way of life and of the Andronovo cultural aspects.)

The fragments found in the Murgab Delta, which were often very small and not particularly characterised from the point of view of shape recognition, come from surface contexts located in takyr interspersed with sand, or sometimes on top of actual aeolian sand dunes. In most cases these fragments are the only type that is present; in some cases they are associated with wheel-made ceramics of the so-called Namazga VI phase; in a very few cases they were found within large built-up areas of the Late Bronze Age.

This essay attempts to deal with the problems related to the interpretation of the presence of the so-called steppe ceramics in the Murgab Delta, and, more generally, to offer an evaluation of the cultural aspects of the Bronze Age in Central Asia, with particular reference to the possible interactions between two communities that were apparently so different from each other.

The interest in the analysis of a wide-spread presence of steppe ceramics influenced the subsequent stage of the researches of the joint Italian-Russian-Turkmenian Archaeological Mission, focusing it on a more in-depth investigation of the presence of not-clearly-recognisable populations whose identity and economy were completely different from those of the predominantly agricultural sedentary communities. The latter civilisation was well known, as a result not only of the surveys carried out for the compilation of the map, but also of the excavations of the nineteen-eighties, which had identified the so-called BMAC, Bactria-Margiana Archaeological Complex (Sarianidi 1976, 1990; Lamberg-Karlovsky 1991; Hiebert 1992). In this connection, it is worth pointing out that the current hypothetical interpretation of the presence of Andronovo pottery within the contexts of the sedentary civilisations of the Late Bronze Age suggested that the settlement crisis was a direct consequence of the arrival

The researches carried out from 1996 to 2002 by the joint Italian-Russian-Turkmen Archaeological Mission to Turkmenistan in order to complete the “Archaeological Map of the Murgab Delta” (Gubaev et al. 1998) were focused on a more detailed analysis of the aspects of the settlement of this region. In particular, an effort was made to attain a chronological and interpretative seriation of the settlement transformation dynamics in the Murgab Delta between the end of the Bronze Age and the initial stages of the Iron Age. In order to achieve a better definition of this transition stage, in which the evidence of sites containing steppe ceramics is included, systematic researches based on intensive surveys were undertaken in the area to the north of Takhirbai depe and in that around the Takhirbai 3 site (Figure 10.1) (NB: The sequence of the settlement pattern in the Murgab Delta was recognised for the Bronze Age as a part of the Bactria-Margiana Archaeological Complex and divided by Sarianidi (1976) into four main

Method of the Research

133

M. Cattani

Figure 10.1 Archaeological map of Murgab Delta with selected evidences of the Late and Final Bronze Age.

the archaeological record. Margiana is characterised by an uninterrupted archaeological landscape, with hundreds of surface sites (mainly pottery scatters) that are a remarkable context for research. Despite this apparent abundance, the archaeological evidence must be spread over a period of at least two millennia, and the individual pieces of information must be decoded on the basis of the postdepositional transformations they have undergone.

different stages: Kelleli, Gonur, Togolok, Takhirbai. The last-named stage is based on finds from the Takhirbai 3 site. The subsequent Iron Age was identified by V. Masson in the excavations at Yaz-depe as phases Yaz I, II and III). The research, moreover, included a methodological approach that was particularly careful to collect and analyse the palaeo-environmental evidence (chiefly traces of the ancient hydrography), endeavouring to reveal a possible association with evidence of the ancient peopling. Excavations and stratigraphic trenches were also carried out in some sites that were considered particularly significant, and an attempt was made to achieve a seriation of the ceramic types and a distribution of the types identified as characteristic of the final phases of the Bronze Age. Finally, the data obtained in the researches that had been carried out mainly from the middle of the nineteen-fifties to the nineteen-seventies were retrieved.

For these reasons, achieving a reconstruction of the natural and anthropogenic landscape of the Murgab Delta in the Bronze Age turns out to be a difficult task. Some observations on the geo-morphological analysis and on the method to be applied, however, are still valid. Cremaschi, in describing the characteristics of the landscape in the Middle and Late Holocene, identified, on the basis of photo interpretation and of a great number of stratigraphic tests, the formation of an extensive, fertile alluvial plain with many active streams during the Bronze Age (Cremaschi 1998). The hypothesis of the existence of soil that was extended for tens of kilometres without great interruptions and could be cultivated by the agricultural communities was repeatedly corroborated by the transects performed in several points of the Murgab

The Environmental Context of Present-day Margiana Most of the possible analyses of the peopling and landscape of the Bronze Age depend on the type and degree of intensity of the surveys, and above all on the visibility of 134

The Final Phase of the Bronze Age and the “Andronovo Question” in Margiana Delta. Wherever systematic, intensive researches were carried out, occupation without any real interruptions was found (see Salvatori, Chapter 5). When interruptions do appear on the surface, it is very likely that this depends on an alluvial sediment that has covered the ancient soil. The examination of modern irrigation canals that sometimes show uninterrupted stratigraphic sections reveals this process more clearly.

farmed during the Bronze Age, and therefore can be ascribed to a more recent period. In some cases we know that the encroachment of sand, which had started already during the Bronze Age, became more massive during the final stages of that period and during the Early Iron Age and since then it has characterised the landscape of the Murgab Delta. Extending Cremaschi’s work, the field research activities focused on the identification of palaeo‑channels, and proposed a chronological attribution on the basis of the association of stratigraphic and population-density data. The reference patterns for the interpretation of the traces seen in the aerial photographs or on the present surface are based on two separate lines of research:

We are far from having completely reconstructed the ancient morphology, and it is likely that the alleged ancient plain was actually more diversified because of undulations created by the course of fluvial ridges of ancient branches of the Murgab, or, at least in the northernmost part, by the vicinity of aeolian erosion escarpments. The stratigraphic analysis shows, for the periods before the Bronze Age, an alternation of sandy deposits and more clayey layers that sometimes are transformed into real pedogenized soils.

a) the presence of contemporaneous anastomosed canals of the individual river branches of the delta;

The attempt at a reconstruction of the palaeo-channels, as it is presented by Cremaschi (1998), identifies a numerous examples of evidence spread over the period of fluvial activity, which we know was still very regular up to the Parthian era.

b) presence of individual river branches to be distributed over a rather long span of time, in which each individual piece of evidence can be ascribed to the archaeological phases of the Middle and Late Bronze Ages and Early and Late Iron Ages.

The landscape of the present-day Murgab Delta is characterised by extensive desert areas, with a parallel series of sandy dune segments that have a north-south orientation and, in the southern part, are interposed among the branches of the endoreic delta of the Murgab River and the increasingly extensive farmed areas (the latter have been recently activated thanks to irrigation canals chiefly fed by the water of the Karakumskij Kanal). Hidden by sand dunes or alluvial deposits, there are the numerous traces of the fluvial branches of the ancient delta, detected by aerial photographs and discussed in the first volume about the “Archaeological Map of the Murgab Delta” (Cremaschi 1998). Describing the characteristics of the landscape during the Middle and Late Holocene, Cremaschi pointed out the presence of an extensive, fertile alluvial plain with a great number of streams that were active during the Bronze Age (Cremaschi 1998: 18-19). This fluvial activity gave rise to a sedimentary aggradation, particularly in the southern part of the delta, up to the area of Takhirbai, with the gradual burying of previous settlements during the Bronze and Iron Ages.

Besides the identification of the palaeo‑channels by means of aerial photographs, the surface survey made it possible to locate and date palaeo‑soils buried in alluvial deposits, areas covered with aeolian sand and many other partial traces of river courses that were often not visible in the aerial photographs. The intensive survey was carried out also along transects that interconnected two or more palaeo‑channels, in order to endeavour to date them on the basis of associated anthropic evidence. In the area where more in-depth investigations were carried out, two main river courses were identified: they are parallel, about 3-4 km from each other, have a meandering course and run south to north (Figure 10.2). The analysis of the peopling revealed that the sites were distributed around these palaeo‑channels with two distinct chronologies. Near the eastern palaeo‑channel, no Bronze Age human presence was detected, sparse traces of the Yaz I period were found (a few ceramic fragments in Site No. 172), whereas there were numerous and sometimes quite extensive evidences that could be ascribed to the Late Iron Age (Yaz III and late Yaz III). Adjacent to the western palaeo‑channel, on the other hand, there were some extensive Bronze Age built-up areas, and a great number of other minor sites.

During the same period, there began the intrusion of aeolian sand, coming from the northern desert area; in the final stages of the Bronze Age this led to the deactivation of the river branches, consequently to the transformation of the general settlement pattern. The encroachment of sand disturbs the visibility of the archaeological record. Where these dunes are present (sometimes they are quite imposing, several tens of metres high, and up to several hundreds of metres long), it is not possible to reconstruct the ancient situation. From a stratigraphic point of view, we can assert that most of the dunes cover soil that was

The course of the palaeo‑channel was reconstructed on the basis of some segments detected by Cremaschi in an aerial photograph, because of a strip of ground formed of vertisoils, visible on the surface on the western side of the Site No. 67 built-up area (Late Bronze Age), in the stratigraphy of the excavation of Takhirbai depe (Site No. 64), and in the stratigraphic verification of Takhirbai 3 (Site 135

M. Cattani

Figure 10.2 Archaeological map of Murgab Delta with ancient river‑beds.

No. 188). It is interesting to re-examine the observation of an artificial canal carried out by Sarianidi before the agricultural transformation (Sarianidi 1990: 55). According to Sarianidi, this canal was visible on the surface as a 25 m ridge starting near Takhirbai 3 and running southwards for about 3 km. In this case it would be an artificial rearrangement of the palaeo‑channel near the large Late Bronze Age built-up area, which became a hanging canal, leading to the formation of a ridge.

habitation did not seem to undergo any interruptions in the Yaz I and Yaz II phases, and was still rather considerable in the Yaz III phase. This reconstruction surmises that there was a natural change in the course of the rivers; but we cannot exclude that in the Achaemenid period an artificial re‑arrangement was performed, in order to create new cultivable land, particularly on the eastern side of the delta.

The hypothesis that we wish to propose is that of recognising these two palaeo‑channels as two different branches of the Murgab River, presumably connected to artificial irrigation canals, and respectively dating back, the eastern one to the Achaemenid, Hellenistic and Parthian Age, and the western one to the Bronze and Early Iron Ages. The western palaeo‑channel remained active, presumably with a senescence stage or as an artificial canal, during the subsequent periods, since in this area the

Types of Late Bronze Age Finds The forms of settlement include “extensive built-up areas”, generally characterised by large unbaked-brick structures that respect the Bronze Age tradition (Takhirbai 3: Site No. 188 and Takhirbai 4: Site No. 67). These sites have an area of several hectares, and usually contain not very high

136

The Final Phase of the Bronze Age and the “Andronovo Question” in Margiana The “scatters of materials” on the takyr surfaces (chiefly wheel-made pottery) are numerous, extensive and in many cases uninterrupted; it is difficult to interpret them. Presumably these materials are connected with the thorough exploitation of the farmed fields, with the effects of manuring and with alluvial dragging from the main sites. Side by side with these sites, there is other evidence of “scatters of ICW tempered ware items”, which are systematically characterised by the absence of the use of unbaked clay bricks, and often associated with small lumps of baked, almost vitrified clay, presumably oven residues. Associated with the course of the western palaeo‑channel, there are Late Bronze Age sites of Takhirbai 4 (Site No. 67) and Takhirbai 3 (Site No. 188). Towards the end of the Bronze Age, the Takhirbai depe complex was built, and the area became one of the most important nuclei of peopling

Figure 10.3 Aerial photo of Takhirbai 4 settlement (Site No. 67).

palace-type buildings, surrounded by smaller dwellings and production structures.

The Takhirbai 4 built-up area (Site No. 67), which had already been identified in V. Masson’s first researches

Unlike other settlement types, such as the great depe of the Parthian-Sasanian Age that characterise the landscape of Margiana, the Bronze Age sites can be qualified as “failed tells”, i.e. as pluristratified sites of short duration, without the construction of foundation platforms and without the progressive construction of elevated buildings. The material culture belongs exclusively to the production of the Bactria-Margiana Archaeological Complex, identified on the basis of the presence of wheel-made purified ceramics. (Note about the use of the phrase “Bactria-Margiana Archaeological Complex” or BMAC: since we must comply with the requirement to distinguish sedentary populations from nomadic bearers of ICW ceramics, we prefer to use BMAC to indicate the sedentary communities. Despite the fact that this phrase is not suitable for describing the historical and archaeological processes of Middle and Central Asia, it is more accurate for Margiana in the Late Bronze Age.) In the surrounding areas, and often also in areas several kilometres away, there are “minor sites”, formed of small built-up centres having a maximum area of about 1 ha, presumably connected with farming activities, but also with production areas, in particular pottery kilns and metalworking plants. These minor sites, which have never been investigated with excavations or systematic researches, were probably meant to control the agricultural production and above all to maintain the activity of the irrigation systems: the latter must have been particularly complex, considering the settlement density and the particular abundance of water obtained from the natural course of the Murgab River. In these minor sites, too, the structural evidence confirms a massive use of clay as building material and the presence of a material culture identical to that of the main sites.

Figure 10.4 Plan and section of Takhirbai 3 excavations (from Masson 1959: 15, Fig.5; 13, Fig.2).

137

M. Cattani

Figure 10.5 Plan of Takhirbai 3 excavations (from Sarianidi 1990: 54, Fig.10.).

Figure 10.6 Andronovo vessel from Takhirbai 3, excavations Sarianidi 1972 (from Kohl 1984: Pl. 23, Fig. 23b).

Figure 10.7 Plan and stratigraphic section of Takhirbai 3 exposed by modern excavations for irrigation ditch.

138

The Final Phase of the Bronze Age and the “Andronovo Question” in Margiana (Masson 1959: 17) is characterised by a superimposed structure, about 2 m high and with an oval layout, elongated on a north-south axis (Figure 10.3). No excavations or other investigations have been carried out on it as yet, and very little can be determined about the architectural structure and settlement sequence of this area. The material culture items collected in the site allow it to be dated to the Late Bronze Age. Further southwards, there is the built-up area of Takhirbai 3 (UTM 41N 420652, 4211587), the most representative site for analysing the peopling in the final stages of the Bronze Age and for investigating the transition to the Iron Age. This area was explored by V. M. Masson between 1954 and 1956 (Masson 1959), with two excavation trenches, in which he brought to light an archaeological stratigraphy approximately 2 to 2.5 m thick. The built-up area was calculated to be about 22 ha, but most of the settlement seems to be buried under aeolian sand or alluvial deposits. The architectural structures and urban configuration, which include a craft area with kilns, appear to be quite similar to those of the other built-up areas of the Late Bronze Age (Figure 10.4). V. I. Sarianidi carried out other researches in 1972 (Sarianidi 1990: 54-55), performing a 10 x 10 m sondage at the centre of the built-up area: it is still visible as a deep, square depression on the ground. Here he found the traces of an unbaked-brick building, associated with wheel-made pottery, typical of the final phase of the Bronze Age, called Takhirbai 3 phase (Figure 10.5). The same excavation unearthed a whole Andronovo vessel (Figure 10.6 from Kohl 1984: Pl. 23b; Sarianidi 1990): its stratigraphic position is uncertain (“in the trimming of south eastern sounding of the excavation”, Sarianidi 1990: 54), but it would be important to determine it, because, if the vessel was under the unbaked-brick structures, it might be one of the most ancient testimonies to the presence of Andronovo material in the Murgab Delta. On the same year, a trench was dug within the palaeo‑channel recognised as a ridge on the surface. The stratigraphy showed an alluvial sedimentation formed of sandy layers mottled in red.

Figure 10.8 Andronovo grave from Takhirbai 3.

The surface collections identified a scatter of materials (chiefly pottery) on a total surface of approximately 80 ha. A more careful evaluation of the types of ceramics collected on the surface led to the identification of an archaeological presence that ranges from the Late Bronze Age to the Achaemenid era, and to the division of the area into several chronologically distinct parts. In particular, at the western end there is a Yaz III pottery scatter, while almost all the eastern side is characterised by Yaz I pottery; the latter indicates the presence of an extensive Early Iron Age built-up area that partly overlaps the Bronze Age one and perhaps is a continuation of it.

The recent survey researches led to the recovery of the Takhirbai 3 area, which the archaeologists previously believed had been completely destroyed by modern cultivations, and to its investigation by means of aerial photographs, surface collections and stratigraphic checks. During the analysis of an aerial photograph, a quadrangularshaped, light-coloured anomaly was detected. This type of reflectance usually corresponds to the presence of clay, and may be interpreted as evidence of an architectural complex similar to the fortifications previously found by excavations in other Bronze Age built-up areas. This surmise is supported by Hiebert’s assertion that “the excavations at Takhirbai 3 appear to be located within a large fortified building complex” (Hiebert 1994: 24).

The surface researches carried out at Takhirbai 3 were started simultaneously with the digging of a new irrigation canal, which unfortunately crossed the archaeological area on its eastern side, but made it possible to evaluate long stretches of the archaeological stratigraphy. The cumulative section that was examined on the basis of 139

M. Cattani

Figure 10.9 Andronovo vessel from the grave of Takhirbai 3.

readings in individual points and in two stretches about 10 m long contains an archaeological deposit whose thickness ranges from 40 cm to 1.80 m, formed of a series of dusty layers, rich in ash and charcoal, alternating with red silty layers (Figure 10.7). There are scantily-attested traces of clay brick walls, while in the northern part, where the Bronze Age built-up area ends, the presence of a waterway, filled with laminar alluvial deposits, was detected. The section readings seem to suggest that the modern canal cut through the ancient waterway diagonally, so it is not possible to precisely evaluate the direction and size of the latter, though it appears to have been 10 m wide.

found: it is one of the rare testimonies to the Andronovo funerary customs in the Murgab Delta (Figure 10.8). The grave, formed of a pit with vertical walls, contained a single skeleton lying curled up on its right side, with its head to the north and looking towards the west, and the fragments of a ceramic vessel with the typical decoration formed of meanders and triangles that is widespread in the steppe world (Figure 10.9) and is different from the more common incised decorations found in the built-up areas. On the basis of the numerous items collected on the surface or in the discharges of soil removed for the construction of the new canal, we endeavoured to work out a typology of the pottery representative of the final stages of the Bronze Age, integrating what had already been proposed

During the stratigraphic exploration, an “Andronovo” grave, partly cut through by the canal excavation, was 140

The Final Phase of the Bronze Age and the “Andronovo Question” in Margiana by Masson’s and Sarianidi’s previous researches. At the same time we wished to tackle the problem of continuity with the previous chronological phase. This problem is particularly significant for the following reasons: 1.

If this phase is regarded, in archaeological literature, as a time of change, it is difficult to establish what was the actual degree of discontinuity and how important these changes were, particularly in relation to the subsequent expressions of the Early Iron Age (Yaz I phase).

2.

The chronological sequence has not been clearly defined yet; the recent acquisition of absolute radiocarbon dates has made it possible, in any case, to determine the period under examination in a more detailed way, and to connect events in the Murgab Delta with other significant cultural expressions of Central Asia.

3.

particularly in relation to the problem of the continuity/ discontinuity between the Late Bronze Age and the Iron Age. The detailed analysis of all the material collected during the campaigns for the Archaeological Map Project is still to be completed. But, from a first evaluation it seems that in the final phase of the Bronze Age there was a strong demographic reduction, although we should be cautious in considering the effective markers from the material culture recognised so far. Only few pottery shapes have been recognised as typical of the Takhirbai phase, but their association with previous types suggests a degree of continuity. A continuity from the Late Bronze Age to the final stage can be surmised also from the presence of Yaz I pottery on the surface of most of the sites (Sarianidi 1981: 180). This might confirm a continuous occupation in the Lower Murgab Delta.

The interaction between the BMAC communities and the populations bearing ICW pottery is one of the most controversial and interesting issues (LambergKarlovsky 2002: 16) for surmising the various configurations of the social and demographic pattern between the Late Bronze Age and the Early Iron Age.

Assuming a general decrease in the range of types, we have tried to recognise the shapes that are characteristic of the Final Bronze Age (Figures 10.10-11):

A recent position assumed by Hiebert considers the phase of Takhirbai not in the BMAC sequence, but in a different period following this culture (Hiebert 1994: 71). This position depends on the scarcity of data included with absolute certainty in this phase, and a final solution will be reached only through new excavations. We wish only to remark that some features continue with the same properties from previous stages: apart from a general demographic decrease, continuity from the Late Bronze Age is supposed in farming economy, in the organisation of water regulation and in the settlement pattern. Considering the size of Takhirbai 3, it seems to have re-established the system of a central place, as compared with the general peer level of the Late Bronze Age. The material culture is marked by few changes in pottery production and especially in the reduction of list of finds. As already specified by earlier authors, a new production of grey or red wheel-made pottery appeared in new shapes. What is noteworthy is the absence of small clay figurines; moreover the occurrence of other items such as seals or stone objects (vessels and spindle-whorls) is reduced. The absence of a ceramic typology relative to the period under examination has considerably affected the chronological attribution of many sites of the Late Bronze Age. The elaboration based on the material collected at Takhirbai 3 is therefore essential for a diachronic recognition of the ceramic production, and, together with the material coming from the Takhirbai depe excavations, has become one of the bases for the dating of the sites,

1.

Large cups on a pedestal with rounded, inflexed rim. The reservoir may be shallow or deep, while the profile is truncated-cone-shaped or rounded. The appearance of this shape characterises a new phase in Bactria.

2.

Large cups with small, almost cylindrical, reservoirs and horizontally everted rim.

3.

Biconical goblets with upper walls that may be rectilinear or concave, and with everted rim. Widely known in Takhirbai 3, they were also present in the burials located in the ruins of the abandoned buildings of the previous period (P’iankova 1993: 110, Fig 3: Togolok 1, Grave 38; Togolok 21, Graves 55, 63, Cenotaph 9; Pyankova 1989: Togolok 21, dwellings of the second period, Fig. 4).

4.

Small bowls with truncated-cone-shaped basin. In some variants, the rims are rounded and everted.

5.

Elongated goblets with “S” profile.

6.

Small cylindrical beaker.

7.

Goblets on a short pedestal.

8.

Globular or ovoid jars with everted rim.

9.

Globular jars with spout (both short and long) on the shoulder.

10. Large jars with incised ornaments on the shoulder, mostly represented by horizontal or wavy lines. This 141

M. Cattani specific ornament is already attested in earlier phases, but it is only with the Takhirbai Phase that it becomes more distinctive. 11. Truncated-cone-shaped pot supports. The Takhirbai 3 pottery shares many shapes with the phases of Molali and Bustan in Bactria. Apart from a few new pottery shapes, the production seems to be a continuation of the previous one and does not represent a real change in the cultural system. More significant changes characterise the beginning of the Iron Age, when the wheel-made production continues, but there appears a new hand-made pottery production with painted decoration. Outside the area of Takhirbai 3, this final phase of the Bronze Age in the Murgab Delta is scantily attested. Some characteristic ceramics of this late phase were found at Togolok 21, in dwellings of the second period and in burials (Pyankova 1989). From this site (“Bol’shoj altar, sealed under secondary floor”: Hiebert 1993: Fig. 2) one C14 dating fits in the period (see the table below). Excavations at Gonur in 1990 revealed “poorly preserved architecture associated with a post-BMAC ceramic assemblage, very similar to that found at Takhirbai 3” (Hiebert 1994: 80).

Figure 10.10 Pottery types of Final Bronze Age from Takhirbai 3 and other sites.

In the recent excavations in Gonur we lost the opportunity to carefully investigate the upper layers, generally located at the top of the stratigraphy, corresponding to a continuity of occupation after the Late Bronze Age. The few available pieces of evidence suggest a continuity of occupation for the main Late Bronze settlements, though they were going through a progressive decay. In connection with the exhaustion of the water-courses or with a precise political strategy, it seems that the Gonur area was abandoned in favour of other branches of the Murgab River, where it was possible to exploit new land with a new irrigation scheme. The main centre of activity and possibly the political hub of the region, moved presumably to a different area, along the eastern branch of Murgab. We cannot exclude the same presence of Final Bronze Age sites along different branches of the river. As ascertained for the later period of Yaz I, at least four different branches of the river were exploited from the western course of the Aravali to the east of Takhirbai. In the sedentary built-up areas of the Namazga (NMZ) V and VI stages, we must point out the presence of coarse, usually undecorated tempered ware produced together with the wheel-made one (P’yankova 1994). In relation to the appearance of ICW, this type of pottery must not be mixed up with the one produced by nomadic communities. The fact that two different types of coarse ware production are attested for the same period also raises the problem of the possible relationships between the two communities that were occupying the same territory and that in this case declared themselves to be completely independent of each other.

Figure 10.11 Pottery types of Final Bronze Age from Takhirbai 3 and other sites.

142

The Final Phase of the Bronze Age and the “Andronovo Question” in Margiana In the Late Bronze Age, the coarse ware production was contemporary with the more abundant purified ceramics, but represented a very small percentage of it (10-20%).

the identification and collection of ICW items, and this prevented a correct evaluation of the actual percentage of this presence with respect to other forms of habitation. In any case, the intensive-research areas highlight a massive presence of camps containing ICW pottery.

Coarse ware is always very compact (chaff-tempered), has a finished surface, often with evident wheel marks, and hues ranging from purple-red to brown. Archaeometric analyses are not available yet, but it is reasonable to suppose that these ceramics were fired at high temperatures in the same kilns that were used for purified ceramics. Kilns dating back presumably to the NMZ VI phase are present in almost all the built-up areas that were identified. They are usually found out of the centre of the built-up area, and seem to be concentrated in real craft districts.

The sites are spread all around main Bronze Age settlements, but some of them are far away in unsettled area, presumably already occupied by desert sand. Their location shows a complementarity of farmers and pastoralists with an economic integration of pasture land assigned or abandoned by farming community to the nomadic tribesmen. (Figure 10.12) The presence of Andronovo pottery within contemporary built-up areas of sedentary communities is attested at very few sites, reported by previous publications:

In the site examined within the sampling area, Takhirbai 4, several areas were identified that contained a great amount of remains of kiln waste and fragments of refractory bricks that attest to the presence of kilns. They are situated to the west (Site No. 302) and to the south (Sites No. 69 and 253) of the built-up area; their specific concentration can no longer be exactly quantified, because of the recent agricultural works that have destroyed surface readability immediately to the south of the built-up area.

South Gonur (Sarianidi 1994: XXXIV): central building room 500: “[..] on the floor [...] near a miniature column, was a sooty vessel with incised decoration – a ceramic typical of the Tazabag’yab steppe nomadic culture of Khorezm.” North Gonur (Sarianidi 1990: 26): “[...] twenty fragments of hand-made pottery of the steppe sphere [...] ten fragments belong to round, coarse ware vessels with everted rim (red or black) with large grog inclusions (Sarianidi 1990: Tab. XII, 14-15). This group of ceramics has incised herringbone decorations on the outside; only in one[...] there are deeply impressed, raised lines[...] of the Andronovo type [...] 5 fragments of hand-made coarse ware of a different type, with not very thick walls [...]. The outer surface has been decorated by means of a comb-shaped tool that forms horizontal herringbone motifs or oblique lines [...]. This category is more similar to the complex of ceramics of the Afanas’evo group.”

The most widespread forms of pottery are large containers, including a jar with globular body, short neck and straight rim, and a truncated-cone-shaped vessel. There are no decorations, except for a very few cases of incised decorations with complex, sometimes figurative, motifs, which were found within built-up areas that can be dated to the Namazga V-VI phases. Catalogue of ICW Sites In the presentation of the data relative to the sites with ICW pottery, the indications given by Masson (1959), Sarianidi (1990) and Masimov (1979) during their researches were considered. Then the new areas identified during the surveys carried out for the compilation of the “Archaeological Map of the Murgab Delta” were added. The distribution of the sites is presumably conditioned by the research strategies that promoted the exploration of the areas involved in the settlement of sedentary farming communities, and in particular by the scarcity and small size of the indicators characteristic of this cultural aspect, which make it much more difficult to detect them. The discovery of ICW pottery in desert environments, which were found sporadically or during survey transects, seems to suggest that there may be many other sites inter-spersed among the sands, in conditions that make it harder to locate them. On the other hand we must admit that within the project of the “Archaeological Map of the Murgab Delta”, once the problems relative to the nomadic presence within the Late Bronze Age peopling had been identified, targeted researches were carried out whose goal was

Gonur North: ceramic scatter interspersed with sand dunes, over an area of 40 x 60 m (Hiebert ˝and Moore 2004: 295). Small lumps of over baked clay and millstones fragment were also found. Associated with the steppe pottery, Final Bronze Age (Takhirbai Phase) pottery was also found. Togolok 1 (P’iankova 1993: 115): “three ceramics from the buildings in the central part of the fortress (Fig. 5.1,2 from Room 20; Fig. 5.5 from Room 14); the rest (Figs. 5.3, 4, 6) was found in the upper part of the discharge near the western front of the defensive wall”. The first type consists of globular vessels with cord decoration impressed along the rim and on the shoulder (Room No. 20). In 1993 P’yankova suggested that these vessels are stylistically similar to the Poltavkino and ancient Petrovko pottery in the northern region of the Caspian Sea, but pointed out that the type of temper was more similar to that of the Andronovo pottery. A second group of nomadic steppe ceramics from Togolok 1 includes slightly restricted jugs 143

M. Cattani

Figure 10.12 Archaeological map of the Murgab Delta. Sites with Incised Coarse Ware (black points).

with a large diameter at the edge and comb impressions (Fig. 4.38). P’yankova ascribes this type of pottery to the first phase of the “timber frame” culture, to the north of the Caspian Sea, and in the Lower Volga Basin (P’iankova 1993).

despite the fact that the excavated area of the settlement did not yield any pottery of this type. Auchin 11: “[...] ceramics of the Bronze Age in the takyr, in which many Andronovo-type pottery fragments have been found [...].” (Sarianidi 1990: 13).

Togolok 3: “Next to some fragments of wheel-made pottery we found some fragments of hand-made pottery of the group of the steppes” (Sarianidi 1990: 46).

Auchin 12: “The pottery, among which there are some Andronovo-type fragments, dates back to the Bronze Age.” (Ibidem: 13)

Auchin 1 (Sarianidi 1990: 10): “[...] from the surface of the built-up area there come some isolated fragments of hand-made pottery, characteristic of the ceramics of the steppe populations [...]”. (Masson 1959: Fig. XI, 4-7, 10). P’yankova (1993: 116) asserts that Auchin-depe yielded a considerable collection of surface finds, fragments of steppe-type pottery (Masson 1959: 44, Figs. 4.1-12, 22),

Adzhi Kui 8: “[...] From the surface of the built-up area there come some isolated fragments of hand-made pottery, characteristic of the steppe population.” (Ibidem: 10) Taip: cf. Masimov 1979: 111-131, Tab. 6, 30-36. Takhirbai 3: cf. Masson 1959: Tab. XI. 144

The Final Phase of the Bronze Age and the “Andronovo Question” in Margiana The exclusive presence of fragments of steppe ceramics, sometimes published with the denomination “Andronovo ceramics”, is pointed out, in addition, by Sarianidi (1990) and Kuz’mina and Lyapin (1984).

be the result of an actual distribution of the artefacts not far from their place of deposition. This mechanism, however, has contributed to a high fragmentation of the pottery remains and to an almost complete destruction of the original surfaces, particularly in steppe ceramics, whose quality is worse than that of the wheel-made ceramics of sedentary populations, both for their fabric and for their firing temperature. The only items found on the surface usually have very compact, almost vitrified tempers, while all the evidence of pottery fired at low temperatures has certainly been lost; the latter is predominant in excavation contexts and is rarely found on the surface, because of the recent exposure.

Gonur 14: “20 x 20 m camp. Andronovo type.” (Sarianidi 1990: 34). Togolok-2: “found some fragments of Andronovo-type pottery.” (Ibidem: 45). Togolok 8: “some fragments of Andronovo-type pottery.” (Ibidem: 46). Togolok 12: “fragments of hand-made pottery of the steppe type.” (Ibidem: 47).

Only in a few cases, where there are takyr, traces of occupation such as remains of baked-clay structures are found in addition to potsherds.

Togolok 13: “some fragments of hand-made pottery of the steppe type.” (Ibidem: 47).

A more detailed survey was carried out in the region around Takhirbai depe, in order to analyse the settlement pattern between the Bronze Age and the Iron Age (Figure 10.13). The landscape next to Takhirbai 4 (Site No. 67) is characterized today by modern cultivated fields and desert zones with takyr and low sand dunes. The use of aerial photographs helped the fieldwork to concentrate on particular areas with a high potential for the observation of archaeological remains.

Togolok 17: “some fragments of hand-made pottery of the steppe type.” (Ibidem: 48). Togolok 31: “hand-made pottery with decorations characteristic of the Andronovo-type steppe tribes.” (Ibidem: 53). Takhirbai 11: “Andronovo-type pottery.” (Ibidem: 57).

Around the built-up area of Takhirbai 4 (Site No. 67), some settlement areas were found that were characterised by the presence (sometimes exclusive) of incised coarse ware (ICW). These areas are either within extensive takyr, where aeolian deflation has left, on the surface, small pottery fragments and a few other artefacts (fired plaster and stone), or in more complex geo-morphological situations characterised by deep hollows covered and/or interspersed with sand dunes. In some cases, the potsherds seem to be located above or within the sand dunes confirming that the process of the aeolian sand encroachment had already started in the Bronze Age (Cremaschi 1998; Salvatori 1994).

Takhirbai 12: “Andronovo-type pottery.” (Ibidem: 57). Takhirbai 13: “Andronovo-type pottery.” (Ibidem: 57). Auchin 7: “... isolated fragments of hand-made Andronovotype pottery.” (Ibidem: 12). Auchin 10: “...isolated fragments of Andronovo-type pottery.” (Ibidem: 12). Auchin 16: “...where the Andronovo-type pottery is the only one that prevails decidedly...” (Ibidem: 13). The recent surveys for the “Archaeological Map of the Murgab Delta”, located a large number of new ICW sites in takyr areas or along sand dune formations.

The archaeological remains of the settlements containing ICW pottery, unlike those relative to other forms of habitation, are sparse and are not focused along water‑courses; on the contrary, they seem to be located in the spaces that are most distant from the potentially cultivable areas, and occupy semi-desertic areas suitable for pastoral farming.

