Iran 07 (1969)

IRAN

Journal of the British Institute of Persian Studies

VOLUME VII 1969

CONTENTS

Page Governing Council .. . . . . vi Director's Report . . . . . . . . . vii Excavations at Tepe Nfish-i Jan, 1967, by David Stronach . . . I La Survie de Shildu et la Route du Khunj-6-Fdl, by Jean Aubin. . 21 Excavations at Siraf: Second Interim Report, by David Whitehouse . 39 The Drhibgird Relief-Ardashir or Shdhpfir ? A discussion in the context

of early Sasanian sculpture, by Georgina Herrmann . . . 63 The Beliefs and Practices of the Ahl-i Haqq of Iraq, by C. J. Edmonds . 89 The

T.hirids and Persian Literature, by C. E. Bosworth . . . 103

Kirman and the Middle East: Paul Ward English's City and Village in Iran: Settlement and Economy in the Kirman Basin, by Brian Spooner and Philip C. Salzman . . . . . 07

Excavations at Biba Jan 1967: Second Preliminary Report, by Clare Goff Meade . .. .

. . . I 15 Prehistoric Copper and Bronze Metallurgy in Western Iran (with special

reference to Lilristin), by P. R. S. Moorey . . . . 131 "Some Ancient Metal Belts"-a Retraction and a Cautionary Note, by

P. R. S. Moorey . . . . . . . . 155 Hinweise und Anmerkungen zu einigen sasanidischen Monumenten, bo

Klaus Schippmann . . . . . . . . . 157 Further Notes on the Shaft-hole Pick-axe from Khurib Makran, by

C. C. Lamberg-Karlovsky . . . . 163 Survey of Excavations in Iran, 1967-68 . . . . . 169

Published annually by

THE BRITISH INSTITUTE OF PERSIAN STUDIES

c/o The British Academy, Burlington House, Piccadilly, London, W.I

Price: ?2 IOs. od.

STATEMENT OF AIMS AND ACTIVITIES

i. The Institute has an establishment in Tehran at which British scholars, men of learning versed in the arts, friends of Iran, may reside and meet their Iranian colleagues in order to discuss with them subjects of common interest; the arts, archaeology, history, literature, linguistics, religion, philosophy and cognate subjects.

2. The Institute provides accommodation for senior scholars and for teachers at British Universities in order that they may refresh themselves at the source of knowledge from which their teaching derives. The same service is being rendered to younger students who show promise of developing interests in Persian studies.

3. The Institute, whilst concerned with Persian culture in the widest sense, is particularly concerned with the development of archaeological techniques, and seeks the co-operation of Iranian scholars and students in applying current methods to the resolution of archaeological-and historical problems.

4. Archaeological excavation using modern scientific techniques as ancillary aids is one of the Institute's primary tasks. These activities, which entail a fresh appraisal of previous discoveries, have already yielded new historical, architectural, and archaeological evidence which is adding to our knowledge of the past and of its bearing on the modern world.

5. In pursuit of all the activities mentioned in the preceding paragraphs the Institute is gradually adding to its library, is collecting learned periodicals, and is publishing a journal, Iran, which is expected to appear annually. The Institute aims at editing and translating a series of Persian texts, the first of which, the Humay-Nama, edited by Professor A. J. Arberry, has already appeared.

6. The Institute arranges occasional seminars, lectures and conferences and enlists the help of distinguished scholars for this purpose. It will also aim at arranging small exhibitions with the object of demonstrating the importance of Persian culture and its attraction for the world of scholarship.

7. The Institute endeavours to collaborate with universities and educational institutions in Iran by all the means at its disposal and, when consulted, assists Iranian scholars with technical advice for directing them towards the appropriate channels in British universities.

MEMBERSHIP OF THE INSTITUTE

Anyone wishing to join the Institute should write to the Honorary Secretary, J. E. F. Gueritz, Esq., M.A., 85 Queen's Road, Richmond, Surrey. The annual subscription for Membership of the Institute is ?I, while the total sum of ?2 los. od. entitles the subscriber to receive the Journal. Application Forms at back of Journal.

The Cambridge History of

IRAN Editorial Board: A. J. ARBERRY, SIR HAROLD BAILEY, J. A. BOYLE, BASIL GRAY, A. K. S. LAMBTON, L. LOCKHART and P. W. AVERY Editorial Secretary For ten years scholars all over the world have been collaborating in the production of what will be the standard history of Iran: an eight-volume survey of Iranian history and culture, and of Iran's contribution to the civilization of the world. The aim has been to provide a collection of readable essays combining new ideas with established facts, rather than a catalogue of information. The essays will cover the religious, philosophical, political, economic, scientific, and artistic elements in Iranian civilization. These volumes will be the first detailed studies of Iran to be published in English.

Volume I: The Land of Iran Edited by W. B. FISHER Professor of Geography, University of Durham

A background study for the whole series. This volume examines the physical environment of Iran: its geography, geology, anthropology, economic life, flora and fauna. It reflects a highly significant trend in modern geographical studies by stressing the influence of this environment on the history and culture of the Iranian peoples. 804 pages, 26 photographs, 151 maps and line drawings

75s. net

U.K. Prices

Volume V: The Saljuq and

Mongol Periods Edited by J. A. BOYLE Professor of Persian Studies, University of Manchester

A survey of the civilizations which flourished in the Iranian region between the eleventh and thirteenth centuries: the rise of the Saljuqs, their defeat by the Mongols, and the subsequent Mongol rule until the emergence of the Timurids. In addition to general history, individual chapters discuss institutions, religious movements, literature, the visual arts, and the exact sciences. 778 pages, 24 photographs, io maps and line drawings

75s. net

Other volumes in the series, in preparation: Volume II Prehistory; the Archaeological Periods; Iran in relation to the Ancient World. Volume III The Seleucid, Parthian and Sasanid Periods. Volume IV The Period from the Arab Invasion to the Saljuqs.

Volume VI The Timurid and Safavid Periods. Volume VII The Eighteenth Century to the Present. Volume VIII Bibliography; Notes on Folklore; Survey of Research; Indices.

CAMBRIDGE UNIVERSITY PRESS

Muslim Intellectual A Study of A/-Ghazali W. M. WATT 224 pp, published, 30s.

The Ghaznavids Their Empire in Afghanistan and Eastern Iran 994-1040 AD C. E. BOSWORTH 330 pp, reprinting, 63s.

Lata'if al-ma'~rif of Tha'5libi The Book of Curious and Entertaining Information C. E. BOSWORTH 164 pp, published, 70s.

The Arabs and the Sudan YUSUF FADL HASAN 300 pp, published, 45s.

Islam and the West NORMAN DANIEL 450pp, published, 45s.

Islam, Europe and Empire NORMAN DANIEL 620 pp, published, 70s.

Islam and Imperialism in Senegal 1847-1914 MARTIN KLEIN 304 pp, published, 50s.

Early Greek Armour and Weapons A. M. SNODGRASS 300 pp, 37 plates, published, 70s.

Ancient Europe A Survey STUART PIGGOTT 365 pp, 84 plates, published, 50s.

The Iron Age in Northern Britain A. L. F. RIVET Editor 165 pp, 14 pages plates, published, 42s.

The Islamic Surveys Series General Editor Professor W. M.Watt

1. Islamic Philosophy and Theology W. M. WATT 219 pp, published, 25s.

2. A History of Islamic Law NOEL COULSON 250 pp, published, 25s.

3. Counsels in Contemporary Islam KENNETH CRAGG 220 pp, published, 25s.

4. A History of Islamic Spain W. M. WATT & P. CACH IA 220 pp, published, 25s.

5. The Islamic Dynasties C. E. BOSWORTH 264 pp, published, 25s.

6. Islamic Political Thought W. M. WATT 200 pp, published, 25s.

7. Religion and Culture in Islamic India and Pakistan AZIZ AH MAD February 1969 25s.

Kin burgA University Press, 22 George Square Edinburgh 8

1969

ARCHAEOLOGICAL TOURS of

IRAN visiting

TEHERAN . ABADAN . SUSA . CHOGA ZANBIL . HAFT TEPE . SHIRAZ PASARGADAE. NAQSH-I-RAJAB. PERSEPOLIS . NAQSH-1-RUSTAM . ISFAHAN RAMSAR . QASVIN . HAMADAN . BISITUN . KERMANSHAH . TAQ-I-BOSTAN

Departures from New York March 25 - April 8 - April 22 - May 6 - September 23 - October 7

GUEST LECTURERS Mr T. S. R. Boase, Hon.D.C.L. (Oxford), Hon.LL.D. (St. Andrews, Melbourne and Rockefeller Institute), Hon.D.Litt.(Durham and Reading), who will accompany Tour No. 11, is a Fellow of the British Academy and from 1947 until 1968 was President of Magdalen College, Oxford. Previously he was Director of the Courtauld Institute of Art and Professor of the History of Art in the University of London; he has travelled over most of the Near East. His publications include the Oxford History of English Art, vols. II and X, and Castles and Churches of the Crusading Kingdom.

Mr John Burton-Page, M.A.(Oxon.), F.S.A., who will accompany Tour No. 9, is Reader in Hindi in the University of London where he is responsible for the teaching of Indian archaeology; he has specialized in the study of Indian Islamic archaeology with reference to the Persian sources and connections. He is Director of the Royal Asiatic Society and a member of the Governing Council of the British Institute of Persian Studies. Has contributed extensively to learned journals and encyclopedias.

Miss Clare Goff, Ph.D., who will accompany Tour No. 7, has done much archaeological fieldwork in Iran and has a wide knowledge of the country. She worked with the British Institute of Persian Studies on the site of the ancient Persian capital of Pasargadae, and is now engaged upon the problem of the famous Luristan culture in central Iran. She has published papers in learned journals.

Miss Margaret Munn-Rankin, M.A.(Oxon.), F.S.A., who will accompany Tour No. 10, is a Lecturer at Cambridge University in the history and archaeology of the ancient Near East, a Fellow of Newnham College, Cambridge, and a member of the Councils of the British School of Archaeology in Iraq and the British Institute of Archaeology at Ankara. She has excavated in Iraq, Syria, Turkey, and Cyprus; has travelled widely in Iran and, amongst other publications, has contributed articles on the history of the Persian Empire to the Encyclopaedia Britannica.

Mr R. H. Pinder-Wilson, M.A.(Oxon.), F.S.A., who will accompany Tour No. 12, is Assistant Keeper in the Department of Oriental Antiquities at the British Museum and is there in charge of Islamic antiquities. He has travelled extensively in the Near East, and has taken part in excavations on the Persian Gulf and in Eastern Anatolia. His publications include books on Islamic Art, and Persian Painting in the Fifteenth Century, and contributions to learned journals.

Mr David Stronach, M.A.(Cantab.), F.S.A., who will accompany Tour No. 8, is Director of the British Institute of Persian Studies at Teheran. He read archaeology and anthropology at Cambridge in 1951-55 and was subsequently a Fellow of the British Institute of Archaeology at Ankara and of the British School of Archaeology in Iraq. He later carried out excavations on the North-West Frontier of Pakistan and has recently been uncovering the ancient Persian capital of Pasargadae. Has accompanied three previous tours.

U.S. $1328 New York/New York 8 page illustrated brochure available from:

W. F. & R. K. SWAN (Hellenic) Ltd. 260-1 (HT.2) Tottenham Court Road, London, W1P OAL, England

IRAN

Journal of the British Institute of Persian Studies

VOLUME VII 1969

CONTENTS Page

Governing Council . . . . . . . . . vi

Director's Report . . . . . . . . . . vii

Excavations at Tepe Nfish-i Jan, 1967, by David Stronach . .I La Survie de Shildu et la Route du Khunj-6-Fal, by Jean Aubin. . 21

Excavations at Siraf: Second Interim Report, by David Whitehouse . 39 The Darabgird Relief-Ardashir or Shahpiir ? A discussion in the context

of early Sasanian sculpture, by Georgina Herrmann . . . 63 The Beliefs and Practices of the Ahl-i IHaqq of Iraq, by C. J. Edmonds. 89 The Tahirids and Persian Literature, by C. E. Bosworth . . . o103 Kirman and the Middle East: Paul Ward English's City and Village in

Iran: Settlement and Economy in the Kirman Basin, by Brian Spooner and Philip C. Salzman . . . . . . . . .o107

Excavations at Bdbi Jan 1967: Second Preliminary Report, by Clare Goff . . . . . . . . . . . II5

Prehistoric Copper and Bronze Metallurgy in Western Iran (with special reference to Lairistdn), by P. R. S. Moorey . . . . . 131

" Some Ancient Metal Belts "-a Retraction and a Cautionary Note, by P. R. S. Moorey . . . . . . . . . 155

Hinweise und Anmerkungen zu einigen sasanidischen Monumenten, by Klaus Schippmann . . . . . . . . . 157

Further Notes on the Shaft-hole Pick-axe from Khurab Makran, by C. C. Lamberg-Karlovsky . . . . . . . . 163

Survey of Excavations 1967-68 . . . . . . . I69

Published annually by

THE BRITISH INSTITUTE OF PERSIAN STUDIES

c/o The British Academy, Burlington House, Piccadilly, London, W.I

BRITISH INSTITUTE OF PERSIAN STUDIES

GOVERNING COUNCIL

President

*Sir MAX MALLOWAN, C.B.E., M.A., D.Lit., F.B.A., F.S.A.