In the first case, the takyr usually correspond to the upper surface of the Bronze Age soil, which, because it has the characteristics of compact clay, has undergone only a slight erosion. The items collected in the takyr, therefore, may belong to original situations existing exactly on the surface of this soil, but may also be the effect of a deflationbasin collection that has concentrated their presence in a secondary position. Many items show evident signs of aeolian wearing and are scattered over extensive areas. It is likely, in any case, that aeolian conveyance cannot have excessively affected their distance from the original context, and that even situations of particular density may

Many other sites containing ICW pottery mixed with wheelmade pottery point to an overlap between the bearers of ICW pottery and the settlement strategies of the sedentary farmers. These are built-up areas where a certain degree of integration between the two communities was reached, or where the ICW-pottery-bearing groups superseded the other groups in the management of the agricultural production, acquiring sedentary characteristics, according 145

M. Cattani cannot be excluded a priori, does not turn out to be the most effective method for explaining the dynamics of the peopling of the steppes. Side by side with the aspects of the investigation into the peopling process and into built-up areas and graveyards, there are several archaeological problems relative to absolute dating: the latter has now been revolutionised by an increasing use of radiocarbon. “The calibrated dates were neither accepted in Germany or Russia because they differed greatly from the historical chronologies of Egypt and the ancient Near East, thus creating a large time gap” (Kuzmina 2002: 121). According several scholars this final phase was fixed between the 18th and 14th century BC. Thanks to recent C14 dating (Table 10.1), a new chronological setting is now accepted by most of the scholars. In this frame, the Late Bronze Age is included between 1800 and 1550 BC, while the final phase of the Bronze Age should be placed between 1550 and 1350 BC. This dating is confirmed by the recent excavation of Takhirbai 1 and by two sites characterised mainly by ICW Andronovo pottery, Sites No. 1211 and 1219. Cultural Setting: the Relationship between ICW and the Andronovo Horizon

Figure 10.13 Topographic survey around Takhirbai 4 (Site No. 67).

From the archaeological evidence that has just been presented, it appears that the phenomenon of the presence of so-called steppe populations in the Murgab Delta during the final phases of the Bronze Age is much more substantial that had been previously supposed. It has become necessary, therefore, to perform a more detailed chronological attribution and to deal with the cultural dynamics that were active in most of Central Asia during the Late Bronze Age.

to a pattern that has been proposed for these communities in Khorezm (Itina 1977). This phenomenon intensified precisely at the beginning of the late phases of the Bronze Age. Among the evidence of ICW-pottery-bearing populations that were in close contact with sedentary populations, or located in potentially cultivable areas, Sites No. 69, 1211 and 1219 were investigated with several trial excavations (cf. Chapter 9). These investigations followed the exploration of a recently published site in the Gonur area (Hiebert and Moore 2004).

Independently from a complete anthropological and cultural identification of these expressions, we also wish to look into the interactions between the various archaeological aspects, and to improve upon the static distinction between sedentary BMAC populations and ICW “nomadic steppe people”. As regards the cultural attribution, besides a general inclusion in the Andronovo cultural horizon, it has been surmised that most of the finds belong to the Tazabag’yab group, attested in the southern delta of the Akchadar’ya in Khorezm. This group, which was identified by Tolstov in 1938, has often been regarded as a western variant of the late Andronovo culture (Alakul’ phase, dating to the second half of the 2nd millennium BC). Some sites of the same group have been detected along the Uzboj and in the delta of the pre-Sarykamish (Tolstov 1962: 59-60).

Chronology of the Finds in the Murgab Delta The definition of the cultural setting and chronological attribution depend on our knowledge of the northernmost regions, where the aspects of the populations that occupied the steppes during the Bronze Age have clearer characteristics. It is not easy to summarize, because there have been more that a hundred years of research, and different interpretations have been proposed by a great number of scholars. An element of discordance that has influenced the interpretations and field researches was the use of the block migration model, which, though it 146

The Final Phase of the Bronze Age and the “Andronovo Question” in Margiana Site Gonur 1 South, Room 178, upper level floor near kiln (Phase 5, Takhirbai ceramics)

Lab. No. Hel-2967

Date bp Cal. Date BC (1σ) 3380±110 1860-1520 (68.2%) 1780-1520 (64.9%)

Cal. Date BC (2σ) 1950-1400 (95.4%)

Gonur 1 South, Room 266, upper level Hel-2970 room deposit. Central building, (Phase 5)

3380±90

1770-1520 (68.2%)

1890-1440 (95.4%)

Takhirbai 1 A5 US 125 Takhirbai 1, US 125 Togolok 21 Site No. 1219, SU 15 Site No. 1211, SU 71 Site No. 1219, SU 4 Yaz-depe, 1989 exc. 4 m below the surface

3375±80 3311±75 3270±40 3300±40 3140±50 2960±140 3120±90

1750-1520 (68.2%) 1690-1510 (68.2%) 1620-1490 (68.2%) 1620-1520 (68.2%) 1500-1310 (68.2%) 1380-1000 (68.2%) 1500-1260 (68.2%)

1890-1490 (95.4%)

GX-20648 Bo-0000 LE-2678 GX-30714 GX-30715 GX-30713 Beta 33566

1690-1440 (91.8% 1640-1440) 1690-1490 (95.4%) 1520-1260 (95.4%) 1500-800 (95.4%) 1610-1120 (95.4%)

Table 10.1 C14 dates from Late to Final Bronze Age in Margiana.

In Khorezm, the previous phase saw the association of Andronovo-culture pottery (Fedorovo phase) with the local, so-called Suyarganskaya culture, which in its early stage was also called “Kamyshli”. Farming with irrigation works is attributed to this phase (Itina 1977: 231).

The pottery finds in the steppes of southern Turkmenistan date to the same period of the excavations in the settlements, with Namazga VI occupation stages: the map published by Kuz’mina shows a homogeneous diffusion along the piedmont of the Kopetdag, and includes the following sites:

According to Tolstov, the formation of the Tazabag’yab group is due to the penetration of populations that mixed some characteristics of the Srubnaya culture with Andronovo ones and integrated with the local Suyargan culture (Tolstov 1939: 174; 1940: 70-71).

Anau: Pumpelly 1908, I. 142, Tab. 15, 7-9. El’ken-depe: Marushchenko 1959: 60, Pl. 10. Tekkem-depe: Ganyalin 1956a: 36; 1956b; Shchetenko and Kutimov 1999.

Kutimov recently proposed an amendment of the chronology of the Tazabag’yab culture, and placed it between the end of the sixteenth century BC and the first half of the fifteenth (Kutimov 2002: 203), highlighting many elements of analogy with the Andronovo world.

The pottery with incised and impressed decoration from Tekkem-depe (Shchetenko and Kutimov 1999) is comparable to that of the Sargar culture (North-western Kazakstan) and more generally to the cordoned pottery typical of the advanced phases of the Bronze Age in the steppes (corresponding to the last quarter of the 2nd millennium BC and called “valikovoj keramika”), whose best-defined expression is in the Begazy Dandybaj culture, attested in all central-northern Kazakstan.

Several scholars believe that the characteristics of the Tazabag’yab group extend into the area of the low Zarafshan (Makhandar’ya), the Murgab Delta and the Tashkent oasis. According to other authors, these cultural expressions have characteristics of their own and should be studied independently.

Conclusions

According to P’yankova, some other finds were connected with the north-western cultures of Poltavkino and with the Petrovo cultures (P’iankova 1993).

The analysis of the landscape and the evaluation of the cultural remains of the final phases of the Bronze Age make it possible to propose a hypothetic reconstruction of the settlement of this region along gravitational axes associated to the water-courses. This habitation is therefore not based on spread-out clusters, as in the Bronze Age settlement pattern organised in “oases” that had been proposed in the edition of the first researches (Sarianidi 1990); and perhaps it is not possible either to accept the hypothesis of a great cultivated plain like the one presumably present in the Middle Bronze Age.

The southernmost attestations of ICW pottery do not go beyond the Kopetdag, and at most penetrate into the valleys of the Tedzhen (Kuin-kuyu) and of the Murgab (Imambaba, Takhta Bazar, Kashan) (Kuz’mina and Lyapin 1984; Kuz’mina and Vinogradova 1986). In Bactria, the southernmost presence is in the oases of Farukhabad, and to the east in the built-up area of Shortugai (Francfort 198˝9).

147

M. Cattani The Murgab Delta had only a few natural water-courses still active in spite of the general process of desertification; they were enriched and presumably organised by a strong control, through artificial canals.

The nomadic component was no longer visible (cf. the barbarians in the Roman Empire). A new power structure was established, in which, among the various components that took part in the new configuration, there prevailed the traditional ones, and presumably a new implantation that acted perhaps as a catalyst and was characterised by painted pottery.

Habitation, therefore, extended in elongated forms, following the still-active Murgab branches with demographic centres whose position was only slightly upstream with respect to that of the centres of the previous Bronze Age phases. This confirms the general tendency to occupy higher ground, less encroached on by sand coming from the north; but at the same time a vanguard of the population were present along some fluvial axes. This is attested by Built-up Area 340 (which can be ascribed to a period from the Late Bronze Age to the full Yaz I phase, with material from the latter period apparently predominant, on the basis of the surface collection), by the sites along the Uch-depe axis, and by Site No. 999.

It would be simplistic to interpret the presence of ICW in Margiana as a result of exchanges or as a proof of exclusive, direct contacts and/or movements between these regions, because the groups settled at Gonur and Togolok were a part of a broader and more general movement that involved Central Asia. Due to the unfortunate, almost total, absence of data about burial rites and of other indicators, very little can be said about the social structure of the Andronovo communities in the Murgab Delta. We must therefore rely on an evaluation of external factors, carrying out, among other things, an analysis of the territory from the point of view of its available resources and as a geographic space, and an analysis of the climatic factors; and we must attempt a general interpretation through a historical and ethnographical comparison between the various nomadicpastoral populations. At this stage of the research there will be more questions than keys to interpretation, and the real achievement will consist in the programming of future, more thorough research.

The occupied zone apparently had very jagged margins, and presumably depended on the organisation of the irrigation canals or on the exclusion of some portions of land that were already abundantly covered with sand. It is precisely in these portions of land, sometimes near the built-up areas, that the camps of the nomadic populations were situated. The considerable proximity of these two forms of settlement seems to suggest that they were socially and economically complementary, and we cannot exclude that the populations traditionally characterised by a pastoral economy took part very actively in the agricultural work.

Despite the fact that it is not possible to ascertain the duration of the presence of these communities side by side with the sedentary settlements, we can assert that there was a precise intention to preserve the formal characteristics (material culture, settlement types), and presumably also the ideological characteristics, that reveal the strong stability of a social group. No mechanisms of acculturation into the sedentary communities or assimilation with them are attested. The repetitiousness of the pottery and of the incised and impressed decorations allows us to classify the communities that were bearers of ICW.

Since there have been no further investigations into the builtup areas of the sedentary communities (e.g. Takhirbai 3), it is not possible to determine the characteristics of the social structure and the changes that had taken place with respect to the better-known situation in the Middle and Late Bronze Age. The extension of Takhirbai 3, though probably overrated because of the enormous dispersion of Bronze Age pottery in a geo-morphological situation mostly buried under alluvial sediments, in any case is indicative of a general continuity in the types of buildings of the built-up areas. In particular, if the traces, detected in the aerial photographs, of a central quadrangular complex higher than the rest of the settlement were confirmed, this structure would not be very different from the previous ones detected at Gonur or Togolok.

The individual economic and productive units, presumably linked to family genealogy, must be identified within tribal groups that had exploited certain environmental conditions and based their economy on animal farming, presumably integrating it with exchanges of various products. Among these products we should not imagine only woollen products or secondary products of herds and flocks, but also metallurgical goods. The latter did not requiree a long stay in the Murgab region, but justified the long routes covered by these communities, at least from central Kazakstan or eastern Uzbekistan, where the copper and tin deposits were being exploited at that time. The close resemblance between the pottery and ornamental motifs of the Tazabag’yab group and the southernmost expressions of the Fedorovo facies allow us to assert that, besides the development of pastoral economy, a strong motivation to

Something else changed towards the end of the Bronze Age: built-up areas on platforms were introduced precisely during this period, and become characteristic of the subsequent Yaz I stage. Sometimes they were surrounded by a broad moat. At the end of the Late Bronze Age, ICW pottery disappeared completely, so it has been surmised that the flow, which had been constant during the previous period, was interrupted: in the north, a similar pottery production went on with regional cultural transformations (e.g. Begazy Dandybaj). 148

The Final Phase of the Bronze Age and the “Andronovo Question” in Margiana long-distance movements (with the partial return of part of the community to the places of origin) was precisely the transportation of metals, which were beginning to be used by the sedentary populations during this period.

d. We can attest to a high degree of interaction and to a perfect complementarity, but we are not able to evaluate it diachronically: it may have been a symbiosis that existed only for a season or a few years.

These long displacements do not necessarily suggest that all the ICW was imported from places distant from the Murgab. We believe, on the contrary, that the enormous quantities of pottery found there may have been produced in the delta area, since the raw materials are not lacking. Obviously adequate analyses are required, but we must not interpret this pottery as an importation or as the personal belongings of immigrants (Hiebert 2002: 243).

The study of the interaction between the Andronovo communities and the sedentary ones is the study of the process through which the two communities influenced each other by means of an exchange that was not only economic, but also involved thoughts, feelings and actions. It was an interdependence in which each action stimulated external reactions and internal readjustments. The idea of an ethnic boundary according to Barth (1964) postulates a social construction (conventional, provisional and contextual) whose purpose is to create a distinction between a group and the other groups with which it interacts. It becomes a mode of communication and contact, limiting the interaction and at the same time promoting it.

As regards what was obtained in the exchanges between the populations of Central Asia and the BMAC sedentary communities, there is no archaeological evidence, except for some vague indications of the presence of wheel-made pottery (Sarianidi 2002: 197; from the Tautara site, Masson 1992: 350; from Kokcha 15 in Khorezm, Hiebert and Moore 2004: 296); so perishable goods seem to be the likeliest option. Considering the behaviour of the Andronovo communities in the Murgab Delta, characterised by a complete lack of any type of acculturation, it is reasonable to imagine a rather closed society that deliberately excluded certain interactions (Lamberg-Karlovsky 2002: 16) and was oriented only to the exploitation of the available economical resources. This attitude may also explain the complete absence of traces in the Iranian area, a region that was beyond what was probably regarded as the limit of the interaction.

The evidence of the material culture of the Andronovo group found in the Murgab Delta shows a homogeneity of production features that persisted throughout its period of presence, from 1800 to 1500 BC. Apart from the introduction of new decorative motifs, the pottery was produced with the same shapes, and no cultural changes emerged, despite the high level of contact with the sedentary civilisation. The latter, on the contrary, showed a series of socio-economic changes between the Taip, Togolok and Takhirbai phases. An examination of the material culture of the ICW-bearing groups (Andronovo) does not reveal phenomena or individual signs of acculturation, of assimilation, or, even less, of cultural amalgamation. The preservation of identity, that is of the set of values, symbols and cultural patterns that are regarded as distinctive are connected with a strong link and continuity of the common origins. The latter is not defined by its primeval contents (blood, language, territory), but is constructed and invoked depending on the circumstances and with different meanings.

The economic basis is not sufficient to define the characterisation of this cultural group. Though the pastoral basis was a decisive motor for the development and formation of the group, there were some adaptation situations where animal farming coexisted, even to a substantial degree, with agricultural economy. In some cases the latter was characterised by canalisation works that perhaps denoted a developed type of irrigative agriculture (Khorezm). In a parallel manner, the actual nomadic aspect tended to disappear in the situations in which agricultural economy made sedentariness necessary.

We are faced, therefore, with a multiethnic society (or rather with a multicultural one), in which the context of an ongoing meeting and interaction between individuals belonging to different cultural traditions was normal rather than exceptional. This society was based on the complementarity of some of the cultural characteristics of the interacting groups, and this complementarity, in turn, was presumably dictated by an economic interdependence.

The boundary between the nomadic character and the sedentary one is not clear-cut or easy to deal with and above all to attest. In a general evaluation of it, the following factors must be considered: a. The sedentariness of contemporary groups in the northern steppes; b. The material culture of the nomadic camps was kept quite distinct from that of the sedentary communities, as regards their characteristics (pottery, decorations, etc.);

From the cultural point of view, the key processes consisted of the preservation of the differences between the separate traditions and ethnic boundaries, against a situation of massive interpenetration and exchange (Barth 1964).

c. This differentiation persisted for a long time in the same context (presumably at least 200 or 300 years);

We cannot exclude the conjecture that there arose a phenomenon opposite to the expected one, i.e. that there 149

M. Cattani was an acculturation mechanism so strong as to cause the Andronovo presence to be continuously integrated into the sedentary civilisation. The traces of the continuity of the presence of camps, in this case, would refer to new “arrivals” of carriers of the original culture from the northern contexts.

Hiebert, F. T. 2002. “Bronze Age Interaction between the Eurasian Steppe and Central Asia”, in Boyle, Renfrew and Levine (eds.) 2002: 237-248. Hiebert, F. T. and K. M. Moore 2004. “A Small Steppe Site near Gonur, in Kosarev, Kozhin and Dubova (eds.) 2004: 294-302. Itina, M. A. 1977. Istoriya stepnykh plemen Iuzhnogo Priaral’ya. Moskva. Kohl, Ph. L. (ed.) 1981. The Bronze Age Civilization of Central Asia. Recent Soviet Discoveries. New York. Kohl, P. L. 1984. Central Asia. Palaeolithic Beginnings to the Iron Age. Paris. Kosarev, M. F., P. M. Kozhin and N. A. Dubova (eds.) 2004. U istokov zivilizazii. Sbornik statiej k 75-letiyu Viktora Ivanovicha Sarianidi. Moskva. Kutimov, Yu. G. 2002. “Nekotorye aspekty razvitiya i absolyutnoj datirovki Tazabag’yabskoj kul’tury yuzhnogo Priaral’ya (po materialam mogil’nika Kokcha 3”, Arkheologicheskie Vesti 9: 189-203. Kuz’mina, E. E. 1971. “The Earliest Evidence of Horse Domestication and Spread of Wheeled Vehicles in Connection with Problem of Time and Place of Formation of Indo-European Unity”. Congrès international des sciences préhistoriques et protohistoriques VIII. Beograd. Kuzmina, E. E. 1985. “Classification and Periodization of Andronovo Cultural Community Sites”, IASCCA Information Bulletin 9: 23-46.

References Cited Askarov, A. 1962. Pamyatniki andronovskoj kul’tury v nizov’yakh Zarafshana.(Istoriya material’noj kul’tury Uzbekistana 3). Tashkent. Barth, F.1964. Nomads of South Persia. The Basseri Tribe of the Khamseh Confederacy. London. Boyle, K., C. Renfrew and M. Levine (eds.) 2002. Ancient Interactions: East and West in Eurasia. Mc. Donald Institute Monographs. Cambridge. Cattani, M. 1998. “Excavations at Takhirbai-depe (Takhirbai-1) (1992-1993). Preliminary Notes”, in Gubaev, Koshelenko and Tosi (eds.) 1998: 97-104. Cattani, M. and B. Genito 1998. “The Pottery Chronological Seriation of the Murghab Delta from the End of the Bronze Age to the Achaemenid Period: a Preliminary Note”, in Gubaev, Koshelenko and Tosi (eds.) 1998: 75-87. Cerasetti, B. 1998. “Prelminary Report on Ornamental Elements of Incised Coarse Ware”, in Gubaev, Koshelenko and Tosi (eds.) 1998: 67-74. Cremaschi, M. 1998. “Palaeohydrography and Middle Holocene Desertification in the Northern Fringe of the Murghab Delta”, in Gubaev, Koshelenko and Tosi (eds.) 1998: 15-25. Davis-Kimball, J., E. M. Murphy, L. Koryakova and L. T. Yablonksy (eds.) 2002. Kurgans, Ritual Sites, and Settlements: Eurasian Bronze and Iron Age. (BAR International Series 890). Oxford. Francfort, H.-P. 1989. Fouilles de Shortughai. Recherches sur l’Asie centrale protohistorique, 2 vols. Paris. Ganyalin A. F. 1956a. Tekkem-Tepe. (Trudy Instituta istorii, arkheologii i etnografii Akademii nauk Turkmenskoj SSR, T. II). Ashkhabad. Ganyalin A. F. 1956b. “Pogrebeniya epokhi bronzy seleniya Yangi-kala”, in Trudy Yuzhno-Turkmenistanskoj Arkheologicheskoj Kompleksnoj Ekspedizii, vol. VII. Ashkhabad: 374-384. Gryaznov, M. P. (ed.) 1966. Andronovskaya kul’tura. Moskva-Leningrad. Gubaev, A., G. A. Koshelenko and Tosi, M. (eds.) 1998. The Archaeological Map of the Murghab Delta. Preliminary Reports 1990-95. Reports and Memoirs. Series Minor III. ISIAO. Roma. Hiebert, F. T. 1993. “Chronology of Margiana and Radiocarbon Dates”, IASCCA Information Bulletin 19: 136-148. Hiebert, F. T. 1994. Origins of the Bronze Age Oasis Civilization in Central Asia. (American School of Prehistoric Research Bulletin 42). Cambridge.

Kuzmina, E. E. 2002. “The Eurasian Steppes: the Transition from Early Urbanism to Nomadism”, in Davis-Kimball, Murphy, Koryakova and Yablonksy (eds.) 2002: 118125. Kuz’mina, E. E. and A. A. Lyapin 1984. “Novyie nakhodki stepnoj keramiki epokhi bronzy na Murgabe”, in Problemy arkheologii Turkmenistana. Ashgabat: 6-22. Lamberg-Karlovsky C. C. 1994. Central Asia in the Bronze Age: Introduction”, Antiquity 68 (259): 353-354. Lamberg-Karlovsky, C. C. 2002. “Archaeology and Language: The Indo-Iranians”, Current Anthropology 43: 63-88. Lamberg-Karlovsky, C. C. 2003. “Civilization, State or Tribe? Bactria and Margiana the Bronze Age”, The Review of Archaeology 24 (1): 11-19. Marushchenko, A. A. 1959. “El’ken-depe. Otchet o rabotakh 1953, 1955, i 1956”, in Trydy Instituta istorii, arkheologii i etnografii Akademii nauk Turkmenskoj SSR V. Ashkhabad: 54-109. Masimov, I. S. 1979. “Izuchenie pamyatnikov epokhi bronzy nizov’ev Murgaba”, Sovetskaya Arkheologiya 1: 111-131. Masimov, I. S. 1981. “The Study of Bronze Age Sites in the Lower Murghab”, in Kohl (ed.) 1981: 194-220. Masimov, I. S. 1982. “Novye pamyatniki rannezheleznogo veka Murgabskogo oazisa”, in Novye arkheologicheskie otkrytiya v Turkmenistane. Ashkhabad: 20-32. Masimov, I. S., S. Salvatori and B. Udeumuradov 1988. “Preliminary Analysis of the Bronze Age Material 150

The Final Phase of the Bronze Age and the “Andronovo Question” in Margiana Collected by the Margiana Archaeological Project and a First Chronological Assessment”, in Gubaev, Koshelenko and Tosi (eds.) 1998: 35-46. Masson, V. M. 1959. Drevnezemledel’cheskaya kul’tura Margiany. (Materialy i issledovaniya po arkheologii SSSR 73). Moskva-Leningrad. Pumpelly, R. (ed.) 1908. Explorations in Turkestan. Washington. Pyankova, L. 1989. “Pottery Complexes of BronzeAge Margiana (Gonur and Togolok 21)”, IASCCA Information Bulletin 16: 27-54. P’iankova, L. 1993. “Pottery of Margiana and Bactria the Bronze Age”, IASCCA Information Bulletin 19: 109127. P’yankova, L. 1994. “Central Asia the Bronze Age: Sedentary and Nomadic Cultures”, Antiquity 68: 355372. Salvatori, S. 1995. “Protohistoric Margiana on a Recent Contribution. Review of IASCCA (International Association for the Study of the Cultures of Central Asia) Information Bulletin, 19, Moscow”, Rivista di Archeologia XIX: 38-55. Salvatori S. 1998. “The Bronze Age Margiana”, in Gubaev, Koshelenko and Tosi (eds.) 1998: 47-55. Sarianidi, V. I., 1972. “Izuchenie pamyatnikov epokhi bronzy rannego z heleza v Severnom Afganistanie”, Kratkie soobshcheniya Instituta arkheologii Akademii Nauk SSSR 132. Moskva: 16-22. Sarianidi, V. I 1975. “Stepnye plemena epokhi bronzy v Margiane”, Sovetskaya Arkheologiya 2: 20-29.

Sarianidi, V. I. 1976. Issledovanie pamyatnikov Dashlinskogo oazisa. (Drevnyaya Baktriya. Materialy Sovetsko-Afganskoj ekspedizii 1969-1973gg., 1). Moskva: 21-86. Sarianidi, V. I. 1981. “Margiana in the Bronze Age”, in Kohl (ed.) 1981: 221-255. Sarianidi, V. I. 1990. Drevnosti strany Margush. Ashkhabad. Sarianidi, V. I. 1994. “Preface”, in Hiebert (ed.) 1994: IXXXXXXX. Sarianidi, V. I. 1998. Margiana and Protozoroastrism. Athens. Sarianidi, V. I. 2002. Margush. Ancient Oriental Kingdom in the Old Delta of Murghab River. Ashgabat. Sarianidi, V. I. 2004. Paper presented to the International Conference Scientific Conference “Turkmenistan as Native Land of Anau Culture and White Wheat (Ak bugday)”, Ashgabat, October 22-23, 2004. Shchetenko, A. Ya. and Yu. G. Kutimov 1999. “Keramika stepnogo oblika epokhi posdnej bronzy Tekkem-depe (yuzhnyj Turkmenistan)”, Arkheologicheskie Vesti 6: 114-123. Tolstov, S. P. 1948. Po sledam drevnekhorezmijskoj zivilizazii. Moskva-Leningrad. Vinogradova, N. M. and E. E. Kuz’mina 1986. “Kontakty Stepnykh i zemledel’cheskikh plemen v Srednej Azii v epohu bronzy”, in Vostochnyj Turkestan i Sredneyaya Aziya v sisteme kul’tur drevnego i srednevekovogo Vostoka. Moskva: 126-151.

151

Chapter 11 An Aspect of the Early Iron Age (Yaz I) Period in Margiana: Ceramic Production at Site No. 999 Gian Luca Bonora* and Massimo Vidale**

Introduction

B. Piotrovskij worked at Gyaur-kala, a 14 ha Hellenistic settlement (Piotrovskij 1949: 35-41; Ershov 1944: 32-33). In this site, the first 3.5 m of stratigraphy above the virgin soil contained material that we now recognise as similar to that of the Yaz II and Yaz III assemblages. The first detailed archaeological investigations on the prehistory of Margiana began in 1946 under the auspices of Yu.T.A.K.E (Yuzhno-turkmenistanskoj arkheologicheskoj kompleksnoj ekspedizii). In 1952 the Takhirbai Oasis was discovered (Masson 1955: 204), and from 1954 to 1956 the XIV brigade of Yu.T.A.K.E., under the direction of V. M. Masson, documented the basic sequence for the Iron Age in Margiana. The settlement pattern in the Murgab Delta was the object of a very extensive archaeological survey, whose results were published in 1959 in a crucial report by V. M. Masson (Masson 1959). In the same year, A. A. Marushchenko published a detailed report on the first excavation trenches made in 1953, 1955 and 1956 at El’ken-depe, in the central piedmont belt of southern Turkmenistan (Marushchenko 1959: 54-109). The researches at El’ken-depe continued in the late nineteensixties under the direction of Kachuris (1967: 335-336). Stratigraphic soundings by G. Gutliev were also carried out at the site of Yasi-depe (near Kaakhka), which yielded materials similar to those from Yaz I-III (Gutliev 1970a: 64-71; 1977: 18-24).

Similarly to what happened for the archaeological study of the beginnings of the Early Iron Age in north-western Iran (Pigott 1980, 1990), the Yaz I-related cultures of western Turkestan were initially defined by specific changes in the material culture and assemblages rather than by the universal and sudden appearance of iron tools per se; at first, this metal was used mainly for ornamental purposes. The date of the Early Iron Yaz cultures in western Turkestan should be more or less contemporaneous with those recognised in north-western Iran, roughly from the middle to the end of the 2nd millennium BC. In this paper, we present the data gathered in a specialised pottery-making site of this period, originally found by M. Tosi and his colleagues in the north-western region of the Murgab Delta, and labelled Site No. 999 (Figure 11.1). On the surface of this site, in about one week of intensive surface analysis, we found no trace whatsoever of iron (objects or slag) and few copper-bronze items (Figure 11.25), although a single find interpreted as a crucible for the melting of copper (Figures 11.24.8 and 11.35) would imply a local smallscale copper-based metallurgy. More generally, our finds (indicators of craft production, the relative distribution patterns, a collection of potsherds) and the results of limited test trenches provide new archaeological evidence and contribute to a better understanding of this important stage of social evolution in central-western Asia

About 20 km north of Ashgabat, the site of Ovadan-depe contains a deposit whose deepest ascertained layers dated back to the early Chalcolithic period. It had been preliminarily investigated by A. A. Marushchenko in the nineteen-thirties and extensively excavated in 1959 by the same scholar. Some pottery fragments having Yaz I features were collected in the trenches and published without further information (Masson 1956b: Tab. XLI). A short survey carried out in 1974 by N. I. Khlopina brought to the identification of some sherds of the Namazga VI period on the surface of this site (Khlopina 1981: 37). Probably Ovadan-depe is the only site in the foothill plain north of the Kopetdag range that attests the Namazga VI-Yaz I stratigraphic sequence. However, from this information, the human occupation of the site during these periods seems to be limited in both duration and size. Only further researches may solve this problem.

MV, GLB. Previous Research The first archaeological work related to Margiana was carried out at the end of the nineteenth century, when archaeological potsherds were collected around Merv (Zhukovskij 1894). The earliest international research project in Margiana was conducted by E. Huntington in 1904 (Huntington 1908: 219-231). In 1916, D. D. Bukinich investigated the site of Gyumish-depe (Bukinich 1924: 98); and as late as 1937 B.

* IsIAO and University of Naples “L’Orientale”, Naples ** IsIAO and Faculty of Conservation of Cultural Heritage, University of Bologna (Ravenna)

About 3 km south of Gyaur-kale, the site of Yashilli-depe was preliminarily investigated by S. A. Ershov in 1939 and 153

G. L. Bonora and M. Vidale

Figure 11.1 Map of the Murgab Delta, with the location of Site No. 999.

1950, and by A. J. Shchetenko in 1967, by mean of a test trench (3 x 2 m) unearthing material belonging to the Yaz II and III period (Shchetenko 1968a: 20, 27-28; 1968b: 348). A few years later, G. Gutliev made a second test trench (4 x 2 m) at the centre of the mound, cutting the stratigraphic layers to a depth of about 6 m. From the 18th to the 13th arbitrary cuts, Gutliev found ceramic sherds belonging to the Yaz I period (Gutliev 1974: 504; 1980: 22)

Ulug-depe, near the mosque of Dushak, is another major site and is noteworthy for its long sequence extending from the Chalcolithic to the Achaemenid period. This site was excavated in 1950 by the “patriarch” of Central Asian archaeology, M. E. Masson (1951: 13-16), and later, in the late nineteen-sixties and early seventies, by a team jointly led by V. I. Sarianidi and I. S. Masimov (Sarianidi 1971: 433-434). These scholars carried out further researches in Margiana, with the help of A. A. Lyapin, in the nineteen-seventies and eighties. Many Bronze and Iron Age settlements and settlement networks (Kelleli, Egri Bogaz, Taip, Adam Basan, Uch-depe, Adzhi Kui, Gonur, Togolok, Takhirbai, Yakiper and Auchin-depe) were found and investigated along the dry courses of the ancient Murgab delta, in the midst of the desert, far north of the classical period boundaries of the Merv Oasis (Masimov 1982; Sarianidi 1981). Each cluster of archaeological sites was thought by the Soviet and Turkmen researchers to represent a single oasis fed by a branch of the Murgab River. Now we regard this view as a rather simplistic projection of the present-day situation into the past. Today the desert areas can be farmed only by means of a dense network of artificial canals implemented through huge labour investments (most probably impossible prior to the Soviet control), that could be sustained only at the cost of

The Yaz I layers at Gosha-depe, also identified by G. Gutliev, were only preliminarily excavated in the nineteenseventies, but the finds were never reported (Gutliev 1974: 504). The site of Garaoi-depe, located 1.5 km north-west of the railway station of Dushak (Gutliev 1982: 33-47) was first investigated in 1970 by means of a topographical plan and a brief surface collection. In 1978, G. Gutliev opened an extensive trench on top of this mound and brought to light a series of mud-brick rooms serving various functions, associated with sherds and other finds dating to the Yaz III period. Some test trenches were made inside the rooms exposed, and this led to the discovery of numerous sherds of Yaz I pottery, wheel- and hand-made, painted and unpainted. 154

An Aspect of the Early Iron Age (Yaz I) Period in Margiana: Ceramic Production at Site No. 999 Archaeological Back-ground

a gradual depletion of the water reserves. The existence of highly concentrated protohistoric and historic sites implies an environmental context that, in the past, differed considerably from that of today.

In Southern Turkmenistan the spread of “coarse” pottery and the small size of many settlements have been traditionally quoted as the proof of a widespread cultural decline in the second quarter of the 2nd millennium BC. But the idea that “coarser” pottery would necessarily imply a state of economic crisis is plainly ridiculous. Moreover, in this huge and little-excavated region, where the correlation between layers and sequences from different and faraway sites is frequently based only on a few potsherds or other strictly selected objects, arguments like these should be considered, at present, highly unreliable.

The “Archaeological Map of the Murgab Delta” project began in 1989, promoted by the former Institute of Archaeology of the Soviet Academy of Sciences of Moscow, the Turkmen State University of Ashgabat and the IsMEO (today I.s.I.A.O.) of Rome, under the joint direction of A. G. Gubaev, G. A. Koshelenko and M. Tosi. Within the framework of the project, between 1989 and 2001 several field campaigns, topographical mappings, excavations and surveys were carried out in various areas and sites of the delta, previously investigated or not (Gubaev et al. 1998). The project is still in progress.

In the second half of the 2nd millennium BC (Figure 11.2), the cultural division between a western area – including south-western Turkmenistan and the Gorgan Plain – and an eastern area – extending from the central part of the piedmont strip north of the Kopetdag Range (Atak) to the

GLB

Figure 11.2 A comparative chronological table for the Late Bronze Age-Iron Age sequences of the area, according to various authors.

155

G. L. Bonora and M. Vidale Bactrian plain, and also including the upper valley of the Atrek River and few other valleys in the Iranian Khorasan (the Kashaf Rud Valley) – became more evident and pronounced.

to some extent only in the late nineteen-fifties in the Yazdepe settlement, seem to represent a substantial break in relation to the previous Final Bronze Age. In particular, the replacement of the old Namazga VI wheel-thrown ceramics by different classes of handmade ware, and the resurrection of painted decoration, although limited to a small percentage of the vessels, seems to suggest a major change in the logic of ceramic production. Nonetheless, some form of continuity seems to be manifest in the ceramics from the Sumbar cemeteries and the Parkhai Iron Age settlements (Khlopin 1975) and could probably be documented on small settlements east and west of Kyzyl Arvat, in the western Kopetdag piedmont strip (Masson 1956b: 423-424). The matches presented by Khlopina, who compared the forms of the Late Bronze Age and Early Iron Age tableware in several settlements of the eastern Kopetdag piedmont strip, do not seem meaningful, and the table published is simplistic. Khlopina herself admits that her table is generic and lacking in details, but at the end she asserts peremptorily that there is a fundamental continuity between the pottery forms compared (Khlopina 1981). It is impossible, in fact, to compare the wheel-made pottery production of the Late and Final Bronze Age with the handmade one of the Yaz I period. E. V. Sajko (1978, 1980, 1982), espousing the same idea without a practical and critical approach, also argued for a continuous evolution of the pottery manufacturing technology. Furthermore, the excavation carried out by V. I. Sarianidi at Ulugdepe, in the western edge of the settlement, supported the idea that Yaz I materials could hardly be distinguished from Namazga VI ones, since they were both contained in two continuous stratigraphic layers, the sixth and fifth, each of them 50 cm deep (Sarianidi 1971: 433-434).

As widely reported by Soviet archaeologists, this pattern may have begun as early as the Middle Chalcolithic period, about 2500-3000 years before the second half of the 2nd millennium BC: the western region of Southern Turkmenistan was associated to grey ware decorated with geometrical incisions, while the eastern region was associated to a bichrome-ware production (patterns in black-dark brown on buff slipped background) and to the so-called Geoksyur ware (bowls, jars and small jars of various forms and sizes bearing geometrical motifs painted in black on a reddish slip). At present, archaeological evidence seems to support the idea of different courses of development during the Early Iron Age as well: in the west, Soviet and French archaeologists have brought to light the remains of the Ancient Dakhistan culture, spreading on the MeshedMisrian Plain, north of the lower course of the Atrek River, up to the south-eastern coasts of the Caspian Sea; in the east, the presence of a series of distinct cultural Early Iron Age complexes, collectively referred to as “Yaz I culture”, on the basis of the excavation of this site-type located in Margiana, is presently reported within a huge territory, from the Ashgabat province to Bactria and beyond. The origin and nature of the Ancient Dakhistan culture are still unclear, but the size of some excavated sites and the huge network of canals and ditches for artificial irrigation make it possible to surmise the existence of a highly organised, maybe politically unified culture (Lecomte 1999; Francfort and Lecomte 2002). The alluvial plain of Dakhistan presents a complex and peculiar geological formation that may have prevented large-scale agriculture without artificial irrigation. The plain appears to have been settled at least from the middle of the 2nd millennium BC by societies that, within an unsuitable territory and in the time-span of about a millennium, built and developed what was probably the largest irrigation network of Central Asia. The traditional view is that these groups, which produced grey ware ceramics, were culturally connected to the so-called Grey-ware cultures of north-eastern Iran and south-western Turkmenistan.

But perplexity and doubts spontaneously grow when we consider these statements, which are supported by excavations based primarily on indistinct arbitrary cuts about 50 cm deep. On the basis of such poor information, one should not conclude that the pottery tradition of Yaz I arose as a “natural” development of the ceramic assemblage of the Late and Final Bronze Age of Southern Turkmenistan and Margiana. Thus, in the light of the archaeological data piled up after 20 years of excavations, the original hypotheses proposed by Masson and Kuz’mina on the origin of the Yaz I painted decoration (and consequently of the overall culture) should be dismiss. V. M. Masson (1956b) and later E. E. Kuz’mina (1971; Kuz’mina and Lyapin 1984) had stressed a “close similarity” between the generic motives of hatched or cross-hatched triangles and bands of oblique lines with crossing segments of Yaz I, and the incised decoration of the Andronovo culture (FedorovoAlakul’ and Alekseevka phases) and of the Srubnaya or Timber-grave cultures of the Steppes cultures. Nonetheless, in the same reports, these authors remarked that the forms of the vessels were completely different. On the whole, a more careful approach to the whole assemblages and less thoroughgoing statements would have been preferable.

GLB The Early Iron Age Settlement Pattern in the Murgab Delta In the chrono-stratigraphical framework of Southern Turkmenistan and Margiana, the period following the Final Bronze Age is the Early Iron Age or Yaz period (Iron Age 1). The layers related to this period, excavated 156

An Aspect of the Early Iron Age (Yaz I) Period in Margiana: Ceramic Production at Site No. 999 In Southern Turkmenistan, assemblages dating from the second half of the 2nd millennium BC were found at Dashli, near Izgant, on the southern mound of Anau (Schmidt 1908), at El’ken-depe (about 15 ha), 20 km south-west of the modern village of Kaakhka (Masson 1951, 1956a; Marushchenko 1959; Khlopina 1981), at Yashilli-depe (Gutliev 1980) and Yasi-depe (Masimov 1968; Gutliev 1970a, b; 1977), both between the villages of BabaDurmaz and Kaakhka, at Ulug-depe (Sarianidi 1971; Masimov 1972), close to the city of Dushak, and in four small settlements, Sarik, Chemche, Kush e Khairak-depe, located in the Meana-Chaacha region, a few kilometres from the Chalcolithic and Bronze Age sites of Altyn-depe and �������������������������������������������������� Ylgynly������������������������������������������� -������������������������������������������ depe�������������������������������������� , on the lower Chaacha stream (Kuftin 1956). According to the most common view, these sites may have been established during the early Yaz I period by descendants of the former inhabitants of Altyn-depe. Farming and irrigation practices evidently continued in the same micro-region. Other four small sites of the same period were discovered near the modern village of Artyk (Shchetenko 1970). All the settlements are distributed in the piedmont belt of the Kopetdag, a situation that recalls the Neolithic-Chalcolithic-Bronze Age settlement patterns.

painted decorative patterns, some stylistic motifs present in the upper Atrek Valley (wavy lines and fish scales) are not attested in Southern Turkmenistan and Margiana, while others (chequers, triangles, lozenges and cross-hatched oblique lines) are common in both regions. The results of the archaeological exploration carried out in the nineteen-thirties in northern Khorasan by C. K. Wilkinson have been published recently (Hiebert and Dyson 2002). During this survey, Wilkinson discovered a prehistoric site named “Nishapur-P”, located in the Rud-e Kal Shur valley of the Northern Iranian Khorasan, a few kilometres south of the Kuh-e Binalud range and some 12 km north-west of the medieval town of Nishapur. The archaeological materials gathered present close analogies with those studied by Schmidt at Anau, southern mound. At Nishapur-P, the Yaz I-related materials were found in three PS, PE and PN soundings: they are formed of a sizeable group of very heavily hand-modelled sandy ware items, with painted motifs. This new recently-published information seems to confirm the presence of two major cultural areas stretching east to west, as previously stated. The presence of Yaz I sherds and the absence of archaic Dakhistan materials in the upper Atrek Valley and in the Nishapur-P layers suggests that this pattern of regional integration is well attested also in the Early Iron Age, and that it continued in later periods.

In this region, Yaz I-related sites generally appear to be smaller, though more numerous, than those of the Late Bronze Age. The settlements are scattered or dispersed across wide landscapes, in close connection with huge deltaic areas or the banks of major rivers.

In summary, according to the published information, Yaz I settlement patterns were dominated by medium-tolarge citadels or fortresses on raised mud-brick platforms with attached settlements and separate small agricultural “manors” or rural living compounds. Their archaeological assemblages contain predominantly hand-made pottery (some with painted decoration), the first iron artefacts, and double-barbed socketed bronze arrowheads, possibly indicating a connection with the northern steppe zone. There are no seals or terracotta figurines.

Pottery dating from the Early Iron Age was found in the upper Atrek Valley, north-eastern Iran, in a few, usually small, sites (XA 27, less than 1 ha., and XA 35), in two slightly larger sites (Tepe Faruj and Tepe Hajji Bibi, both about 1 ha), and in two other larger settlements (Tepe Yam, 6 ha and Tepe Shirvan, 3 ha) ����������������������� (Venco Ricciardi 1980; Kohl et al. 1982). These ���������������������������������������� sites, concentrated in the widest part of the valley, seem to cluster around the main site, Tepe Yam, that may have acted as a central place for this valley. The same socio-political picture is probably true also for El’ken-depe (as regards the Kaakhka area in Southern Turkmenistan), for Yaz-depe (in the western lower Murgab) and for Kuchuk-depe and Kyzyl-depe (in northern Bactria). The lack of carefully gathered data and information does not permit a similar hypothetical reconstruction of the political settlement system for the Dushak and Meana-Chaacha areas.

In Margiana, some centres, instead of being abandoned, moved farther south, or shrank in size. In this region, the geographic spread of the Early Iron Age settled area is impressive, and large sites are still on record (for example: Yaz-depe, 16 ha; Aravali, 7 ha; Kyzyl-depe, 22 ha). We must also allow for the fact that Early Iron Age sites, being on average smaller than their forerunners, are harder to locate on the surface. Only in exceptional cases sites as small as the Kyzylcha manors, surrounding the major settlement of Kyzyl-depe, in Northern Bactria, appear to have been preserved and have been positively identified on the surface. The change, in short, seems to have taken place in the socio-political sphere, with the rise of some

At the same time, we assume that each of these siteclusters grouped around a central place could represent a local variant of the Yaz I culture with distinctive and individual characteristics. For instance, in the upper Atrek Valley pottery commonly includes black-on-red slipped ware and occasionally plum-red-on-buff slipped ware. In Southern Turkmenistan, in the Kaakhka region and in Margiana for example, buff slipped ware predominates, while the red slipped one is more rare. Concerning the

 A first article on the work carried out at Anau south in 1994-2000 (Hiebert 2003: 226) was announced in the last-published issue of the journal IRAN (vol. 41) . But this text has probably not been completed yet. Thus, regrettably, the only book available concerning and depicting the Yaz-I material culture at Anau South is still Schmidt 1908.

157

G. L. Bonora and M. Vidale semi-autonomous potentates or decentralised regional centres over a wide geographical area, rather than brought about by a sudden and massive incursion of nomads from the steppes or the Iranian plateau. In particular, the packed presence of Yaz I settlements in the westernmost section of the eastern Murgab Delta (erroneously defined in previous publications as “Merv Delta Oasis”) seems to imply a noticeable cultural and political vitality and a strong economical growth in this region.

a section visible through a ditch (the dating, in any case, is a post-quem one). At present two major canals actually run parallel to each other from west-south-west to west-northeast for about 3 km. Small irrigation ditches or aryk have been traced departing from these canals at sharp angles. Canal 1 is 3.4 m wide and 1.2 m deep at its beginning, and 5.1 m wide and 0.8 m deep in its middle course, directly north of the site (Lisizina 1965; Lisitsina 1969). To return to the Early Iron Age, it was only when the political élite had the power of controlling and managing thousands of workers that large and powerful rivers such as the Zarafshan or the Amudar’ya could be tackled with, and the large deltaic plains, the surrounding alluvial lands, together with a longer series of ecological micro-regions (plains, low hills, marshes, terminal lakes, riverine banks and forests, artificial canals) could be fully exploit. The ethnographic, geographic and archaeological Khorezm mission led by Tolstov and Andrianov (Andrianov 1969; Itina 1977: 13-22; Tolstov 1948a, 1948b) showed that the settlements of the Kamyshli culture (late 3rd-early 2nd millennium BC) were located on upper alluvial terraces of the Akchadar’ya, and that the exploitation of the inner deltaic flatlands began around the middle of the 2nd millennium BC, thanks to the construction of canals and drainage works in the terminal branches of the delta. These works were ascribed to a “medium investment” technology definitely exceeding the range of the primitive hydraulics envisaged by Tolstov and Andrianov. Starting from the early 1st millennium BC, the economic growth and social evolution of the Central Asian societies of the Iron Age led to the construction and management of monumental hydraulic works. Incidentally, it is worth pointing out that the coincidence between the spread of fortified citadels on raised platforms and hydraulic investments in the agricultural territory represents, in the present picture, a perfect match with the core of Karl Wittfogel’s “hydraulic hypothesis” on the evolution of oriental despotism.

GLB Growth of Social Complexity: Was Wittfogel Right? The material culture unearthed at some Yaz I sites, on the other hand, shows innovations and noticeable improvements in subsistence and technology. Hydraulic resources available in medium- and large-sized river branches or in the inner endoreic deltas were exploited on a larger scale, and apparently in a more efficient manner, suggesting a final “conquest” of the riverine ecosystems. This becomes evident when we compare the locational strategies of the early Iron Age with those of the Early Chalcolithic, Chalcolithic and Bronze Age settlement networks. The sites were formerly located along the Kopetdag piedmont plains, particularly on the alluvial fans of minor streams and in the marshy environments of the terminal branches of the Tedzhen. In Tadzhikistan, along the middle Zarafshan valley, the centre of Sarazm (mid 4th-mid 3rd millennium BC), is an eloquent example of the absolute scarcity of resources; the local river, the Zarafshan, could not be fully exploited by the local communities. As a matter of fact, Sarazm is not located on the alluvial floodplain or in the deltaic fan, but on an alluvial terrace 100 m above the present river bed, within a narrow valley, and near the piedmont fan of a small stream (Lyonnet 1996). Although many authors agree (admittedly, on little archaeological evidence) that Neolithic and Chalcolithic farmers may have constructed small irrigation systems as far back as the late 5th millennium BC, these communities lacked the labour force, the surplus and the political inputs for damming and locking the water of the large rivers with large earthworks and canals running across long distances. According to traditional Soviet views, before the Iron Age hydraulic projects and canalisation were limited to small-scale local works, requiring limited investment of labour (“primitive hydraulic system”, in the MarxistLeninist lexicon of Tolstov and Andrianov: see Francfort and Lecomte 2002 as well). This hypothesis was also supported by Lisizina’s research into the “Geoksyur oasis” (i.e. the Tedzhen endoreic delta) in the 4th millennium BC (Namazga III). Here, the construction of irrigation canals was documented through the study of aerial photographs and excavations over an area of 400 sq m surrounding the site of Geoksyur 1. It is well known that the date of the irrigation network was apparently established by the discovery of a Geoksyur female terracotta figurine found in

In Margiana, the case seems to be quite similar to what was observed in archaic Dakhistan. The groups that reorganised the farming and settlement patterns across the Murgab Delta did not simply adapt to a new ecological reality, but planned and created new and more favourable subsistence opportunities by means of appropriate hydraulic works. In fact, “[...] Associated with meandering palaeo-channels, there are palaeo-channels which are straight […] interpreted as artificially modified river-beds or man-made canals. They are located in the belt of the Bronze Age sites and slightly north of it. Furthermore, they are directly linked to the Bronze Age sites and therefore may be considered contemporary to them” (Cremaschi 1998: 17). Current studies on the shifts and meandering activities of the Murgab Delta and on the attendant settlement changes from the Late Bronze Age to the 1st millennium BC point to the variation, in time, of the agricultural core areas. 158

An Aspect of the Early Iron Age (Yaz I) Period in Margiana: Ceramic Production at Site No. 999 Iron Age, starting around 300 BC, that they called “Late Yaz III”. GLB Monitoring Craft Production: Ancient Kilns in the Soviet Archaeological Literature The study of craft production played a prominent scientific role in Soviet archaeology as a key for pinpointing the development of social stratification and the emergence of class-based societies. It also directly reflected the Marxist concern with production and the belief in technical evolution as a fundamental motor of social change. Following a paradigmatic procedure, Soviet scholars constantly attempted to demonstrate that pottery-firing kilns were the object of continuous technical evolution from the Neolithic to the Iron Age. Their conclusions, in several cases, were impaired by what, by today’s standards, would be considered poor excavation techniques, by a total lack of attention to site formation processes (resulting in a general confusion between potsherds abandoned contemporarily with the activity of the kilns and pottery dumped at later times into their cavities) and by the highly selective publication of the associated finds. Moreover, because of the ideological and highly idealistic focus on the kilns as productive tools, their maps and sections did not achieve an effective rendering of the structural elements, and completely failed to record the associated stratigraphies (internal and external). Nonetheless, the quantity of unearthed data partly made up for these drawbacks. I. Masimov’s studies on the pottery-firing kilns of the Middle Bronze Age at Altyn-depe (in the so-called “Craftsmen’s quarter”) suggested that there were basically three types of kilns: circular, rectangular and spread out with a long entrance (Masimov 1971, 1973, 1976). Although our experience in South Asia suggests that perhaps these differences were partly due to different rates of erosion (Miller 1997), at least some of the infrastructures were two-tiered vertical kilns, with elongated oval shapes in horizontal section and the lower combustion chamber separated from the upper firing chamber by a perforated floor or grid.

Figure 11.3 Topographic map of the Iron Age site of Churnok, with the location, towards south-west and south, of peripheral potters’ quarters (from Masson 1959: 18, Fig. 7).

S. Salvatori’s reconstruction of the evolution in time of the Murgab settlement patterns between the Middle Bronze Age and the Early Iron Age identifies the Late Bronze Age as “[...] a moment of strong crisis, probably in the institutional frame [...] causing the disaggregation of the political system, with a beheading of the highest hierarchic settlement rank [....] Gonur north, whose centrality is no longer acknowledged [...]” (Salvatori 2000). GLB, MV Chronology It is well known that the basic prehistoric sequence for the foothill plain north of the Kopetdag range, from the Early Chalcolithic to the Late Bronze Age, was established in 1952 by B. A. Kuftin on the basis of 5 stratigraphic soundings carried out at Namazga-depe, the main settlement in the piedmont strip of southern Turkmenistan (Kuftin 1956: 260-290). The first Iron Age sequence for Margiana and Southern Turkmenistan was established by V.M. Masson after the excavation of Yaz-depe. G. A. Koshelenko (1985) and I. M. Diakonoff (1985) later proposed some minor changes. M. Cattani and B. Genito (1998: 75) correctly asserted that “[...] a revision of that chronological sequence, nevertheless, appears more and more necessary today[...]”. After the analysis of a substantial amount of potsherds datable from the Late Bronze Age to the ends of the Late Iron Age, they identified a new phase of the

Some Middle-to-Late Iron Age kilns were excavated at Churnok (Figure 11.3), the southernmost settlement in the Takhirbai Oasis (Masson 1959: 65-66). Churnok is located 4.5 km north of Garry Kishman. The site extends from south-east to north-west for about 400 m. The mounds are rather low, and much of the site is marked by clusters of potsherds loosely scattered on the takyr. In the northern part of the site, a hummock having a roughly egg-shaped contour is about 6 m above the surrounding plain; towards the south-east; on the surface of a separate elongated mound, a series of pottery firing kilns was visible (two of them were excavated in 1955). The topography of the mound suggests a separate, peripheral ceramic production area or quarter. The kilns are reported as structurally identical but 159

G. L. Bonora and M. Vidale

Figure 11.4 Drawings of the structure of Iron Age pottery-firing kilns from Uch-depe (left) and Churnok (right) (from Masson 1959: 67, Fig. 17).

having different sizes (Masson 1959: Fig. 27.2); one, in the map, has a diameter of about 2 m, while the second and largest one reportedly reaches the considerable diameter of 4.5 m. The combustion chambers are bell-shaped, and that of the largest kiln is at least 2.75 m deep. They are sunk into the ground, and were internally lined with mud bricks and plastered with a thick layer of chaff-tempered clay. The walls and bricks of the first kiln (Figure 11.4, right), looked compact and tough. The fuel canal or mouth, giving access to the deep-sunk, bell-shaped combustion chamber, was slightly lowered at the centre, and perhaps had been originally covered by a small arch-like element, as attested by similar kilns (Yacobson 1941). At the top of the combustion chamber, through the floor of the upper firing chamber there ran seven flues, arranged in a radial fashion and interpreted as openings or “vents for air supply” (our understanding of the Russian text is poor, but the map clearly shows that these flues brought the hot combustion gases into the upper firing chamber). Other two vents ran into the floor of the firing chamber. The vents were built with “clay blocks” and covered with large mud bricks measuring 40 x 20 x 10 cm; when they were found, they were filled with soft earth. Some of the bricks in the collapsed part had curved concave sides, and this suggested that they had been used in a dome-like structure. The upper firing chamber had a round plan and could be conveniently covered by a temporary dome.

the report, perhaps connected the combustion chamber to the exterior of the kiln, a circumstance difficult to be accounted for. The fuel canals or mouths of the two kilns have different orientations, possibly suggesting that the kilns had been not contemporary, or were used in alternate seasons, with different prevailing wind directions (but the fact that in Margiana the wind blows decidedly and constantly from the north contradicts this idea). Both the kilns were well planned and constructed; it is difficult to explain the depth of the combustion chambers, which were apparently lacking in an adequate supply of air (but we must remember the problem with the structure of the second kiln); the excavators thought that the explanation might be that the fuel (dry wood) was piled into these chambers in substantial amounts. Charcoal remains in the first kiln were identified as saksaul. In the sandy filling within both kilns some pieces of fired bricks, probably unintentionally baked while firing pottery, were gathered. The associated sherds indicated a dating within the Yaz III period. Surface evidence of iron slag and limited test trenches in the same area suggested that iron work was also practised in that location, but the actual craft area was not identified (Kohl 1982). The southernmost site of the Takhirbai Oasis, Churnok, seems to have been mainly a centre for ceramic production, and if the report is reliable, the huge size of the kilns, greater than any known Bronze or Iron Age examples, suggests an incomparably larger production volume (Sarianidi 1957).

The second kiln, a few metres away (not illustrated), was equipped with eleven flue-holes. Here we are faced with a problem, because only one of the flues, according to the excavators, ran through the floor from the combustion cavity into the firing chamber; the others, according to

Other two kilns were excavated in the northern extension of Uch-depe. One, smaller, was better preserved: the walls of the firing chamber survived to a height of 0.6 m (Figure 11.4, left; Masson 1959: Fig. 27, 2). The cylinderlike combustion chamber, sunk into the ground as in the 160

An Aspect of the Early Iron Age (Yaz I) Period in Margiana: Ceramic Production at Site No. 999 Churnok examples, was crowned by a multiple series of radiating vents, formed of fragments of fired bricks and surrounded by them. The vents, at the upper edge of the combustion chamber, were arranged into three concentric circles, connecting the upper and lower chambers. The outermost circle contained sixteen large flue-holes; then there was an intermediate circle of other sixteen smaller holes, and finally an innermost circle of seven holes of the same size. The arrangement of the flue-holes is very ingenious, because the larger flue-holes run along the walls of the vault, the point of maximum heat dispersal, and the other circles of holes ensure a very uniform spread of the heat in the firing environment. The flues, according to the report, were coated or filled with “clay pellets”, possibly elements used for regulating the air flows in the firing chamber; the outer circle vents were made of peculiar hand-formed clay rings (Masson 1959: Tab. 27.16). These clay rings had been observed in previous surveys of early sites in Margiana (Masson 1955: 242). The rings, during the firing, perhaps were also used as supports for small containers. Hand-made clay balls with the imprint of bases and rims show that while the kiln was being loaded they could be used as supports for the unfired vessels. The upper firing chamber was covered by a dome-like temporary element. The fuel canal or mouth was made of bricks (36 x 12 cm; 32 x 12 cm): three superimposed rows of them survive. A large potsherd near the mouth had perhaps been used to replace a lid or door. The associated potsherds belonged exclusively to small-sized containers, cups, cylindrical vessels and beakers, whose shapes were well represented at Yaz-depe in Yaz III levels.

is described as formed by unpainted light-slipped bowls, dishes and cooking pots. A more detailed description of the Yaz-I pottery was made possible by the findings of the Uch-depe 10 site. The pottery assemblages from Uch-depe 10 were divided first of all into two main groups: wheel-thrown (about 27.5%) and hand-made (72.5%). The wheel-thrown vessels were further divided into two classes: I) burnished grey-black ware; and II) plain thrown ware. The former (about 35% of the wheel-thrown production) appears is grey or black in its fracture surfaces, with clear, evident traces of interior burnishing. The temper always contains crushed potsherds (hereafter called grog) that have been finely ground. The quality of this class is considered high. It includes large bowls with a slightly inward- or outward-projecting rim (Masimov and Udeumuradov 1984: respectively, Figs. 4.5-7-8 and 4.11). Some bowls have a thin horizontal projecting ledge below the rim (Ibidem: Fig. 4.2). Other shapes are deep open bowls (Ibidem: Fig. 4.1), small pots (Ibidem: Figs. 4.6 and 10), large jars with horizontal plastic strips applied to the shoulder (Ibidem: Fig. 4.3) and sub-cylindrical containers (pots or beakers) with a slightly outward-projecting rims (Ibidem: Fig. 4.9). Some forms in this class have lugs (Ibidem: Figs. 4.12 and 13), often applied horizontally just below the rim, or vertically on the shoulder. Similar forms were found also at Yaz-depe (Masson 1959: Tabs. XXI, XXII), in the Dal’verzin site, in the Fergana valley (Zadneprovskij 1962: Figs. 6.11 and 13) and in other secondary sites. The second class, that of plain ware, is reddish, pinkish or buff in its fracture surfaces, while on the outside it has a lighter hue (yellowish to pale green). It is heavily tempered with coarsely ground grog. The vessels, in this class, are larger-sized: jars with outward-projecting rim (Masimov and Udeumuradov 1984: Fig. 5.1-7-9 and 10), sometimes rounded (Ibidem: Fig. 5.3-4, 13-14). These vessels sometimes have thin plastic strips running horizontally below the rim (with rounded or square sections, outwardprojecting) (Ibidem: Figs. 5.13-14, 6.2). There are also high globular pots with truncated-cone-shaped neck (Ibidem: Fig. 5.6 and 11), small pots with cylindric neck (Ibidem: Fig. 5.12), low truncated-cone-shaped bowls (Ibidem: Fig. 5.8), shallow dishes (Ibidem: Fig. 5.5), and flat dishes with lateral lugs (Ibidem: Fig. 5.15). These types are good matches with other contemporary productions. Lids and dishes of the same shape were found in the layers of the second period of Kuchuk-depe in Uzbekistan (Askarov and Al’baum 1979: Tabs. 10.6, 8.11-12, 11.24-25) and in the lower levels of Yaz-depe (Masson 1959: Tab. 23.3-4). On the same archaeological levels, vessels with thin clay strips running horizontally below the rim, both with semicircular and rectangular sections, were found (Askarov and Al’baum 1979: Tab. 14.4-9).

GLB, MV Yaz I Pottery: Preliminary Descriptions As we have already stated, the “typical” Yaz-I type pottery was been simply described as hand-made, in the wake of the idea of a complete break with previous Bronze Age traditions. In the oldest layers of the Yaz I culture only 5.5% of the vessels was made on a potter’s wheel; later the figure reached 14.5%, but even this is a very small proportion. It should be noted, however, that the small percentage of wheel-made pottery has very few direct connections with the preceding Final and Late Bronze Age pottery. Hand-made pottery may be divided into three sub-groups. The first is painted ware, mostly in the form of bowls and globular-shaped pots. The decorative patterns are described as predominantly geometric, often rather intricate, and applied with reddish-brown or often greenish paint. The variety of patterns at the site of Yazdepe is much more limited than at El’ken-depe and Ulugdepe, where there are many instances of peculiar, entirely new designs. The second pottery group comprises the grey ware, often carefully burnished, in the shape of bowls and jars with looped handles at the rim, while the third group

Hand-made ware is distinguished by pinkish-buff colours; in 23.8 % of the cases the clay paste is fine, without visible 161

G. L. Bonora and M. Vidale Class Arbitrary cuts and period XI –Yaz I-II XII – Yaz I XIII - Yaz I XIV – Yaz I XV – Yaz I XVI – Yaz I

Hand-made Painted Ware

Hand-made Unpainted Ware

Grey Burnished Ware

Wheel-made Ware

6% 18% 25% 20% 24% 22%

42% 50% 54% 52% 59% 60%

8% 4% 3% 6% 4% 9%

44% 28% 18% 22% 13% 9%

Table 11.1 Percentage of the different types of ware recovered from Yaz I arbitrary cuts at Yashilli-depe.

inclusions; 9.5 % of the potsherds contain a finely ground grog, while the remaining 66.7 % of the sherds contains a medium-to coarse grog component. In their turn, these sherds have been further divided into two sub-classes, called I) painted and II) unpainted ware (Masimov and Udeumuradov 1984: Fig. 6).

percentages have not been published, but the description is very clear: “[…] the ratio in the percentage identified between those two classes [hand-made and wheel-made] is more close to the similarity of approximately half quantity each other, thought it is not supported by extensive excavation activities, but only by a surface collection […]” (Ibidem: 76).

Painted ware is uncommon (about 4-5%). There are brown or red drawings on a lighter background. The designs include friezes of triangles filled with thin parallel strokes (Ibidem: Fig. 6.13 and 14) or with an oblique hatching (Ibidem: Fig. 6.15). Friezes are sometimes bordered by horizontal bands, and often run along the shoulder of the vessel. This is the oldest feature of the painted pottery of Yaz-depe (Masson 1959: Tabs. XVII.14, 19; XIX.7; XX.22) as well as of the early pottery of Tillya-depe, in northern Afganistan (Sarianidi 1972, 1977: 108-110). The painted pottery from Uch-depe 10 also includes cups (Masimov and Udeumuradov 1984: Fig. 6.13) and wide pots with a short neck (Ibidem: Fig. 6.15). The unpainted pottery largely shares the same shapes, for example some small jars and pots (Ibidem: Fig. 6.1-8, 10) and bowls, both truncated-cone-shaped and sub-cylindrical.

While the published information is consistent in presenting some basic features of the Yaz I assemblages (for example, the basic features of painted ware, and their association with grey and coarser ware characterised by plastic decorations in the form of cords or sheets of clay and lugs applied on the outer surfaces), we feel that a major source of confusion consists in the diagnosis of wheel-throwing and hand-making techniques. After a careful scrutiny of the collections from Site No. 999 (see below), we actually feel that the greatest majority of the vessels was manufactured by combining coiling and wheel throwing, a basic composite ceramic technique widespread in the 3rd and 2nd millennia BC from India to the Mediterranean Sea. As stated above, the clear-cut classes proposed in the past, where the two principles are opposed, instead of being combined, are probably quite arbitrary and incorrect.

The small proportion of wheel-made ware in the Yaz I period is well attested also by the ceramic assemblage unearthed at Yasi-depe, in the test trench published in 1977. In fact, Gutliev describes the wheel-made sherds as “singular and rare finds”, while hand-made vessels are the great majority of the pottery corpus (Gutliev 1977). The same situation is attested at Yashilli-depe, as confirmed by Table 11.1, which shows the percentages of different ware types recovered from the arbitrary cuts dating from the Yaz I period (Gutliev 1980).

GLB A Short Survey of Site No. 999 Site No. 999 (Figures 11.5-7) came to the attention of Maurizio Tosi and Sandro Salvatori because on the surface it showed a substantial presence of archaeological indicators of pottery production, and a ceramic assemblage clearly datable as Yaz I (i.e., as stated above, second half of the 2nd millennium BC). In September 2002, on request of Maurizio Tosi and Sandro Salvatori, Massimo Vidale carried out a short surface survey of the site. The principal activity areas on the surface (defined on the basis of the surface clustering of craft indicators) were mapped and described. Two very limited test trenches were excavated, in order to observe the relationship between the surface and sub-surface record, and to establish the local archaeological sequence. A substantial number of potsherds, mainly formed of rim sherds and a few other diagnostic pieces

A completely different picture (probably incorrect or unrealistic, as a result of a poor understanding of the technical attributes that distinguish coiled from wheelthrown pottery) comes from the statistics of the technological classes in the test trench made in 1970 at Yasi-depe by G. Gutliev. Here, in the lower Yaz I arbitrary cuts (from the 9th to the 4th), the percentage of wheel-made pottery is 59%, while that of hand-made pottery is 41% (Gutliev 1970a). These unexpected values, however, probably match with those obtained at Takhirbai (Cattani and Genito 1998). The 162

An Aspect of the Early Iron Age (Yaz I) Period in Margiana: Ceramic Production at Site No. 999

Figure 11.5 Topography and surface distribution of the craft activity areas at Site No. 999.

Figure 11.6 Three dimensional contour map of Site No. 999.

163

G. L. Bonora and M. Vidale The canals excavated for farming activities must be kept constantly clean, or they soon become clogged. Around the site there are also some takyr, stretches of alluvial plain that emerge from the movement of the sand dunes, and are marked on their surface by light-coloured alkaline films. These natural basins, in spring, serve as seasonal water reservoirs and attract grazing animals and their predators. They also support temporary camps for caravans and nomadic herders. If a few potsherds were found in the site they might suggest a human presence from the beginning of the 2nd millennium BC (Figure 11.24.1-3). Intensive surface surveys carried out by walking through the site and mapping the historical evidence, with the support of low altitude aerial vertical photographs and satellite imagery, pointed to the presence of no less than two paleochannels at the north-western edge of the site. Another possible paleochannel (not visible in the satellite images) seems to connect the settlement of Adam Basan 1 to Adam Basan 5 or to Site No. 1047. This connection is suggested by a series of archaeological spot areas evidently arranged in a straight line; in this light, the channel might even be interpreted as an artificial feature. At the same time, nonetheless, we are not sure that this orientation might not simply be the result of an off-site walking transect moving south-north from Adam Basan.

Figure 11.7 Site No. 999, view of the general area from the centre of the site.

(brown ware, grey burnished ware, painted buff ware) was systematically collected and recorded, and is a part of the present report. It will be used for providing, together with the information already published on the subject, a first ceramic reference for the archaeology of the Yaz I complexes of the Murgab. These activities brought to light new and substantial evidence on the technology of potterymaking at the beginning of the Iron Age in Margiana. This evidence includes new information on the ceramic production cycles, ranging from the procurements of raw material, the preparation of grog-tempered mixtures, the forming techniques (coiling and slab construction, associated or not to the use of the potter’s wheel), to the surface finishing and above all the firing; the latter was illustrated by the recovery of a well-preserved potteryfiring kiln whose construction employed ingenious and previously unsuspected manufacturing techniques. Finally, the report discusses the implications of the presence of a specialised ceramic manufacturing centre like this in the overall frame of the Yaz I settlement pattern in the Murgab Delta.