Vice President

Professor A. J. ARBERRY, M.A., Litt.D., D.Litt., F.B.A.

Members

Professor Sir HAROLD BAILEY, M.A., D.Phil., F.B.A. R. D. BARNETT, Esq., D.Litt., F.B.A., F.S.A.

*Sir MAURICE BOWRA, M.A., D.Litt., Litt.D., LL.D., F.B.A. Professor J. A. BOYLE, B.A., Ph.D.

MICHAEL BROWNE, Esq., Barrister-at-Law

JOHN BURTON-PAGE Esq., M.A., F.S.A. Professor W. B. FISHER, B.A., D. de l'Univ., F.R.A.I.

Dr. ILYA GERSHEVITCH, M.A., Ph.D., D.Litt. BASIL GRAY, Esq., C.B.E., F.B.A.

Professor A. K. S. LAMBTON, O.B.E., D.Lit., Ph.D. Professor SETON H. F. LLOYD, C.B.E., M.A., F.B.A., F.S.A., A.R.I.B.A.

LAURENCE LOCKHART, Esq., Litt.D., Ph.D., F.R.Hist.S. RALPH H. PINDER-WILSON, Esq., M.A., F.S.A.

BASIL W. ROBINSON, Esq., M.A., B.Litt. *Sir MORTIMER WHEELER, C.H., C.I.E., M.C., T.D., D.Lit., F.R.S., F.B.A., F.S.A.

Professor R. C. ZAEHNER, M.A., F.B.A.

Hon. Treasurer

Sir JOHN LE ROUGETEL, K.C.M.G., M.C.

Hon. Secretary

JOHN E. F. GUERITZ, Esq., M.A.

Joint Hon. Editors

Mrs. LUKE HERRMANN, D.Phil., F.S.A. Professor C. E. BOSWORTH, M.A., Ph.D.

OFFICERS IN IRAN

Director

DAVID STRONACH, Esq., M.A., F.S.A.

Assistant Director

DAVID BLOW, Esq., B.A.

c/o The British Academy, Burlington House, Piccadilly, LONDON, W. I

P.O. Box 2617, Tehran, IRAN

*Denotes Founder Member

DIRECTOR'S REPORT

November Ist 1967 to October 31st 1968

Guests Those staying at the Institute since November Ist have included the following: Miss Susan Russell

(University of Sydney); Miss Winnifred Morgan (University of California, engaged in Armenian studies); Professor Charles Adams (reviewing plans for McGill University's proposed Institute of Islamic Studies); Mr. G. H. R. Wright (Institut Frangais d'Archeologie, Beirut); Professor Roger Savory (University of Toronto, engaged in historical research); Mr. T. Unestram (University of Uppsala); Professor George Dales (University Museum, Pennsylvania, passing through); Professor Sir Max and Lady Mallowan; Sir Mortimer Wheeler; Dr. and Mrs. W. Lerouge (Ghent University, members of the Belgian Archaeological Expedition to Liiristan); Professor and Mrs. Henry T. Wright (University of Michigan), Mr. Robert Bettarel (University of Michigan) and Mr. Robert C. Gibbs (University of Maryland) members of the Farukhabdd Expedition; Mr. William Sumner (University of Pennsylvania, engaged in an archaeological survey of the Kfir river basin, Fars); Mr. R. F. Wye (studying Persian); Professor L. de Meyer (Ghent University, Delegation Archeologique Frangaise); Mr. Vincent Costello (University of Durham, making an urban study of Kash~n); Dr. Olu? Arik (Dil ve Tarih-Cografiye Fakiiltesi, Ankara University), Dr. A. D. H. Bivar (School of Oriental and African Studies, University of London), Lady Charlotte Bonham Carter, Miss Mary Burkett (Director, Abbot Hall Art Gallery, Kendal), Mrs. Rachel Maxwell-Hyslop (Institute of Archaeology, University of London), Dr. G6anil Oney (Dil ve Tarih-Cografiye Fakiiltesi, Ankara University), Mr. Ralph Pinder-Wilson (Department of Oriental Antiquities, British Museum), Dr. Philip Tyler (Department of History, University of Western Australia), Mr. Christopher Walker (Department of Western Asiatic Antiquities, British Museum), and Mr. M. E. Weaver (Middle East Technical University, Ankara) all of whom attended the Fifth Congress of Iranian Art and Archaeology; Sir Maurice Bowra (lecture tour on behalf of the Institute and the British Council); Professor and Mrs. Richard Ellis (Yale University); Mr. and Mrs. Philip Salzman (McGill University, engaged in ethnographic research in Baluchistan); Professor and Mrs. R. H. D. Slater (Harvard University); Dr. and Mrs. Robert Thomson (Harvard University, engaged in Armenian Studies); Dr. and Mrs. Louis Dupree (en route to Afghanistan); Miss Theresa Goell (visiting museums and archaeological sites); Miss Margaret Hall (Victoria and Albert Museum, studying textiles); Professor Carl Lamberg-Karlovsky, Mr. J. Humphries, Mr. Richard Meadow, Mr. and Mrs. Arthur Bankoff, Miss Jane Britton and Mr. Peter Dane (Harvard University Expedition to Tepe Yahyd); Dr. Robert Stigler, Mr. Newbold Noyes, Mr. Richard Davis, Mr. Howard Hecker, Mr. William McClintock and Mr. Terence McClintock (Columbia University Expedition to Azerbaijan); Dr. Joel Kraemer (Yale University, investigating material on the Muslim philosophers); Professor Robert H. Dyson, Jr., Dr. Oscar White Muscarella, Mr. and Mrs. Christopher Hamlin, Miss Mary Voigt, Miss Gayle Weaver, Miss Mary Virginia Harris, Miss Mary Sherman Parsons, Miss Elizabeth Stone, Mr. Bennet Bronson and Mr. Arthur Smith (IjasanlGi Project); Mr. and Mrs. Charles Burney (University of Manchester), Miss Leila Khalidy, Mr. Malcolm Stephenson and Mr. Timothy Dickinson (Excavations at Haftavn Tepe, Azerbaijan); Mr. Brian Clark (Dept. of Geography, University of Durham, engaged in urban research); Professor and Mrs. Saul Weinberg (University of Missouri); Dr. Clare Goff Meade, Miss Judith Pullar, Mr. Robert Hurst, Mr. Nigel Palmer, Mr. Stephen Mitchell, Mr. Julian Baldick, Mr. Hans Smits and Mr. Jeffrey Frye (Excavations at BThi Jan Tepe in Lflristan); Mr. and Mrs. Richard Tapper (University of London, engaged in social anthropological research in Afghanistan); Mr. Robert Spertus (University of Michigan, engaged in historical research); Miss Valrae Reynolds and Miss Sandra Leff (Institute of Fine Arts, New York); Dr. John Gurney (Oriental Institute, Oxford); Dr. Schuyler van R. Camman (University of Pennsyl- vania); Mr. and Mrs. David Pearl (Cambridge University, en route to Pakistan to commence studies

vii

in Islamic Law); Dr. Klaus Schippmann (University of Gottingen: archaeological research in the Bakhtiari region); Mr. John Wertime (Princeton University: study of Safavid history); Mr. David Gye, Mr. Ian Durie, Mr. Guy Williamson, and Mr. Richard Beauchamp (Cambridge University Expedition to Q6mis); Mr. Theodore H. Wertime, Dr. R. Tylecote, Dr. Frederick Matson, Dr. Beno Rothenberg, Dr. Radomir Pleiner, Dr. Robert Brill and Mr. Sam Bingham (Smithsonian Pyrotechno- logical Expedition to Afghanistan, Iraq and Turkey); Mr. Robert Hillenbrand (University of Oxford, investigating Islamic tomb towers in Mazandaran); The Rev. Harry Williams and Mr. Graham Storey (University of Cambridge, visiting); Mr. Peter Heyworth and Mr. John Warrack (reporting the Shirtz Festival of Arts); Professor and Mrs. James Barr (Dept. of Near Eastern Studies, University of Manchester); Mr. Leonard Harrow (University of Edinburgh, studying Persian); Dr. Pierre Verrin (Directeur, Centre d'Archdologie, Madagascar, studying late Islamic pottery); Miss Cynthia Roberts (University of Durham, studying Persian); Dr. and Mrs. David Whitehouse, Mr. and Mrs. Frederick Aldsworth, Miss Enid Parson, Miss Sonya Sprinthall, Miss Diana Clegg, Miss Rosemonde Nairac, Mr. Peter Donaldson, Mr. Edward Harris, Mr. Clifford Long, Mr. Ronald Shoesmith, Mr. Gerald Dalby, Mr. Giles Scholl, Mr. Jan Roberts, Mr. Nicholas Lowick and Mr. Edward Hans (Sirif Expedi- tion); Professor Richard N. Frye (Harvard University); Mr. and Mrs. Dale Eickelman (University of Chicago, studying Shi'ism in Iraq); Mr. William Royce (Princeton University, engaged in historical research); Professor and Mrs. Frank Hole (Rice University Expedition to Deh Luran); Mr. John Taylor (Selly Oak Colleges, Birmingham); and Mr. Michael Rogers (Center for Arabic Studies, American University in Cairo).

Visit of Professor Sir Max Mallowan

Early in January Sir Max and Lady Mallowan returned to Iran for a period of three weeks. During the course of his visit Sir Max delivered lectures on Elamite history and archaeology at the Institute and also at each of the British Council Centres in Tehran, Isfahdn and Meshed.

Visit to Sirdf Towards the middle of January Dr. David Whitehouse and his team were visited by Sir Mortimer

Wheeler, Sir Max and Lady Mallowan and Mr. Stronach. Many separate facets of the work were reviewed and, with reference to further campaigns, it was agreed that one of the Expedition's most rewarding tasks would be to clear a still larger part of the residential quarter close to the Great Mosque.

Congress of Iranian Art and Archaeology Two members of the Governing Council, Mr. Basil Gray and Mr. Ralph Pinder-Wilson, were

amongst the large number of British delegates that attended the Fifth Congress of Iranian Art and Archaeology, held in Tehran at the invitation of the Ministry of Culture, from April i i th to I8th. At the close of the Congress Mr. Basil Gray was elected to succeed Professor Arthur Upham Pope as President, and it was agreed that the Sixth Congress should be held at Oxford in four years' time.

Sir Maurice Bowra

After delivering the Institute's inaugural lecture on Edward Fitzgerald almost seven years ago, Sir Maurice Bowra returned to Iran in mid-April as a guest of both the Institute and the British Council. In his lecture at the Institute Sir Maurice took as his subject the father of British Oriental studies, Sir William Jones. Members of the Fifth Congress contributed to the discussion that followed and equally lively interest was shown in Sir Maurice's eloquent lecture at the British Council entitled " Omar, Fitzgerald and Graves ". In addition to lecturing at Tehran University on "Some Peculiarities of English Poetry ", Sir Maurice also spoke at Isfahan and Shiraz Universities.

V111iii

Further Lectures at the Institute " Forts, Peasants and Nomads in South-East Persia " by Dr. Brian Spooner on January 6th. " Contemporary Studies of Religion: Some Recent Developments " by Professor R. E. L. Slater,

Emeritus Professor of Comparative Religion at Harvard University, on May 25th. Professor Hussein Nasr, Dean of the Faculty of Letters at Teheran University, kindly took the chair.

Director

Mr. Stronach left Iran in early November, returning again at the end of December. During his visit to the United Kingdom he lectured at Sheffield, Cambridge and London universities on various aspects of the Institute's recent archaeological work in Iran. At the Fifth Congress of Iranian Art and

Archaeology in April he read a paper entitled " Excavations at Tepe Nfish-i J~n and Shahr-i Q5mis: New Light on Media and Parthia ". In common with other Institute directors in Tehran, Mr. Stronach will serve on the Executive Committee of the forthcoming Sixth Congress.

Assistant Director

Following his farewell lecture on January 6th, Dr. Spooner left Iran, first to return to St. Anthony's College, Oxford, and later to take up his present position at the University of Pennsylvania.

Mr. David Blow, a doctoral student at the School of Oriental and African Studies, London Univer-

sity, who first began his studies in history at the University of Cambridge, has been appointed as Dr.

Spooner's successor.

Excavations

Thanks to the handsome co-operation of all who contribute to our annual Survey of Excavations it is probably unnecessary to add further notes on last summer's excavations in Iran. The Institute has extended special help, of course, to those expeditions from the United Kingdom currently working at Sirdf, Bdba Jan Tepe and Haftavdn Tepe and the success of each of these projects is a matter for

gratification. Although no fresh excavations took place at either Tepe Ntish-i Jan or Shahr-i Q5mis during the

past year, a generous subvention from Mr. Ian Bowler, O.B.E., the President of I.M.E.G., was of great assistance in allowing a small team of geography and engineering students from Cambridge University to complete a number of surface studies at the latter site. At the same time we must also place on record our deep gratitude to H.E. Mr. Mehrddd Pahlbod, the Minister of Culture, and his staff for their

continuing interest in the work of the Institute-an interest most recently and most generously affirmed by the Ministry's munificent grant towards the cost of a new museum and temporary expedition house at Sirif.

Fellows

Mrs. Ruth Whitehouse and Mr. Hugh Ainsley, the Institute Fellows for the year 1967/8, both visited the Institute at the close of the second season of excavations at Siraf, Mr. Ainsley staying on to

complete a number of plans and drawings before his return to London. The past month has also seen the arrival of two of the Institute's Fellows for the current session, Mr. Andrew Williamson, who will be making a survey of early Islamic trade routes in southern Iran, and Mr. Anthony Hutt, who will be gathering material for a study of Ghaznavid architecture.

Cambridge History of Iran

In completing these pages, reference must be made to the celebrations that accompanied the publication of volumes I and V of the Cambridge History of Iran in October 1968. At a reception arranged under the auspices of the University of Tehran, at which both Professor Fazlullah Rezi, the Chancellor of the University, and Professor Hussein Nasr, the Dean of the Faculty of Letters, welcomed the new

ix

publication, Sir Frank Lee, the Chairman of the Syndics of the Cambridge University Press, was able to enlarge on the general concept of the History and the plans that have been laid for the remaining six volumes. In further lectures delivered at Tehran University and at Pahlavi University, Shiritz, Professor Sir Harold Bailey spoke on " A Vision of Aryashayana " while Professor Sir Max Mallowan spoke on " The First Steps in Iranian Archaeology ".

Finally, at a meeting at the Institute at which Sir Max Mallowan took the chair, Sir Harold Bailey read a paper entitled " Saka Studies: Recent Contributions to the History of the Ancient Kingdom of Khotan ". Most happily the text of this distinguished contribution will appear in Iran VIII.

x

1

EXCAVATIONS AT TEPE NUSH-I JAN, 1967

By David Stronach

Following the excavations of the British Institute of Persian Studies at Pasargadae,1 at least a part of the Institute's efforts has been directed towards a complementary search for Achaemenian, and more particularly Median, remains in the central Zagros region. In the early stages of this programme I was fortunate enough to be able to join Dr. T. Cuyler Young in several exploratory journeys, one of the most productive of which led to the discovery of Tepe Nish-i Jan (Fig. I) in March 1965.

'Tehran CASPIAN

IRAN SEA LAKE URMIA

SHASANL0L

* ZIWIYE

I" A

0 HAMADAN (ECBATANA)

/ GODIN TEPE *

) BISITUN * * TEPE NUSH-I JAN

(. BABA JAN TEPEM W

. 150

1.,., ' iKM.

Fig. I. Tepe NzMsh-i JIn and neighbouring sites located in western and north-western Iran.

Our route had taken us to the edge of the flat Malayer or Jowkar plain, some 6o km. south of Hamadin and 20o km. west of Maliyer, when the distant outline of Tepe Naish-i Jan first came in view. The unusual height of the site automatically suggested a natural feature, but on closer inspection it became apparent that a thick, artificial deposit had heightened and softened the outline of a bold rock outcrop (P1. Ia). The possibility of finding a Median settlement in this situation-not unlike that of the seventh century B.c. hill-top fortress of Ziwiye-had never seemed very remote and it was not too great a surprise when several surface sherds began to recall others known from Ziwiye. Within easy reach of

1 D. Stronach, Iran I (1963), pp. 19-42; Iran 11 (1964), pp. 21-39, and Iran III (1965), pp. 9-40.

2 JOURNAL OF PERSIAN STUDIES

Hamadan such prominent signs of seventh century construction could only be expected to tell us more about the Medes themselves, and, with little evidence for any major re-occupation, we soon felt that few other sites of limited size would offer more promising prospects for an investigation, principally concerned with the Medes.

Actual excavations at Tepe Nfish-i J~n began in August 1967 and lasted for just under two months. The work was conducted by the writer, assisted by Mrs. Stronach, Dr. A. D. H. Bivar, Dr. Oscar Muscarella, Mr. Michael Roaf, Mr. Andrew Williamson, Mr. Ian Herring and Miss Susan Bird. Mr. Ali Sarfaraz lent us his fine skills and companionship as the Representative of the Archaeological Service, while both Dr. W. Kleiss, Second Director of the German Archaeological Institute and Mr. Geoffrey Hewitt, A.R.I.B.A., were able to offer us valuable help in planning the site and its structures.

Generous grants towards the cost of the work were received from the Metropolitan Museum of New York, the Oriental Institute of the University of Chicago and the British Institute of Persian Studies.

At the same time our warm thanks are due to H.E. Mr. Mehrddd Pahlbod, Minister of Culture, to Mr. A. Pourmand, Director-General of the Iranian Archaeological Service, and to Professor Ezzatullah

Negahban, Director of the Archaeological Museum, for the constant co-operation of the Archaeological Service. In addition, we owe a further debt to the Director of Education in Hamadan for permission to live in a schoolhouse near the site.

The text of the present interim report is divided between a brief outline of the history of the Medes and a second section covering the results of the past season's work.

THE MEDES Of all the early first millennium inhabitants of the central Zagros the Medes are probably the best

documented in existing sources. They find a place in the Annals of each Assyrian king from Shalmaneser III (859-824 B.C.) to Assurbanipal (669-626 B.c.); their vital role in the overthrow of the Assyrian Empire in 612 B.c. is undisputed; and they are known to have ruled quite as wide a portion of the Near and Middle East as either the Babylonians or the Lydians during the succeeding six decades. Moreover, when Astyages, the last of the Median royal line, was defeated in battle by Cyrus the Great in 550 B.C. it was still the combined strength of the " Medes and Persians " that created and sustained the far-flung Achaemenian Empire.

The Medes-and the Persians like them-can be counted among the many waves of Indo-European invaders who reached the Iranian plateau from the north or north-east. At the beginning of a long and

complex process the first, still undifferentiated Indo-Iranian tribes may have reached the northern borders of Iran not long after 2000 B.C.2 but as the bearers of locally distinct customs, such as extra- mural burial, they or their more strictly " Iranian " successors cannot be said to reveal their permanent presence on the plateau before the fourteenth century B.c.

The points of entry used by these newcomers are still uncertain. No specific invasion routes have been positively identified in the course of recent archaeological work in the Caucasus, Soviet Central Asia or Afghanistan and, in Iran itself, most scholars have only begun to explore the implications of the second millennium attacks on Tepe Hissar and the subsequent appearance of pottery very like that of the Bronze Age north-east in both western and southern parts of the country.3

Young4 has reminded us that the " Eastern Grey Ware Culture " or the " Gurg?n Culture " of the north-east could have been associated with the still undifferentiated Indo-Iranians, while Deshayes5 would bring the Medes to the same region at the time that the local grey ware makes its first appearance. But both these tempting suggestions must be viewed with caution. The grey ware culture north of the Elburz is not by any means an intrusive phenomenon: many of its ceramic features can be traced back to the Chalcolithic painted pottery of the late fourth millennium B.c. and those changes that do occur

during the third millennium never seem to suggest anything save a smooth progression from a more

2 Cf. R. A. Crossland, " Immigrants from the North ", CAH I

(1967), ch. XXVII, pp. 31 f. * See especially R. H. Dyson, " The Archaeological Evidence of

the Second Millennium B.c. on the Persian Plateau ", CAH II (1968), ch. XVI, pp. 4 f.; and T. Cuyler Young,

" The Iranian Migration into the Zagros ", Iran V (1967), p. 24.

4 T. Cuyler Young, op. cit., p. 31, note gI. 5 J. Deshayes, " Turang Tepe et la plaine de Gorgan A l' ge du

Bronze ", Archeologie Vivante I, no. I (1968), p. 38.

EXCAVATIONS AT TEPE NUSH-I JAN, 1967 3

primitive stage of culture." The only mildly discordant note in the relevant sequence at Turang Tepe is the brief appearance-possibly a little after 3000 B.c.-of a brown-on-red rather than a black-on-red

painted pottery, the shapes of which clearly resemble those of contemporary grey burnished forms.7 As Deshayes himself has indicated,8 this may only be a transitional ware, born at a time when the established black-on-red painted pottery of the Gurgan plain was already beginning to fall out of fashion.

At the other end of the scale the difficulties take on another form and we seem to see too many extreme, indeed improbable changes. If we accept, for example, the proposition that the third and early second millennium inhabitants of the Gurgan plain were Iranian settlers of long standing, we have to ask why they should have deserted most of their hitherto prosperous settlements, why they should have

sought to destroy certain of their own southern outposts (such as Tepe Hissar), and, last but not least, why they should have rejected their ancestral burial habits ?9

Needless to say, the real crux of the problem is the extraordinary diffusion of north-eastern grey ware or, more strictly, that of various pottery forms that can only be descended from this parent tradition, for, with the discovery of evolved " Khorvin " types at Turang Tepe,10 more weight than ever has been placed on the possible Iranian origin of the Gurgan Culture. Yet it must be stressed that this same ceramic evidence is capable of being interpreted in different ways. If the Indo-European invaders of the second millennium came without elaborate cultural equipment of their own, and if they then divided the native north-eastern women amongst themselves-probably also adopting small numbers of the local male population-they would have created at least minimal conditions for the survival of a number of the arts and crafts of the early north-east. By carrying these same captive groups with them, they most probably worked a permanent change in their own tastes, not to mention those of others that they encountered.

Even in the latest levels at Turang Tepe (where Professor Deshayes has only begun to explore the

post-Hissar and post-Shah Tepe " III C2 " phase) these same circumstances could well account for the presence of evolved pottery forms not unlike those of Khorvin and other western sites. Turang Tepe was far more important than most local mounds-it was in all probability the Bronze Age capital of the

Gurgdn plain-and it could always have survived long enough to presage important trends in ceramic

production that are now familiar from various Iron Age settlements south of the Elburz. Also, if it should seem surprising that the pottery fashions of the north-east should have retained their popularity almost wherever the invaders went, it can only be said that the quality of the Gurgan pottery probably aided the process. No potters in any other part of the country were able to outshine this single regional product, then at its best after centuries of refinement."

Further promising support for a north-eastern or eastern invasion route, as opposed to a north- western one, is also to be found in other domains. Linguistic research leaves no doubt whatsoever of the broader bonds that tie eastern and western Iranians together, and, if we take only the immediate affinities of Old Persian, a south-west Iranian dialect, and Sogdian, a north-east Iranian dialect, it is manifest that both are closely related tongues that must have remained in contact during the period of their formation.12

Historical sources also point in the same direction. Apart from various hints of an early eastern connection in the works of Iranian, Greek and Roman authors,13 the Assyrian Annals offer their own

6 V. E. Crawford, " Beside the Kara Su ", Bull. Met. Mus. of Art

21 (8), (1963), pp. 271-2. 7 J. Deshayes, op. cit., p. 37. 8 J. Deshayes, op. cit., loc. cit.

9 Even if it should be contended that the Bronze Age inhabitants of the Gurgan plain were part of a long established Indo- Iranian community that found itself displaced by the Iranian peoples as such, it would seem most difficult to account for this last fundamental distinction between two closely related groups.

10 J. Deshayes, op. cit., p. 38. 11 At the same time, of course, the original monochrome tradition

began to weaken in its turn at the beginning of the first millennium B.c. Almost as soon as the Medes or elements of another closely related people succeeded in penetrating eastern Lfirist~n-an area with its own age-old taste for polychrome effects-we find the newcomers beginning to add painted designs to their once monochrome vessels, often with motifs culled from the native culture (cf. Clare GoffMeade, " Lfiristin in the First Half of the First Millennium B.C.", Iran VI (1968), p. I26).

12 W. B. Henning, " Mitteliranische ", in B. Spuler, et al., Handbuch der Orientalistik IV, i (1958), pp. 105-o8.

13 R. N. Frye, The Heritage of Persia (1963), PP. 36-49; T. Cuyler Young, op. cit., p. I I.


explicit evidence for a close relationship between the Iranians located on the eastern and western sides of the Iranian plateau in the opening centuries of the first millennium. In a series of references the Medes appear in the Annals as " the distant Medes ", " the mighty Medes of the east ", the Medes " whose country lies far off", " the distant Medes on the borders of Mount Bikni " (Mount Demdvend), and the Medes who live near the salt desert.14

Terms of this kind can hardly be taken to refer to individual tribes within the Median confederation, but at least the reference to Mount Bikni is particularly telling in the present context. Even if we do not know precisely how far or how often the Assyrians campaigned on the plateau, it is still clear that beyond a certain point in the Zagros, and continuing at least as far eastwards as Tehran, the Assyrian forces met nothing but people who professed to be Medes.