GLB, MV Surface Survey and Activity Areas The extension of Site No. 999 is about 120 x 120 m. The topography is low and irregular, and in the flat surrounding landscape it is not easy to achieve a constant proper orientation. The settlement rose on a natural bank north-west of a takyr. The present topography reaches a maximum elevation of about 1.8 m at the northern limit of the site; immediately to the west and north of the takyr there are two minor mounds spreading from the northern slope, reaching a maximum height of about 1.5 m. A wide, low erosion basin from northwest to southeast separates them. A slightly S-shaped track runs along the erosion basin or gully, and crosses the site in the same direction. The topmost surfaces, as a rule, are free from grass and bushes, as well as from artefacts. The central gully separates two different parts of the site (Figures 11.6-7).

MV The Site Site No. 999 is a low, small mound on the north-western fringe of the ancient delta of the Murgab. It is located some 17 km north-north-west of Uch-depe and a few kilometers north-east of the Late Bronze Age site of Adam Basan 1, near the paved road that connects the modern “Zelinnij” and “Moskva” villages (formerly sovkhoz) to the “Kompressor” industrial plant, built between the archaeological cluster area of Taip-depe and Kelleli. In the dissected landscape of the terminal riverine branches and their recent transformations, Site No. 999 is almost invisible even at a short distance. It is surrounded by a nullah: small sandy dunes, uninhabitable reedy marshes and dense tugai bushes spreading along the edges of the marshes. Here occupation and soil exploitation are difficult.

On the north-western and western slopes of the area, Activity Areas (hereafter AA) 1-6 were distinguished by a continuous spread of fragments and gravel-sized crumbles of yellowish chaff-tempered walls and grid-like perforated slabs from kilns. Within these surface concentrations, some points showed dense clusters of yellowish crumbles, pieces of walls and grids. These clusters were 1/1.5 m wide, and thick enough to suggest, as a first impression, that they covered some small infrastructure in situ. The impression of small ovens or kilns was also supported by 164

An Aspect of the Early Iron Age (Yaz I) Period in Margiana: Ceramic Production at Site No. 999 Class

Dishes

Basins

Brown-buff Grey burnished Painted buff

+ -

+ -

Bowls (hem.) + + +

Bowls (tc) + -

Small-to-medium Jars or Pots + + +

Medium-to-large Jars or Pots + +

Very Large Jars + -

Table 11.2 Distribution of pottery classes through functional types.

the apparently small size of the perforated slabs observed and collected on the surface. AA 1, in particular, contained three of these clusters, and at its northern limit a fragment of a possible copper melting crucible, distinguished by the mark of a detached vertical handle, was found (Figures 11.24.7 and 11.40). AA 4 and 5 also contained secondary clusters of fragments of overfired bricks and thick overfired kiln walls; AA 6 was entirely formed of a cluster of these overfired ceramic pieces.

chronological-cultural marker of Yaz I complexes across a wide geographical area. In our sample, it is distinguished not only by painted designs that, with few exceptions, are rather repetitive and standardised, but also by the use of fine clays with scarcely visible inclusions or no inclusions at all. When present, inclusions were described, after a simple visual inspection, as a sandy component, or, in other cases, as finely crushed ceramics or fine grog temper. While in the case of sand this tempering component may have been deliberately added or (more likely) may have been a natural component of the local clay, the use of grog would be an explicit cultural choice; in painted buff ware ceramics, nonetheless, the actual presence of finely ground ceramic tempers, for the moment, is only a working hypothesis.

In contrast, along the eastern and south-eastern slopes, down to the edge of the southern takyr, AA 7-10 contained very few chaff-tempered walls and grid-like perforated slab fragments, but quite a number of large overfired kiln blocks. These areas also included the most important surface concentrations of potsherds. Some of the ceramic fragments in these AA (both painted buff ware and grey burnished ware) were overfired to green-black and in some cases slightly warped, as expected in firing misproducts. Thus, although it appeared highly likely that specialised pyrotechnological activities had been carried out in the site (including the production of at least two ceramic classes), we wondered about the precise meaning of the differences observed in the surface AA. These questions were solved later, in our opinion, by excavating two test trenches (see below), one in the western slopes and one at the centre of the site, where the central pillar and an arm of a wellpreserved kiln were discovered on the very day in which Vidale had planned to leave the site (Figure 11.41).

Besides painted buff ware, the ceramics visible on the surface of Site No. 999 fall into other two classes, defined here as brown-buff ware and grey burnished ware. Brownbuff ware, as a class, includes various types of products, distinguished generally by oxidising firing conditions and variable textures. As a rule, brown-buff ware has grog temper components of various sizes and shapes, ranging from large pottery fragments, described as coarse rounded grains, polyhedral particles or even large laminar pottery flakes, to (less frequently) finely ground particles that are very difficult to recognise with the naked eye. The shape of the brown-buff ware containers is sometimes rough, and the surface finish is coarse. A particular category of this brown ware consists of a series of sub-globular-tocylindrical coiled jars with lugs or plastic cords stuck to the shoulders, whose grog temper is distinctively coarse. Brown-buff ware may partially merge into grey ware, particularly when vessels of this type were fired in a very small space, but grey burnished ware forms an easily recognisable class on its own. Many of these vessels are distinguished by the use of fine clay without visible inclusions (although some specimens seem to include grog temper components of variable sizes), are manufactured with great care, and their external surfaces are carefully polished with hardstone polishers. The table below (Table 11.2) shows the typological division of these three fundamental classes at Site No. 999, into the most relevant functional pottery types.

MV Ceramics One of the purposes of the surface research at Site No. 999 was the gathering of a good, homogeneous sample of the Yaz I painted buff ware production. Consequently, in the course of various days, Vidale obtained a complete collection of all the rim sherds of this ceramic class, integrating the sample with a few wall fragments that had particularly relevant painted designs. Other ceramic classes (see below) were not systematically sampled, and, although Vidale attempted to collect a reasonable range of typological variations, the collection and the catalogue presented here cannot be regarded as an exhaustive description of their variability.

Empirically, it seems that painted buff ware and grey burnished ware were relatively fine and carefully manufactured products, whereas brown-buff ware was the ordinary pottery to be used in everyday domestic work; but this impression may well be subjective. Painted buff ware

Painted buff ware is highly visible and easily recognisable, and has been widely used by former researchers as an useful 165

G. L. Bonora and M. Vidale

Figure 11.8 Site No. 999, surface ceramics. Brown ware, grogtempered dishes.

and grey burnished ware (particularly the latter), at any rate, are limited to vessel shapes that may have been used for serving and consuming drinks and/or food, the vessels of the group called “small-to-medium jars and pots”, in absence of other candidates, may have been used as beakers. In contrast with the lavish painted designs of painted buff ware, in brown-buff ware the painted decoration is limited to rare instances of coarse bands painted along the rims, and grey burnished ware was left entirely unpaint bowls and small-to-medium jars and pots, the most common categories, are reproduced in the three basic varieties of pottery: may be the three ceramic classes expressed not only different activities but also different social ranks or statuses. The following catalogue of ceramics and other finds represents the totality of the materials collected by Vidale on the surface of Site No. 999.

Figure 11.9 Site No. 999, surface ceramics. Brown and greyish ware, grog-tempered to finer mixtures, hemispherical bowls or basins with upper S-shaped wall.

1) Fragment of a large, possibly hemispherical bowl or basin with an upper S-shaped wall. Brown ware possibly tempered with a medium-fine grog fraction. Colour: 10 YR 6/2. Made by coiling. Site ����������������������������������� No. 999, surface. ������������ 2) Fragment of a large hemispherical bowl with an upper S-shaped wall. Grog-tempered greyish-brown ware (grog particles are polyhedral). Colour: 10 YR 7/2. Made by coiling. Site No. 999, surface. 3) ������������������������������� Fragment of a hemispherical bowl with an upper S-shaped wall. Grog-tempered brown ware. Colours: 5YR 6/4 (exterior); 10 YR 8/3 (interior). Made by coiling. ���������������������������������������� Site No. 999, surface. 4) ����������������� Fragment of a hemispherical bowl with an upper S-shaped wall. Grogtempered brown ware (polyhedral grog particles). Colour: 2.5 Y 8/2. Apparently formed of large flat slabs pressed one onto the other, perhaps inside a mould. �������������� Site No. 999, surface. �������������������������������������������� 5) Fragment of a hemispherical bowl with an upper S-shaped wall. Brown ware, fine, with few evident inclusions. Colour: 7.5 YR 7/4 (exterior, oxidised); 10 YR 8/4 (interior, reduced). Site ����������������������� No. 999, surface.

MV Catalogue of Pot-sherds Collected on the Surface Figure 11.8 Brown-buff ware grog-tempered dishes. 1) Fragment of a thick, low dish. Coarsely grog-tempered ware with granular inclusions. Colour: 10 YR 7/4. Formed of slabs. ���������������������������������������������� Site No. 999, AA. 10. 2) ������������������������ Fragment of a thick, low dish. Coarse grog-tempered ware, possibly using large particles of fired plaster as temper. Colour: ������������������������ 10 YR 7/3. Site ����� No. 999, AA. 1. ����������������������������������������� 3) Fragment of a thick, low dish. Coarse grog-tempered ware with granular inclusions. Fragment of a thick, low dish. Coarse grog-tempered ware with granular inclusions. Colour: 7.5 YR 5/4 (interior reduced). Slab construction. Site ������������������������������������ No. 999, AA. 4. ��������������� 4) Fragment of a thick, low dish. Coarse grog-tempered ware with large polyhedral, angular inclusions. Colour: 2.5 Y 8/4. Formed of slabs and probably shaped on a wheel. Grey burnished ware, unrestricted and restricted hemispherical to subglobular bowls and basins (various sizes).

Figure 11.10. Brown-buff ware, rim fragments of jars or large restricted basins. 1) Fragment of a jar or large basin with thick walls and an upper S-shaped wall. Grog-tempered brown ware (grog particles are granular). Colour: 7.5 Y 8/4. Formed of coils and slabs. ��������������������������������������������������� Site No. 999, surface. 2) ���������������������������� Fragment of a restricted basin. Grog-tempered brown ware (the grog particles are granular). Painted decoration: a thick coarse line traced on the rim. ����������������������������������������������� Colour: 2.5 Y 8/4, pigment 7.5 YR 5/4. Coiled. �������� Site No. 999, surface. 3) Fragment of a restricted basin. Grog-tempered brown ware (the grog particles are granular but large laminar pottery flakes are also present). Colours: 10 YR 8/4. Formed of coils and perhaps slabs. ��������� Site No. 999, surface.

Figure 11.9. Brown-buff and greyish ware, grog-tempered to finer mixtures, hemispherical bowls or basins with an upper S-shaped wall. 166

An Aspect of the Early Iron Age (Yaz I) Period in Margiana: Ceramic Production at Site No. 999

Figure 11.10 Site No. 999, surface ceramics. Brown ware, rim fragments of jars or large restricted basins.

Figure 11.11. Brown-buff ware, large bowls, hemispherical to truncated-cone-shaped. 1) Fragment of a large hemispherical bowl with S-shaped upper contour and horizontal semi-circular handle applied to the rim. Fine clay with scarcely visible inclusions. Colour: 10 YR 8/3. Coiled. Site No. 999, surface. 2) Fragment of a large hemispherical bowl with an external protruding horizontal ridge at the rim. Fine, compact clay with scarcely visible inclusions. Colours: 10 YR 7/4. Formed of coils, wheel-finished, lower external surface trimmed on a potter’s wheel. ����������������������������������� Site No. 999, surface. ������������ 3) Fragment of a large hemispherical bowl with an external protruding horizontal ridge at the rim. Fine, compact clay with scarcely visible inclusions. Colours: 10 YR 8/4 (exterior), 7.5 YR 8/6 (interior). Formed of coils, wheel-finished, lower external surface trimmed on a potter’s wheel. ����� Site No. 999, surface. �������������������������������������� 4) Fragment of a large truncated-coneshaped bowl, slightly S-shaped. Fine, compact clay with scarcely visible inclusions. Painted decoration: a coarse, thick band traced on the mouth. �������������������� Colours: 10 YR 8/4, pigment 5 YR 5/4. ������������������������������������� Formed of large coils exposed by the breakage of the vessel. ����������������������������������� Site No. 999, surface. 5) ������������ Fragment of a large truncated-cone-shaped bowl, slightly S-shaped. Fine clay, possibly mixed with very fine grog inclusions. Colour: 2.5 Y 8/4. Formed of carefully joined coils. ����� Site No. 999, AA. 5. �������������������������������������� 6) Fragment of a large truncated-coneshaped bowl, slightly S-shaped. Fine clay with scarcely visible inclusions. Colour: 10 YR 7/4. Coiled. Site No. 999, surface. 7) Fragment of a large truncated-coneshaped bowl. Coarse grog-tempered brown ware. Colour: 2.5 Y 8/4. Formed of large, badly joined coils. Site ��������� No. 999, surface.

Figure 11.11 Site No. 999, surface ceramics. Brown ware, large bowls, hemispherical to truncated-cone-shaped.

tempered brown ware (the grog particles are rounded to polyhedral). Colour: 10 YR 8/4. Made by coiling. ����� Site No. 999, surface. 3) ��������������������������������������� Fragment of a medium-sized jar with restricted mouth and rounded everted rim. Coarse grogtempered brown ware (the grog particles are laminar). Colour: 7.5 YR6/4. Plastic decoration: a strip of clay with impressions on the shoulder. Coiled. Site No. 999, surface. 4) Fragment of a medium-sized jar with restricted mouth and rounded everted rim. Coarse grog-tempered brown ware. Colour: 10 YR 8/4. Plastic decoration: a strip of clay with impressions continuously running with an irregular, wavy movement along the shoulder. Coiled, possibly also with use of slabs. ����������������������������������� Site No. 999, surface. ������������ 5) Fragment of a small-to-medium sub-cylindrical jar. Coarse grogtempered brown ware with large inclusions. Colour: 2.5 Y 6/2. Plastic decoration: a horizontal strip of clay with rhythmic impressions. Site ��������������������������������� No. 999, AA. 4. ������������ 6) Fragment of a large ovoid or sub-globular jar with simple everted rim. Coarse grog-tempered brown ware with polyhedral inclusions. Colour: 10 YR 8/4. Made by coiling, possibly combined with slab construction. Plastic decoration or lug: a horizontal clay strip on the shoulder with triangular section. Site ���������������������������������������������� No. 999, surface. ����������������������� 7) Fragment of a large ovoid or sub-globular jar with simple everted rim. Coarse grog-tempered brown ware with laminar flaked inclusions.

Figure 11.12. Brown-buff ware, coarsely grog-tempered, large-to-medium-sized sub-cylindrical-to-globular jars, some with plastic decoration. 1) Fragment of a large jar with restricted mouth and everted rim, thickened at the joint with the shoulder. Coarse grog-tempered brown ware (grog particles are laminar). Colour: 7.5 YR 8/4. Made by coiling. ��������� Site No. 999, surface. ������������������������������������������� 2) Fragment of a large jar with restricted mouth and thick, rounded everted rim. Coarse grog167

G. L. Bonora and M. Vidale 10 YR 8/4. Coiled. Site No. 999, surface. 5) Fragment of a small-sized globular pot or jar. Buff ware, fine clay, possibly with very fine grog inclusions. Colour: 10 YR 8/3. Coiled. Site No. 999, surface. 6) Fragment of a smallsized globular pot or jar. Buff ware, fine clay, possibly with very fine grog inclusions. Colour: 10 YR 8/4. Coiled. Site No. 999, surface. 7) Fragment of a small-sized globular pot or jar. Painted buff ware (one line on the mouth), fine clay, perhaps with finely ground grog. Colour: 10 YR 8/4 (the pigment is not determinable). Formed of coils. Site ����� No. 999, surface. 8) �������������������������������������� Fragment of a small-sized globular pot or bowl with raised horizontal ridges. Buff ware, fine clay, possibly with very fine grog inclusions. Colour: 10 YR 8/3. Possibly wheel-thrown. Site ���������������������� No. 999, surface. Figure 11.14. Grey burnished ware basins and bowls. 1) Fragment of a large hemispherical basin with flat everted rim. Grey burnished ware, without macro inclusions. Colour: 2.5 Y 7/2. Formed of coils and/or slabs. Site ��������� No. 999, surface. 2) ����������������������������������������� Fragment of hemispherical bowl with a S-shaped upper contour. Grey burnished ware, fine clay with sandy or fine grog (?) temper. Coiled. Site No. 999, surface. 3) Fragment of a large hemispherical bowl with S-shaped upper lip. Grey burnished ware with macroinclusions such as grog-temper pottery chunks (polyhedral) and laminar ceramic flakes. �������������������������������� Colour: 2.5 Y 5/2. Coiled. Site No. 999, surface. 4) ����������������������������������������� Fragment of a restricted sub-globular jar or pot (?). Grey burnished ware, fine clay with sandy or fine grog (?) temper. Colour: ������������������������������������ 2.5 Y 5/2. Coiled. Site No. 999, surface. ����������������������������������������� 5) Fragment of a small sub-globular bowl with slightly S-shaped upper wall thin upper S-shaped wall. Grey burnished ware, fine clay without evident inclusions. 6) Fragment of a small sub-globular bowl with slightly S-shaped upper wall thin upper S-shaped wall. Grey burnished ware, fine clay with a sandy component. Colour: 2.5 Y 7/2. 7) Fragment of a small grey burnished ware bowl or pot. Fine ware free from evident inclusions. Colour: 2.5 Y 6/2. Coiled. Site No. 999, surface.

Figure 11.12 Site No. 999, surface ceramics. Brown ware, coarsely grog-tempered, large-to-medium-sized sub-cylindrical to globular jars, some with plastic decoration.

Colour: 2.5 Y 8/4. A tongue-shaped lug on the shoulder. Site No. 999, surface.

Figure 11.15. Grey burnished ware, small-to-mediumsized globular pots or jars.

Figure 11.13. Buff to brown ware, fine clay with scarcely visible inclusions, small-to-medium-sized sub-globular pots or jars.

1) Fragment of a small-to-medium-sized globular pot or jar. Grey burnished ware, fine clay without evident temper or inclusions. Colour: 2.5 Y 7/2. Formed of coils. Site No. 999, AA. 10. ��������������������������������� 2) Fragment of a small-to-mediumsized globular pot or jar. Grey burnished ware, fine clay without evident temper or inclusions. Colour: ������������������� 10 YR 7/3. Site No. 999, AA. 10. 3) ���������������������������������� Fragment of a medium-sized subglobular pot or jar. Grey burnished ware, fine clay without evident temper or inclusions. Colour: 2.5 Y 5/0. Over fired, possibly a kiln reject. Site ������������������������������������ No. 999, AA. 4. ��������������� 4) Fragment of a medium-sized sub-globular pot or jar. Grey burnished ware, fine clay without evident temper or inclusions. Colour: 5Y 6/2. Made by coiling. Site ���������������������� No. 999, surface.

1) Fragment of a medium-sized sub-globular pot or jar. Buff ware, fine clay, possibly with fine grog temper. Colour: 10 YR 8/3 (possibly a painted buff ware vessel, with vanished decoration). Coiled. Site No. 999, surface. 2) Fragment of a medium-sized sub-cylindrical pot or jar. Painted buff ware (decoration largely vanished). ���������������������������� Colours: 2.5 Y 8/2, pigment 10 YR 5/4. ���������������������������������������������� Coiled. Site No. 999, surface. 3) Fragment of a small-sized globular pot or jar. Fine clay with a sandy component. �������������������������������������������������� Colour: 2.5 Y 8/4. Coiled? Site No. 999, surface. 4) Fragment of a small-sized globular pot or jar. Buff ware, fine clay, possibly with very fine grog inclusions. Colour: 168

An Aspect of the Early Iron Age (Yaz I) Period in Margiana: Ceramic Production at Site No. 999

Figure 11.13 Site No. 999, surface ceramics. Buff to brown ware, fine clay with scarcely visible inclusions, small-to-medium-sized sub-globular pots or jars.

Figure 11.14 Site No. 999, surface ceramics. Grey burnished ware, basins and bowls.

169

G. L. Bonora and M. Vidale

Figure 11.15 Site No. 999, surface ceramics. Grey burnished ware, small-to-medium-sized globular pots or jars. Figure 11.16 Site No. 999, surface ceramics. Grey burnished ware, small-to-medium-sized globular pots or jars.

Figure 11.16. Grey burnished ware, small-to-mediumsized globular pots or jars. 1) Fragment of a small-sized globular pot or jar. Grey burnished ware, fine clay with fine grog temper. �������� Colour: 5Y 6/2. Site No. 999, surface. ����������������������� 2) Fragment of a smallsized globular pot or jar. Grey burnished ware, fine clay without visible inclusions. Colour: �������������������������������� 2.5Y 6/2. Site No. 999, surface. 3) �������������������������������������������������� Fragment of a small-sized globular pot or jar. Grey burnished ware, fine clay without visible inclusions. Colour: 2.5Y 7/2. Coiled. Site ������������������������� No. 999, AA. 10. ��� 4) Fragment of a small-to-medium sized globular pot or jar. Grey burnished ware, fine clay without visible inclusions. Colour: 5Y 6/2. Coiled. Site No. 999, surface. 5) ������������ Fragment of a small-sized or miniature globular pot with two incised lines on shoulder. Grey burnished ware, fine clay, probably with fine grog temper. Colour: ������������������������������������� 5Y 6/2. Site ��������������������� No. 999, AA. 4. 6) Fragment of a medium-to-large globular pot or jar. Grey burnished ware, fine clay with coarse, large grog temper. Colour: 2.5Y 7/2. Coiled. Site No. 999, surface.

1) Fragment of a large globular jar or pot. Painted buff ware, fine clay free from evident inclusions. Painted decoration: horizontal lines at the base of the mouth and what is left of a large double (?) zigzag on the shoulder. Colours: 10 YR 7/4, pigment 10 YR 6/4. ����������������� Formed of coils; the mouth is continuously worn to a flat surface, so this vessel must have been covered with a wooden or pottery lid. ����������������������������������������������������� Site No. 999, AA. 4. �������������������������������� 2) Fragment of a large globular jar or pot. Painted buff ware, fine clay with possible fine grog particles. Painted decoration: parallel horizontal lines at the mouth and set of oblique lines on the shoulder. Colours: 5 YR 4/6, pigment 10 YR 8/4. ����������������� Coiled. ��������� Site No. 999, AA. 4. 3) ����������������������������������������� Fragment of large hemispherical bowl. Painted buff ware, fine clay with sandy or fine grog (?) temper. Painted decoration: on the upper exterior, frieze of hatched triangles alternating with two dots joined by parallel strokes. Colours: ��������������������������������������� 2.5 Y 7/4, pigment 10 YR 8/4. Site No. 999, AA. 10. ���������������������������������� 4) Fragment of hemispherical bowl with slightly S-shaped upper profile painted decoration: on the upper exterior, an angular hatched pattern joined to the

Figure 11.17. Painted buff ware, large globular jars or pots, hemispherical bowls. 170

An Aspect of the Early Iron Age (Yaz I) Period in Margiana: Ceramic Production at Site No. 999

Figure 11.17 Site No. 999, surface ceramics. Painted buff ware, large globular jars or pots, hemispherical bowls.

rim. ���������������������������������������������������� Colours: 2.5 Y 8/4, pigment 10 YR 8/4. ������������� Coiled. Site No. 999, surface. 5) Fragment of small hemispherical bowl. Painted buff ware, fine clay with sandy or fine grog (?) temper. Painted decoration: on the upper exterior, coarse strokes of paint hang from the rim. Colours: �������������������� 2.5 Y 7/4, pigment 10 YR 5/3. ��������������������������������������� Over fired, possibly a kiln reject. 6) Fragment of a small-unrestricted bowl. Painted buff ware, fine clay without visible inclusions. Painted decoration: on the upper exterior, triangle- or lozenge-based pattern with hatched filling. �������������������������������� Colours: 10 YR 8/3, pigment 7.5 YR 5/6. ���������������������������������������������� Coiled. Site No. 999, surface. 7) Fragment of an unrestricted bowl. Painted buff ware, grog-tempered. Painted decoration: on the upper exterior; a cross-hatched pattern creates lozenges that are alternatively solid and filled with dots. 8) Fragment of a small hemispherical bowl with vertical handle. Painted buff ware, fine clay without visible inclusions. ��������������������������������������� Colours: 7.5 YR 7/6, pigment 5 YR 4/3. Coiled. Site No. 999, surface.

Figure 11.18 Site No. 999, surface ceramics. Medium-sized globular jars or pots.

decoration: two lines on mouth, large zigzag band of five parallel lines, irregularly cross-hatched, below. ��������� Colours: 10 YR 8/4, pigment 10 YR 5/4. Formed �������������������������� of coils. ��������� Site No. 999, surface. ������������������������������������������� 3) Fragment of a medium-sized globular jar or pot. Painted buff ware, fine clay, finely grog-tempered and/or with a sandy component. Painted decoration: one line running on the mouth, large attached triangle (?) with inner cross-hatching. Colours: ���������������������������� 10 YR 8/3, pigment 5 YR 4/2. Coiled. ������������������������������������������������ Site No. 999, surface. 4) Fragment of a medium-sized globular jar or pot. Painted buff ware, fine clay free from evident inclusions. Painted decoration: two lines on mouth, unidentified pattern (lozenge or triangle) with inner cross-hatching. Colour: ������������������������������ 10 YR 7/4, pigment 10 YR 6/4. Site No. 999, surface. ������������������������ 5) Fragment of a mediumto-small-sized globular jar or pot. The rim is missing. Painted buff ware, fine clay free from evident inclusions, with a sandy component. Painted decoration: two or more lines on mouth, zigzag band of four or more lines below; in the triangular field between the zigzag band and the upper lines a design formed by two dots joined by two oblique parallel segments. ����������������������������������� Colours: 10 YR 8/4, pigment 7.5 YR 4/4. Coiled. ����������������������������� Site ��������������������� No. 999, AA. 4.

Figure 11.18. Painted buff ware, medium-sized globular jars or pots. 1) Fragment of a large-to-medium-sized globular jar or pot. Painted buff ware, fine clay free from evident inclusions, with a sandy component. Painted decoration: two lines on mouth, zigzag band of four parallel lines, partially cross-hatched, below. Colours: 10 YR 8/3 (the pigment is faded away). Coiled. Site No. 999, surface. 2) Fragment of a medium-sized globular jar or pot. Painted buff ware, fine clay free from evident inclusions. Painted 171

G. L. Bonora and M. Vidale Figure 11.19. Painted buff ware, medium-sized globular jars or pots. 1) Fragment of a medium-sized globular jar or pot. Painted buff ware, fine clay free from evident inclusions. Painted decoration: two horizontal lines irregularly painted on the mouth, cross‑hatched triangle (?) and parallel lines hanging from the lower one. Colours: ���������������������������������� 7.5 YR 8/3, pigment 5 YR 4/3. ����������������������������������������������������� Some faint colour traces suggest that the vessel was originally painted with at least three colours. Coiled. ����� Site No. 999, AA. 4. ��������������������������������������� 2) Fragment of a medium-sized globular jar or pot. Painted buff ware, fine clay free from evident inclusions. Painted decoration: two lines on the mouth, double zigzag (?). ���������������������������������������� Colour: 10 YR 8/4. Site ��������������������� No. 999, AA. 4. 3) Fragment of a large globular jar or pot. Painted buff ware, fine clay free from evident inclusions. Painted decoration: a triangle or lozenge-based pattern, poorly preserved, with inner partitions filled by thick paint strokes. ��������������� Colours: 2.5 Y 6/2, pigment 5 YR 4/2. The ���������������������������������� sherd is over fired and might be a kiln waste piece. ������������������������������������ Site No. 999, AA. 4. 4) ��������������� Fragment of a medium-sized globular jar or pot. Painted buff ware, fine clay free from evident inclusions. Painted decoration: one or two horizontal lines irregularly painted on the mouth, three oblique parallel lines on the shoulder, together with a parallel chain of solid triangles pointing downward, framed by two lines. Colour: ��������������������������������������������� 10 YR 8/3. Site No. 999, surface. ��� 5) Fragment of a medium-sized globular jar or pot. Painted buff ware, fine clay free from evident inclusions. Painted decoration: two horizontal lines irregularly painted on the mouth, two oblique cross‑hatched lines on the shoulder. Colours: 10 YR 8/3, pigment 10 YR 5/4. ������������� Coiled. Site No. 999, surface. 6) Shoulder fragment of a mediumsized globular jar or pot. Painted buff ware, fine clay free from evident inclusions. Painted decoration: one or two horizontal lines running between mouth and shoulder, with an attached zigzag band whose interior is divided into lozenges, alternatively empty and cross-hatch. Colours: ������������ 10 YR 7/4, pigment 5 YR 5/4. Site No. 999, surface.

Figure 11.19 Site No. 999, surface ceramics. Medium-sized globular jars or pots.

buff ware, fine clay with very fine grog inclusions (?). Painted decoration: two lines on mouth, irregularly crosshatched triangle (?) below. Colours: 7.5 Y 5/4 (exterior), 10 YR 6/4 (interior), pigment 5 YR 5/4. Coiled. Site ��������� No. 999, AA. 4. 5) ������������������������������������������� Fragment of a medium-sized globular jar or pot. Painted buff ware, fine clay with very fine grog inclusions (?). Painted decoration: two lines on mouth, unidentified cross‑hatched design below. �������������� Colours: 5 YR 7/3, pigment 2.5 YR 4/6. ������������������������������� Coiled. Site No. 999, surface. 6) Wall fragment of sub-globular jar or bowl. Painted buff ware, fine clay with very fine grog inclusions (?). Painted decoration: a cross‑hatched lozenge alternatively filled in solid and with dots. ����������������������������������� Colours: 10 YR 8/4, pigment 2.5 YR 5/4. Site No. 999, surface.

Figure 11.20. Painted ����������������������������������������� buff ware, medium-sized globular jars or pots. 1) Fragment of a medium-sized globular jar or pot. Painted buff ware, fine clay free from evident inclusions. Painted decoration: two horizontal lines irregularly painted on the mouth, unidentified cross‑hatched pattern below. ��������� Colours: 7.5 YR 8/4, pigment 2.5 YR 4/4. Coiled. ���������������������� Site �������������� No. 999, AA. 4. ���������������������������������������������� 2) Fragment of a medium-sized globular jar or pot. Painted buff ware, fine clay with medium-sized grog particles (?). Painted decoration: unidentified cross‑hatched angular motif. Colour: 5 YR 6/6 (exterior), 10 YR 7/3 (interior), pigment 5 YR 4/3. Coiled. Site ������������������ No. 999, AA. 10. ��������������������������������������������������� 3) Fragment of a medium-sized globular jar or pot. Painted buff ware, fine clay with very fine grog inclusions (?). Painted decoration: two lines on mouth, hatched triangle (?) below. ����������������������������������������� Colour: 2.5 Y 7/6. Coiled. �������������� Strongly over fired, most probably a kiln reject. ������������������������ Site No. 999, AA. 4. 4) ��� Fragment of a medium-sized globular jar or pot. Painted

Figure 11.21. Painted buff ware, medium-sized globular jars or pots. 1) Fragment of a small-to-medium-sized globular jar or pot. Painted buff ware, fine clay with sandy temper. Painted decoration: two horizontal lines irregularly painted on the mouth, zigzag band made of four parallel lines. Colours: 5 Y 7/3, pigment 7.5 YR 7/4. Probably ����������������� coiled. Over fired, most probably a kiln reject. ������������������ Site No. 999, AA. 172

An Aspect of the Early Iron Age (Yaz I) Period in Margiana: Ceramic Production at Site No. 999

Figure 11.20 Site No. 999, surface ceramics. Medium-sized globular jars or pots.