The contrast with conditions to the west of the main Zagros peaks is marked. As the Annals remind us, the Medes were only gradually able to extend their influence in the tangled terrain of the western Zagros where they not only had to contend with Assyrian and even Elamite alliances, but also with the tough, independent spirit of the Zagros peoples themselves.

In short, several separate sources allow us to reconstruct at least part of the main course of the Median, if not yet the entire Iranian, invasion. Moving from east to west, possibly both north and south of the Khurasdn highway, many of the Medes can be presumed to have settled south of the Elburz, near Ray, Qazvin and Siveh and, above all, near Hamadan, the future capital. But, once at Hamadan, the great mass of Kiih-i Alvand must have seemed to symbolize the multiple obstacles, physical and political, that lay in the west; thus only certain more intrepid groups would seem to have pushed further westwards in the first instance, while many others can be assumed to have spread north- westwards, and perhaps also south-eastwards, along the eastern edge of the Zagros alignment. In keeping with this pattern, it was probably only the fierce southern thrusts of the Iranian Scythians of the mid-seventh century and the ultimate collapse of Elam in 640 B.C. that hastened the Medes' eventual supremacy throughout the central and northern Zagros.15

It might be added that the role of the Persians in these last regions had little significance. Neither those from the land of Parsua, south-west of Lake Urmia, who receive their earliest mention in the Annals in 843, eight years before the Medes, nor those from more southerly areas of Kurdistan, can be said to have played anything more than a strictly local part.16 In brief, the majority of the Persians look as if they may never have attempted to invade the north-west angle of Iran, despite a probable political vacuum in that region immediately after the fall of the Mitannian kingdom,"1 and it is preferable to argue, pending still further work in the field, that the main body must have moved southwards along various tracks located between the Kavir and the Zagros, not to mention certain other more arduous routes possibly located east of the Kavir.18

Finally, before we leave the problem of the original Iranian migration and the many facets that any such movement must have had, we should perhaps recall the historical parallel that Herzfeld19 has drawn with the much later movements of the Iranian Saka. " After a short migratory period in what is now Russian Turkestan, they entered Iran, in about 130 B.c., through the only natural gate of the highland, near Sarakhs." In remarkably short order they " overran the whole newly established [Parthian] empire. Groups of them separated from the main body and successfully founded the Saka dynasty of Adiabene ... east of the Tigris. They may have founded also, at the same time-between 128 and I25 B.c.-the dynasty of Characene, modern Muhammera at the Persian Gulf. After a few

14 E.g. D. D. Luckenbill, Ancient Records of Assyria and Babylonia (1926-27) (henceforward LAR) I, paras. 784, 795 and 81 -I12; II, paras. 23, 54, 79, 82, 118, 137, 238, 432, 519, 540 and 566. See also T. Cuyler Young, op. cit., p. 13.

15 See also Edith Porada, Ancient Iran, the Art of Pre-Islamic Times (1965), p. 137; and T. Cuyler Young, op. cit., p. 13 f.

16 Both Young (ibid., p. 17) and Porada (op. cit., p. 137) remind us that the countries named " Parsua " may have been established at the same time by different groups of the same people. Thus those communities located in the north-west need never

have served as the fountain-head for a north to south migration through the length of the Zagros.

17 R. H. Dyson, op. cit., p. 31.

18 These last routes gain at least some support following the recent discovery of beak-spouted, grey ware vessels of early first millennium date at Tepe Yahyd-a mound that straddles an important route from Bampfir to Firs. Information kindly supplied by the excavator, C. C. Lamberg-Karlovsky.

19 E. Herzfeld, Iran in the Ancient East (0941), p. 191.

EXCAVATIONS AT TEPE NUiSH-I JAN, 1967 5

years of anarchy ... Mithradates II ... restored order, allowing the Saka to settle in the south-east of the plateau ", from whence they passed to India, founding " a short-lived empire, which extended as far as the gates of Delhi and Bombay ". As Herzfeld indicates, there are not a few arresting parallels including a probable north-eastern invasion route and the westward thrust of several smaller groups who chose to detach themselves from the main horde.

To return to the story of the Medes themselves, they were possibly no more than a loose confederation of tribes as late as the middle of the eighth century.20 Not all of these are thought to have been Iranian in origin and much of the population probably continued to speak either Guti or Lullubi dialects well into the seventh century.21 Yet it seems to have been during the second half of the eighth century that the Medes began to think of larger unions.

From the records of Sargon's eastern campaign in the year 715 we hear for the first time of a Mannaean or more probably Median chieftain, Daiaukku, who opposed the Assyrians and who appears to have been prominent enough to merit exile in Syria.22 Herodotus also seems to refer to the same local ruler when he describes the rise of Deioces, " a clever man . . . enamoured of sovereignty " who at length " united the Median nation " under his dominion.23

The real powers that were once wielded by Deioces/Daiaukku are far from certain-just as we can hardly say how much of his work outlived his own stewardship in a far from settled period. But whether or not Deioces succeeded in creating a united Media, and whether or not it was he who founded Hagmatana-modern Hamadan, called Agbatana or Ecbatana by the Greeks-we must acknowledge the fascination, and perhaps even the archaeological value of a passage in which Herodotus describes the successive tasks that Deioces set his people on his accession. " He bade them build him houses worthy of his royal power, and arm him with a bodygyard: the Medes did so; they built him great and strong houses at what places soever in the countrythe showed them, and suffered him to choose a bodyguard out of all their people. But having obtained the power, he constrained the Medes to make him one stronghold and to fortify this more strongly than all the rest. This too the Medes did for him: so he built the great and mighty circles of walls within walls which are now called Agbatana. This fortress is so planned that each circle of walls is higher than the next outer circle by no more than the height of its battlements; to which end the site itself, being on a hill in the plain, somewhat helps, but chiefly it was accomplished by art. There are seven circles in all; within the innermost circle are the king's dwelling and the treasuries; and the longest wall is about the length of the wall that surrounds the city of Athens. The battlements of the first circle are white, of the second black, of the third circle

purple, of the fourth blue, and of the fifth orange: thus the battlements of five circles are painted with colours; and the battlements of the last two circles are coated, these with silver and those with gold."24

In the same year that Sargon appears to have humbled Deioces the Assyrian ruler probably also settled a number of captive Jews in Media,25 presumably in various local townships since the Bible makes specific if perhaps over-generous reference to the settlement of the exiles in the " cities " of Media.26

From the Medes' point of view, Sargon's presence in the east may have had its compensations: both the older, more tightly organized states of Urartu and Manna were shorn of their original power, leaving the way open for further Iranian expansion. To begin with, however, the paramount might of

Assyria overshadowed all else. Sennacherib drew rich tribute from the " distant Medes "27 and, during the first years of Esarhaddon, three Median princes took the exceptional step of travelling to Nineveh to seek Assyrian aid against local rebellions at home.28 In 676 Esarhaddon's troops reached

20 Cf. Herodotus i. Io0. For comments on the six tribal names provided by Herodotus-those of the Busae, the Paretaceni, the Struchates, the Arizanti, the Budii and the Magi-see A. Christensen, " Die Iranier ", Kulturgeschichte des Alten Orients, Handbuch der Altertumswissenschaft III (1933), pp. 233 f.

21 I. M. Diakanov, Istoria Medii (1956), ch. 2. 22 LAR II, paras. x2 and 56. 23 Herodotus i. 96; i. Io1. (Quotations drawn from the Loeb

Classical Library edition.)

24 Herodotus i. 98.

25 I. M. Diakanov, op. cit., pp. 210o-I2. 26 II Kings 17: 6; 18: 11.

27 D. D. Luckenbill, The Annals of Sennacherib (0924), p. 133; LAR II, paras. 238, 282 and 432.

28 S. Smith, CAH III (1954), p. 82; D. J. Wiseman, " The Vassal Treaties of Esarhaddon ", Iraq XX (1958), pp. i-ioo; LAR II, para. 519.


Mount Bikni and over the next two or three years still other punitive raids struck far to the east.29 Then for reasons that are still obscure, but which would seem to be connected with more spirited local

leadership, we suddenly observe the first hints of an eastern threat to Assyria itself. From 674 onwards the Assyrian omen texts begin to refer to a certain Kashtariti, a chief of the

Kar-Kassi in the central Zagros, who apparently sought to unite Mannaean, Median and freshly arrived Cimmerian forces in a broad, anti-Assyrian alliance.30 In Kashtariti we may perhaps see the throne name31 of the Phraortes of Herodotus, the son, or more probably the grandson, of Deioces-if this single family succeeded in retaining its pre-eminence in the face of Assyrian displeasure. But the time for a concerted attack on Nineveh was not propitious: during the last years of his reign Esarhaddon was at least as interested in the renewal of his eastern, as his western, treaties32 and in the end Kashtariti's ambitious design probably collapsed before it could take concrete shape. What is to be seen though is an end to the annual raids of a few years before, and in 659, in response to various Median assaults in Manna,33 we ultimately find the Assyrians fighting vigorously to retain their hold over the rich Mannaean homeland. Indeed, despite Assurbanipal's claim to have also punished the Medes in this

campaign,34 it is not without interest to note that this may have been one of the last occasions on which the Assyrian army chose to confront the Medes on their own territory.

In keeping with these events, it is not at all unlikely that the battle-hardened Kashtariti/Phraortes should have turned his attention to southern Iran and that he should, for a time at least, have made the Persians " subject to the Medes ".35 Somewhat less convincing is Herodotus' claim that he died in a second crusade against the Assyrians after a reign of only twenty-two years36 for, whatever the manner of his death, it is hard to suppose that it occurred c. 653 B.C.37-a date that would place his son, Cyaxares, an improbable sixty-nine years on the throne.

In claiming twenty-eight years for the dramatic Scythian interregnum at a point early in the reign of Cyaxares, Herodotus also implies that this long occupation had little lasting effect on the Medes'

triumphant progress.38 But the actual date of the episode is far from firmly established-it might conceivably follow the fall of Nineveh39-and the seemingly precise figure of twenty-eight years possibly only refers to the active years of a warrior rather than to any carefully measured, historical

period.40 From the little we know of Cyaxares, the founder of the Median Empire, he was a notable leader.

Herodotus calls him far more warlike than any of his ancestors and asserts that he was the first commander in the distant east to abandon the use of tribal contingents and create a disciplined army with distinct companies of spearmen, archers and cavalry.41 Moreover, as the first of the Iranians to extend his own rule beyond the western limits of the Zagros, he probably set vital precepts for not only his own line but also that of the Achaemenians as well.

In seeking the final overthrow of Assyria, Cyaxares looked beyond the resources of the plateau; and in Nabopolassar, the founder of the Neo-Babylonian dynasty, he found a substantial ally. As the

Babylonian Chronicle indicates42 there were two major invasions of Assyria; in 614 B.c. when the Medes advanced on Nineveh from the direction of Kirkuk and then marched south along the Tigris to take Assur; and in 612 B.C. when, supported by their Babylonian and new-found Scythian allies, they again attacked, and finally captured Nineveh. The fury of these campaigns needs little stress, but in the first attack on Nimrfid-which probably took place during the advance on Assur43-the Iranians

took special pains to destroy the copies of the treaties that had once bound their princes to the Assyrian

29 LAR II, paras. 54o and 566; W. Culican, The Medes and Persians ( 965), p. 44.

so Cf. CAH III, loc. cit.

31 A suggestion that I owe to Richard Frye. 32 See especially M. E. L. Mallowan, Iraq XVIII (1956),

pp. 12-14; and D. J. Wiseman, Iraq XX, p. 3 f.

33 LAR II, paras. 851-3; W. Culican, op. cit., p. 46. 34 CAH III, p. 119. 35 Herodotus i. 102. 36 Ibid., loc. cit.

37 Cf. W. Culican, op. cit., p. 50.

38 Herodotus i. io6. 39 E. Porada, op. cit., p. 138. 40 K. Jettmar, " Ausbreitungsweg und sozialer Hintergrund des

eurasiatischen Tierstils ", Mitteilungen der Anthropologischen Gesellschaft in Wien XCII (1962) (Festschrift Franz Han'ar), p. 185-

41 Herodotus i. 103. 42 D. J. Wiseman, Chronicles of Chaldean Kings (1956), pp. 14-15. 4' David Oates, Iraq XXIII (I96I), p. 9.


throne. Breaking into the royal scribal stores, they apparently transported the clay tablets to the long throne room of the Nabu temple, smashed them to the ground, and lit a great fire over them.44

From here onwards the last decades of independent Median rule can be passed in rapid review. Possibly because of difficulties with his powerful Scythian allies, we hear virtually nothing of Cyaxares or his army between 6Io, when he joined Nabopolassar in a joint attack on the Assyrians posted near IHarrdn, and 590 when, in concert with the Babylonians' own southern offensive against Palestine, we find the Median king striking westwards across Anatolia to the line of the Halys river.45 There he met Alyattes of Lydia in what soon devolved into a war of attrition, reportedly five years in length (590-585 B.C.). In the end, with neither side able to claim any material advantage, the two monarchs agreed to an armistice. Babylonian and Cilician mediators were brought in,46 the Halys was fixed as the boundary between the two states and the two royal houses were joined in marriage, Alyattes offering his daughter's hand to Astyages, the son of Cyaxares. When Astyages succeeded in 584, therefore, a balance had been struck between the three major powers of the Near East-Babylonia, Lydia and Media-and for a time at least the embattled Medes must have known comparative peace.

To the east, where we have little definite information, the Medes may have exercised some measure of control over a number of smaller Iranian states, each with a king of its own. If Henning's attractive thesis is correct, the great Iranian prophet, Zoroaster or Zarathu'tra, should find his place in this milieu-as should his protector, Kavi Vistaspa, who probably emerges as the last ruler of the Khwarez- mian state of Mary and/or Herat in the first half of the sixth century B.C.47 Equally the brief statement of Nicholas of Damascus, claiming that Cyrus received the submission of the Hyrcanians, Parthians, Saka and Bactrians at the beginning of his reign48 would seem to show that these same peoples had already come to accept Median supremacy. But without doubt the most important local ruler after the " king of kings " at Ecbatana was the leader of the Persians, a man who already ruled most of south- western Iran. Throughout almost all the final years of Media's greatness this prominent figure was Cambyses I (600-559 B.C.), the son of Cyrus I. Although subject to the Medes, Cambyses seems to have given no sign of resenting his position and, perhaps as a mark of his fidelity as well as his rank, he was permitted to marry Mandana, the daughter of Astyages. Cambyses' own son by this marriage was Cyrus II or Cyrus the Great (559-530 B.C.).

Cyrus' temperament was not one to brook a subordinate role and within four years of his accession we find him entering into an alliance with Nabonidus of Babylon against the Medes. In taking such a step Cyrus may have detected the first hints of weakness in Astyages' leadership both at home and abroad. In 554, for example, the no longer vigorous Babylonians were able to evict the Medes from IHarrin without provoking any earnest counter measures. But for all this the Median army was still a formidable military machine and there can be little doubt that the Persians were forced into a most desperate defence, possibly over a period of several years, before they achieved a decisive victory in 550 B.c.

Neither the Babylonian Chronicle49 nor the Abbu Habba cylinder (foretelling the overthrow of Astyages)50 can be said to offer any real clue as to where the climactic battle was fought. The Chronicle does state, however, that Astyages took the offensive. He " collected his army and marched against Cyrus king of Anshan to [conquer him] ". But the assault was of no avail; Astyages' army mutinied and he himself was captured and given up to Cyrus.51

Amongst Classical authors, Herodotus52 and Polyaenus53 talk of two encounters between the opposing armies, while Justin54 and Nicholas of Damascus55 appear to refer to at least two Persian reverses before the tables were turned. Beyond this, both Strabo and Polyaenus agree that the final engagement was fought at or near Pasargadae, in Persian territory. As Strabo writes in explicit,

4" M. E. L. Mallowan, op. cit., pp. I f.

45 CAH III, p. 214- 46 Ibid., pp. 512-I3. 41 Cf. W. B. Henning, Zoroaster: Politician or Witch-Doctor? (1951),

p. 43. 48 Cf. A. W. Lawrence, Herodotus, Rawlinson's Translation Revised

and Annotated (I935), note on p. 76.

49 S. Smith, Babylonian Historical Texts, p. II5- 10 Ibid., p. 44. 5x Ibid., p. I15-

52 Herodotus i. 127 and 128.

53 Strateg. VII.

54 I. 6. "5 Fragmenta 66.

2A


succinct terms: " Cyrus held Pasargadae in honour because he there conquered Astyages the Mede in his last battle, transferred to himself the empire of Asia, founded a city, and constructed a palace as a memorial to his victory."56

Herodotus tells us that Cyrus spared Astyages, keeping him at the Achaemenian court until his death.57 Moreover, by presenting himself as the heir of Astyages and treating his new subjects with every mark of respect, Cyrus laid strong claim to the hard-won gains of the Medes themselves.58

THE EXCAVATIONS The mound of Nfish-i Jdn rises to a maximum height of nearly 37 m. with almost equally steep

contours on all sides (Fig. 2). The summit is surprisingly restricted, running to little more than 8o m.

A 8 C DEIF G H I J K/L M !N O P 0 R S T U!V W!X /z

AVERAGE LEVEL OF PLAIN AND

A7M 1

ILis

4 1

-- *--- 4

I- - - - - \ ii /

iI'i''-' 15

IIIt

1 10 tE

15 4 -"4 --4_ I i

0 10 3-0_ METRES \\\

Fig. 2. Site Plan.

in length and 30 m. in breadth. Although still other parts of the site may demand exploration in the future, all our present work has been confined to a 5 m. grid, 130 x 75 m. in area (Fig. 2). Within this compact space at least three distinct Median structures have been identified, and the plans of two of them (Fig. 3) already provide an intriguing preview of the quality and interest of Median mud-brick architecture.

Since the sequence of occupation already seems to be clear in its main outlines, it may be worth attempting to sketch the history of the mound in chronological order, beginning with events in the Median period (Naish-i Jan I) and following with notes on both the Achaemenian (Nfish-i Jan II) and Parthian (Ntish-i Jan III) periods. 6 Strabo, XV. 3. 8. 5 Herodotus i. I30.

*8 The above account omits almost all reference to either the

religion of the Medes or the organization of the Median state. The reason will be plain: here, as in so many other spheres, we

depend on future excavations for enlightenment.

Pl. Ia. Tepe Nash-i Jan from the south. Pl. Ib. A detail of the alternate layers of mud and stone that were used to cap the lower stone fill of the Central Building.

P1. Ic. The interior of the Central Building, showing the east end of room i. The original floor lies I m. beneath the level of the ranging-rod.

Pl. Id. The northern and north-eastern walls of room I at the close of the excavations.

P1. Ha. A detail of the recessed crosses and blind windows Jbund in the north-east corner of the hall. The right-hand window measures I - 35 m. in

height and 96 cm. in width.

PI. IIb. The strongly battered, southern wall of room I, showing the door and window found near the middle of this otherwise blank surface.

Pl. IIc. The southern end of the Fort's long guardroom (room 22). Pl. IId. One of three triangular-headed niches discovered in room 22. (The scale measures 25 cm. in length.)

P1. IlIa. The ramp inside the Fort looking north. The doorway from the guardroom appears on the right.

Pl. IIIb. A view looking down the ramp towards the doorway to room 23. The silver hoard was found in the left-hand corner, immediately to the left of

the ranging rod.

P1. IIIc. The central pier in room 20. On the left, the ramp gives way to all that remains of the steps leading to the second storey.

Pl. IIId. Three broken mud-brick struts from the former ceiling of the ramp.

Pl. IVa. The secondary staircase in room 23. Pl. IVb. A view from room 26 looking through successive doorways to room 23.

PI. lVc. A detail of the doorway between rooms 25 and 23. Pl. IVd. Part of the corbelled ceiling at the west end of room 23. On the

right is the only window, set at the apex of the original ceiling.

P1. Va. A detail of part of the ceiling. The plaster on the left has been cut

away to reveal the steps of the brick corbel. Pl. Vb. Long mud-brick struts from the upper fill of room i8. The ledge

from which they seem to have once sprung can be seen on the left.

Pl. Vc. Part of the western wall of the Fort, showing four arrowslots and- on the extreme right-a single window.

P1. Vd. A detail of the narrow window looking into room 23. The

adjoining arrowslot contains traces of a secondary blocking.

Pl. VIa. Part of the curved, mud-brick wall enclosing the south-east corner of the Central Building. An adjoining buttress of the Fort appears on

the right.

P1. VIb. The weathered north-east corner of the Central Building traced down to bedrock. The triple mud-brick courses on the left appear to mark the

beginnings of the bastion.

Pl. VIc. Room II : looking north-west towards the only doorway. Pl. VId. A second view of room i i, two secondary skins of mud-brick mask most of the buttressed and recessed facade of the original Fort. An

earthbaulk appears on the extreme left.

Pl. VIIa. Rooms 6 and 7, looking east from room 8. Pl. VIAb. A small hearth of Parthian date rests above a substantial Median wall with the remains of a corbelled niche. Square F io east, Levels i and 2.

Pl. VIIc. A vertical view of room 40 (square F io west, Level 2). Note the large and small beam-holes in the right-hand wall.

P1. VIId. The bronze bowl containing the Niush-i Jdn silver hoard. One side of the brick that originally concealed the bowl appears on the right.

Pl. V

IIIa. Tw

o double spiral beads from the P

eriod I (Median) silver hoard.

The upper exam

ple m

easures 4 cm. across, the low

er, 5" I cm. across.

Pl. V

IIIb. T

wo further double spiral silver beads (4" 3 and 5

o cm. in w

idth respectively)

; a quadruple spiral bead adheres to the larger specim

en.

Pl. IXa. Three quadruple spiral silver beads, the two largest of which measure 2- 6 cm. across.

Pl. IXb. Three silver coils with tapered ends. Each coil measures 2- 6 cm. in diameter.

Pl. IXc. A less regular silver coil with four loops. Maximum diameter 4 cm.

Pl. IXd. A silver ear-ring or pendant. Length 2"

2 cm.

Pl. Xa. Three silver bars from the hoard. The largest example measures 9g I cm. in length and weighs ioo 8 gm.

Pl. Xb. A bronze elbow fibula with cross-hatched incisions on each arm. From the floor of room 6. Length

4"3 cm.

Pl. Xc. A bronze head qf the demon Pazuzu, an evil genius reputed to bring fever and sickness. From the floor of room 4.

Height 3'3 cm.

PI. XIa. A two-handled jar of fine buff ware, fired pale red near the base. Period I

(see also Fig. 6, 8).

Pl. XIb. A fragmentary bowl of Period III cinnamon ware. Diameter 22 cm.; height 7 cm. From surface deposits located above the Fort.

Pl. XIIa. Part of a pottery handle in the shape of a duck's head. Length 5" 2 cm. Period I.

P1. XIIb. A rim fragment of a burnished grey ware bowl with a horizontal handle and two applied knobs. Period I (see also Fig. 6, 4).

Pl. XIIc. Rim sherds from three cinnamon ware bowls. Diameter of each 20 cm. From surface deposits located above the Fort.

EXCAVATIONS AT TEPE NUfSH-I JAN, 1967 9

N1ish-i Jdn I The architectural remains of this earliest period almost certainly reflect a longish period of occupa-

tion, possibly extending from the second half of the eighth century to the first half of the sixth century. From the pottery recovered to-date, the bare rock surface of the hill was not occupied before the eighth century or, if it was, the original deposit was cleared in its entirety so that each of the main Median structures could be founded on bedrock.

10~1

S_ -

"

.../

t

HIG LEELDORWY

77

24

PRIMARY WALLS OF CENTRAL BUILDING AND FORT

ZZSECONDARY BRICKWORK ADDED TO THE ABOVE7

BRICKWORK ENCLOSING THE CENTRAL BUILDING

05 10 15O20OME TRES

GH LEVEL WALLS OF LATE MEDIAN DATE

1 ;ii ' .:. .

1

... ...? ...... i7?h- .

-. .l -". 1

_;: I.:.:!.7:-:':; :."...4 • :W :: :: : . -." :-. • .. •- ,,.. • , , "

... .--.N ,

,~ ~ ~ ~ ~:??..?

...•?i:•?::i 7 '~?. ...- ... ;SY • •

u

,? :.-: ..... , . ...:, ,,

i .:.? .... :

.•• . r

. ;i..• : l

i 22

' , ': 1:':...i. .':T"fb •..fi;'•+# XJ J

• •"t PIMAY W LLSOF ENTALBUIDIN AN FO T !'" ..