4. ������������������������������������������������������� 2) Fragment of a small-to-medium-sized globular jar or pot. Painted buff ware, fine clay without visible inclusions. Painted decoration: thick line on mouth, double zigzag band made of five cross‑hatched lines. ��������������� Colours: 10 YR 8/4, pigment 5 YR 5/4. Coiled. ��������������������������������� Site ������������������������� No. 999, AA. 10. 3) ��� Fragment of a small-sized globular jar or pot. Painted buff ware, fine clay without visible inclusions. Painted decoration: two lines on mouth, zigzag line (?). Colours: 7.5 YR 8/4 (exterior), 5 YR 7/6 (interior), pigment 2.5 YR 5/4. Coiled. Site ����������������������������������������������� No. 999, AA. 4. �������������������������� 4) Fragment of a small-tomedium-sized globular jar or pot. Painted buff ware, fine clay without visible inclusions. Painted decoration: two lines on mouth, zigzag line (?). Colour: 5 Y 6/3 (pigment not determinable). Over fired, probably a kiln reject. Site No. 999, AA. 3. 5) ������������������������������������ Fragment of a small globular jar or pot. Painted buff ware, fine clay with very fine grog particles (?). Painted decoration: parallel horizontal lines at the mouth, hut-like pattern below (“tectiform” covering

Figure 11.21 Site No. 999, surface ceramics. Medium-sized globular jars or pots.

a grid-like design). Colours: 10 YR 8/4, pigment (faintly preserved) 10 YR 6/4. Coiled. Site No. 999, surface. 6) Fragment of a small-to-medium-sized globular jar or pot. Painted buff ware, fine clay without visible inclusions. Painted decoration: zigzag band made of three parallel lines enclosed within a horizontal band or frieze (running on the external shoulder). ������������������������������� Colours: 2.5 Y 7/4, 2.5 Y 5/2. Over fired, probably a kiln reject. Figure 11.22. Painted buff ware, medium-sized subglobular jars or pots. 173

G. L. Bonora and M. Vidale 1) Fragment of a medium-sized sub-globular jar or pot. Painted buff ware, fine clay without visible inclusions. Painted decoration: two lines on mouth; below there hangs a pattern probably formed of three triangles joined at their tips, with internal cross-hatching (?). ������������ Colours: 10 YR 7/3, pigment 5 YR 4/4. ����������������������������� Coiled. Site ��������������������� No. 999, AA. 5. 2) Fragment of a medium-sized sub-globular pot or bowl (?). Painted buff ware, fine clay without visible inclusions. Painted decoration: cross‑hatched triangles and/or lozenges (?). ������������������������������������������������ Colours: 10 YR 8/3, pigment 10 YR 5/4. Site No. 999, surface. ������������������������������������������ 3) Fragment of a small-sized sub-globular or sub-cylindrical jar or pot. Painted buff ware, fine clay without visible inclusions. Painted decoration: zigzag cross‑hatched band made of three lines. Colours: ������������������ 5 Y 5/3, pigment 5 Y 4/2. Strongly ��������������������������������������������� over fired, probably a kiln reject. Site No. 999, AA. 3. ��������������������������������� 4) Fragment of a small-sized subglobular or sub-cylindrical jar or pot. Painted buff ware, fine clay without visible inclusions. Painted decoration: one or two lines on the mouth, three oblique lines running in a horizontal frieze or band. ���������������������������� Colours: 10 YR 7/4, pigment 5 YR 5/4. ��������������������������������������������� Site No. 999, AA. 4. ������������������������ 5) Fragment of a mediumsized sub-globular jar or pot. Painted buff ware, fine clay without visible inclusions. Painted decoration: two lines running on the mouth, a zigzag band of four lines (partially cross-hatched) below. Colours: ������������������������������ 10 YR 8/4, pigment 5 YR 5/4. ���������������������������������������������� Coiled. Site �������������������������������������� No. 999, AA. 1. ����������������� 6) Fragment of a miniature globular jar or pot. Painted buff ware, fine clay without visible inclusions. Painted decoration: a horizontal line below the mouth. Over fired, probably a kiln reject. Site No. 999, AA. 3. Figure 11.23. Painted buff ware, small globular jars or pots. 1) Fragment of a small globular jar or pot. Painted buff ware, fine clay with very fine grog particles (?). Painted decoration: two lines running on the mouth. Below probably there was attached a cross‑hatched lozenge (?), alternatively filled with dots and cross-hatching. ��������� Colours: 2.5 Y 8/4, pigment 5 YR 5/4. Site ������������������������ No. 999, AA. 4. 2) Fragment of a small globular jar or pot. Painted buff ware, fine clay without visible particles. Painted decoration: one line on the mouth, and below a frieze with a zigzag band of five lines occupying the upper part of the vessel. ��������� Colours: 2.5 Y 8/4, pigment 5 YR 4/6. Coiled. �������������������������� ������������������ Site No. 999, AA. 4. 3) ���������������������������������������������������� Fragment of a small globular jar or pot. Painted buff ware, fine clay free from evident inclusions. Painted decoration: two lines running on the mouth, and below a zigzag band with four or more parallel lines, partially cross-hatched. Colours: ������������������������������������������� 7.5 Y 7/3, pigment 5 YR 5/4. Site ����� No. 999, AA. 4. 4) �������������������������������������������� Fragment of a small globular jar or pot. Painted buff ware, fine clay free from evident inclusions. Painted decoration: two lines running on the mouth; below there hangs a partially cross‑hatched triangle. Colours: 10 YR 8/4 (exterior), 10 YR 8/3 (interior), pigment 2.5 YR 4/4. Coiled (?). Site No. 999, surface. 5) Fragment of a small globular jar or pot. Painted buff ware, fine clay with very fine grog particles (?). Painted decoration: one line

Figure 11.22 Site No. 999, surface ceramics. Medium-sized subglobular jars or pots.

running on the mouth, and a relatively large band or frieze, occupying the upper part of the vessel, with a zigzag band of four lines. �������������������������������������� Colours: 10 YR 8/3, pigment 5 YR 5/4. Coiled. Site ��������������������� No. 999, AA. 4. Figure 11.24. Other types of ceramic finds and varia. 1) Fragment of a Late Bronze Age (Namazga VI) carinated bowl or trumpet-shaped vessel (cf. Hiebert 1994: Fig. 6.11). Very fine clay without visible inclusions. Colour: 5 YR 7/6. Wheel-formed coils. Site ������������������������������������ No. 999, AA. 1. 2) ��������������� Fragment of an Andronovo-like incised vessel. Grog-tempered brown ware. Colour: ����������������������������������������� 2.5 Y 5/2. Site No. 999, surface. 3) Fragment of an Andronovo-like incised vessel. Grogtempered brown ware. ��������������������������������� Colour: 2.5 Y 6/2. Site No. 999, surface. �������������������������������������������������� 4) Fragment of a rough hemispherical bowl. Coarse grog-tempered brown ware with large inclusions, both laminar and granular. Colours: 10 YR 8/4 (exterior), 5 YR 6/6 and 5 YR 8/4 (interior). Coiled. Site No. 999, surface. 174

An Aspect of the Early Iron Age (Yaz I) Period in Margiana: Ceramic Production at Site No. 999 Buff ware, fine clay with sandy or fine grog (?) temper. Colour: 2.5 Y 8/4. Formed of one or two wheel-shaped coils. �������������������������������������������������� Site No. 999, surface. 10) ��������������������������� Fragment of a globular miniature pot. Fine buff ware without visible inclusions. Colours: 2.5 Y 8/4-8/6 (exterior), 5 YR 6/6 (interior). Wheel-thrown and trimmed on the lower half. ��������� Site No. 999, AA. 5. ������������������������������������������ 11) Fragment of a miniature hemispherical bowl with a slightly thin S-shaped upper contour. Fine buff ware without visible inclusions. Colour: ������������������� 10 YR 8/2. Site No. 999, surface. 12) ���������������������������������� Fragment of a small-to-mediumsized jar on a raised foot. Grog-tempered ware with large inclusions. Colour: 10 YR 6/3. The foot was formed of a single coil. Site ����������������������� No. 999, surface. Figure 11.25. Small finds. During the survey and excavations, Vidale also collected some lithic and metallic objects. They are reproduced in Figure 11.25. The following stone and copper/bronze items were found in Site No. 999: 1) Fragmentary grinding stone in vacuolar basalt, found near the edge of the kiln. 2) Polisher-hammerstone in an unidentified black metamorphic rock (a kind of quartzite?) with highly polished and pecked surfaces. 3) Probable ceramic polisher in grey white-banded marble, found in the bottom of the combustion chamber of the kiln (with fine angular

Figure 11.23 Site No. 999, surface ceramics. Painted buff ware, small globular jars or pots.

5) Fragment of a cylinder-like vessel resembling a crucible, but free from metallic residues in the interior. Colour: 2.5 Y 7/2-6/2. Grog-tempered ware, with inner reduced core. Site No. 999, AA. 8. ������������������������������������ 6) Fragment of a potter’s bat. Many other specimens of this type were visible on the surface. Coarse grog-tempered brown ware with large inclusions, both laminar and granular. Colour: 2.5 Y 8/2. Formed of slabs. Site ����������������������������������������������������� No. 999, AA. 8. 7) �������������������������������� Fragment of a copper-melting crucible (?) with vertical, raised handle (missing). Coarse grog-tempered ware with large granular inclusions. The interior is burnt with heat-spalling fractures and may have had metallic crusts. The exterior is reduced and slightly over fired. ����������������������������������������������� Colours: 2.5 Y 7/4-7-6, burnt areas 2.5 Y 6/2. 8) Fragment of a miniature truncated-cone-shaped bowl. Fine clay without visible inclusions. Colour: 10 YR 8/2. Formed of a wheel-shaped coil and a slab. Site �������������� No. 999, surface. ����������������������������������������������� 9) Fragment of a miniature hemispherical bowl.

Figure 11.24 Site No. 999, surface ceramics. Other types of ceramic finds, varia.

175

G. L. Bonora and M. Vidale polishing planes). 4) Probable ceramic polisher made out of a square-sectioned blocklet of red jasper, with fine polishing planes visible, in the picture, as dotted surfaces. 5) Fragmentary phallus-shaped polisher (?), in a fine-grained grey metamorphic stone. 6) Fragmentary disc-shaped bead in an unidentified greenish translucent stone (moss-agate or perhaps jade?). 7) Fragmentary knife or razor in copperbronze. 8) Fragment of a pin (?) in copper-bronze. MV Comparisons and Some Chronological Implication An overview of the painted motifs at Site No. 999 (Figures 11.26-27) shows a substantial convergence, in terms both of stylistic features and frequency, with contemporary ceramic assemblages in the Murgab Delta and the piedmont strip. There are few elements in common with Kuchukdepe, in northern Bactria, as well as with the Iron Age sites in Fergana. Many of the painted buff ware potsherds collected and drafted by Vidale have the same repetitive, standardised designs, while others are poorly preserved and difficult to reconstruct. For these reasons, here, we have considered only a limited, suitable part of the record. The motif in Figure 11.17.2 is common at Yaz-depe: for a good example, see a medium-sized globular jar, formed of coils, with a dark brown paint on a buff background in Masson 1959: Tab. XX.1, that was found in excavation II, second arbitrary cut.

Figure 11.25 Site No. 999, surface. Small finds.

A different decorative pattern with dots joined by segments appears also in two other sherds of Site No. 999 (Figures 11.17.3, 11.18.5); but so far this is a unique piece of evidence in the Murgab assemblages. A sherd bearing a very similar pattern was found in the upper Atrek Valley. But here only one big dot was drawn between a thick horizontal line on the rim and a frieze of cross-hatched triangles (Venco Ricciardi 1980: Fig. D.6). Zigzag bands formed of at least three parallel lines, totally or partially cross-hatched, below two lines running on the mouth, appear on our sherds in Figures 11.18.1, 11.19.5, 11.20.1, 11.22.3 and 5. At Yaz-depe, excavation II, this pattern is rather frequent: it was found in almost all the upper arbitrary cuts dug by V. M. Masson (one from the second cut, one from the third, two from the fourth: Masson 1959: Tabs. XVII.5, 8, 13; XVII.7). It is present also at Garaoidepe, where it was found in one example (Gutliev 1982: Fig. 1.3).

The motif in Figure 11.17.7 (a cross-hatched pattern creating lozenges, alternatively solid and dotted) is similar to that of Figure 11.20.6. In Figure 11.23.1 the dot-filling is alternated with cross-hatching, while in Figure 11.19.6 the cross-hatched filling is alternated with no filling at all. The basic alternatives are lozenges filled with dots and black-and-white patterns. The first solution appears eight times at Yaz-depe. Three cases are reported from the second arbitrary cut of excavation II, one from the fourth; and four from the eleventh arbitrary cut of excavation III (Masson 1959: Tabs. XVII.1, 2; XVIII.8; XIX.9; XX.14, 19-20). One example is known at Uch-depe 10 (Masimov and Udeumuratov 1984: Fig. 1.10) and two specimens are reported at Yasi-depe, both of them from the eleventh arbitrary cut (Gutliev 1970a: Fig. 22.21-22). In turn, the second alternative appears in two variants: solid filling versus empty and half-filling versus empty. The first variant appears on four sherds at Yaz-depe, all from different arbitrary cuts (second cut, excavation II; eleventh, thirteenth and fourteenth cuts, excavation III: Masson 1959: Tabs. XVIII.2, XIX.5 and 11, XX.8) and also at Garaoi-depe on a very small fragment (Gutliev 1982: Fig. 2.13). The second variant is encountered on five specimens (Masson 1959: Tabs. XVII.1 and 2, XVIII.8, XIX.9, XX.14). Interestingly, the same design is also found in the small site of Ovadan-depe, north of Ashgabat (Masson 1956b: Tab. XLI.8).

A similar design is formed of a set of parallel lines, all or in part cross-hatched and running below the two lines of the mouth. There are minor variations in the number of lines or number and position of the cross-hatching. In our specimens in Figures 11.18.2 and 11.21.2, the band has five lines, entirely cross-hatched. In Figures 11.18.5, 11.21.1 and 6, 11.23.2 and 5, a band with four or five lines, cross-hatched only in the upper or lower part, can be seen. The first two cases are matched at Yaz-depe (third arbitrary cut, excavation II: Masson 1959: Tab. XVII.19), and at Yasi-depe (eighth cut: Gutliev 1977: Fig. 4.40). The 176

An Aspect of the Early Iron Age (Yaz I) Period in Margiana: Ceramic Production at Site No. 999 incorrect interpretations. This pattern, on the contrary, is not reported at Yasi-depe and El’ken-depe. The design in Figure 11.18.3 can be reconstructed as a frieze of large triangles with an inner net-like filling. Most probably the item shown in Figure 11.20.5 also bore this design; while in the one in Figure 11.23.4 the inner net filling is only partially drawn. At Yaz-depe, this design is encountered in three sherds found from the eighth arbitrary cut of excavation III to the twelveth (Masson 1956b: Fig. 23.2 and 14; Idem 1959: Tabs. XIX.7, XX.5 and 9). At Uch-depe 10 there is a single similar piece of evidence (Masimov and Udeumuradov 1984: Fig. 6.15); other similar designs come from El’ken-depe, 2nd architectural layer (Marushchenko 1959: Tab. XVI.5), from Yasi-depe, 8th arbitrary cut (Gutliev 1970a: Fig. 22.18 and, probably, 19) and from Garaoi-depe (Gutliev 1982: Figs. 1.29 and, probably, 2.24). This pattern was also found in the upper Atrek Valley (Venco Ricciardi 1980: Fig. D.3).

Figure 11.26 Site No. 999, surface ceramics. Examples of painted buff sherds.

A possible variant of the same motif (a frieze of large triangles with oblique filling: Figures 11.20.2, 11.17.3) is not documented at El’ken-depe, and appears at Yaz-depe in a single published specimen (eleventh arbitrary cut, excavation III - Masson 1959: Tab. XX.22). At Uch-depe 10, there are several similar cases, but only two have been positively ascertained (Masimov and Udeumuradov 1984: Figs. 1.4, 6.14). The others, poorly preserved, are doubtful (Ibidem: Figs. 1, 6, 9, 11, 12). The unicum hut-like design in Figure 11.21.5, finally, is quite anomalous. Five vessels have applied plastic decoration. Three show impressed horizontal clay strips (Figures 11.12.3, 5; 6 has a unmodified strip with triangular section, while Figure 11.12.7 has a tongue-shaped lug on the shoulder). The first three are comparable to sherds found at Yaz-depe. Figure 11.12.3 is very similar to a medium-sized jar with restricted mouth and rounded everted rim from excavation II, without a vertical context (Masson 1959: Tab. XXIII.8); Figure 11.11.1 resembles two jars of the same type from excavations II and III (the second coming from the 11th arbitrary cut) (Masson 1959: Tabs. XVII.2, XXVIII.6); Figure 11.12.5 is similar to a jar from the same context (Masson 1959: Tab. XXIII.9). As a rule, in both sites (Site No. 999 and Yaz-depe) this type of plastic decoration is strictly associated to restricted medium-sized jars having a sub-globular body, round everted rim, and probably a flat base. Another find from Uch-depe 10 may perhaps support this impression, but the drawing is not clear (Masimov and Udeumuradov 1984: Fig. 5.14).

Figure 11.27 Site No. 999, surface ceramics. Painted buff sherds and a specimen of burnished grey ware (first from left) with evident signs of overfiring and incipient warping.

others are comparable to a painted sherd from Uch-depe 10, surface (Masimov and Udeumuradov 1984: Fig. 1.2), with two specimens from Yaz-depe, excavation II (second and fourth arbitrary cuts, Masson 1959: Tabs. XVII.4, XVIII.5) and with one from Garaoi-depe (Gutliev 1982: Fig. 4.25). The poor preservation of the items shown in Figures 11.21.4 and 11.22.4 does not permit a reliable reconstruction, but they probably show another variant of the zigzag band pattern. Here the zigzag, framed inside a thick line on the mouth and a thinner line on the shoulder, has three parallel lines but no cross-hatching. Besides, in the item shown in Figure 11.19.4 this design is combined with a chain of solid triangles pointing downward, framed by two lines. At Yazdepe, the same design appears in three sherds, respectively found in the second cut of excavation II (band of six lines) and in the thirteenth arbitrary cut of excavation III, with a four-line band (Masson 1959: Tabs. XVII.6, 14, XIX.6). At Uch-depe 10, many fragments are apparently similar to our materials, but their fragmentary state may lead to

Table 11.3 highlights the available stratigraphic columns reporting the arbitrary cuts used in the various excavations and test trenches in southern Turkmenistan and Margiana for this period. The dark cells represent the contexts where positive comparisons could be established. It seems clear that Site No. 999 was contemporary with the levels exposed at Yaz-depe in excavation III (eleventh - thirteenth 177

G. L. Bonora and M. Vidale cuts) and in excavation II (second - fourth cuts). For Uchdepe 10 and Garaoi-depe we have no precise stratigraphic indications, but we may assume that in the Early Iron Age they belonged to the same cultural area as Yaz-depe and Site No. 999.

additional temper some laminar pottery flakes detached with this method. This technology may actually help us to explain the presence, on the surface of many sites of the Murgab Delta, of thousands and thousands of ceramic disks (round re-utilised chipped potsherds).

For Bactria, no positive comparison with the Early Iron Age pottery from Kuchuk-depe is possible (Al’baum 1969; Askarov 1982). Here, in the Sherabad Valley, the decorative patterns (friezes of solid-filled triangles below the rim or on the shoulder, and a few oblique or vertical strokes hanging from the rim) are more simple and standardized than those described for Margiana and the piedmont plain north of the Kopetdag.

Vidale analysed the ceramics collected on the surface of Site No. 999, observing their surfaces and fractures under incident light, and registering, whenever possible, the discontinuities and interfaces (created by macropores, voids and macro-inclusions) in their sections. This analysis, in a considerable number of cases, made it possible to record the sections of the vessels showing the outline of the coils and/or slabs used in their construction. Wheel-throwing, on the other hand, was surmised on the basis of the absence, in the fracture sections, of clear evidence of coils or slabs, or of their modification, during their mouldable stage, under strong rotational effects (in addition to the observation of more traditional technical attributes, such as regular parallel turning striae, reduced thickness, and decrease in thickness from the base towards the rim). While vessels manufactured with coils and thinned and/or regularised on a potter’s wheel are not common, but present in the Yaz I assemblage, entirely wheel-thrown vessels are very rare, and in some cases they may belong to different chronological horizons. Nonetheless, the many bat fragments (pieces of flat pottery cakes or disks, Figure 11.24.6) observed on the surface of Site No. 999 suggest that protohistoric potters made regular use of the potter’s wheel. These bats, in fact, are used still today by South Asian (as well as by western) potters for centring the clay on the wheel and removing the formed vessels at the end of the throwing. Figures 31-33 respectively show the sections obtained for specimens of brown-buff, grey burnished and painted buff ware.

GLB Ceramic Technology The sections of the two test trenches (see below) show that the site rose on a natural elevation near the banks of a swampy depression, probably corresponding to an ancient ox-bow meander. The natural substratum is formed of a massive clayey deposit with prismatic structure. Perhaps the ceramic manufacturing site exploited the local clays; the pit visible in section below the kiln and the mud brick structure may have been a clay-digging trench, but at present this is only a hypothesis. At any rate, presumably the potters had selected a favourable setting that ensured easy access to clay, water and fuel. The basic materials (clay or silt and possibly sand, if ceramics required the artificial mixture of different types of natural sediments) were evidently mixed with artificial temper, to a large extent consisting of flaked, crushed and ground potsherds. In the ceramics of the Late Bronze Age, formed of coils and thinned and regularised on a potter’s wheel, as well as in earlier productions, as far as we know these inclusions are not evident, but they become easily recognisable starting from the middle of the 2nd millennium BC, with Andronovo-related productions (Figure 11.24.2 and 3). Recycling ancient ceramics for obtaining artificial tempers is quite a specific technique, based upon a particular mindset, and it would be extremely rewarding to achieve an analytic definition of the precise period of this innovation in Margiana and southern Turkmenistan, and to understand whether and to which extent this technique was related to forms of nomadic exploitation of the territory. On the surface of some Yaz III sites of the Murgab Delta (second half of the 1st millennium BC), characterised by massive, homogeneous, well-fired pottery, we identified some large or thick potsherds that were evidently used as cores for detaching laminar pottery flakes. The cores were thick conical lugs, large rims, or, more commonly, potsherds flaked around the edge until they took on a rounded contour and became too small to be handled easily. We also found some peculiar, coarse chaff-tempered sub-cylindrical vessels that had evidently been manufactured using as an

In brown-buff ware products (Figure 11.28) we observe the greatest technical variation. Dishes and potter’s bats (Figure 11.28.11) were evidently made by superimposing clay slabs and shaping them on the wheel. Basins and large hemispherical bowls seem to have been shaped by simultaneously combining coil construction with wheelthrowing (Figure 11.28.2-7, 10); it is also possible to recognise a quite unusual technique, i.e. the manufacturing of a pre-form with upper S-shaped rim, followed by the application of a last vertically shaped coil on top, resulting in a distinctive type of bowl where the original S-shaped projection becomes a horizontal ledge (Figure 28.6, 10). In general, we find, as a mental template (in the sense of van der Leew and Pritchard 1984), the creation of a rim with a single final, narrow coil (Figure 11.28.6, 7, 10), a feature that is repeated in other pottery forms (see below). In hemispherical bowls, the coils are shaped, joined and thinned on the wheel by squeezing and pulling or stretching down the coil on the internal surface (Figure 11.28.2), or by thinning the upper ends of the coil and carefully coating the lower coil on its internal and external sides (Figure 11.28.3-4). In some anomalous specimens, 178

An Aspect of the Early Iron Age (Yaz I) Period in Margiana: Ceramic Production at Site No. 999

Site

Yaz-depe

Yaz-depe

Yaz-depe Yaz-depe

Jassy-depe Jassy-depe

Yashylly-depe Uch-depe 10

Excavation Arbitrary Cuts I II III IV V VI VII VIII IX X XI XII XIII XIV XV XVI XVII XVIII

TT 2

Exc. 4

Exc. 3

TT 1970

TT 1977

TT 1980

Exc. 1981

Yaz III Yaz II Yaz II Yaz I Yaz I Yaz I Yaz I Yaz I Yaz I

Yaz III Yaz III Yaz III Yaz III Yaz II Yaz II Yaz II Yaz I Yaz I Yaz I Yaz I Yaz I Yaz I NMG VI-Yaz I

Yaz III Yaz III Yaz III Yaz III Yaz IIb Yaz IIb Yaz IIb Yaz IIb Yaz IIa Yaz IIa Yaz IIa Yaz IIa Yaz I Yaz I Yaz I Yaz I Yaz I Yaz I

Yaz I Yaz I Yaz I Yaz I Yaz I Yaz I Yaz I Yaz I

Yaz III Yaz III Yaz III Yaz III Yaz II Yaz II Yaz II Yaz II Yaz II Yaz II Yaz I-II Yaz I Yaz I Yaz I

Yaz III Yaz III Yaz II-III Yaz II Yaz II Yaz II Yaz II Yaz I Yaz I Yaz I

Yaz III Yaz III Yaz II Yaz II Yaz II Yaz II Yaz II Yaz I Yaz I Yaz I Yaz I Yaz I

Exc. 2

Yaz I Yaz I Yaz I Yaz I

Table 11.3 Here are reported the available stratigraphic columns with the arbitrary cuts used in the various excavations and test trenches in southern Turkmenistan and Margiana for the Iron Age period. The dark cells represent the contexts where positive comparisons could be established between the surface collections at Site No. 999 and other sites.

Figure 11.28 Sections showing the forming techniques in brown-buff ware.

179

G. L. Bonora and M. Vidale coiling seems to be barely modified (or not at all) by turning (Figure 11.28.16). The hemispherical basin of Figure 11.28.1 is clearly formed of large slabs and most probably within a hemispherical mould (a larger basin or part of a hemispherical jar may have been used for this purpose). In truncated-cone-shaped bowls (Figure 11.28.9) the intervention of the wheel is far less evident. Restricted forms such as small-to-medium, medium-to-large and large jars or pots are evidently built with slab-construction processes, apparently without a major use of centrifugal force (this does not exclude the use of the wheel as a plain rotating device for aiding the processes). The rim often consists of a final, small-sized coil.

ware. Significantly, these additional technical steps are encountered in the two classes possibly used for serving and consuming drinks and/or food, and therefore more probably connected to status display and/or other representative functions. Brown-buff vessels in general show no evidence of particular finishing techniques, with the exception of rare instances of rim painting. Burnishing was carried out carefully, covering a good percentage of the surface of the vessels, with small polishers made out of hardstones. A specimen in red jasper was found on the surface near a cluster of kiln refuse and the excavated kiln (Figure 11.25.3); a second ceramic polisher in blackand-white striped marble was recovered in the filling of the mouth of the same kiln (Figure 11.25.4), suggesting that finishing (and perhaps forming) operations were carried out not too far from the kiln firing area. Both these stone polishers show clear surface facets produced by friction with the surface of unfired grey ware vessels. In the red jasper one, in particular, the wear surfaces are so discontinuous (touching the stone surface and each other with well-defined angles) as to suggest that the burnishing of the coils may have been performed on the vessels while they were revolving. The painting has been described, as a general rule, as monochrome, but the find of a single globular vessel with red-brown and white designs makes it possible to surmise that other vessels may have been bichrome as well, since the white pigment was unstable. During the firing, the pigments used in painted decoration responded to the uncompleted oxidising conditions prevalent in the final step of manufacturing; thus, in the finished products their colours vary from red to brown and to darker greenish and blackish hues.

In grey burnished ware (Figure 11.29), small-to-medium jars or pots (the most common form on record) appear to have been manufactured with a careful coil construction process. The joints between the coils have quite simple concave and rounded interfaces, with limited pullingstretching lateral effects. This suggests a careful, relatively time-demanding process unaided by the use of a potter’s wheel. The orientation of the different parts of the vessel wall seems to dictate the prevalent direction of the observed side-stretching of the clay of the coils: where the wall opens up, the coil base is slightly stretched down towards the inside, while where the same wall narrows down, the coil base is stretched towards the outside. The prevalent globular shapes of the vessels of this class make us wonder if they could have been also slightly thinned, while they were being manufactured with coils, by means of a paddleand-anvil technique; but anvils are not on record, and the outer surface does not show any evidence of paddling. The few open forms, such as the bowl or basin of Figure 11.29.8, might have been manufactured with coils but also with parallel slabs. Figure 11.29.8 is a single specimen, but it suggests, on the whole, that the overall shape of the vessels, together with an established technical tradition, may have determined the precise technique to be used in the various cases.

Many fragments of painted buff containers and a few greyburnished potsherds found on the surface were classified as kiln rejects. In the excavated kiln, the latest firing were evidently performed in slightly reducing conditions, as shown by the surface features of the last layers of plaster applied to the inner walls. On the whole, although no misfired specimens were found inside the lower chamber of the kiln, it is quite likely that vessels of these two ceramic classes were baked in infrastructures quite similar to the main kiln excavated by us (see below). If this is true, for grey burnished ware the potters were able to master the performance of their kilns differently; it is also possible that they used a different type of firing infrastructure for this purpose.

Finally, while painted buff ware vessels (Figure 11.30) frequently show, in section, the shapes of the original coils, the joints are different: the extremities are thinner and in general more pointed, with longer base stretching features (the direction of the stretching depends on the opening or closing of the shape, as in grey burnished ware). In this case, too, the rim often consists of a final, small-sized coil, corresponding to the slight outward projection visible in most of the globular or sub-globular forms.

Summarising, in terms of labour investments, grey burnished ware vessels appear to be the most costly items, followed by painted buff products. Firing in much smaller kilns may have been symbolically associated to metals and to the purifying effects of smoking; and this may be related to the waterproofing effects of burnishing. The three classes we described are partially distinguished in terms of forms and function, and well distinguished as far as basic forming and finishing techniques are concerned. This does not necessarily imply that different individual

On the whole, these modifications of the coil structure suggest, as a first hypothesis, that coiling, as in part of the bowls and basins in brown-buff ware, was performed simultaneously with wheel-throwing. This information is summarised in the following table (Table 11.4). Surface finishing techniques include burnishing, in grey burnished vessels, and painting, in painted buff 180

An Aspect of the Early Iron Age (Yaz I) Period in Margiana: Ceramic Production at Site No. 999 potters produced different ware classes. Nonetheless, in the case we described, specific techniques were selectively used for specific types of vessels, and this suggests the existence of different specialised forms of know-how, a condition compatible with relatively advanced forms of labour division (by sex, age, and possibly status) within the community where the vessels were produced. MV Excavation Trenches: a Trashing Ground and a Pottery-firing Kiln As reported above, the surface of the site showed a clear contrast between its eastern slopes, occupied by relatively large clusters of fired bricks, broken overfired kiln blocks and linings, ceramics and a few polished stone tools, and its peripheral western sloping area, where these indicators were quite rare and broken perforated slabs similar to fragments of kiln grids were the most common find. In the latter type of surface assemblage, grid fragments were accompanied by large amounts of kiln linings and slightly oxidised plaster pieces breaking into countless small, pale yellowish plaster particles and into plaster powder. Our first deduction, as stated above, had been that these perforated slabs or grid elements belonged to small ovens, perhaps portable, and that the areas of maximum concentration of these elements on the surface were actually the site of the prevalent use of these hypothetical small ovens, in the frame of an unknown pyro-technological activity. For this reason, Salvatori asked Vidale to make, in a couple of days, a limited test trench on the spot, in order to see what these “ovens” actually were. Vidale selected a reduced area that included a large stretch of chaff-tempered plaster wall and a concentration of smaller fragments of the same material, and excavated a first trench, 1 x 0.5 m, reaching the natural soil about 50 cm below the present surface (Figure 11.31). The Stratigraphic Units encountered (hereafter SU) are visible in the section and described as follows:

Figure 11.29 Sections showing the forming techniques in grey burnished ware.

SU 101. Natural substratum: compact sandy clay with polyhedral structure, colour 10 YR 6/3. SU 102. Sandy clay of a similar colour, but rich in vertical mottling characterised by a greenish colour (10 YR 5/6). The mottling is interpreted as root casts marked by reduction processes in a partially anaerobic aquatic environment such as a swamp (probably covered with reeds). Thus, SSUV 101 and 102 may represent an ancient natural soil profile evolved in situ in the context of the bank of an ancient deltaic branch and/or a natural depression with stagnant water. SSUV 103, 104. Tabular, wavy lenses of silt with homogeneous, very fine sand, prevalent colour 10 YR 7/4. These layers are devoid of any artefact and were probably deposited after a large-scale alluvial activity.

Figure 11.30 Sections showing the forming technique in painted buff ware.

181

G. L. Bonora and M. Vidale SSUV 105, 106. Layers of loose silt, colour 10 YR 6/4, both of them rich in an ashy component and charcoal bits, and containing a few potsherds and bones and several fragments of chaff-tempered clay or plaster, generally in the form of kiln linings and/or broken perforated kiln fragments. The horizontal interface between these layers was particularly rich in plaster pieces. SU 107. Fine sandy sediments, colour 2.5 Y 7/2, with thin hydromorphic sheets and small clusters of chaff-tempered plaster pieces in various stages of local decay. Our original impression was not confirmed, as no structure or infrastructure was brought to light. At least for the very limited cleared ground, the hypothesis that in this part of the site there had been craft installations employing small portable ovens or other kinds of hearths with a specialised function was incorrect. The sequence, on the contrary, showed that the area had been a low, deserted natural elevation in the form of a clay bank near a swampy depression, later invested by large-scale alluvial layers that left sheets of silty sand in the location. Possibly, as already stated, potters had selected this spot because the clay substratum offered reasonably good raw material (but this idea needs to be analytically confirmed). The anthropic deposits encountered in this test trench are the output of processes of dumping, erosion, re-deposition and sedimentary evolution of residues of kiln firing activities (in the first place kiln linings and wall fragments, broken grid fragments and ash). In other words, perforated kiln grid pieces were abundant on the tested spot not because the kilns or ovens were installed and used in the place, but because this particular location had been used for a long time by the potters as an open-air trashing ground, widely exposed to atmospheric and other disturbance agents. This trench, however, yielded no information on the nature of the pyrotechnological activities performed in Site No. 999, or on the nature and function of the enigmatic perforated grid slabs so common in this mound. The solution came from a second, larger test trench, opened at the centre of the mound, where the erosion due to trampling and to the passage of vehicles on the main track in the core of the site had exposed a wall and the top surface of the central pillar of what turned out to be an exceptionally well-preserved ceramic kiln (Figure 11.36). The results of this second limited excavation are reported in the plan and section Class Brownbuff Grey burnished Painted buff

Dishes, bats SL+W

Basins

-

M? C+W -

-

-

Figure 11.31 Plan and section of a test trench excavated in the north-western slopes of Site No. 999.

of Figures 11.32 and 11.33, in the Harris Diagram of Figure 11.34 and in the enclosed list of SU. Details of the excavation may also be checked in the pictures of Figures 11.37-43. SU 1. Natural substratum: compact sandy clay with polyhedral structure, identical to SU 101, with traces of the same greenish stains or mottling caused by vertical root casts. The top of this layer seems to have been truncated by the subsequent sandy layers, SU 2 and 3. SU 2. Medium-to-fine sand, clean and homogeneous, 2.5 Y 6/4. SU 3. Silt with homogeneous, very fine sand, 10 YR 7/4. These layers are free from anthropic inclusions and probably represent a flood deposit. The top of SU 1 may have been eroded and removed by the deposition of these sandy alluvial strata. SU 4. Negative interface: excavation of a trench later filled with SU 5-8 (see below). SU 5. Loose, fine homogeneous silt (10 YR 6/4) with plaster crumbles and clay or sand particles excavated and

Bowls (hem.) M C+W C (SL?)

Bowls (tc) C (+W??) -

Small-to-medium jars or pots SL

Medium-to-large jars or pots SL

Very large jars

C

-

-

C (W?)

-

C (W?)

C (W?)

-

SL

Table 11.4 Preliminary assessment of technical variations in the shaping techniques of the main pottery forms in brown-buff, grey burnished and painted buff ware, Site No. 999. C = manufactured with coils; M = manufactured in a mould; SL = manufactured with slabs; W = manufactured or regularised on a potter’s wheel.

182

An Aspect of the Early Iron Age (Yaz I) Period in Margiana: Ceramic Production at Site No. 999 SU 21. Probable building or platform in mud-bricks, exposed only in one corner. Stratigraphically in phase with the small SU 12 kiln.

scattered from the SU (1) substratum. Some of the clay particles are fired and oxidised. SU 6. Loose, fine homogeneous silt (10 YR 6/4) similar to SU 5, but without plaster particles and abundantly scattered with charcoal chips.

SU 12. Small kiln filled with broken plaster pieces and collapsed linings: its mouth is in the same direction as that of the large SU 14 kiln. No parts of this feature had been left undisturbed in place. The abandonment-destruction interface of this small pyro-technological infrastructure is conventionally identified with its secondary filling.