?;".' .", SECONDARY~

~~~~~~~.:. BRCWR DDDT H AOE[""

"..

L••.I

~~: BRCKOR ENLOIG HECNTALBILIN

Vi/ HIGH LEVEL

DOWALL OF AEMDA AE ?y...

Fig. 3. Plan of the Central Building and adjoining Fort.

The Central Building Towards the middle of the mound, where the contours were visibly higher than elsewhere, most of

the past season's work came to be concentrated on a still baffling construction which we shall refer to as the Central Building.59 Possibly once lozenge-shaped in plan, with stepped inner and outer wall faces (Fig. 3 and Pls. Ic-IIa), this unique mud-brick construction appears to have had at least two opposed rooms separated by a straight partition wall. The partition wall itself contains a broad door with a wooden lintel and at least one tall internal window 90 cm. deep (P1. IIb).

Within the partly excavated principal chamber (room I) floor level was only reached at two points: at the end of the east bay where a single, sandy floor deposit yielded a thin scatter of seventh century buff-ware pottery and at one corner of the north bay, where a deep pocket of ash was found. However,

59 For previous references see Iran VI (1968), p. 162; Antiquity XLII (1968), p. 98; and D. Stronach, " Tepe Nfish-i Jdn: a

Mound in Media " in Bull. Met. Mus. of Art, November 1968, p. 177-86.


in approaching floor level over a wide area the excavations have exposed a series of unusual wall decorations, including recessed crosses, square " scaffold holes " and two types of blind window (Pls. Ic-IIa). The crosses only occur above the broad windows found on the stepped side walls, while the scaffold holes appear in at least two of the bays, either above a narrow type of false window or above a plain rectangular recess. There is no proof that either the crosses or the square scaffold holes were ever used to take beam ends. In the case of the scaffold holes this must remain a definite possibility, however. Those that can be seen at the back of each bay are matched by others in the adjoining side walls, suggesting that an elevated wooden platform could have once stood at each of the room's three corners. Yet another distinctive feature is the long corner pilaster found between the two blind windows at the north-east angle of the hall (P1. IIa).

Unless we look far ahead to the very different architectural traditions of the Sasanian period,60 the lozenge-shaped plan of the Central Building is unparalleled elsewhere, and it seems safe to assume that it stems from a local building tradition of which we still know very little. As we might expect, the closest analogies to the plan and appearance of room I come from Phase III (Level Vb) at Bbdt Jan Tepe, some Ioo km. to the south-west. In the " Painted Chamber" of this neighbouring eighth century settlement we again find an oddly shaped room with its longest dimension parallel to the entrance- wall, and a deep, central bay opposite the door. In addition, this same room appears to have made prominent use of both corner pilasters and carefully matched recessed windows.61

The less lofty walls at Bdbd Jan make it impossible to say whether or not similar recessed crosses were employed in the " Painted Chamber ", but at least we do find upright or diagonal cross motifs on most of the tiles that fell to the floor from some higher position.62

The discovery that two large columns once stood near the middle of the " Painted Chamber " is also of interest, since we may yet find that the roof of room I had to be provided with similar support.

Turning to wider parallels for these characteristic Median features, it is possible to cite various neighbouring areas that used the blind window to advantage. In ninth to seventh century Assyria such windows may not have been entirely unknown;63 in Urartu they were particularly popular, occurring not only in mud-brick but perhaps in stone as well;64 and in Achaemenian Fars a well-known form of Urartian blind window is probably represented in both the Zenddn-i Sulaiman at Pasargadae and the Ka'ba-yi Zardusht at Naqsh-i Rustam.65 The closest parallel to the type of blind window found at

Nfsh-i Jan comes from a surprisingly remote source, however: the fifth century temple at Marib in the Yemen.66

In both the stone windows from Marib and the mud-brick examples from Niish-i Jan we see a similar double " dentil cornice " stepping downwards and inwards. Both forms suggest a wooden prototype, and at Marib at least the whole lower part of each window is thought to imitate a wooden lattice.67

To judge from other indications of rigorous symmetry, room I must have had four windows with six, as opposed to four, mock beam ends and two windows of the narrower form. But since the only excavated example of the slender form occurs at the end of the narrow east bay, and the only box-like niche at the head of the still smaller north bay, the selection of such decorative elements was most probably dictated by the form of the room rather than by specific ceremonial considerations.

From what has been found so far, therefore, it is still too early to attempt to define the function of the Central Building and we can only conclude by describing what is known of its later fate. To begin with the whole edifice was filled with small stones up to a height of6 m. Such stones were not merely thrown

60 Cf. R. Ghirshman, Iran: Parthians and Sassanians (1962), p. 139. 61 See " Survey of Excavations ", Iran VII (1969), pp. 169-70. 62 Clare Goff Meade, " Excavations at Bdba Jan 1967: Second

Preliminary Report ", Iran VII (1969), pp. 128-9. For the appearance of the cross motif on the earlier painted pottery of Baba Jan B, see also Clare Goff Meade, " Lfiristin in the First Half of the First Millennium B.c.", Iran VI (1968): p. I 18.

63 Note especially certain internal, elevated niches with double reveals that occur at Nimrfid. D. Oates, op. cit., pl. VIb.

64 R. D. Barnett, Iraq XVI (1954), fig. I; and D. Stronach, " Urartian and Achaemenian Tower Temples ", JNES XXVI (1967), pp. 284-5-

65 D. Stronach, loc. cit. 66 G. Van Beek in R. L. Bowen and F. P. Albright, Archaeological

Discoveries in South Arabia (1958), pp. 223 f. and pls. I65. I66 and 168.

67 G. Van Beek, ibid., p. 223.

EXCAVATIONS AT TEPE NTSH-I JAN, 1967 11

in from above: they were placed in position with great care so that no part of the original structure would be damaged in any way. The large and small chips used in this operation were all obtained locally-the stone being identical with that of the main Niish-i J~n hill and that of various outcrops near it. When the pure shale fill had reached a point within 2 m. of the top of the present walls, it was capped by a series of alternate bands of mud and shale followed by a thick protective seal of mud-brick

(P1. Ib). This last cap not only covered the area of the large triangular room but (to add to our difficulties in the early phases of the excavation) it also ran over the original walls of the entire structure.

Beyond such direct measures to fill and seal the interior of the building, the south side of this once free-standing structure was surrounded by a long curved " bastion " of mud-brick at least 7 m. high; its east side was concealed by the construction of the neighbouring Fort; and at least part of its stepped north face was hidden behind other secondary walls (Fig. 3)- It is possible, of course, that the Fort may have been built for the protection of this distinctive monument: but once again this last point can only be left open. The Fort

As can be seen from Fig. 3, this second major structure possesses a straightforward ground plan composed of a guardroom, and adjoining ramp and staircase, and four long parallel magazines. Approaching 22 x 25 m. in size,68 the outer walls are marked by six projecting buttresses on the short sides and seven on the long sides. The only entrance lies between two standard buttresses in the middle of the east wall.69

Unfortunately the main door has suffered much from subsequent disturbance, and the door jambs themselves stand little more than I m. in height. The full depth of the doorway has not, as yet, been examined and it is not known if the inner jambs survive.70

Although the northern half of the guardroom (room 22) is also damaged, the southern half remains in excellent preservation (P1. IIc). Here we see three triangular-headed niches, each roofed with pairs of

opposed bricks (P1. IId), and at least two raised hearths, where the soldiers on guard duty presumably warmed themselves during the long winter months. For reasons that are still obscure large parts of the floor of this room were covered with water-worn pebbles; these lay directly on the floor, under the mud-brick collapse that constitutes the remainder of the fill.

West of the guardroom a broad, but now poorly preserved, door opens upon a gently sloping ramp (P1. IIIa). Most of the original ceiling of the ramp has long since disappeared, but near the head of the first straight section we were able to reveal a series of parallel mud-brick struts at each side of the

passage. At a distance such struts only appear to possess a slight curve (P1. IIIa) while on closer inspection it can be seen that they were not only curved but angled as well (P1. IIId). Those bricks immediately above the struts appear to represent part of a corbelled vault that has again slipped downwards from a higher level (P1. IIIa).

In room 20 the ramp progresses round three sides of a solid mud-brick pier before giving way to a staircase on the fourth side (P1. IIIc). Each corner of the pier leans outwards to take the thrust of the now vanished corbelled vault, while the diagonal, broken surface along the upper right-hand face

(P1. IIIc) most probably marks the zone where the roof members once joined the central block. Returning down the ramp-or turning left out of the guardroom-we come to the door leading to

the first of the four magazines (P1. IIIb). The near left-hand corner of this same opening produced the richest find of the first season: the silver hoard discussed on pp. 15-16. As can be seen from Fig. 3 and P1. IVa, the southernmost magazine (room 23) did not remain a simple storeroom, but instead suffered various changes. Probably following a partial collapse in the area of the ramp most of room 23 was filled with mud-brick and a narrow path was left for a new staircase to the first floor.

6s If we measure from buttress to buttress. The dimensions are in fact slightly irregular, since the west end of the Fort is visibly wider than the east end (Fig. 3).

69 For at least an impression of the original appearance of the Fort-complete with a strong external batter-compare our own half-excavated structure with the bronze model of a square fortress recovered from eighth or seventh century Sfisa.

Edith Porada, " Battlements in the Military Architecture and in the Symbolism of the Ancient Near East ", Essays in the

History of Architecture Presented to Rudolf Wittkower (1967), figs. Io a-d.

70 All short dotted lines in Fig. 3 represent restored wall-faces where the present work has not yet confirmed the line of the walls.

3


With reference to the remaining magazines, each seems to have been systematically emptied before the Fort was abandoned. But for all this these tall chambers still recall something of their original appearance (Fig. 4) with a string of more of less intact doors opening between them (P1. IVb).71 The most complete doorway is perhaps that between rooms 25 and 23, where the reveals approach 2 m. in height and where the springing of the inner arch remains perfectly sound (P1. IVc). In such narrow doors it is difficult to detect any unusual mud-brick members, and it looks very much as if a series of standard bricks, measuring 40 x 25 x 13 cm., were each pitched at a slight angle to form a low triangular vault.

NORTH SOUTH

"'- ,. ... . :.

. . . ? . .

PARTHIAN

.TAU PERIOD Zil

-R"'ASH_.LL

FALLEN MUD BRICK

MEDIAN PERIOD

I- w

//,,

120

??I: "'|?

" "

F EMUD-BRRICK WALL AND

________________________ I FLOOR OF POEM 18

0 2 IMETRES.

...... -

:!!.. ///

.:

:

??

PER101

?/1/

,/l,

I ~ ~ ~ ~ ~ ? "/

// ,

`U- :ilKWL N [ ~ ~ ~ ~ ~ ~ r

......?,- -- ....

o i 2 MET RES I I /

Fig. 4. P 9, East Section.

We already know of the existence of three windows in the west wall of the Fort (Fig. 3) and, in one case, in room 23, the window itself is bordered by the remains of a corbelled vault (P1. IVd). From such details it is evident that each magazine was lit by a single western window, set at or near the apex of the

ceiling. Indeed, the head of such openings appears to have passed only just below the level of the first floor (P1. IVd).

71 The two central magazines, 25 and 26, each share the same dimensions. While the average width of the two chambers works out at 2-34 m., the original length can be calculated at i2 6o m. and the maximum height at 5.60 m. (Contrast the

measurements for room I: from east to west, I - .20 m.; from north to south, 6.90 m.; original height unknown, but reaching at least 8 m.)

EXCAVATIONS AT TEPE NfSH-I JXN, 1967 13

In shape, the windows appear to have been short and squat on the outside (Pl. Vd) and appreciably longer on the inside, where a downward extension ensured an adequate opening below the line of the vault. On the inside, for example, the narrow aperture looking into room 23 measured 80 cm. in height with a maximum width of just over 20 cm.

As far as the vaulting system is concerned, a tolerably complete picture emerges from rooms 23 and 18. In room 23 we still find parts of the original corbel standing in situ, with the underside covered with plaster and a narrow ledge appearing at the junction of the long side wall and the corbelled ceiling (P1. Va).

For a time the ledge beneath the corbel remained a mystery; but then in cleaning the upper fill in room I8 we found a second series of mud-brick struts (P1. Vb) not unlike those first observed above the ramp, and, as the work progressed, it became apparent that each had slipped from a distinct ledge 4 - 80 m. above floor level (P1. Vb). Thus to summarize what is known of the Fort's mud-brick ceilings so far: the main weight was born by a standard form of corbel, presumably rising to a sharp triangle, while, possibly for decorative purposes only, the underside of the corbel was hidden by a series of long, parallel mud struts or beams. The latter may have consisted of identical " half beams " that met at the centre of each room or they may have been made in one piece; in either case the ends rested on a narrow ledge at the top of each side wall. The exposed plaster in P1l. Va shows no clear impress from the back of any strut and it is certainly hard to suppose that these relatively soft, slender objects were ever meant to bear more than their own weight. Moreover, from their very shape it looks as if they were designed to create the illusion of a barrel vault as opposed to a pointed vault. But it must be added that the evidence in question still only comes from several separate parts of the Fort and that the final verdict may well rest with a wholly intact vault, still to be recovered from the west end of either room 25 or 26 (Fig. 3).

The tall external walls of the Fort are of most interest for their multiple arrowslots (P1. Vc). These are found on all sides of the Fort save along part of the west wall, where the proximity of the Central Building presumably made them unnecessary (Fig. 3). Otherwise (at least between corner buttresses) we find two arrowslots to each recess and one to each buttress. Most examples only stand to half their original height, if this, but in one case the external elevation is complete (Fig. 5 and P1. Vc). This same loophole-the northernmost example from the west wall-also proves that those slots that survive today were each manned by an archer standing on the first floor. In the partly restored section in Fig. 5 the external height of 2 - I m. is taken to be uniform throughout.

Although one cannot be very certain about the nature of the Fort's defences above this last point, it would seem not unlikely that a second tier of arrowslots sloped down from the open roof above the first floor rooms and that a " parapet walk " provided access to the crenellations-let alone any more lofty additions to the buttress-towers.

As in the Neo-Assyrian defences at Assur,72 the individual arrowslots appear to have been crowned by two angled bricks forming an almost arrowshaped, triangular cap. More curious is the distinct taper of the

Nfish-i Jan slots which can hardly have been anything but a hindrance to the archers, particularly when it came to firing downwards at close-range targets.

Before leaving the Fort, it is of interest to note that many first floor walls appear to have been directly mounted over those of the ground floor. Moreover, all such elevated wall fragments look relatively slim-as if the builders had made a conscious effort to lighten the weight of the upper storey.

Additions to the Central Building and the Fort The nature of the additions made to the Central Building suggest that none of them were introduced

until work on the adjoining Fort had been completed. But equally, the absence of any arrowslots immediately beside the Central Building (Fig. 3) can be used to show that those who built the Fort were already aware of the plan to construct a solid " bastion " of mud-brick round a large part of the older, tower-like structure (P1. VIa). In probing down to bedrock between the two main buildings (P1. VIb) it is possible to see the beginnings of the curved bastion starting opposite the end of room 26. However,

72 E. Porada, op. cit., fig. I I.


AI

PLAN SECTION A-A A 'V -V-

EXISTING SILL

ELEVATION SECTION

PLAN SECTION A-A

0 1

I I METRES

Fig. 5. An arrowslot from the west wall of the Fort.

the bastion was not the only filling device that was contemplated, and either immediately after it had been constructed, or at some more distant remove, a mass of shale was laid over the earlier brick blocking (P1. VIb) until, in the end, two of the Fort's windows were also sealed.

Further to the north, many of the Fort's other appointments were ignored in the course of time: new skins of mud-brick were added to the outer wall (P1. VId) and entirely new rooms, such as Nos. Io and I I, were added to the original complex (P1. VIc). In all probability the size of the Fort grew to be too small, and the original design came to be absorbed in a still larger construction. It is conceivable that this new, larger unit was again equipped with its own defences, although confirmation of this will have

EXCAVATIONS AT TEPE NUJSH-I JAN, 1967 15

to wait for a further season. One possibility, of course, is that the upper walls of the original Fort were still allowed to out-top those that were added to the north, east and south-so that in the end the Fort came to provide an inner set of battlements reminiscent of those ascribed to the seven-walled citadel of Ecbatana.

In marked contrast to these quite substantial additions, a series of cramped, irregular rooms (Nos. 4-8) were also erected-possibly at a still later time-on a broad platform immediately outside the line of the bastion. Such chambers possess few features of architectural note, and they are probably of most value for the quality of the small objects that have been found inside them.73

Typical of the rooms in question are the thin, sometimes irregular walls shown in Pl. VIIa. This same photograph also illustrates a narrow doorway cut into the upper brickwork of the bastion-an area where we may well find similar stores or workrooms in the course of future work.

Soundings at the West End of the Mound Although this part of the site has seen very little excavation, the Median structures from this quarter

are again of interest. In square F Io we were fortunate enough to discover part of a large, plastered wall with a tall form of corbelled niche on one side (P1. VIIb) and a small, plastered room, No. 4o, on the other (P1. VIIc). In this last unit all the main features would seem to point to a small, domestic chamber very different in scale and function from anything found in the other primary structures.

While some speculation must surround the curious oblong hole-complete with a brick cover-in the middle of the floor, most attention will be focused on the beam sockets at the top of the west wall (P1. VIIc). The latter include three square holes arranged in the form of a rough triangle with, above them, the marks of many smaller timbers that probably formed the actual ceiling.

Such evidence for wooden roofing techniques is of particular value. At the partly contemporary site of Godin Tepe, 50 km. to the west, we already possess the ground plan of a magnificent thirty-columned throne hall4 and, if we should learn more of the standard beam arrangements used in seventh century Media, it should no longer be very long before we are able to discuss monumental Median, as well as monumental Achaemenian roofing methods.

One of the odder features of room 40 is its present state of preservation, for, while the western limit of the room has been perfectly preserved, the remainder has been neatly cut away, and the above- mentioned, plastered cross-wall substituted instead (P1. VIIc). The reasons for this change of plan may well be of some interest, but for the moment it is another attractive prospect: if this secondary cross-wall should continue down to bedrock and if it should be planned, photographed and finally removed, might we not find immediately behind it, the entire section of a two- or even three-storey Median dwelling that had been split in two while still in excellent condition ? The floor of room 40 is quite high enough to suggest that this could be the case-even if future excavations should prove this particular hope to be unfounded.

The Hoard One of the expedition's most fortunate discoveries came at the close of the past season while work

was being completed in the eastern rooms of the Fort. On lifting an incomplete brick from the sloping floor at the south-east corner of the ramp (P1. IIIb) it was found that this same brick had been used to conceal a plain, hemispherical bowl filled with more than 200 small silver objects (P1. VIId).

Such a hiding place could never have been chosen unless the Fort had fallen out of use and we may assume that the brick itself was selected from the debris that had already begun to collect on the floor. In the conditions such a method of burial was quick and practical-with an element of simplicity that deceived other intruders.

The objects themselves include double and quadruple spiral beads (Pls. VIIIa and b and IXa); small coils with tapered ends, each 2 6 cm. in diameter (P1. IXb); a larger, less finished form of coil (P1. IXc); a small ear-ring or pendant composed of several separate globules (P1. IXd); six silver bars of varying size, three of which are shown here (P1. Xa); more than ten small cubes that appear to

7" See p. i6 f. '* T. Cuyler Young, Iran VI (1968), p. 161.


have been cut from similar bars; and well over a hundred scraps of metal-each apparently fragments of ingots or of other once-finished objects.

Given the presence of these last fragments, it seems most probable that we are dealing with part of a silversmith's stock-in-trade. The jewellery is all of silver, without any composite elements, and any cache that was being stored for its intrinsic value would almost certainly have included other cherished materials, particularly gold if not also agate and carnelian. But at the same time it is possible that the bars, cubes, and coils from the hoard were more than convenient lumps of raw material or simple forms of jewellery; although further careful studies are required, many of these items may take their place as samples of local currency-each with a purchasing power equivalent to their own weight.?5

It should perhaps be added that both the double and quadruple spiral beads add a curiously archaic element to this relatively late collection. The large double spiral pendants belong to a relatively rare form with a long, horizontal sleeve that is otherwise only known from the late third or early second millennium levels of Tepe Hissir IIIB76 while the small quadruple beads provide much the latest parallels for both those recovered from Marlik Tepe7" and those said to come from Ziwiye.78

Bronze Objects Although relatively few iron or bronze objects were found in the course of the excavation, three

bronze items from the rooms south of the bastion each look as if they might represent western products. These include an elbow fibula with cross-hatched incisions on each arm (P1. Xb), a pendant head of the demon Pazuzu with the perforated top knot common to such objects (P1. Xc) and a short pin with a roll-headed top. All these objects find parallels in seventh century Nimriad79 and we may wonder if part of the more commonplace booty from the sacked cities of Assyria was not brought to Niish-i Jan prior to its subsequent dispersal.

Pottery The study of Median pottery is still necessarily at an early stage and, in the present context, it may

be most useful to illustrate, both in line drawings and photographs, certain typical shapes that have been encountered in the present work (Figs. 6, I- I and 7, 1-3 and Pls. XIa and XIIb). As still further Median pottery is published from both Bbat Jan Tepe and Godin Tepe a most useful corpus should emerge, with all that this implies for the dating of Ziwiye and other sites. For the moment, however, most of the pottery from Nfish-i Jan can only be ascribed to the single bracket of the seventh century.80

Chronology Apart from the dating evidence referred to above, a large fragment of wood-perhaps part of a

beam-from the lower stone fill of room I has yielded a carbon 14 date of 723 -?220 B.C.81 Such a date may well accord with the early life of the Central Building, unless further excavations should produce unexpected new evidence. As for the Fort itself, its bricked-in walls clearly never suffered from anything like such long exposure as those of the Central Building (P1. VIb) so that it may only have come into being after 700 B.C. with its own later additions following soon afterwards.

The reasons for the abandonment of the mound cannot be divorced from Tepe Nfish-i Jan's still uncertain, central function. Nevertheless, the site undoubtedly came to acquire a certain military significance in the course of time, and both the collapse of Assyria and the gradual erosion of Scythian power may have led to the desertion of a number of strongpoints, at least where these lay near the heart of Media's extensive territory.

7" Of those excavated objects illustrated in the present report, I owe Pl. XIIc to Mr. M. Rustami and Pls. VIII, IX, X, XIb and XIIa and b to Mr. Arthur R. Smith.

7s E. F. Schmidt, Excavations at Tepe Hissar Damghan (I937), pl. LV, H 2389.

" E. Negahban, A Preliminary Report on the Marlik Excavation (1964), pl. VIc.

78 W. Culican, " Spiral-end Beads in Western Asia ", Iraq XXVI (1964), p. 39 and pl. VIIIa (where incidentally the form of the

beads remains closer to those from Marlik than those from

Nfish-i Jan). 79 D. Stronach, " The Development of the Fibula in the Near

East ", Iraq XXI (0959), pl. LI, 5; M. E. L. Mallowan, Nimrud and its Remains I (1966), p. Ix9; and D. Stronach, " Metal Objects from the 1957 Excavations at Nimrud ", Iraq XX (1958), pl. XXXVI, 3.

80 See also the Catalogue of Pottery on p. 20.

81 Information kindly supplied by Dr. J. E. Noakes, Oakbridge Associated Universities.


5 6

2 32

7 8 9

IIi -

10 12

13

15

7j

Fig. 6. Bowls and jars of Median date (Nos. i-.i)

with four " cinnamon ware " bowls of Parthian date (Nos. I2-15).


Fig. 7. Additional pottery of Median date.

EXCAVATIONS AT TEPE NUfSH-I JAN, 1967 19

Nash-i Jdn II After an interval of at least a century we find traces of a brief reoccupation of the site in Achaemenian

times. High floors from this period have been found in squares S 9 and T 9, complete with various types of small jars that are also known from the Achaemenian levels at Bdbd Jan Tepe.8A However, the settlement never grew beyond a few structures placed at the east end of the mound and nothing approaching monumental construction is represented.

Nash-i Jdn III Finally, following a much longer break in occupation, both the eastern and western ends of the

mound were resettled in the Parthian period. Beaten earth floors, irregular tauf walls, mud-ringed hearths and numerous pits (Fig. 4 and P1. VIIb) all combine to suggest a flurry of village activity that lasted for not much more than fifty years. At the end of this period, which was probably one of relative unrest, the mound was again deserted, no doubt as the inhabitants moved back to more convenient if less defensible homes at the level of the plain.