SU 7. Loose sandy silt (10 YR 7/3) including tiny bone chips and plaster crumbles. SU 8. Fine sand, slightly silty (10 YR 7/3), with diffuse plaster particles (diameters averaging a few mm).

SU 13. Trench for construction of SU 14 kiln.

SU 9. Fine sand, similar to SU 8 (10 YR 6/3), but including large pieces of fired plaster (> 10 cm) scattered on a horizontal surface.

SU 14. Pottery-firing kiln with double chamber, quadrangular plan, central pilaster and suspended floor made of five arms in star-like arrangement.

SU 10. Loose silty sand (10 YR 7/3) with inner lightcoloured lenses of pure sand. Plaster particles are more rare.

SU 15. Earliest abandonment deposits (lenses of fine aeolian sand, ash) forming the deposition surface of the main collapse layers (large chunks of plaster, plaster crumbles, fragments of bones) deposited in the lower (or combustion) chamber of the SU 14 kiln.

SU 11. Trench for the construction of the small SU 12 kiln on top of SU 10.

Figure 11.32 Plan of the potter’s kiln excavated at the centre of Site No. 999.

183

G. L. Bonora and M. Vidale

Figure 11.33 Section of the potter’s kiln excavated at the centre of Site No. 999.

Figure 11.35 Two fragments of a metallurgical crucible, from the surface of AA 1 (see also Figure 11.26.7). Figure 11.34 Harris Diagram of the excavation in Figures 32 and 33. See also the list of the Stratigraphic Units (SU) in the text.

184

An Aspect of the Early Iron Age (Yaz I) Period in Margiana: Ceramic Production at Site No. 999

Figure 11.36 Identification, on the surface, of the central pillar and an arm of the kiln grid.

Figure 11.39 Section excavated next to the kiln, down to virgin soil.

Figure 11.37 A stage in the excavation of the main trench.

Figure 11.40 A detail of the kiln wall at the joint with one of the suspended arms.

Figure 11.38 Surface of the kiln after the removal of the inner collapse layer. The expansion of the trench in the lower left of the picture corresponds to the location of the fuel canal (see map in Figure 11.36).

Figure 11.41 Detail of another suspended arm (lower left) and of the collapse area of the missing arm, with its broken pieces in the filling.

185

G. L. Bonora and M. Vidale SU 16. Layer of charcoal and ash in a sandy matrix (10 YR 4/2), dumped into the mouth of the kiln. Very fine blackish sand with plaster pieces and crumbles, fragments of perforated slabs from the kiln grid, bone fragments, a few large pieces of charcoal. SU 17. Layer with a very fine sandy matrix (10 YR 7/3) almost entirely formed of detached kiln linings and fragments of perforated slabs from the kiln grid (primary collapse of the structural elements of the mouth?). SU 18. Homogeneous fine sand, with scattered granules of chaff-tempered clay or plaster, colour 10 YR 7/3. The plaster pieces come from linings and broken perforated grid elements, and belong to one or more horizontal collapse planes. SU 19. Main layers of destruction, collapse and filling of the combustion chamber: formed of large chunks of plaster, large linings with variable effects of exposure to heating, pieces of a broken suspended arm, broken perforated grid slabs, plaster crumbles, fragment of bones, a few potsherds. SU 20. Fine compact sand on the surface, colour 2.5 Y 7/2, including (on top) some hydromorphic sheets and sandy silt lumps from decayed mud-bricks, and covering the topmost part of the secondary filling of the kiln. SSUV 22. and 23. Describe a later, not datable round pit cutting through the edge of the kiln and breaking part of its linings, and its loose filling.

Figure 11.42 A fragment of a perforated grid slab lying horizontally near the fuel canal (see also Figure 11.14).

MV Discussion The sequence and items found in the kiln complete the reconstruction provided by the first trench, and add further details to it, solving many of the doubts expressed above. The natural substrata (basal compact sandy clay, Figure 11.40) and the layers of fine sandy silt deposited on top at a later time − SU 1-3 − tally with what was unearthed at the base of the trench excavated in the trashing ground. In the area of the kiln, the first evidence of human intervention is a pit − SU 4 − later filled with a sequence of silt, sandy silt and sand, including partly-fired angular particles of the sandy clay of SU 1, partially fired, charcoal bits and pieces of chaff-tempered clay (Figures 11.37 and 40). Though, given the clay nature of the substratum, one is obviously tempted to interpret this pit as a trench for excavating the clay needed by the potters, its filling layers clearly show that, in the nearby areas, ovens or kilns had already been built and were being used. When the trench was completely filled, the SU 9 horizontal layer could be related to an episode of destruction-rebuilding of the kiln or kilns erected near the trench. SU 10 is a thick layer of loose silty

Figure 11.43 Detail of the sub-cylindrical “pillar” made of spirally formed slabs and sunk vertically to the right of the mouth of the kiln, immediately outside its wall. The technical meaning of this feature (not completely excavated) is unknown at present.

186

An Aspect of the Early Iron Age (Yaz I) Period in Margiana: Ceramic Production at Site No. 999 sand apparently covering a large portion of the centre of the mound. Less rich in plaster particles, this layer may be the transformation of an alluvial layer, but the evidence is inconclusive. What is certain is that SU 10 marks a temporary suspension of the local pyro‑technological activities. At a later time, the area came to be occupied by a mud-brick platform or structure, SU 21, and by a small oven or kiln, SU 11 and 12, about 60 x 60 cm, having a ovoid or sub-triangular ground layout, near the corner of the same building. A minor expansion probably marks the location of the mouth (in this case, the smaller kiln would have the same orientation as the larger kiln built later in the same area). Since this small kiln was completely destroyed and the cavity was filled with its collapse elements in a secondary context of deposition, its function and technical details are matter of guess. Nonetheless, a careful scrutiny of the overfired linings gathered in SU 11 showed the presence of the outer lining of a possible central pillar having, on top, a diameter of about 10 cm; from this top there protruded a set of horizontal joined slabs forming a suspended plane broken at its outer edge. This plane, conserved for a maximum length of about 30 cm, had no holes, and if any holes were present, they were undoubtedly arranged at the periphery of the suspended plane, which was not preserved.

lower or combustion chamber (for keeping the fuel) that in all likelihood was partially sunk into the ground, a central pillar supporting a star-shaped structure of five plastered clay bars or arms, and an upper firing chamber protected by a vault for the vessels to be fired. The latter has not been preserved, but it can be reconstructed in a fairly detailed manner on the basis of the examination of the materials recovered in the secondary filling of the combustion chamber. Its interpretation as a pottery-firing infrastructure depends more on the general context of the site and on the surface finds than on the features and associations of the kiln itself. Pottery manufacture is actually suggested by the large number of vitrified kiln blocks visible on the surface, by the presence of chaff-tempered flat dishes or bats on the surface near the locations of the kilns, and by the find, already reported, of at least two small polishers in marble and jasper, which had clearly been used for smoothing and polishing ceramic surfaces (both found in association with the excavated kiln). The kiln excavated in Site No. 999 is formed, in its lower or combustion chamber, of a simple, shallow trench which had been excavated in the ground, and whose walls were carefully plastered (at least two or three times) with layers of chaff-tempered clay. In some points, particularly in the rear of the combustion chamber, the last plastering left patches of overfired clay that retained the deep parallel grooves left by the hands of the potters. The lower chamber is quadrangular or trapezoidal, slightly larger at the rear: the sides measure 1.30 x 1.20 m (Figure 11.32). The height of the combustion chamber, measured from the floor to the base of the suspended arms, is only 25 cm (Figure 11.33). apparently in marked contrast with the deep, bell-shaped combustion chambers of the Iron Age kilns from Uch-depe and Churnok. The rear wall is slightly convex, while the sides, on the contrary, are slightly convex. The fuel canal or mouth is a squarish, shallow depression of about 70 x 70 cm, originally flanked by two convex plastered wall segments jutting out approximately 40 cm.

Later, after the small kiln had been abandoned and destroyed, a larger and better-preserved pottery-firing kiln was built on the edge of the platform or building. The technical details of the infrastructure are reported in detail in the next part of the report. The cross-section shows the stratigraphic relationships with the smaller kiln SU 12, the earliest abandonment and collapse layers on the pavement of the combustion chamber (the mouth of the kiln seems to have its own collapse deposits), and the final destruction and collapse materials filling the combustion chamber. The latter are almost entirely formed of kiln linings, fragments of perforated grid slabs, and broken pieces of one of the suspended arms of the firing chamber (Figures 11.39, 42-43: see below). Probably they also include part of the ephemeral structure of the vault, but, with only a few exceptions, these collapsed features are hardly recognisable in the filling. It may be relevant to note that the secondary filling of the combustion chamber did contain a very small number of potsherds, but none of them could be considered part of the ceramics originally fired in the kiln. The sequence is concluded by an undatable pit cutting through the edge of the kiln, SU 22, 23, and by the present surface, SU 20, also containing the decay products of the nearby platform or building.

Vidale excavated the kiln on the basis of his field experience in South Asia, assuming that the bottom of the fuel canal was approximately at the same level as the inner combustion chamber; for this reason, he stopped his excavation as soon as he reached the first coherent layer suggesting a solid trampling surface rich in ash, charcoal and relevant ceramic indicators such as perforated grid slabs. Later, he realised that ceramic kilns in Central Asia dating from the Bronze and Iron Age are regularly reported as having a deep cylinder- or bell-shaped cavity in front of their mouth, often more than 1 m below the surrounding surfaces. While these fire-pits are an interesting problem (they should involve a considerable waste of heat, and in our experience have no obvious technical explanation) there is the possibility that our kiln too was originally equipped with a similar fire-pit, and that Vidale’s excavation actually stopped at the uppermost and upper filling layer.

GLB, MV The Kiln: Technical Features The kiln (Figures 11.32-33, 39, 41-42) is interpreted as a simple double-chambered or vertical kiln, equipped with a 187

G. L. Bonora and M. Vidale The central pillar is a rough cylinder spreading out at the top, with a diameter of about 20 cm at the base and about 25 at the top. The pillar is formed of a pile of modelled and pressed sandy clay lined with the usual chaff-tempered plaster (here, too, longitudinal hand and finger impressions were visible). The most evident technical feature is the star-shaped structure formed of five clay bars or arms, completely lined with chaff-tempered plaster, that depart from the central pillar to reach the inner walls of the kiln (Figures 11.32, 41-42). The arms are about 40 cm long; they have a distinctive trapezium-shaped section, with the greater base measuring 14 cm, an equivalent height and a smaller base exceeding 8-10 cm; so they are rather thin, probably in order to reduce the waste of heat. Four of the arms were still preserved in their place. One was broken (Figure 11.43), and its fragments had been dumped into the combustion chamber together with other residues; this accident was actually a lucky event, because the study of these fragments allowed us to reconstruct the forming technique of the bars. Instead of using an inner wooden armature, the potters prepared large slabs of a very fine light-grey clay without inclusions, and rolled these slabs into long bars with a sub-triangular section (the section of the core of the arms is close to a equilateral triangle) by means of a double or single spiralling technique (Figures 11.44.1-3; 11.45). These triangular-section bars were later coated with one or two layers of heavily chaff-tempered plaster, turning their triangular section into a trapezoidal one (Figures 11.44-45), and finally fixed between the central pillar and the upper edge of the lower chamber. But how did the perforated slabs fit in this technical process?

perforated in the plastic state both while they were being formed by pressing them against flat surfaces and while they were being applied to the bars. The holes are always conical; they were made quickly, with rapid movements, and rarely pass through the entire thickness of the slabs; they seem to stop at the end of the piece, as if the punch had encountered a solid obstacle (a plane surface such as the ground or perhaps the palm of the potter’s hand). In some instances these holes simply fail to reach the opposite side of the slab. As a rule, the perforations have an angle of about 45° with respect to the surfaces of the slabs. It is not clear whether this angle is a clever device for diverting the hot flames (for example, towards the walls of the kiln?) or more simply a by-product of the craftsperson’s posture while crouching over the kiln under construction. It is evident, in conclusion, that the distinctive broken perforated slabs that had attracted the attention of the explorers are components of ephemeral perforated floors or grids that had to be built and dismantled after each firing cycle. As far as we know, these unusual, efficient techniques for manufacturing the bars and grid were unknown in the technical repertory of South and Central Asia, and are described here for the first time. If the slab-made grid was a transient feature, this was probably true also of the vault. The use of non-permanent vaults made with earth, fuel, potsherds and plaster linings, to be reconstructed with each firing, is a recurrent but unsubstantiated hypothesis in the field of protohistoric pyro‑technology. In this case, we recovered at least three potsherds that, bearing considerablel traces of soot and/ or a thin plaster coating on their surfaces (interior and exterior) most probably belonged to the last-constructed vault of the kiln (Figure 11.49). Thus potsherds were used by the Yaz I potters of Site No. 999 for building both the suspended grids and the vaults of the kilns. Our data do not allow a precise estimate of the height or volume of the firing chamber, and the reconstruction, in this case, is completely conjectural.

The study of the pieces of perforated grids found on the surface and in the context of the kiln (Figures 11.47-48) showed that these slabs, probably in the form of small, irregularly rounded or ovoid patches of chaff-tempered clay, were shaped one by one by the potters, as needed, partially superimposed on each other to form a multiple stratified floor, and applied between the five bars or arms. Some of the pieces recovered from the collapse layers of the combustion chamber actually show the point of application of these multiple layers or slabs on the top part of the bars; after the contact (where the angle of the upper part of the trapezium is preserved, as in Figures 11.46 and 11.44.4), in most cases the slabs, because of their weight, took on a concave, sagging profile.

Another remarkable and still mysterious technical feature of this kiln was discovered on the last day of excavation. Where the mouth opens outside, forming a projecting wall, exactly at the corner, outside the wall of the kiln, there came to light a vertical-oblique bar or pillar made of clay, still sunk into the ground in its original position (Figure 11.44). This pillar has a round section, a maximum diameter of 14-15 cm at its lowermost exposed point and of 9 cm at the broken top, and a surviving length of more than 35 cm. Like the bars supporting the grid, this pillar had been made out of a large, spirally-rolled slab of clay. Its base jutted out below the wall and floor of the kiln. Most probably, another similar pillar stood in a symmetrical position on the opposite side of the mouth. Time constraints and the fear of destroying the structure by undercutting it prevented Vidale from further exposing the base of this strange feature. These pillars (or this pillar), which are another element of considerable technical ingenuity, must

The technical features of the perforated slabs found in the kiln (Figures 11.43, 47-48) and part of the specimens in Figure 11.46 correspond closely to those found on the surface. The thickness ranges from 0.6-7 cm to more than 3 cm, the estimated size of the individual pre-formed slabs from 5-6 cm of diameter, in the round ones, to more than 20-25 cm, in the elongated patched ones. The diameter of the holes, generally related to the thickness and maximum diameter of the slabs, ranges from around 1 cm to 2.5 cm or more. Our data suggest that these slabs (sometimes enclosing potsherds as reinforcing elements) were 188

An Aspect of the Early Iron Age (Yaz I) Period in Margiana: Ceramic Production at Site No. 999

Figure 11.44 Site No. 999, fragments of the missing suspended arm of the kiln. 1 and 3 were both formed by rolling thick slabs of a relatively pure light grey clay, forming two spirals and firmly pressing them together into a almost equilateral triangular section. Above the bars thus formed, the potters applied two or more regular layers or slabs of chaff-tempered clay, bringing the triangular form to a trapeze-like section (Figure 11.44.2). Figure 11.44.3 is another arm fragment showing a simpler spiral-like section. The three arm fragments come from the collapse layers inside the excavated part of the combustion chamber. Figure 11.44.4 is a reconstruction of the overall structure of the arm, changing from a triangle at the core to a trapeze with the application of two chaff-tempered coatings; to the left, the joint with the perforated slabs of the grid can be seen (as reconstructed on the basis of the two pieces of the following figure).

Figure 11.46 Site No. 999, fragments of roundish to oval-shaped perforated slabs from dismantled kiln grids. Coarse chafftempered clay, largely overfired or partially vitrified. Figures 11.46.2, 4, 6, 7, 8, 10, 13 were found in the collapse filling in the lower or combustion chamber of the kiln; Figures 11.46.3, 11, 14 in the same collapse layers, but near the mouth. Figure 11.46.9 is a surface find from AA; Figures 11.46.3, 12 from AA. 4. Note in particular the finger impressions in Figure 11.46.9, showing how the piece, a small slab, was formed and perforated holding it in one hand, and in Figure 11.46.10, where a potsherd was included between different slabs or layers of chaff-tempered clay. Figures 11.46.3, 5 and 13 suggests the existence of three roundish slabs with a single central hole; Figure 11.46.14 is a roundish or ovoid slab, and has three holes near its centre and other four peripheral holes. Figures 11.46.3, 6, 8, 12, 14 are slightly concave, while Figures 11.46.1 and 7 were perhaps formed by pressing the clay against a flat surface.

have had a precise function, perhaps that of supporting an arc-shaped element on the mouth (?) but the question, for the time being, remains unanswered. MV Conclusions Within a radius of 50-60 km from the ceramic production centre of Site No. 999, there are some relatively large centres such as Uch-depe and Churnok, both settled in the later Iron Age, and Adam Basan-1; it is not easy to understand

Figure 11.45 Two grid fragments showing the joint between the upper surface of the suspended arms and the edges of the perforated slabs of the grid (see also Figure 11.47.4).

189

G. L. Bonora and M. Vidale the Site No. 999 (which, considering the surface evidence and its size may supposedly contain the remains of, about 15 or 20 similar infrastructures) should not necessarily be considered a large industrial centre active for centuries: it may also have been, a craft compound settled for a couple of generations by one to three families of potters. We know for sure that the kiln area was regularly cleaned and maintained, since the western part of the site, at least judging from the surface and from the results of a single, small test trench, was probably an open trashing ground. On the basis of the limited size of our trenches and of the small ceramic assemblage collected on the surface, it is impossible to be more precise; but the stratigraphy around the kiln suggests two and possibly three different phases of occupation that may have been continuous or separated by chronological gaps. Interestingly, the first evidence on the site suggests that the first inhabitants were potters; the fact that at a later time the site, after a flood, contained a considerable building, and that craft activities did not stop, may imply a relatively high degree of socio-economic integration of the potting community into its overall social context (it appears that we are not dealing with a group of nomadic craftsmen occasionally exploiting an exposed clay bed). Unfortunately at present it is not possible to ascertain whether the occupation of the site was permanent or periodical, or whether the manufacture of ceramics was a seasonal activity.

Figure 11.47 Fragments of chaff-tempered perforated slabs from the fillings of the kiln.

Figure 11.48 Fragments of chaff-tempered perforated slabs from the fillings of the kiln.

At any rate, the site was involved in the production of comparatively fine ceramics (only for the medium-tolarge brown-buff jars, so far, we lack positive evidence of local production). There is the possibility that these vessels, mainly used as table display ware, were used by relatively high social ranks. The capacity of the firing chamber itself was probably not great, and this is another factor that suggests a fairly accurate production. There are some indications that the three ceramic classes found at the site were formed with different manual techniques, aided or not by the use of the potter’s wheel; so there may have been some form of technical specialisation. Finally, as regards the firing technology, both the inner horizontal bars and the outer pilasters of the kiln are an intelligent, creative technical adaptation to the need of manufacturing strong, resistant structural or architectural elements using only a fine variety of the local clays. Social integration, production of vessels possibly aimed at satisfying the demand of different social ranks, technical specialisation and an efficient technology resulting from a sound, longestablished tradition appear to be quite plausible in the stratified and partly centralised social configuration of the early Iron Age in central Asia. But, for the time being, we had better wait for the acquisition of a more substantial archaeological background, and we are the first to regard our “conclusions” as a set of preliminary working hypotheses.

Figure 11.49 Three potsherds found in the collapse layer that filled the firing chamber of the kiln. The interior is coated with a film of chaff-tempered clay, both blackened by soot and reduced (left-hand specimen) or oxidised (right-hand specimen). These potsherds most probably belonged to the last temporary vault of the kiln. Together with the grid piece in Figure 11.51.10, they are an eloquent example of the potter’s universal propensity to recycle potsherds as building material in many different ways.

which centre or which micro-region was actually supplied with vessels by the potters. Site No. 999 was probably active for a relatively short period within the second half of the 2nd millennium BC. Ethnoarchaeological experience suggests that a kiln like the one we excavated may have been used for a time span of a few years. As a consequence,

GLB, MV 190

An Aspect of the Early Iron Age (Yaz I) Period in Margiana: Ceramic Production at Site No. 999 References Cited

Gutliev, G. 1982. “Raskopki Garaoi-depe” in Novye arkheologicheskie otkrytiya v Turkmenistane 1982: 3347. Hiebert, F. T. 1994. Origins of the Bronze Age Oasis Civilization in Central Asia. Cambridge. Hiebert, F. T. 2003. A central Asian Village at the Dawn of Civilization, Excavations at Anau, Turkmenistan. University Museum Monograph 116, University of Pennsylvania Museum of Archaeology. Philadephia. Hiebert, F. T. and R. H. Dyson jr. 2002. “Prehistoric Nishapur and the frontier between Central Asia and Iran”, Iranica Antiqua XXXVII: 113-150. Huntington, E. 1908. “Description of Kurgans of the Merw Oasis”, in Pumpelly (ed.) 1908: 217-232. Itina, M. A. 1977. Istoriya stepnykh plemen Yuzhnogo Priaral’ya. (Trudy Khorezmskoj arkheologoetnograficheskoj ekspedicii, vol. X). Moskva. Kachuris, K. 1967. “Raskopki na El’ken-depe v Iuzhnoy Turkmenii”, Arkheologicheskie otkrytiya 1966: 335336. Khlopin, I. N. 1975. “Poselenie epokhi bronzi Parkhaidepe”, Kratkie soobshcheniya Instituta arkheologii AN SSSR 142: 116-121. Khlopina, L. I. 1981. “Namazga-depe and the Late Bronze Age of Southern Turkmenia”, in Kohl (ed.) 1981: 3560. Kohl, Ph. L. (ed.) 1981. The Bronze Age Civilization of Central Asia. Recent Soviet Discoveries. New York. Kohl, Ph. L. 1982. Central Asia. From Beginnings to Iron Age. Paris. Kohl, Ph. L., R. Biscione and M. L. Ingraham 1982. “Implications of Recent Evidence for the Prehistory of North-Eastern Iran and South-Western Turkmenistan”, Iranica Antiqua XVI: 185-204. Koshelenko, G. A. (ed.) 1985. Drevnejshie gosudarstva Kavkaza i Sredney Azii. Moskva. Kuftin, B. A. 1956. “Polevoj otchet o rabote XIV otryada Yu.T.A.K.E. po izucheniyu kul’tury pervobytnoobshchinnykh osedlo-zemledel’cheskikh poselenij epokhi medi i bronzy v 1952g.”, Kratkie soobshcheniya Instituta arkheologii AN SSSR VII, Ashkhabad: 260290. Kuz’mina, E. E. 1971. “O nekotorykh aspetakh problemy kul’turnykh i etnicheskikh svyazej Srednej Azii i Irana v epokhu pozdnej bronzy i rannego zheleza” in Isskustvo i arkheologiya Irana. Moskva: 119-123. Kuz’mina, E. E. and A. A. Lyapin 1984. “Novye nakhodki stepnoj keramiki epokhi bronzy na Murgabe” in Problemy arkheologii Turkmenistana, Ashkhabad: 621. Lecomte, O. 1999. “Vehrkana and Dehistan: Late farming Communities of South-west Turkmenistan from the Iron Age to the Islamic period”, Parthica 1: 137-169. Lisizina, G. N. 1965. Oroshayemoye zemledeliye epokhi eneolita na yuge Turkmenii. Moskva. van der Leew, S. E. and A. C. Pritchard 1984. The Many Dimensions of Pottery: Ceramics in Archaeology and Anthropology. Amsterdam.

Al’baum, L. I. 1969. “K datirovke verkhnego sloya poseleniya Kuchuktepa”, in Istoriya material’noj kul’tury Uzbekistana, vol. 8. Tashkent: 69-79. Andrianov, B. V. 1969. Drevnie orositel’nye sistemy Priaral’ya. Moskva. Askarov, A. A. 1982. “Kuchuk-tepe. Spätbronzezeitliche and ältereisenzeitliche Siedlung in Süd-Uzbekistan”, Beitrage zur allgemeinen und vergleichenden Archëologie 4: 1-22. Askarov, A. A. and L. I. Al’baum 1979. Poselenie Kuchuktepa. Tashkent. Bukinich, D. D. 1924. Istoriya pervobytnogo oroshaemogo zemledeliya v Zakaspijskoj oblasti v svyazi s voprosom o proiskhozhdenii zemledeliya i skotovodstva. Kholpovoe delo, No. 5. Ashkhabad. Cattani M. and B. Genito 1998. “The Pottery Chronological Seriation of the Murgab Delta from the End of the Bronze Age to the Achaemenid period: a Preliminary Note”, in Gubaev, Koshelenko and Tosi (eds.) 1998: 75-88. Cremaschi, M. 199., “Palaeohydrography and Middle Holocene Desertification in the Northern Fringe of the Murgab Delta” in Gubaev, Koshelenko and Tosi (eds.) 1998: 15-25. Diakonoff, I. M. 1985. “Media”, in ������������������ Gershevitch (ed.) 1985: 36-148. Ershov, S. A. 1944. “Arkheologiya v Turkmenskoj SSR za 20 let” in Izvesti TurkmenFAN SSR 2-3. Ashkhabad: 27-35. Fano Santi, M. (ed.) 2000. Sudi di archeologia in onore di Gustavo Traversari. Roma. Francfort, H-P. and O. Lecomte 2002. “Irrigation et société en Asie Centrale des origins à l’époque achéménide”, Annales. Histoire, Sciences sociales, 57 année, num. 3: 625-663. Gershevitch, I. (ed.) 1985. Cambridge History of Iran, vol. II. Cambridge. Gubaev, A. G., G. A. Koshelenko and M. Tosi (eds.) 1998. The Archaeological Map of the Murgab delta, Preliminary Reports 1990-95. Reports and Memoirs, Series Minor, III. IsIAO. Rome. Gutliev, G. 1970a. “Raboty na poselenii rannezheleznogo veka Yasi-depe i Baba-Durmaza”, Karakumskie drevnosti III: 64-71. Gutliev, G. 1970b. “Yasi-depe – Novij pamyatnik perioda Yaz-depe”, Arkheologicheskie otkrytiya 1969: 423424. Gutliev, G. 1974. “Raskopki na poselenii Yashilli-depe i Gosha-depe”, Arkheologicheskie otkrytiya 1973: 504. Gutliev, G. 1977. “Raskopki poseleniya rannezheleznogo veka Yasi-depe v Kaakhkinskogo raione”, Karakumskie drevnosti V: 18-24. Gutliev, G. 1980. “Stratigraficheskij shurf Yashilli-depe i Gyaursa”, Novye arkheologicheskie otkrytiya v Turkmenistane 1980: 21-34.

191

G. L. Bonora and M. Vidale Pigott, V. C. (ed.) 1990. The Archaeometallurgy of the Asian Old World. University Museum Symposium Series, Volume VII. Philadelphia. Piotrovskij, B. B. 1949. “Razvedochnye raboty na Gyaur-Kala v Starom Merve”, in Materialy YuzhnoTurkmenistanskoj Arkheologicheskoj Kompleksnoj Ekspedizii 1. Ashkhabad: 35-41. Pumpelly, R. (ed.) 1908. Explorations in Turkestan. Washington. Rice, P. and D. Kingery (eds.) 1997. Prehistory and History of Ceramic Kilns. Columbus. Sajko, E. V. 1978. “K izucheniiu teplotekhniki keramicheskogo proizvodstvo epokhi bronzi v Iuzhnom Turkmenistane”, Karakumskie drevnosti VII: 87-93. Sajko, E. V. 1980. “ Spezial’nye glinyanie massi v praktike Iuzhno-Turkmenistanskikh masterov epokhi bronzi” in Novie issledovaniya po arkheologii Turkmenistana. Ashkhabad: 120-145. Sajko, E. V. 1982. “Glini goncharov IuzhnoTurkmenistanskikh poselenii epokhi bronzi”, in Novie issledovaniya po arkheologii Turkmenistana, Ashkhabad: 139-144. Salvatori, S. 2000. “Mappa archeologica del Murgab: distribuzione insediamentale dal Bronzo medio al periodo Sasanide. Elementi per lo studio della formazione delle società complesse”, in Fano Santi (ed.) 2000: 751-772. Sarianidi, V. I. 1957. “���������������������������� Keramicheskie pechi drevnej Margiany”, in Kratkie soobshcheniya Instituta istorii material’noj kul’tury 69: 72-77. Sarianidi, V. I. 1971. “Issledovanie sloev rannezheleznogo veka na Ulug-depe”, Arkheologicheskie otkrytiya 1970: 433-434. Sarianidi, V. I. 1972. Raskopki Tillya-tepe v Severnom Afganistane. (Materialy k arkheologicheskoj karte Severnogo Afganistana, I). Moskva. Sarianidi, V. I. 1977. Drevnie zemledel’zy Afganistana. Moskva. Sarianidi, V. I. 1981. “Margiana in the Bronze Age”, in Kohl (ed.) 1981: 165-193. Shchetenko, A. Ya. 1968a. “Raskopki pamyatnikov epokhi eneolita i bronzovogo veka v Kaakhkinskom rajone”, Karakumskie drevnosti I: 18-29. Shchetenko, A. Ya. 1968b. “Raboti otryada dva v Iuzhnoj Turkmenii”, Arkheologicheskie otkrytiya 1967: 347348. Shchetenko, A. Ya. 1970. “Raboti artikskogo otryada Karakumskoj ekspedizii”, Arkheologicheskie otkrytiya 1969: 415-417. Schmidt, E. 1908. “Archaeological Excavations in Anau and Old Merv”, in Pumpelly (ed.) 1908: 83-216. Tolstov, S. P. 1948a. Drevnij Khorezm. Moskva. Tolstov, S. P. 1948b. Po sledam drevnekhorezmijskoij civilizacii. Moskva. Venco Ricciardi, R. 1980. “Archaeological Survey in the Upper Atrek Valley (Khorassan, Iran). Preliminary Report”, Mesopotamia XV: 51-72.

Lisitsina, G. N. 1969. “The Earliest Irrigation in Turkmenia”, Antiquity XLIII: 279-288. Lyonnet, B. 1996. Sarazm (Tadjikistan) Ceramiques. Chalcolithique et Bronze Age. (Mémoires de la Mission Archeologique Française en Asie Centrale, Tome VII). Paris. Marushchenko, A. A. 1959. “El’ken-depe. Otchet o rabotakh 1953, 1955, i 1956”, in Trudy Instituta istorii, arkheologii i etnografii Akademii nauk Turkmenskoj SSR, vol. V. Ashkhabad: 54-109. Masimov, I. S. 1968. “Izuchenie pamyatnikov epokhi bronzy i rannego zheleza v rajone stanzii BabaDurmaz”, Karakumskie drevnosti II: 25-29. Masimov, I. S. 1971. “Raskopki keramicheskikh pechej epokhi bronzy Yuzhnogo Turkmenistana”, Arkheologicheskie otkrytiya 1970: 432-433. Masimov, I. S. 1972. “Izuchenie keramicheskikh pechej epokhi bronzy na poselenii Ulug-Depe”, Karakumskie drevnosti IV: 35-46. Masimov, I. S. 1973. “Raskopki keramicheskikh pechej epokhi bronzy na Altyn-depe”, Arkheologicheskie otkrytiya 1972: 482-483. Masimov, I. S. 1976. Keramicheskoe proizvodstvo epokhi bronzy v Yuzhnom Turkmenistane. Ashkhabad. Masimov, I. S. 1982. “Novie pamyatniki rannezheleznogo veka Murgabskogo oazisa”, in Novie arkheologicheskie otkritiya v Turkmenistane: 20-32. Masimov, I. S. and B. N. Udeumuradov 1984. “Novye materialy po rannezheleznomu veku nizovij Murgaba”, in Masson and Atagarryev (eds.) 1984: 12-27. Masson, M. E. 1951. “Iuzhnoj Turkmenii Arkheologicheskoi Kompleksii Ekspediziya v 1947 g.”, in Trudy YuzhnoTurkmenistanskoj Arkheologicheskoj Kompleksnoj Ekspedizii, vol. II. Ashkhabad: 7-72. Masson, M. E. 1955. “Kratkaya khronika polevykh rabot Yu.T.A.K.E. za 1948-52 gg.”, in Trudy YuzhnoTurkmenistanskoj Arkheologicheskoj Kompleksnoj Ekspedizii, vol. V. Ashkhabad: 197-249. Masson, V. M. 1956a. “Pervobytno-obshchinnyj stroy na territorii Turkmenii”, in Trudy Yuzhno-Turkmenistanskoj Arkheologicheskoj Kompleksnoj Ekspedizii, vol. VII. Ashkhabad: 233-259. Masson, V. M. 1956b. “Raspisnaya keramika Iuzhnoj Turkmenii po raskopkam B. A. Kuftina”, in Trudy Yuzhno-Turkmenistanskoj Arkheologicheskoj Kompleksnoj Ekspedizii, vol. VII. Ashkhabad: 291-373. Masson, V. M. 1959. Drevnezemledel’cheskaya kul’tura Margiany. (����������������������������������������� Materialy i issledovaniya po arkheologii SSSR, vol. 73). Moskva-Leningrad. Masson, V. M. and E. Atagarryev (eds.) 1984. Turkmenistan v epokhy rannezheleznogo veka. Ashkhabad. Masson, V. M. and V. I. Sarianidi 1972. Central Asia. Turkmenia before the Archaemenids. London Miller, H. M.-L. 1997. “Pottery Firing Structures (Kilns) of the Indus Civilization during the Third Millennium B.C.”, in Rice and Kingery (eds.) 1997: 41-71. Pigott, V. C. 1980. “The Iron Age in Western Iran”, in Wertime and Muhly (eds.) 1980: 417-462. 192

An Aspect of the Early Iron Age (Yaz I) Period in Margiana: Ceramic Production at Site No. 999 Wertime, T. A. and J. D. Muhly (eds.) 1980. The Coming of the Age of Iron. New Haven. Yacobson, A. L. 1941. “Goncharnie pechi srednevekovogo Khersonesa”, in Kratkie soobshcheniya Instituta istorii material’noj kul’tury X. Moskva: 54-59. Young, T. C. Jr. 1965. “A Comparative Ceramic Chronology for Western Iran, 1500-500 BC”, Iran 3: 53-85.

Young, T. C. Jr. 1967. “The Iranian Migration into Zagros”, Iran 5:11-34. Zadneprovskij, Yu. A. 1962. Drevnezemledel’cheskaya kul’tura Fergany. Moskva-Leningrad. Zhukovskij, V. A. 1894. Razvaliny starogo Merva. Moskva.

193

Chapter 12

Iron-working and Ceramic Recycling on the Surface of a Late Iron Age Fort at the North-eastern Fringe of the Murgab Delta. Massimo Vidale, Enrico Battistella and Giuseppe Guida Introduction

example the site of Churnok, the southernmost of the Takhirbai area, 4.5 km north of Garry Kishman, contained pottery-firing kilns (Masson 1959: 65-66). The associated sherds suggested a date within the Yaz III period. Surface evidence of abundant iron slag and limited test trenches in the same area suggested that some form of iron working was also practiced on the spot (see Kohl 1981). Other Iron Age kilns were excavated in the northern extension of Uch-depe; the ceramics (small-sized containers, cups, cylindrical vessels and beakers) could be compared with pottery from Yaz-depe, Yaz III levels. Small dumps from specialized semi-precious stone bead making activities (mainly lapis lazuli, turquoise and/or apatite: Vidale et al. 1998) were detected on the surface of Takhirbai 1 and dated to the same general time-span. Thus, crafts not directly monitored by the élite could have been decentred and independently performed in peripheral villages, while other activities having a more strategic interests for the same elites could have been temporarily hosted and performed within central fortified buildings; at least, this might be the first working hypothesis. But what about iron working? Was it controlled by the state, tied to the urban markets or freely performed at the level of each village and rural community? In future, by answering similar questions we might be able to widen and detail the picture of the organization of craft production among the leading centres and the peripheral areas in the (proto)history of this early state in the delta.

In spite of a century of field research, the seminal work by V. M. Masson and his colleagues and the new substantial evidence from our own work, the Middle and Late Iron Age of the Murgab Delta remains poorly known. This is mainly because, at present, the pottery sequence, as observed from the new surface collections and from the Takhirbai 1 trenches, has been studied and published only in a very preliminary form (Cattani and Genito 1998; Genito et al. 1998; Bonora and Vidale, Chapter 11), and the changing settlement patterns for the second half of the 1st millennium BC are known only in a very broad outline. The prevailing opinion is still that the Yaz I-related archaeological complexes and sites, at the beginning of the 1st millennium BC belonged to one or more Iranian-speaking tribal groups, that were capable, in a few centuries, of organizing themselves into a state. The contemporary settlements in the westernmost section of the eastern Murgab suggest a local polity centred on a large settlement cluster (Sites No. 211, 212 and 213) functioning as a capital, dominating smaller villages, some inhabited by specialized craft groups (Bonora and Vidale, chapter 11; Vidale 2006) and surround by standard-sized agricultural lands. We know that in the following Yaz II-Yaz III phases the delta was less densely populated and the political centre may have shifted to the area of the fortified site of Erk-kala. This movement may have been anticipated by a general change of the hydrographic system in the south-eastern part of the delta. At the same time, some scholars connected the abandonment of the walls of Yaz-depe and the destruction of Aravali-depe to the military conquest by the army of Darius I mentioned by the Bisutun inscription (Genito 1998a, 1998b).

More explicit is the evidence of a precise military organization of some sites at the easternmost edge of northeastern Murgab (Figure 12.1). Here the project located a line of fortified sites (namely Sites No. 172, 173, 174, 215, 237 of the general survey) spaced about 1-1.5 km apart, the first 4 preliminarily interpreted as camps or “fortresses” (Genito et al. 1998). This planned defence line (Figure 1.8) might ideally be extended south, towards the so-called and presumably later Antiochus Wall in its north-south oriented tract at the extreme eastern fringe of the ancient cultivable areas (Ibidem: Fig. 1). The surface ceramics were ascribed to Yaz III and thus, hypothetically, to the 5th3rd centuries BC. A dried-up stream bed meandering west of the forts’ line follows their general orientation and might have provided water to the contemporary settlements. The presence of a defensive boundary wall was confirmed by a test trench excavated at one of these sites (Site No. 215). The surface topography suggested a fortification oriented south-north with sides measuring about 160 x 160 m, with

After that time, presumably, the delta might have been part of the Persian satrapy until the foundation of Alexandria Margiana in 328 BC, although, besides this information, there is no historical and evidence of a direct military or political interference of the Achaemenid State in the Murgab Delta (Genito 1998b). The common pottery, in fact, belongs to a local unbroken technical tradition and there is no evidence of a clear-cut Persian influence. At any rate, the settlement pattern of the later Iron Age suggests the integration of the delta in a single state-level polity, perhaps under a fluctuating control by the Achaemenid kings. The available record shows that some Yaz III sites were deeply involved in specialized craft activities. For

195

M. Vidale, E. Battistella and G. Guida

Figure 12.1 Map of the north-eastern fringe of the Murgab Delta with the location of Site No. 172 (from Genito 1998c: Fig. 1).

196

Iron-working and Ceramic Recycling

Figure 12.2 A view of the central plateau of Site No. 172 with its central square ruin. In the foreground, near the recording sheets, a surface cluster of iron-working indicators. Figure 12.3 Topographic map of Site No. 172 (by G. L. Bonora and Putzolu), with 50 cm contour lines.

possible entrances on the northern and southern sides and possible round towers at the corners. In this case the wall had been made with pressed clay blocks, and in its eroded form had a width of 5-6 m and a remaining height of little more than 1 m. According to the exposed sections, the site was interpreted as a seasonal fortified camp rather than a permanent fortress.

a steep mound having a square ground plan. This ruin clearly represents what is left of a main tower or “castle” built in mud bricks, with sides measuring about 30 x 30 m, oriented from south-east to north-west. The tower rises about 5 m above a lower mound that forms great part of the actual surface of the site; its top is flat and measures about 10 x 5-7 m. Below, the most evident topographic feature is a scarp running from south west to north-east, about 150 m long. This scarp rises with a steep slope of about 3-3.50 m from the surrounding flats. Most probably it is what remains of the southern mud-brick wall of the fort. It has no apparent gaps, and at its ends there are two cone-shaped, eroded hummocks that we interpreted as the remnants of two angular towers. This wall shows the same orientation of the inner major building. Less recognizable are the north-western extensions of the wall, respectively northwest, north-east and south-east of the inner building. The topographic contour lines, at any rate, suggest the presence of shorter orthogonal wall tracts at north-east and southwest of the main mound. The north-eastern corner of the fort seems to be completely eroded, and the north-western one could have been flattened by the same processes into a faint slope. A gap in the north-eastern short side might indicate the location of a gate. If our interpretation is correct, the fort might have been rectangular or trapezoidal in ground plan, protected by walls measuring about 150 x 120 m, with round towers at the corners, at least two of which still recognizable, and a high, massive rectangular building in its north-eastern corner. In terms of size and overall spatial patterning, the fort seems comparable to those previously described on the base of surface evidence and excavated at Site No. 215. Outside the wall, particularly at south, large dumps and scatters of ceramics might indicate further extensions of the settlement. The whole settled area might have measured about 2 ha; about

Site No. 172 (Figure 12.2) was found few kilometers south of Site No. 215 and appeared broadly similar. It attracted our attention because of the obvious presence on surface of large amounts of iron slag, suggesting, at first sight, that the “fort” could have also sheltered the production of weapons. One week of surface analysis and sampling allowed M. Vidale to produce a preliminary map of the “activity areas” unearthed by erosion inside and outside the fortified compound, and to gather some artefacts for future analytical studies. In this paper, we present some comments on the site’s topography and a preliminary rendering of its fortified structures, a description and general distribution of craft indicators on the surface of the site, some hypotheses on the iron-working sequences performed in the fortified compound, based upon a first archaeometallurgical evaluation, and, finally, a graphic reconstruction of a type of craft installation presumably used in the site. Fortifications and Craft Areas as Observed on Surface The following descriptions and maps are based upon an original topographic recording carried out on October 10th, 1999 by Gian Luca Bonora and Cristiano Putzolu. The contour lines have intervals of 50 cm (Figure 12.3). In the low, monotonous surrounding landscape, where the lightcoloured depressions of the takyr are bounded by sanddunes and by the low reliefs of tamarisk trees, Site No. 172 is dominated by its central elevation (Figure 12.2), 197

M. Vidale, E. Battistella and G. Guida

Figure 12.4 Map of the site showing the location of the Activity Areas (AA) discussed in the text.

activity areas (designated AA, see Figure 12.4), four of which outside the south-eastern ramp or wall (AA 2, 3, 4, 6) possibly marking the location of ancient dumps flattened and scattered by two thousand years of erosion. Area 11, an isolated, small cluster of iron slag in an ashy matrix outcropping on surface, was originally interpreted as a smithing deposit eroded in situ, but a limited subsurface excavation did not confirm this idea and the area was considered as another small-scale dump disturbed by intensive weathering. Even if the surface clusters of iron working slag and other craft indicators within the fortified compound (1, 1A, 5, 5A, 7-10) are homogeneous and substantial, there is no archaeological proof that their locations exactly mark metallurgical furnaces, fireplaces or even large eroded dumping grounds still partially embedded in the buried deposits. Nonetheless, even if we are dealing with secondary or even tertiary deposits, it is

1/3 of the site, roughly speaking, is covered by dumped iron slag. Another craft activity, discussed in greater detail at the end of this paper, is the systematic chipping of Late Iron Age potsherds for the purpose of obtaining grog, or ceramic flakes or powders to be used as temper in some form of ceramic production. This industry cannot be restricted to specific locations of the mound, nor can it be precisely dated. The left-overs of this activity are mostly round exhausted potsherds, found on this and other sites of the Murgab Delta by the thousands, across ruins of many different ages. Iron-working Activity Areas During five days of field work on the surface of Site No. 172, Vidale could single out eleven different craft 198

Iron-working and Ceramic Recycling highly probable that their occurrence within the fort depends upon the activity, in the general area, of such productive installations. Vidale also observed that across wide areas of the fort burrowing animals exposed to surface patches of sediments very rich in ash, charcoal and clay bits reddened by fire, showing that the site still contained layers and deposits having meaningful relationships with the surface archaeological record of pyrotechnological activities. The most common indicators, across the detected iron working activity areas, are various types of iron slag, hammering scales and vitrified linings interpreted as plaster walls from smithing pits. We therefore enclose a short description of AA 1-11, their general context and the craft indicators identified on surface. For each activity area, craft indicators are listed in order of relative frequency.

m from south to north, with a maximum width of 10 m. The artefacts recognized on surface were: small shapeless fragments of iron slag; fragments of concave ceramic linings, vitrified and iron slagged; many potsherds; many shapeless pieces of thick, coarse ceramic containers with abundant chaff temper; and a fragment of an unidentified iron artefact. AA 4 This activity area is the bed of a long, winding gully (a literal “iron slag stream”) joining along the foot of the southern ramp AA 2, 3 and 6 (see below). It transports down slope an artificial black gravel formed only by iron slag spherules, small drops and drop fragments, and other iron slag pieces with diameters below 1 cm, mostly round in contour. Such a peculiar iron working sediment was produced by the hydraulic-gravitational selection, after rains and temporary mud flows, of the inclusions in the layers forming on surface AA 2 and 6. Other indicators were very rare or absent.

AA 1, AA 1A A dense cluster of ceramic residues and small-sized iron slag on the top of the southern ramp, near the south-eastern corner of the fort. AA 1 runs without gaps on the top of the ramp or wall for about 10-15 m, in a long and narrow strip, with a maximum width of 5 m. Craft indicators were listed and described as follows: shapeless slag pieces, showing gas bubbles on the fracture surfaces; fragments of iron slag drops; fragments of concave ceramic linings, slagged on the inner side, with iron-bearing crusts rich in bubbles; abundant iron slag spherules; many hammering scales of variable thickness; small iron slag drops. Other artefacts were a fragment of a copper-bronze ring, a fragment of an iron sheet, a three-flanged copper/bronze arrowhead, and a piece of a thick, coarse ceramic container with abundant chaff temper. Immediately north of AA 1, on the top rear of the hypothetical corner tower, AA 1A was described as a weaker cluster of the same materials. It occupied a subtriangular space extending about 20 x 20 m.

AA 5, AA 5A The cluster of metallurgical indicators labelled as AA 5 was located on a slightly elevated surface of the inner plateau, between the corner “tower” and its AA 1A and the main hummock. It had various types of ceramic and iron slag fragments turned into small fragments (mostly wavering around 1-2.5 cm in size). The indicators (always in order of relative frequency) were described as: concave ceramic linings, vitrified and slagged on the inner side; many pieces of thick, coarse ceramic containers with abundant chaff temper, yellowish on the interior and red outside (from 2 to 4 cm in size); iron slag pieces, shapeless and with polyhedrical fractures; broken iron slag drops; hammering scales (less common); two pieces of sandstone grinding stones. AA 5A, like AA 1, was identified on top of the southern ramp, and on the base of its distribution could be considered as the southern limit of AA 5, as well as the south-western extension of AA 1. It was distinguished by the very small size (below 1 cm) of its surface inclusions. These were listed as follows: shapeless iron slag bits; hammering scales; pieces of drops and spherules; pieces of vitrified linings with iron slag; pieces of coarse chafftempered ceramics. A three-flanged copper/bronze arrowhead was also found here.

AA 2 Beyond the southern scarp or wall, starting from the foot of its slope, AA 2 is without doubt the output of a severe washing-off process of the dumps from the metallurgical workshops within the fort, more precisely in the southeastern corner of the compound, and visible on surface in AA 1-1A. The post-depositional displacement produced on surface a wide fan-like scatter measuring not less than 30 x 20 m. AA 2 contained the following craft indicators: various types of iron slag pieces, from small to large; fragments of concave ceramic linings, vitrified and slagged on the inner side, 4-8 cm large; medium-sized iron slag drops; iron slag spherules; pieces of thick, coarse ceramic containers with abundant chaff temper (from small to large); a fragment of an unidentified iron artefact.

AA 6 Another surface scatter of craft indicators outside the wall, at the foot of the slope, measuring about 30 x 30 m, perhaps the extension of AA 2. This context was distinguished by a large amount of ceramic finds, as one would expect in a large urban dump. On the other hand, some of the observed sherds, given their technical features, might have been used in some specific pyrotechnological activity. We enclose the list of the recovered craft indicators: large

AA 3 Another secondary surface cluster, gathering objects eroded and scattered outside the fort. AA 3 runs for 35-40 199

M. Vidale, E. Battistella and G. Guida AA 10

fragments of heavy chaff-tempered coarse ceramic walls, about 4-7 cm thick, having thick square or pointed edges, perhaps belonging to parts of ovens or furnaces rather than to portable containers; among them, a rim of a cylinderlike vessel, tempered with flat and sharp pottery flakes; a fragment of a coarse dish with inward-projecting walls; fragments of large concave ceramic linings, vitrified and iron slagged; medium to large-sized iron slag fragments; small to medium-sized fragments and crumbles of iron slag; small amounts of iron slag drops and pellets.

Very similar to the AA we just described, both in terms of surface morphology and associated items, AA 10 had a round shape, with a diameter of about 20 m. It might cover the north-western corner (and possibly the corresponding “tower”) of the fort. The indicators are: iron slag pieces identical to those of AA 9, small to medium-sized, occurring

AA 7 It was found on the top of the southern ramp, near the south-western corner of the “tower”, where it represents the ideal prosecution of AA 1 and 5A. The surface distribution is sub-triangular, and measures about 7 x 10 m. On average, its inclusions are smaller and less dense than those formerly described. The craft indicators in this context are: small iron slag fragments, not determinable; small pieces of thick, coarse ceramic walls with abundant chaff temper; iron slag spherules; medium-sized fragments of concave ceramic linings, vitrified and iron slagged on the interior; hammering scales; fragments of sandstone grinding stones; a bent iron sheet and a point fragment in the same metal, both heavily corroded; many chipped ceramic disks probably used as grog cores; unidentified pieces of polished stone objects; a piece of a bronze ring and a three-flanged copper/bronze arrowhead. AA 8 North-west of AA 7, in the central plateau, this site, very rich in pottery, is a secondary cluster of small-sized objects, among which: small unidentified fragments of iron slag; drops and pellets of iron slag; fragments of concave ceramic linings, with inner iron slagged surfaces; coarse ceramic walls with abundant chaff temper; chipped ceramic disks or grog cores. AA 9 Near the south-western corner of the inner plateau, AA 9, rich in potsherds like the former one, appeared as a large elongated surface (about 40 x 20 m), with a main axis from west to east. The most common craft indicator are medium-sized massive, heavy pieces of slag full of gas bubbles, very rich in iron. Other indicators on record are: many hammering scales; large coarse ceramic pieces with abundant chaff temper, flat in section and similar to bricks; iron slag spherules; fragments of concave ceramic linings, with inner iron slagged surfaces; iron slag drops; some fragments of iron artefacts, perhaps finished and heavily corroded; pieces of grinding stones in sandstone; two lapis lazuli flakes. Two fragments of small glass vessels came also from the same spot.

Figure 12.5 Site No. 172, surface. Tapping slag samples collected from the central plateau of the site. Numbers refer to the samples’ identification.

200

Iron-working and Ceramic Recycling

Figure 12.7 Site No. 172, surface. Examples of hammering scales.

Figure 12.6 Site No. 172, surface. Examples of iron slag drops and spherules.

Figure 12.9 Site No. 172, surface. Fragment of vitrified clay lining with iron slag, bearing the imprint of the mouth of the tuyere.

Figure 12.8 Site No. 172, surface. Fragment of vitrified clay lining with iron slag, bearing the imprint of the mouth of the tuyere.

201

M. Vidale, E. Battistella and G. Guida

Figure 12.10 Site No. 172, surface. Two fragments of vitrified clay linings with iron slag, from the central cavity of smithing pits. Note the concave profile and the inner slag encrustation in form of a “step”, just above the contact with the tuyere.

within the limits of the area in small isolated clusters; small fragments of concave ceramic linings, with inner iron slagged surfaces; iron slag spherules and drops; chipped ceramic disks or grog cores; fragments of grinding stones; few fragments of coarse ceramic walls tempered with flat pottery chips. Within the limits of AA 10 was observed a main cluster of chaff-tempered clay pieces, some of which over fired, about 3 m in diameter, surrounded by scatters of iron slag particles. This may be possible evidence of a buried pyrotechnological infrastructure.

Figure 12.11 Graphic reconstruction of the inner cavity of four smithing pits, showing the angle of the imprint of the tuyere’s mouth. Note also, in shadow, the thickness of the slag encrustation and the renovated plastering.

The Archaeological Record of Iron Working: Spatial Patterning and General Remarks

AA 11 A patch of grey, ashy soil measuring about 50 x 50 cm, originally interpreted, as stated above, as the surface evidence of a smithing fireplace eroded on the spot. Subsurface testing dismissed this idea. The indicators associated on surface were: not determinable small pieces of iron slag; iron slag spherules and drops; hammering scales; small vitrified, iron-slagged plaster lining fragments; fragments of coarse chaff-tempered ceramic walls, in various states of oxidation or reduction (red, yellow, grey).

The associations of craft indicators so far described are summarized in Table 12.1. Heavy iron slag pieces with prismatic fractures (Figure 12.5) are mostly found in the western plateau locations (AA 9, 10, 5 and 5A), together with more ubiquitous indicators such as hammering scales, shapeless and smaller iron slag bits, slag drops and spherules. The other activity areas on

202

Iron-working and Ceramic Recycling Activity Areas (AA) heavy prismatic slag ceramic walls with chaff shapeless slag fragments slag drops slag spherules hammering scales concave plaster linings iron objects ceramic walls with grog grinding stones grog cores ashy matrix

9 ** ** * * * ** * * -

10 ** ** * * * * * * * -

5A * * ** ** ** ** * -

5 * ** * * * * ** * -

4 ** ** ** -

1A * ** ** ** ** -

1 * ** ** ** ** * -

2 * ** ** ** ** * -

3 ** ** ** * -

6 ** * * * ** * -

7 ** ** ** * * * * ** -

8 * ** ** ** * * -

11 * ** ** ** ** * **

Table 12.1 Site No. 172. Associations of craft indicators observed on surface in the surveyed AA.

the eastern part of the plateau show a consistent association among relatively large amounts of slag drops, spherules (Figure 12.6), hammering scales (Figure 12.7) and vitrified concave linings (Figures 12.8-11). Hammering scales were not observed in AA 2, 3, 6, but most probably this is due to the fact this indicator, being flat and light, was probably completely removed by rain and by the consequent washing off processes. Another indicator affected by local sedimentary formation, as discussed above, is the visibility on surface of an ashy matrix closely associated to the iron residues (AA 11). Our interpretation is that the western sites were involved in the smelting of iron mineral (using the bloomery process), together with some smithing operations (the preliminary refining of the blooms?), while the eastern sites, more regularly including fragments of concave slagged or vitrified plaster linings and fragments of finished or unfinished iron artefacts, hosted the bulk of the smithing activities. Among the few pieces of finished (or unfinished, it is hard to say) fragmentary iron artefacts there are pieces of sheet-like and bar-like objects (Figure 12.12). Some of them could have been lost or discarded while forging sets of similar objects. We can recognize a fragment of a sickle blade (Figures 12.12e, 12.13) and what seems the end of a pointed tool or weapon having a rectangular section (Figure 12.12d). Shapeless slag fragments are generally small and light particles, thus it is not surprising to see larger amounts of these particles in the smithing areas. Fragments of chafftempered ceramic materials occur in both locations, and we may preliminarily conclude that this was the base building materials of smelting furnaces and smithing forges. Other indicators, such as grinding stones fragments, pieces of grog-tempered vessels and the possible disk-shaped grog cores, cannot be attributed to one or another technical context. This interpretation includes the identification of the heaviest, bulky iron slag pieces common in AA 9, 10, 5 and 5A as tapping slag (see among others Bachmann 1982; Allen 1986; Pleiner 1993; Giardino et al. 2002), to be distinguished from the lighter slag formed during the smithing stages (McDonnell 1984; Vidale et al. 2004).

Figure 12.12 Fragments of finished or unfinished iron objects found on the surface of Site No. 172 together with the ironprocessing indicators. B is the point of a pick-like tool or weapon; e is a piece of a sickle blade.

Figure 12.13 Detail of the fragment of sickle blade drafted in Figure 12.12e.

203

M. Vidale, E. Battistella and G. Guida Some of the pieces of this type we observed, in fact, show evidence of surface flow and of slow cooling wrinkles, sometimes in superimposed pourings. A single slag piece looks like a small round plano-convex cake with what remains of a pouring flow at the edge, suggesting that the tapped slag might have flowed above a structural floor and gathered into a small round cavity. Empirically, such slag pieces feel heavy, and their iron content appeared much higher. In contrast, smithing slag in most iron-age contexts comes in form of shapeless drops, small slag pellets. Even more distinctive are small to medium-sized plano-convex cakes or cake fragments (whose formation process is still unclear) or “cake rings” hollow at the bottom, formed around the edge of the smithing pit above the extremity of the bellow. In Site No. 172, both forms were not clearly identified, and we have to suppose that most of the smithing slag cakes, to a large extent, had been turned into the mass of small iron slag crumbles recorded at almost every location but particularly at west.

furnace smelting process as well, and vitrified linings, in some cases, sometimes may be very ambiguous. Fortunately, this latter is not our case, and the vitrified linings found on surface can be positively related to iron smithing (Figure 12.12). They demonstrate that the forges or smithing hearths had a central concave pit about 15 cm wide and 10-12 cm deep, lined with silty-sandy slightly refractory sediments or plasters that judging on the base of the superimposed layers could be renovated two or three times. The clay of the linings is light grey-pale yellow (5 Y 7/2-7/3) in the parts less exposed to vitrification, and olive grey-dark olive grey (5Y 5/3, 3/2) in the syntherized surfaces. Some of the fragments are heavily vitrified, covered with a dark glassy slag rich of bubbles. In some specimens, where the heat was stronger and the slag thicker, it retained the negative imprint of a cylindrical object, about 3 cm wide, that probably marks the contact with the end of the tuyere. Presently there is no evidence of terracotta or stone tubes, and for the moment we may think that the tuyere itself was in iron. The tuyere entered the cavity of the fire pit from below, near the bottom or at midst height (Figure 12.12a-c) sometimes with a distinctive steep angle with the main axis of the pit (Figure 12.12a-b). This position of the tuyere probably explains the formation, immediately above the contact point, of a step-like, irregular iron slag formation that we will later discuss in greater detail (in the same Figure 12.12b and d).

Nonetheless, some slag pieces, with a maximum diameter of 5-6 cm, are flat on top and roughly cone-shaped below; they actually resemble very irregular plano-convex cakes, and although their structure is apparently identical to that of the tapping slag pieces we analyzed (see below) their form is compatible with the lower cavity of the smithing pits we reconstructed (Figure 12.11). Their interpretation remains questionable. On the other hand, light, bubble-rich iron slag crusts are visible, often as residues of peculiar, irregular “steps” on the interior of the concave plaster linings detached, we think, from the fire pit of the forges (Figures 12.10-11). These “slag steps” growing at midheight within the plastered smithing pits do represent a variation of the ring slag type.

Analytical Tests The analytical procedures were those normally followed for a preliminary investigation of this types of metallurgical residues. The pieces of slag (smelting and smithing) and some other iron objects were cut with a diamond rotating saw. One or both surfaces were polished with sand paper and diamond paste up to 6 mµ. The sections were observed with a metallographic microscope and then with a Scanning Electron Microscope (SEM), and the miscrostructures observed and described. Various slag samples and all the unfinished or finished iron objects we cut were subjected to X-ray fluorescence analysis to determine their elemental components (EDXRF- energy dispersive X-ray fluorescence) (qualitative analysis). Three fragments of tapping slag were cut, polished and analyzed with a SEM and its microprobe, detailing the inner microstructure of the slag and its chemical variability (thus obtaining a preliminary set of quantitative analysis). Finally, six fragments (including some sample of smelting and smithing slag, hammering scales and slag drops) were also finely powdered and analyzed by the means of Xray diffraction (XRD) for determining the newly formed crystalline phases on record.

It is well known that distinguishing bloomery and smithing slag is often difficult, particularly when interpretation is based upon archeometallurgical evidence alone. In most cases, at any rate, the most reliable conclusions come from a careful evaluation of the general archaeological context, when laboratory studies provide validating arguments. Although the record of iron working at Site No. 172 is a surface one, the general pattern of co-occurrence among the most important indicators is rather coherent and it is broadly confirmed, as we shall see, by the analytical tests we performed in Rome. As far as smithing activities are concerned, the small plano-convex slag cakes may be part of the general picture, but the indicators more generally accepted for this metallurgical stage are hammering scales (films of high-temperature iron oxide detached from the surface of the hot-hammered blooms and semi-finished objects during the forging stages), tiny slag drops and pellets (the syntherized silicatic fraction extracted by heavy hammering and fast cooling from the forged items), and concave, vitrified and slagged plaster linings (detached from the interior of the smithing pits or fireplaces at the end of a cycle of use, maintenance or destruction). Slag drops and pellets, nonetheless, may be a by-product of the bowl204

Iron-working and Ceramic Recycling Internal Structure of Smelting Slag, Slag Drops and Hammering Scales

temperature iron oxide usually found in protohistoric iron working residues) tightly packed within a silicatic matrix (the grey background). As is usually observed in these cases, the silicatic matrix, at a closer look, is formed by crystals having two basic shades of grey (in this cases, too, interpreted as iron (?)-calcium-potassium-aluminum silicates). The well developed iron oxide dendrites indicate a slow cooling of the thick slag flow.

Although the observed samples showed the same components (silicatic fractions and iron oxide dendrites), their inner structures is quite variable from point to point and often quite complex. Sample 4 (from AA 10), sample 5 (from AA 10) and sample 65, gathered on the surface of AA 9, were observed both at the metallographic and microscope and at SEM, and interpreted as a piece of a tapping slag. Sample 5 (Figures 12.13-14) is almost entirely formed by a two-component silicatic fraction. The lighter-colored phase appears in form of well-developed star-like dendrites with orthogonal arms on a darker phase background. In both cases the microprobe suggests that we are dealing with two different iron (?)-calcium-potassiumaluminum silicates. Near the edge of the slag piece, there are some light-coloured needle-like formations (Figure 12.15), possibly representing other dendrites of the same type visible in the core but cooled at an earlier stage and not completely formed. The almost total absence of iron oxide formations shows that in this case the slag tapping process had been carried out in a highly proficient way. In both sample 4 (Figures 12.16-17) and sample 65 (Figures 12.18-20) the section shows the well known orthogonal dendrites of iron oxide (light-coloured, preliminarily identified as wüstite – FeO - the most important high-

In sample 4, highly visible are small pellet-like, rounded formations almost entirely formed by wüstite branches. Other wüstite formations are visible outside the rounded particles, freely embedded within the silicatic matrix. The iron pellets are probably of the same type of those progressively packed, in the frame of the smelting process, into the blooms. In sample 65, the two light-gray and dark-gray silicatic fractions are equally in evidence, the former ones assuming the shape of well developed dendrites (Figures 12.19-20). Wüstite dendrites are interspersed within the silicatic matrix. In some areas they have irregular, arborescent patterns, while in other spots they have a clear orthogonal (cubic) spatial setting. The chemical composition of these slag samples is detailed in the next section. Sample 1 (Figure 12.21), taken from a slag drop, in contrast, shows the incomplete development of the same

Figure 12.14 Slag sample 5, SEM section showing star-like silicatic dendrites.

205

M. Vidale, E. Battistella and G. Guida

Figure 12.15 Slag sample 5, SEM section with silicatic phases and needle-like formations.

Figure 12.16 Slag sample 4, SEM section with rounded particles made of wüstite formations within silicatic phases.

206

Iron-working and Ceramic Recycling

Figure 12.17 Slag sample 4, SEM section with detail of wüstite formations embedded within the usual silicatic phases.

Figure 12.18 Slag sample 4, SEM section with wüstite formations in two silicatic phases.

207

M. Vidale, E. Battistella and G. Guida

Figure 12.19 Slag sample 65, SEM section with scarce wüstite formations within well crystallized silicatic phases.

Figure 12.20 Slag sample 65, SEM section with wüstite formations arranged according a cubic pattern within well crystallized silicatic phases.

208

Iron-working and Ceramic Recycling Quantitative Chemical Analysis In the following Table 12.2 we report the microprobe analysis of the general composition of the tapping slag, as well as the characterization of the two silicatic fractions (the light and the dark ones) and of the iron oxide formations. We counted eight elements (oxigen, magnesium, aluminum, silicon, phosphorus, potassium, calcium and iron). Carbon was not included, even if we cannot exclude that some carbonate was present. Phosphorus, in every analysis, is present as a minor trace. Values in Table 2 represents weight percentages normalized after the eight elements taken into account. “Id” gives the sample number, while the number after the slash (/) gives different analyzed spots in the same slag piece. We compare the general composition of samples 4 and 5 with the chemical characterization of the two silicatic phases (the dark-grey and the light-grey) and with the measurements taken into the wüstite branches.

Figure 12.21 Metallographic section of a slag drop showing poorly developed wüstite formations within a silicatic matrix. 50 X.

EDXRF Analysis We analyzed twenty four objects, from smelting slag fragments gathered at the western location to the main types of smithing indicators and finished or unfinished iron artefacts. Some samples were analyzed more than once, i.e. in various spots of the observed microstructures. In these cases, tests labelled b and following were performed on oxidized patches. Although EDXRF measures a relatively large area of the samples, it is less sensitive than SEMmicroprobe, and the results are obviously qualitative and not quantitative (Table 12.3). The iron smelting slag fragments we analyzed show the consistent association of iron and calcium (this latter was the second element in the SEM-microprobe tests of Table 12.2, if silicon and oxigen are not counted). Interestingly, in the slag drops the same association is less strong, and this suggests that calcium in the former type is not entirely due to post-depositional contamination from the soil. One of the tapping slag fragments had distinctive the peak of arsenic. The hammer scale, analyzed in three spots, turned out to be made of pure iron. In contrast, the possible unfinished objects we tested, besides minor traces of calcium and strontium, showed occasional traces of lead, copper and arsenic. These secondary contaminants, if they are not due to depositional environment, might be well explained by the forging, together with the iron products, of objects in lead and copper-bronze (a common find in Iron-Age smithing contexts).

Figure 12.22 Metallographic section of a hammering scale showing iron dendrites packed by the blows into polygonal grain structures within a silicatic matrix. 50 X.

dendrites of wüstite caused by a very fast cooling, when the slag fell on the ground or anyhow at a distance from the source of heat. The same basic structure (scarce and poorly developed wüstite dendrites within a silicatic matrix) was observed in other slag drops as well. We also sectioned and observed with the microscope four hammering scales. In these cases (Figure 12.22) wüstite dendrites are abundant, the silicatic matrix is scarce, and the wüstite crystals appear compacted under the smithing blows, less recognizable and partially transformed into polygonal structures. Preliminary mineralogical identifications of the silicates and oxides present in the slag samples were provided by the XRD tests (see below).

XRD Analysis We sampled, powdered and analyzed six samples: two fragments of tapping slag (four, from AA 10, and sixty five from AA 9); one piece of smithing slag detached from the

The iron objects we cut were completely corroded and had no surviving trace of the original metallic structures, nor, therefore, of the smithing techniques. 209

M. Vidale, E. Battistella and G. Guida Id/Spot 5/1 5/2 5/3 5/4 4/1 5/1 5/2 5/3 4/4 65/1 5/3 5/4 4/2 4/3 4/4 65/1 65/2 4/2 4/3 4/3 4/4 65/1 65/2

Description general general general general general dark phase dark phase dark phase dark phase dark phase light phase light phase light phase light phase light phase light phase light phase Fe dendrites Fe dendrites Fe dendrites Fe dendrites Fe dendrites Fe dendrites

O 39.32 38.52 38.11 41.92 33.91 40.49 39.27 37.58 36.33 37.52 38.00 41.99 36.38 34.07 37.89 35.99 36.22 33.83 36.39 35.35 35.00 31.72 32.27

Mg 1.79 1.67 1.74 1.88 1.42 1.93 2.33 1.83 1.09 0.99 1.45 2.00 1.61 1.35 1.47 1.55 1.98 1.24 0.99 0.85 1.09 1.08 0.79

Al 4.07 4.10 4.02 4.88 2.67 4.36 3.88 3.93 3.52 5.03 4.30 5.01 2.40 3.14 2.80 2.34 2.02 2.74 2.61 2.89 2.33 2.46 2.02

Si 20.47 20.41 20.53 24.89 8.52 22.0 21.12 19.45 9.16 15.78 20.59 24.99 11.51 9.49 9.87 14.13 13.83 8.65 8.52 7.66 9.38 9.86 9.18

P 0.23 0.22 0.21 0.23 0.08 0.23 0.16 0.21 0.19 0.51 0.28 0.27 0.31 0.26 0.31 0.21 0.32 0.10 0.18 0.22 0.17

K 1.92 1.81 1.98 2.32 0.75 2.27 1.33 1.52 1.55 3.04 2.18 2.45 1.26 0.92 1.16 1.12 0.92 0.96 1.02 1.07 1.27 1.16 1.00

Ca 10.33 10.21 10.58 10.83 7.56 11.37 12.28 10.19 10.24 6.75 8.70 10.85 12.09 7.71 8.99 8.46 3.53 8.27 6.56 7.93 8.89 4.26 2.34

Fe 21.86 23.04 22.82 13.06 45.08 17.33 19.64 25.30 37.93 30.38 24.50 12.44 34.43 43.05 37.87 36.09 41.31 44.53 43.60 44.15 41.46 49.23 51.88

Total 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00

Table 12.2 Microprobe analysis of the general composition of the tapping slag, with characterization of the two silicatic fractions (the light and the dark ones) and of the iron oxide formations. Eight elements (oxigen, magnesium, aluminum, silicon, phosphorus, potassium, calcium and iron) are counted. Values are weight percentages normalized after the eight elements. “Id” are the sample numbers; the number after the slash (/) refers to different analyzed spots in the same slag piece.

interior of a vitrified lining (fifty four, from AA 5); one hammering scale (from AA 1); three light, hollow slag drops (18 and 18a, both from AA 7).

Hammering scale: as expected, the XRD pattern, besides some quartz from the soil, shows the exclusive occurrence of wüstite.

Tapping slag samples: as often happens with this type of iron working indicator, both samples showed the cooccurrence of more mineral phases whose peaks are to a large extent superimposed and difficult to disentangle. In one of the samples we positively identified as a major component wüstite and probably kirschsteinite – CaFeSiO4 – a calcium iron silicate. Other possible matches might be magnesioferrite – MgFe2O4 – magnesium iron oxide, or as an alternative trevorite - (Ni, Fe) Fe2O4 – another iron (nickel) oxide. But we might be dealing with other compounds as well. In the other sample, very similar in shape to the former one, we detected wüstite, fayalite – Fe2SiO4 – iron silicate, and maghemite – γ-Fe2O3 – gamma iron oxide. Fayalite and kirschsteinite are two of the silicates to be searched within the composite gray glassy matrix of the slag.

Slag drops, probably formed after smithing operations: one of the samples showed the associaton of quartz, wüstite and kirschsteinite, thus suggesting that the drop formed and fell during a preliminary smithing stage. The other shows quartz, wüstite, another component possibly identified as rozenite -FeSO4.4H2O – iron sulfate hydrate, perhaps present as an alteration product, and traces of an artificial high temperature potassic feldspar. The hypothesis of a sharp difference between the ironrich, massive slag of the western AA and those common in the spots of the eastern plateau, in this light, in spite of the compositional variability we have observed, is fully confirmed. Iron Production at Site No. 172. Hypothetical Reconstruction

Slag from the interior of the smithing lining: the XRD pattern, as expected, is to a large extent amorphous, due to the presence of syntherized glassy phases. We saw the peak of quartz – Si02 – silicon oxide, and perhaps some peaks ascribed to fassaite – Ca(Mg,Fe, Al)(Si, Al)2O6 – a calcium magnesium iron aluminum silicate.

The evidence thus suggests that people brought into the fort some type of iron ore. This ore was processed within the compound. This is interesting, because at that time the metal ores were more commonly smelted near the extraction 210

Iron-working and Ceramic Recycling Sample Id 39a

Type iron smelting slag

Fe X

Ca X

Pb

39b

iron smelting slag

X

X

40a

iron smelting slag

X

X

40b

iron smelting slag

X

X

41a

iron smelting slag

X

X

41b

iron smelting slag

X

X

42

iron smelting slag

X

tr.

43a

iron smelting slag

X

X

43b

iron smelting slag

X

X

06

iron smelting slag

X

X

05

iron smelting slag

X

X

67

iron smelting slag

X

X

07

iron smelting slag

X

X

65

iron smelting slag

X

X

01

slag drop

X

tr.

02a

slag drop

X

02b

slag drop

X

X

03

slag drop

X

X

04a 04b

slag drop slag drop

X X

X X

04c

slag drop

X

X

05a

hammering scale

X

05b

hammering scale

X

05c

hammering scale

X

sl74

unfinished iron object

X

slb74

unfinished iron object

X

X

sl72

unfinished iron object

X

X

sl61

unfinished iron object

X

sl71

unfinished iron object

X

sl33

unfinished iron object

X

tr.

sl84

unfinished iron object

X

tr.

nn

sickle blade

X

tr.

sl 56

unfinished iron object

X

tr.

tr.

sl 75

unfinished iron object

X

X

tr.

Cu

As X X

tr.

tr.

tr.

tr.

tr.

tr.

tr.

tr.

tr.

tr.

Table 12.3 EDXRF analysis of twenty four objects, from smelting slag fragments to the main types of smithing indicators and finished or unfinished iron artefacts. Some samples were analyzed more than once, i.e. in various spots of the observed microstructures.

areas, in order to minimize the high transport costs. Presumably, the smelting stages were performed within the fort because of security reasons. As sulfur, according to our analytical results, is consistently absent from the Xray ED spectra, sulphides such as pyrite and chalchopyrite are unlikely candidates for the ore itself. The tapping slag samples, in contrast, at the microprobe showed among other elements peaks of aluminum, and this rather indicates that weathered deposits rich in oxides such limonite, goethite and the like could have been the locally exploited source of iron. The ore was smelted within some furnaces located

in the western part of the fortified settlement and carried out on the spot a preliminary hammering of the blooms thus obtained. We may imagine – admittedly, in absence of any material proof – that these furnaces were the truncated cone-shaped, small and tall infrastructures commonly used for this purpose in many pre-industrial contexts. AA 10, as stated above, may include the surface record of one of these furnaces completely collapsed and crumbled in its original location. At the base, these furnaces should have had an opening for tapping out the molten slag, while the blooms formed at a higher level. Possibly, the tapped slag 211

M. Vidale, E. Battistella and G. Guida

Figure 12.23 Hypothetical reconstruction with the location of iron smelting and smithing installations within the fort.

was gathered on prepared clay surfaces or cavities. The furnaces might have been erected in an open ground, and flanked by spaces where craftpersons could smash and grind the ore and the charcoal. Nearby there might have been smithing fireplaces for the preliminary reduction by hot hammering of the coarse iron blooms. We expect that the large amount of chaff-tempered clay fragments, variously affected by firing and visible on surface came from the destruction of similar furnaces. It seems that the forges, or part or them, rose in another part of the compound, along the inner face of the southern wall, as in the design of Figure 12.23 (where the crenellated walls follow the reconstructions commonly proposed for Central Asian palaces and forts, are purely conjectural and have no further implication).

distance from the fire pit, debouching at its bottom. But in our cases the tuyere entered the pit on the side of the plastered cavity, rather than from its lower bottom. This is why we have preliminary hypothesized a hearth with a raised bench and a central pit, with the bellows in a slightly lower position, and the tuyere sloping upward. Aligned along the southern wall, in front of the main inner building or “castle”, the forges could have conveniently dumped immediately beyond the wall large amounts of industrial debris. The present wide extent of these trash deposits reflects more the intensity of the action of the natural agents than iron working, but one gets, at any rate, the impression of a substantial amount of production. Little is known about the types of products manufactured on the fort. We even ignore if the iron artefacts, or a part of them, can be positively considered as examples of the local products. In this case, the sickle blade of Figure 12.13 might suggest the local manufacture of agricultural tools; the point and the fragments of laminar objects are more ambiguous.

In our reconstruction, the forge (Figures 12.24-25) is isolated and shaded by a light roofing, but this is only a guess. The remnants of these installations in contemporary South Asia may be commonly found within closed and sheltered spaces but, as a rule, near entrances that provided light and air. Usually Iron Age forges are reconstructed as open hearths with a central fire pit, although the contemporary iconography from classical Greece consistently depicts tall, chimney-like furnaces certainly used as forges (of which, curiously enough, we have no archaeological evidence). The imprint of the tuyeres in the linings of Site No. 172 might indicate a setting of the bellows similar to that commonly applied today by ironsmiths in South Asia: the bellows, built with skin and wood, is suspended with poles and rope and its nozzle enters the clay floor at some

Recycling of Pottery by Chipping. The Evidence of Grog Production Iron Age sites in this region, since the early 1st millennium BC, are studded with unbelievable numbers of recycled or modified potsherds. In most cases they appear as disks with trimmed or better knapped edges, ranging in diameter from 2 to 5-6 cm. These objects are so abundant that one, right or wrong, gets the impression of a continuous, uniform distribution across the surface of whole sites. 212

Iron-working and Ceramic Recycling

Figure 12.24 Hypothetical reconstruction of the type of smithing heart used in Site No. 172. The location of the bellows follows the angle with the inner pit visible in Figure 12.11.

Figure 12.25 Another hypothetical reconstruction of the smithing installations of Site No. 172.

213

M. Vidale, E. Battistella and G. Guida Generally speaking, similar disks are not uncommon in many archaeological cultures and contexts (even if in much lesser amounts). Nobody really knows their function, and usually are referred to as “counters”, “tokens”, “game pieces” and the like; or, sometimes, they are interpreted as tools for scraping or lids for small vessels. The surface of our site, too, was very rich in this enigmatic artefact (Figures 12.26-27, 30). Looking more carefully to the surface, Vidale observed many instances of potsherds whose edges had been systematically chipped, sometimes in a bifacial fashion. Sometimes, the pottery knappers had expediently used the peculiar shape of some potsherds: for example, a conical lug had been perimetrally flaked along the edges using the original inner surface of the vessel as striking platform; in another case, a potsherd belonging to the mouth of a restricted jar had been rhythmically and intensively flaked along the rim (Figure 12.29), the lip and its lower fracture surfaces. In another case, a thick rim was transversally broken and the fracture surface was used as a striking platform for detaching a series of ceramic blades from the round perimeter of the rim itself (Figure 12.28). The analogy with flint knapping in some cases seems undeniable, and might suggest that the pottery knappers had or maintained a general understanding of the principles of flint working.

Figure 12.26 Site No. 172, surface. Small pottery disks found on the surface of the site; some have polished edges. This type of wear on the edge is rare and limited to the smallest specimens of this type.

Inspecting the more common ceramic disks, we found that they probably represent the residual cores from which

Figure 12.27 Site No. 172, surface. A pottery disk obtained by the means of a complete chipping of the edges, most probably for the production of grog. Figure 12.28 Site No. 172, surface. A pottery “blade core” obtained by chipping a thick rim of a large jar (cfr Figure 31. 3).

214

Iron-working and Ceramic Recycling

Figure 12.29 Site No. 172, surface. Another jar rim systematically flaked on the rim and the fracture edge, in bi-lateral fashion.

Figure 12.31 Site No. 172, surface. A sherd of a coarse vessel evidently tempered with large, unmodified pottery flakes (from AA 6).

standardized, flat pottery flakes were serially detached. In a random sample of about a hundred disks from the surface of Site No. 172, we found that only a very small percentage of the ceramic disks (2-3% of the total) had wear traces on the edges to support the hypothesis of scraping tools, and they were the smallest ones (some of them appear in Figure 12.26). Clearly, the greatest majority of the disks had been used simply as cores, to be discarded when exhausted, i.e. when the solid geometry of the core did not allow any more a satisfactory handling. And obviously there was no scarcity of raw material. What could be the purpose of this mysterious activity? The only possible explanation is ceramic recycling for the production of grog temper. This component of ancient ceramic raw materials is notoriously difficult to recognize, even in thin sections, particularly when the recycled pottery is finely ground. But in our case there is at least one category of vessels that for some reasons was definitely made with a clay tempered with large amounts of flat pottery flakes knapped from older potsherds. For some reasons, this type of vessel required large flat flakes and not ground pottery powders (Figure 12.31). These vessels are coarse vessels or dishes with a cylinder-like body, whose complete form is still unknown. On the surface of Site No. 172, they were found in AA 6 and 10, and they might well have had some kind of craft function. Figure 12.31 shows potsherds of these coarse containers; one can easily see, in the breaks, how flat pottery flakes had been used as a temper component without further reduction. At present, there is no way to assess the precise chronology of this peculiar phase of occupation. It could belong to a late moment of the Yaz III period or even to a later time; also, it is really impossible to ascribe it to a resident population or, on the contrary, to possible temporary occupations by nomads, strictly aimed at exploiting this residual technical resource.

Figure 12.30 Site No. 172, surface. Various “pottery cores” exhausted after intensive flaking for the production of grog.

215

M. Vidale, E. Battistella and G. Guida

Figure 12.33 Buff ware from the surface of Site No. 172, Yaz III period. Restricted forms.

the greatest majority of the surface sample (Figures 12.3234); 2- Coarser chaff-tempered ceramics, hand-made (sometimes with coils). These might belong to containers having specialized craft functions (crucibles?), or they might be components of larger craft infrastructures such as fireplaces or metallurgical ovens (Figure 12.35a-b); 3The above described grog-tempered cylinder-like vessels, a very specialized type of unknown function (Figures 12.31 and 12.35c-g). Buff Ware Figure 12.32 Buff ware from the surface of Site No. 172, Yaz III period. Open forms.

All the potsherds in this class are fine wares, without visible inclusions, formed or shaped on the potter’s wheel by highly competent craftspeople (small vessels being directly thrown on the wheel, larger containers made with coils and later fashioned on the same wheel). They were fired at medium-high temperature in uniform and well controlled atmospheres. The most common colours range from very pale brown (10 YR 8/3) to light yellowish brown (10 YR 6/4). The outer surfaces sometimes develop a lighter colour, most probably due to the use of slightly salinized clays. When vessels were fired in reduced atmospheres and at relatively high temperatures, the pottery became greenish-brown to pale yellow (5Y

Pottery Collected on Surface The surface of Site No. 172 was not very rich in wellpreserved potsherds. Vidale selected, drafted and briefly described the best specimens, as a contribution to the growing data base on the ceramic sequence of the delta and for the purpose of chronology. The sampled forms belong to three distinctive empirical groups: 1- a local buff ware including containers of everyday use, amounting to 216

Iron-working and Ceramic Recycling

Figure 12.34 Buff ware from the surface of Site No. 172, Yaz III period. Bowls.

Figure 12.35 Coarse wares from the surface of Site No. 172, Yaz III period (?). a-b. chaff-tempered ware; c-g. grog-tempered wares.

7/2-7/3). When strongly oxidized, the vessels assumed a reddish-yellow colour (7.5 YR 6/6). lower ridge. 10148. Fragment of a hemispherical bowl (?) with a rim thickened on the exterior surface, slightly projecting outward, underlined by a thick, rounded lower ridge. The upper and lower part of this rim is marked by a deep horizontal groove. 10074. Fragment of a truncatedcone shaped bowl with a slightly convex upper profile, and a rim thickened on the exterior surface; a horizontal ridge runs at the base of the rim. 10077. Fragment of a hemispherical bowl (?) with a rim thickened on the exterior surface, symmetrical, and a flat, slightly convex upper surface. 10118. Fragment of a truncated-cone or hemispherical bowl, having a slightly thickened rim, rounded outside, with a slight saddle-shaped depression on top. 10039. Fragment of a truncated-cone bowl with a straight upper wall, having a slightly thickened rim, rounded outside, and a slightly pointed extremity on the interior, at the point of horizontal tangency. 10075. Fragment of a truncatedcone bowl with a slightly S-shaped upper wall, having a slightly thickened rim, rounded outside. 10040. Fragment of a hemispherical bowl, having a slightly thickened rim,

The short description we enclose starts from open forms and move to progressively restricted ones; at the same time, it proceeds from the simple contours and ends with the more complex shapes (complexity being an empirical function of the number of angle or inflexion points and the degree of elaboration at the rim). Thus, we begin with simple open bowls and end with necked jars. Each design is accompanied by basic information and short comments on the basic geometry of the potsherd and the original form of the vessel, on the shape of the rim and other attributes. The numbers that identify each find are those of the general inventory of the Italian Archaeological Mission in Turkmenistan. Figure 12.32: 10149: Fragment of a truncated-cone or hemispherical bowl with a slightly S-shaped upper wall, having a rim thickened on the exterior surface. 10125: Fragment of a truncated-cone or hemispherical bowl with a slightly S-shaped upper wall, having a rim strongly thickened and rounded, symmetrical, underlined by a sharp 217

M. Vidale, E. Battistella and G. Guida projecting outside and inside, and an upper sloping surface with a slight saddle-shaped depression.

flat, outward projecting rim having a sub-triangular section. 10153b. Fragment of a small jar with a short concave neck and a simple rim projecting outwards, slightly thickened on the interior surface. 10151. Fragment of a restricted jar with a short concave neck and a slightly thickened, symmetrical rim, projecting outside.

Figure 12.33: 10041. Fragment of a coarse globular jar, pot or bowl with a rounded shoulder and a simple everted rim. A horizontal ridge runs below the rim. 10047. Fragment of a hemispherical or sub-globular bowl with a slightly thinned rim, underlined by a horizontal groove. 10078. Fragment of a restricted jar or pot with a hemispherical or sub-globular upper part. The rim is rounded and projects outwards, The shoulder surface is fluted with two parallel, sharp ridges divided by a large convex groove. 10082. Fragment of a restricted jar or pot with a rounded, outwards projecting rim. On the shoulder runs a horizontal sharp ridge. 10083. Fragment of a restricted jar with a outwards projecting rim, flat on top, having a sub-triangular section. On the shoulder runs a broad horizontal ridge, sub-triangular in section too. 10076. Fragment of a restricted jar or pot with a rounded, outward projecting rim. 10075. Fragment of a restricted jar with a straight oblique shoulder, and a thickened, outward projecting rim ending into a pointed lip. 10145. Fragment of a restricted jar or pot with a thick, rounded, outward projecting rim. The potsherd retains at the lower extremity part of a sharp horizontal ridge. 10146. Fragment of a restricted jar or pot with a thick, outward projecting rim having a sub‑triangular section, ending into a slightly pointed lip. The upper surface is convex. 10150. Fragment of a restricted jar or large pot with a thick, rounded, outward projecting rim. 10124. Fragment of a coarse globular jar, pot or bowl with a simple rim, smoothed at the corners. Most probably built with large coils. 10156. Fragment of a restricted jar with a hemispherical or sub-globular upper part. The rim is thickened outside, with a sub-rectangular smooth section. The upper part of the rim has a strong wear trace caused by the friction of a pottery or wooden lid. 10157. Fragment of a restricted jar with an hemispherical or sub-globular upper part. On the shoulder there is a wide flat ridge. The rim has a sub-triangular smooth section, and a slight saddle on top. 10155. Fragment of a small jar or bowl with a hemispherical or sub-globular upper part and a thick rim, rounded on the interior surface and strongly projecting outwards, with a sub-triangular smooth section. 10154. Fragment of a small jar or bowl, with a rim thickened on the exterior and having a sub triangular section. 10181. Fragment of a restricted jar, pot or bowl with a hemispherical or sub-globular upper part. The rim is thickened outside, flat, with a sub-rectangular smooth section. 10153. Fragment of a sub-globular jar, pot or bowl with a pointed, outward projecting rim. 10046. Fragment of a coarser restricted jar with rounded shoulder and a slightly thinned, outward projecting rim. Most probably built with coils. 10116. Fragment of a coarse restricted jar with rounded shoulder and a slightly thinned, outward projecting rim. Most probably built with coils. 10117. Fragment of a jar with a concave neck and a rounded, downward projecting rim. 10045. Fragment of a small jar with a short concave neck and a simple rounded rim. 10122. Fragment of a jar with a sub-cylindrical neck and a

Figure 12.34: 10054. Fragment of a small hemispherical bowl with a simple pointed rim. 100123. Fragment of a small hemispherical bowl with an upper straight wall and a simple pointed rim. 10048. Fragment of a small hemispherical bowl or pot, with a rim slightly thickened on the exterior surface and a sub-triangular smooth section. 10050. Fragment of a small hemispherical bowl with a simple pointed rim and a horizontal ridge near the point of maximum expansion. 10049. Fragment of a small hemispherical-biconical bowl with a simple pointed rim and a horizontal ridge near the point of maximum expansion, marking also the corner point of its carinated contour. 10080. Fragment of a small hemispherical bowl with a simple pointed rim. 10053. Fragment of a small hemispherical bowl with a simple pointed rim. 10044. Fragment of a small hemispherical bowl with a simple pointed rim and a double horizontal ridge near the point of maximum expansion. 10079. Fragment of a small hemispherical bowl with a simple pointed rim and a horizontal ridge near the point of maximum expansion. 11151. Fragment of a small hemispherical bowl with a simple pointed rim and a horizontal ridge near the point of maximum expansion. These buff ware potsherds may be compared with previously published materials of Yaz-related culture complexes, particularly with the Yaz II and III horizons. In detail, our fragments Figure 12.33: 10078, 10082, 10083 (restricted jars or pots with rounded everted rim and fluted shoulder) resemble some Yaz II potsherds published in Cattani and Genito 1998 (Tabs. 6, 9-11, even if in these specimens have sub-triangular rims, rather than rounded; see also a Yaz III jar Ibidem: Tab. 11, 20). Fragment 10083 may also be compared with a specimen from the Iron age levels of Mundigak (Casal 1961: Fig. 121, 643). All the small hemispherical restricted bowls with pointed rim and horizontal ridges near the point of maximum expansion from Site No. 172 (Figure 12.34: 10123, 10044, 10048, 10049, 10050, 10054, 10079, 10080, 11151) are comparable to Yaz III bowls found in other sites of the delta (Cattani and Genito 1998:1140; Tab. 8, 2, and Tab. 11, 4, but in this case the bottom of the bowl is probably different) as well as to some Mundigak specimens (Casal 1961: Fig. 124, 680). The truncated cone-shaped bowls from Site No. 172 have a generic resemblance with similar vessels from the Yaz III and particularly Late Yaz III sites surveyed in the delta and published by Cattani and Genito (1998) in Tabs. 7, 10 and 11. In particular, compare our sherds Figure 12.32: 10040 with their Tab. 7, 5 and our Figure 12.32: 10149, 10125, 10148, 10074 with Tabs. 11, 1 and 2; or necked jars Figure 12.33: 10117, 10045, 10122 with Tabs. 11, 8-11. Always 218

Iron-working and Ceramic Recycling Francovich R. (ed.) 1993. Archeologia delle Attività Estrattive e Metallurgiche. Firenze. Fontana, M. V. and B. Genito (eds.) 2003. Studi in Onore di Umberto Scerrato per il suo settantacinquesimo compleanno. Napoli. Genito, B. 1998a. “The ������������������������������������� Iron Age in the Merv Oasis”, in Gubaev, Koshelenko and Tosi (eds.) 1998: 89-95. Genito, B. 1998b. “The ����������������������������������� Achaemenids in the History of Central Asia”, in Gubaev, Koshelenko and Tosi (eds.) 1998: 149-158. Genito, B. 1998c. “Trial-Trench ����������������������������������� at Site No. 215”, in Gubaev, Koshelenko and Tosi (eds.) 1998: 125-135. Genito, B. 2003. “Al di là dei confini degli imperi iranici e oltre…”, in Fontana and Genito (eds.) 2003: 403-429. Genito, B., R. Castelli and P. Mozzi 1998. “Trial ����������������� Trench at Site No. 215”, in Gubaev, Koshelenko and Tosi (eds.) 1998: 125-135. Giardino, C., G. Guida, A. Migliarelli and M. Vidale 2002. “La documentazione archeologica dell’antica lavorazione del ferro”, in D’Amico (ed.) 2002: 639656. Gubaev, A., G. A. Koshelenko and M. Tosi (eds.) 1998. The Archaeological Map of the Murgab Delta, Preliminary Reports 1990-95 Personal Memoirs. Series Minor, III. IsIAO, Rome. Kohl, Ph. L. (ed.) 1981. The Bronze Age Civilization of Central Asia. Recent Soviet Discoveries. New York. Masson, V. M. 1959. Drevnezemledel’ceskaja kul’tura Margiany. (����������������������������������������� Materialy i issledovaniya po arkheologii SSSR, vol. 73). Moskva-Leningrad. McDonnel, J. G. 1984. “The Study of Early Smithing Residues”, in Scott and Cleere (eds.) 1984. ��������� Belfast: 47-52. Panaino, A. and A. Piras (eds.) 2006. Proceedings of the 5th Conference of the Societas Iranologica Europea held in Ravenna, 6-11 October 2003. ������� Milano. Pleiner, R. 1993. �������������������������������������� “The Technology of Iron Making in the Bloomery Period. A Brief Survey of the Archaeological Evidence”, in Francovich (ed.) 1993����������� : 533-560. Scott, B. G. and H. F. Cleere (eds.) 1984. The Craft of the Blacksmith. ������� Belfast Vidale, M. 2006. “Technology ���������������������������������� and Decoration of Jaz I Painted Buff Ware Pots as Observed at Site M-999 (Murgab Delta, Turkmenistan)”, in Panaino and Piras (eds.) 2006����������� : 293-303. Vidale, M., P. Bianchetti and M. Cattani 1998. “A Semiprecious Stone-working Area Dating to the Late Iron Age on the Surface of Takhirbai-depe” in Gubaev, Koshelenko and Tosi (eds.) 1998: 105-113. Vidale, M., P. Bianchetti and G. Guida 2004. “Forge Nepalesi contemporanee. Livelli di specializzazione e processi formativi delle scorie di forgiatura”, in Barogi and Lugli (eds.) 2004�������� : 75-89.

to Yaz III point our fragments of restricted sub-globular jars or pots 10156 and 10081, very similar to a specimen in Genito et al. 1998: Fig. 4, 8780, from the nearby Site No. 215. In summary, the pottery from Site No. 172 indicates a Yaz III and Late Yaz III context, and preliminary suggest a dating within the 4th-3rd centuries BC. Coarse Chaff-tempered Containers and Walls Figure 12.35: a. A fragment of a restricted jar with subcylindrical neck, distinguished by a rough pointed rim; b. A fragment of a truncated-cone shaped bowl (?), formed by pressing the coarse clay into a mould, a process often reserved, in the protohistory of South Asia, to the fabrication of ceramic lids. The rim is modified by a series of plastic impressions. The wall retains a hole made immediately after moulding. Coarse Grog-tempered Ware Figure 12.35: c. Fragment of vessel or dish having a sub-cylindrical body contour, slightly restricted, with a simple rim flat on top. Hand-built with coils. d. Fragment of vessel or dish having a sub-cylindrical body contour, slightly restricted, with a simple rim flat on top. Hand-built with coils. e. Fragment of base of a vessel or dish having a sub-cylindrical body contour, hand-built with coils. f. Fragment of base of a vessel or dish having a sub-cylindrical body contour, hand-built with coils. g. Hypothetical reconstruction, after non-refitting potsherds, of the general shape of this type of grog-tempered vessels. References Cited Allen, J. 1986. �������������������������������������� “Interpretation of Some Roman-British Smithing Slag from Ayre in Gloucerstershire”, Journal of the Historical Metallurgical Society, 20/2: 97-104. Bachmann, H. G. 1982. The Identification of Slag from Archaeological Sites. (Occasional Papers, 6, Institute of Archaeology, University of London). London. ������� Barogi, M. and F. Lugli (eds.) 2° Convegno Nazionale di Etnoarcheologia. Atti del Convegno. ��������� Mondaino: Casal, J.-M. (1961). Fouilles de Mundigak, 2 vols. Paris. Cattani, M. and B. Genito 1998. ��������������������������� “The Pottery Chronological Seriation of the Murgab Delta from the End of the Bronze Age to the Achaemenid Period: a Preliminary Note”, in Gubaev, Koshelenko and Tosi (eds.) 1998: 75-87. D’Amico, C. (ed.) 2002. Atti del II Congresso Nazionale di Archeometria, Bologna, 29 Gennaio-1 Febbraio 2002. Bologna

219

Chapter 13 An Egyptian Vessel at Site No. 203 Sabina Malgora

Introduction One fragment of a basalt vessel with serpentinite inclusions (Inventory No. 826), Site No. 203. The study of this fragment (Figures 13.1-3) gives rise to problems and obstacles that are not easy to overcome. Stone Vessel They are one of the most frequent elements in Egyptian grave goods inventories. The most ancient examples were found in Lower Egypt, near the site of Merimde Beni-Salame, and are datable to the beginning of the period called “Merimde” (4800-3500 BC). During the subsequent Predynastic Period, technique improved and production increased sharply; during the Late Predynastic, particularly in Upper Egypt, there was such an increase in the production of stone vessels as to almost exceed that of clay vessels. This reflected a change in the trend of the demand of the élite consumers who required luxurious, exotic products for their funerary outfits. The stone vessel production technique attained very high standards during the First Dynastic Period (approximately 3000-2700 BC) and also during the Old Kingdom (2700-2200 BC), when productivity was extremely high (the corridors of the Step Pyramid of Djoser at Saqqara yielded about 10.000 stone vessels). After the Ancient Kingdom, this production decreased dramatically.

Figure 13.1 Fragment from a basalt vessel.

Figure 13.2 Drawing of the fragment.

Basalt The use of basalt was intensive during the Predynastic Period (Table 13.1). This material was utilised during the First Dynastic Period, and was still being utilised, though rarely, during the Third and Fourth Dynasties. Subsequently, during the Fifth and Sixth Dynasties, its use was limited to small vessel models included in “sets” that contained the miniature versions of the implements required for the mouth-opening ritual. After the Old Kingdom, examples of basalt vessels are quite rare: we should mention two pieces preserved at the British Museum: a vessel fragment with the inscription “Tuthmosis I” (BM 43405), whose origin is unknown, and a mortar (BM 32181), which has been dated, on the basis of its shape, to a period subsequent to the New Kingdom. Figure 13.3 Graphic reconstruction of the vessel shape.

221

222

4000 B.C.

3500 B.C.

Naqada II

Naqada III

3000 B.C.

Early Dynastic

2700 B.C.

Old Kingdom

I Inter.

2100 B.C.

Middle Kingdom

Table 13.1 Use of materials in Egypt (Aston 1994; Lilyquist 1995; Nicholson and Shaw 2000).

Alabaster Amethyst Amphibolite Andesite porphyry Anhydrite Anorthosite gneiss Basalt Carnelian Chert Diorite gneiss Dolomite Gabbro Granite Granodiorite Hematite Hornblende diorite Lapis lazuli Limestone Limestone breccia Malachite Marble Mica schist Metasiltstone Obsidian Quartz crystal Rose quartz Travertine Tuff Serpentine Siliceous sandstone Steatite

Naqada I

II Inter

1550 B.C.

New Kingdom

1070 B.C.

III Inter

712 B.C.

Late Period

GrecoRoman

395 B.C.

S. Malgora

An Egyptian Vessel at Site No. 203 The Egyptian word for “basalt” is unknown. Basalt outcrops are present from Abu Za’bal, north of Cairo, down to the latitude of Assiut, to the south, and from a location west of the oasis of Baharya to the Gabala Plateaus. The most important deposits, however, are near the Nile: on the eastern bank in front of Samalut; east of Cairo towards Suez, in a scattered fashion; near Abu Za’bal; and along an arc starting from a location west of the plateau of Abu Roash towards south-east, up to Gebel Qatrani, north of the Fayum.

It is an open-shape, large bowl. AP (aperture diameter): 24.3 cm. MBD (maximum body diameter): 25.9 cm. H (height): estimated at 18 cm. VI (Vessel Index = MBD/H x 100)= ? Rim : ?

Literature

Bottom: not preserved.

In Egyptological literature there does not exist a book that comprehensively classifies all forms of stone vessels.

In any case, the identification of the shape of this vessel, obtained by means of comparisons, is approximate; only a series of analyses of the material will be able to confirm or refute its origin and identification.

An excellent work is Ancient Egyptian Stone Vases: Matrials and Forms, by (Aston 1994), to which I will refer in this text.

The shape, on the basis of the part that has been preserved, does not seem to be comparable with those of vessels dating from the Third Intermediate Period or Late Period, i.e. from the period to which Site No. 203 has been ascribed.

There do exist some works that deal with individual types of vessels, covering a time span that includes the Predynastic Period and the Old Kingdom, but, surprisingly, there are no studies of the types of stone vessels dating from a period after the Old Kingdom. There are some corpora where the artefacts found in a grave or excavation are included, but where the vessels are not organised or discussed in terms of shapes and relative classifications.

The shape may correspond to No. 104 in Aston’s classification, which it resembles. The dating of this shape is Dynasties 1-3 (3050-2575 BC). References Cited

Our fragment, unfortunately, does not have any inscription (an important element in the analysis of this type of artefact), so we are compelled to date it on the basis of its shape and material.

Aston, B. G. 1994. Ancient Egyptian Stone Vases: Materials and Forms, Heidelberg. Lilyquist, P. 1995. Egyptian Stone Vessels Khian through Tuthmosis IV, New York. Nicholson, P. T. and L. Shaw, 2000. Ancient Egyptian Materials and Techniques. Cambridge-New York.

Shape: it is not possible to determine whether the artefact had handles or a spout (these elements are useful in the determination of the shape).

223

BAR S1806 2008

Centro di Ricerca per le Tecnologie Multimediali Applicate all’Archeologia – Università di Bologna

THE BRONZE AGE AND EARLY IRON AGE IN THE MARGIANA LOWLANDS

Dipartimento di Archeologia – Università di Bologna

SALVATORI & TOSI (Eds)

Istituto Italiano per l’Africa e l’Oriente

B A R

The Archaeological Map of the Murghab Delta Studies and Reports

Series Editors: Annageldy Gubaev, Gennady A. Koshelenko and Maurizio Tosi

Volume II

The Bronze Age and Early Iron Age in the Margiana Lowlands Facts and methodological proposals for a redefinition of the research strategies Edited by

Sandro Salvatori and Maurizio Tosi with the editorial collaboration of

Barbara Cerasetti

BAR International Series 1806 2008