No coins have been found so far, but the pottery has proved to be unexpectedly fine. Apart from the discovery of several glazed bowls, one of which shows concentric grooves on the inside of the base, the site has also produced a thin, fine, dense pottery with a grey core and a reddish to yellowish brown surface that may be termed " cinnamon ware " (Fig. 6, 12-15 and Pls. XIb and XIIc).

Cinnamon ware has been known for some time in fact, although none appears to have been published from any other context. At least one stratified sherd from one of the first seasons of excavation at Seleucia-on-the-Tigris has been dated to between 140 B.c. and 40 A.D.83 while still others from later seasons may be expected to supplement this information." Recent surveys and excavations in Iran have also helped to define the region in which this central Zagros ware was most at home. Apart from appearing at many sites on the natural highway between Kirminshdh and Hamaddn,85 the area concerned appears to have included the Bujnurd plain in the south,86 the Mahi Dasht plain in the west,"8 the Arak plain in the east88 and the distant religious centre of Takht-i Sulaimdn in the north, where the first sherds of this type were only unearthed during the summer of 1968.89

Without question, the most remarkable feature of this pottery"9 is the distinctive quality of the ware -a quite singular quality that may yet allow us to plot most settlements of Parthian date over a generous area of western Iran.

82 Information kindly supplied by Clare Goff Meade. 83 Sherd with accession number 33265 (excavation number

F2693) from the Kelsey Museum of Archaeology, Ann Arbor. Information kindly provided by Clark Hopkins and Louise A. Shier.

84 Personal communication from Frederick Matson. 85 Personal observation. 86 Information kindly supplied by T. Cuyler Young, Jr.

87 Information received from Clare Goff Meade and T. Cuyler Young, Jr.

88 Personal observation.

89 Information kindly supplied by Rudolf Naumann. 90 The samples illustrated in the present paper only include

certain of the more common types of bowl encountered at Nfish-i Jan; small jars and pilgrim flasks are among other shapes attested.


CATALOGUE OF POTTERY IN FIGURES 6 AND 7

Fig. 6 Nos. I-II all date from Period I.

I. Core: pinkish-buff, fine, with white grits; surface inside: creamy buff, smoothed; surface outside: light-pinkish-brown changing to pale buff near rim, burnished on rim and body. Wheelmade. One-third missing. Diameter: 16 cm.; Height: 8.5

cm. F Io west, room 40. 2. Core: pinkish-buff, fine, with barely visible inclusions; surface inside: creamy buff, burnished near rim;

surface outside: light pinkish-brown changing to pale buff on upper body, burnished on both base and body. Wheelmade. Half missing. Diameter: 18 cm.; Height: 8 cm. Provenance as above.

3. Core: pinkish-buff, medium fine; surface: pale buff, burnished inside and outside. Wheelmade. One- third missing. Diameter: 14 cm.; Height: 6-5 cm. Provenance as above.

4. Core: grey-black, medium-fine; surface: same, finely burnished. Wheelmade. Base and half of rim missing. Existing rim attests one horizontal handle and two small applied knobs (P1. XIb). Diameter: 14 cm. S 9, room I I, upper brick collapse.

5. Core: reddish-brown, medium fine; surface inside: same, with horizontal burnish marks; surface outside: reddish-brown to pale buff, with diagonal burnish marks sloping to left and right in alternate areas. Wheelmade. Diameter: 18 cm.; Height: io cm. F Io west, room 40.

6. Core: reddish-brown, fine; surface: light reddish-brown to buff, burnished on upper body. Wheelmade. Diameter: 20 cm.; Height: 9 cm. Provenance as above.

7. Core: reddish-brown, medium fine, with mica grits; surface inside: same, finely smoothed near the rim; surface outside: creamy buff with red patches, burnished with long, horizontal strokes. Wheelmade. Rim diameter: Io cm.; Height: 9"3

cm. N 9, room I, floor. 8. Core: pinkish-buff, fine, with white grits; surface inside: same, smoothed; surface outside: upper body

buff, lower body pale red, burnished. Wheelmade. Rim diameter: 12 cm.; Height: 11 .4 cm. N 11, from fill of arrowslot.

9. Core: reddish-brown, slightly coarse, with black and white grits; surface inside: warm buff, finely smoothed except near base; surface outside: reddish-brown to buff, with well-marked wheel ridges. Rim diameter: 10o.5 cm.; Height: 14-'5 cm. S 9, room i i, upper brick collapse.

Io. Core: light grey, slightly coarse; surface same, with horizontal scraped and burnished lines. Wheelmade. Base and part of rim missing. Diameter: 20 cm. M 8, room I, lower stone fill.

i i. Core: brown, medium fine, with mica grits; surface: light red to light reddish-brown, burnished on rim and exterior of body. Wheelmade. Body not complete. Diameter: 22-5 cm. F Io west, room 40.

12. Core: grey in the middle and brown to reddish-brown towards the surface, very fine in texture; surface: reddish-brown, smoothed. Wheelmade. Concentric grooves inside upper body. Diameter: I8 cm.; Height:

5"5 cm. P 9, Level Ib and P 9, Level Ib, pit i. Period III (Parthian).

13. Core: orange-brown, very fine; surface same, smoothed. Wheelmade. Concentric grooves inside upper body. Diameter: 18 cm.; Height: 5"5

cm. Provenance as above. 14. Core: grey in the middle and reddish-brown towards the surface, very fine; surface: reddish-brown

darkening to brown in patches, smoothed. Distinct wheelmarks on upper body. Broken and mended in

antiquity. Diameter: I8 cm. P 9, Level Ib. Period III (Parthian). 15. Core: grey in the middle and reddish-brown towards the surface, very fine; surface: red to reddish-

brown, smoothed. Wheelmade. Diameter: 20 cm.; Height: 7 -2 cm. Provenance as above.

Fig. 7 All vessels date from Period I.

I. Core: buff, medium fine; surface inside: pale yellowish-brown, smoothed near rim; surface outside: same, showing horizontal burnish marks. Wheelmade. Upper elements now restored as in lower left- hand view. Maximum diameter: 23 cm.; Restored height: 35'3 cm. R 12, Level 2, from mud-brick collapse.

2. Core: brown to reddish-brown towards the outside surface, rather coarse; surface inside: same, smoothed; surface outside: same, burnished. Wheelmade. Height: 2 i5 cm. R I2, Level 2, from mud-brick collapse.

3. Core: buff, medium fine, with white grits; surface inside: same, smoothed; surface outside: same, scraped and burnished. Wheelmade. Partly restored; end of trefoil spout slightly damaged. Rim diameter: 24 cm. N Ix, Level 2.

21

LA SURVIE DE SHILAU ET LA ROUTE DU KHUNJ-O-FAL

Par Jean Aubin

I

Voyageant dans le Golfe Persique au debut du XIIIe siecle, le lexicographe arabe Ydqit constatait de ses yeux la desolation de Siraf, dont il est le premier a mentionner le nom vernaculaire, Shildu, qu'il entendit de la bouche des marchands. Lorsqu'il ecrit, dans l'article " Sirdf" de son dictionnaire

geographique, qu'y subsistent " les vestiges d'un bel edifice et d'une elkgante mosquie cathedrale " et

que l'endroit " n'est plus habit6 que par de pauvres gens (sa'dlk) que seul y attache l'amour du pays natal ",1 on peut conclure, et on a conclu, a la disparition definitive de l'activite de Sirdf.

Les premieres observations des savants britanniques qui fouillent le site invitent a ne pas forcer outre mesure le contraste entre l'6poque de splendeur de Sirdf et le temps de son obscure survie. Au d6but du XIIIe siecle, notent-ils, Siraf recevait encore de petites quantitis de poterie lustr6e en

provenance de Syrie ou d'Iran du Nord, et une des belles demeures de la periode faste etait toujours habitee.2 Plus tard meme, au XIVe ou au XVe siecle, une mosquie nouvelle y fut construite.3

La chute de Sirdf, assurement, ne fut pas aussi brutale, et les causes accidentelles ne furent pas aussi de?cisives de sa ruine qu'on l'a pretendu, - le declin s'etala sur un siecle et demi.4 Apres les crises du XIe siecle, qui la firent descendre au rang de port local desservant les regions avoisinantes, Sirdf ne recouvra jamais sa prospirite d'antan, contrairement a ce que Badger a soutenu, en s'aidant d'une

argumentation aujourd'hui depassee.5 Il est vrai neanmoins que les sources 6crites, si pauvres soient- elles en informations sur l'histoire tardive de Siraf, corroborent les deductions des archeologues. Le

temoignage de Yaqilt n'est pas irrivocable. II a d'ailleurs exprime dans un autre de ses ouvrages une

impression qui vaut d'etre relevie, car elle est legerement moins d6favorable que celle qu'on cite d'ordinaire: " Sirif est un bourg (bulaid) sur le bord de la mer, en Fdrs. Je l'ai vu. Il s'y trouve les

vestiges d'un idifice ancien et d'une belle mosquee cath6drale, mais il est maintenant en majeure partie en ruine."6 Bourg decadent, plut6t qu'asile de pauvres hares, tel parait bien avoir et6 Sirif au XIIIe siecle, off le cabotage entretenait un petit trafic,7 of les pfcheries de perles fixaient un artisanat specialise et attiraient les courtiers,8 et of on assiste a une mise en valeur du terroir. Alors que les descriptions

1 Yaqt, Mu'jam al-bulddn, ed. Wiistenfeld, III, p. 211-212. La traduction p6jorante de sa'dlik par "voleurs " (Eghbal, radgdr 11/2, p. 14) me semble d6mentie par le contexte. Etymologie populaire de la forme persane " Shirab ", cf. al- Tha'alibi, 6d. Zotenberg, p. 167, et Ydqfit, l.c. Pour une etymologie philologique, cf. Eghbal, i.c., p. 8, et W. Eilers, Archiv Orientilnt XXII (1954), P. 320-321. Il est reconnu depuis Tomaschek (1890) que "Siraf" est la forme arabisde.

2David Whitehouse, " Excavations at Sirdf. First interim report ", Iran VI (1968), p. 12 et p. 21. Aux indications bibliographiques donn6es A la note 13, on ajoutera: Alastair Lamb, " A visit to Siraf ", Journal of the Malaysian Branch of the R.A.S. XXXVII (1964), pp. i-9. En persan, un article ancien et sans originalit6 (de Abbas Eghbal, non sign6), " Sir~f-i qadim ", Yddgdr II/2 (1324 s./1945), PP. 7-18, dont partie se retrouve dans le travail de jeunesse du meme auteur (mais publi6 en 1328 s./1949 seulement),

Mu.tdla'dti dar bab-i...

Khalyj-i Fdrs.

3 Whitehouse, op. cit., p. 5. Sir Aurel Stein, Archaeological Reconnaissances in North-Western India and South-Eastern Iran,

(Londres 1937), p. 206, datait la mosqu6e du XVe ou du XVIe siecle.

* Dans un pr6c6dent article, " La ruine de Siraf et les routes du Golfe Persique aux XIe et XIIe siecles ", Cahiers de Civilisation Midiivale 11/3 (1959), PP- 295-301, j'ai 6voqu6 la d6cadence de Sirdf en fonction de l'6volution d'ensemble de l'Iran du Sud. J'ai depuis sugg6r6 qu'il y avait eu permanence du trafic maritime (" Y a-t-il eu interruption du commerce par mer entre le Golfe Persique et I'Inde du XIe au XIVe siecle? ", Studia XI (Lisbonne 1963), Pp. 65-I7I), et

o cette occasion

cess6 d'admettre la datation des instructions nautiques sirdfies propos6e par Sauvaget.

5 G. P. Badger, History of the linams and Sayvids of Oman (Londres 1870), notes pp. 409-420 (retour des gens de Qais a Sirdf, localis6 A Chirfi).

6 Yqfit, Irshdd al-arib, 6d. Beyrouth, VIII, p. 145-

7 Sirdf est inclus dans la liste des iles et des localit6s du Golfe 6num6r6es dans la relation de l'ambassade du roi de Ceylan a

Qala'fin (Quatremere, Mimoires giographiques et historiques sur

l'Egvpte II, p. 284). 8 Cf. " La ruine de Sirf ", p. 300oo, note 32.


anciennes mettaient l'accent sur l'absence de ressources agricoles et de vegetation, les textes de l'6poque mongole font 6tat de l'existence de points d'eau, de jardins et de terres en culture.9

Si modeste soit-il par rapport au passe, le r81e que Siraf continue de jouer apres sa decheance lui

permet de retrouver, du XIVe au XVIe siecle, Qais une fois detruite, une animation relative, a laquelle Shiraz mise a part et Kdzirin exceptie, n'atteignent aucune des cites du Fdrs qui avaient ete florissantes

at l'poque bouyide. L'evolution geopolitique du Fars meridional entre le XIIe et le XVIe siecle ne saurait s'expliquer sans tenir compte de la survie de SirMf. Avant d'en discuter, rappelons quelques notions methodologiques10 propres at guider une recherche de geographie historique:

a) Les noms de ville iraniens designent le territoire dont la ville est chef-lieu aussi bien que ce chef- lieu lui-meme (le frangais, disposant de l'article defini, exprime facilement la distinction, par exemple entre Sirdf, la ville, et le Sirif, son district).

b) Les agglomerations ayant une fonction iconomique, leur developpement est command6 par celui des dessertes routieres; la courbe de la prosperite urbaine est fonction du trace des itineraires commerciaux.

c) Contr1le de l'appareil politique et contr6le de l'Fconomie sont le double objet de l'ambition des clans politiques; geographie historique et histoire sociale sont complkmentaires: l' tude du reseau routier est indissociable de celle des groupes de pression.

II

Dans les textes posterieurs a Ydqiit la forme litteraire arabisde, Siraf, et la forme populaire d'usage courant, Shildu, se rencontrent concurremment, mais on n'en peut collectionner qu'un nombre

d'exemples infime. Le vocable, en effet, s'il subsiste pour d6signer la localit6, disparait comme nom de district, ce qui est un indice tres net de la regression du centre urbain. La terminologie geographique et administrative a recours disormais pour denommer l'ex-territoire du Sird f, a un autre terme, jusque-la non attestS, mais des lors represente avec constance: le nom de Ftl. La chaine montagneuse qui domine Siraf, et que la tradition geographique ancienne appelle JibMl Jammn1 (du nom du canton de l'hinterland de SirMf) ou JibMl Sir•f•2 est appelke JibMl FMl dans les instructions nautiques de Ibn Majid,13 et encore connue sous une forme alt6ree de ce nom par les marins modernes,14 alors que l'agglomeration de FTl se trouve dans les terres, B l'est du massif montagneux, a deux journees de marche de Siraf. De l'emploi du nom de FMl, une lettre de Rashid al-Din offre un cas remarquable. Le conseiller des Ilkhans 6numere une s6rie de ports et de districts c6tiers qui le fournissent de produits exotiques, dont FdI et la principaut6 de Qais (mulk-i Kish).15 Puisque le bourg de FMl n'est pas au bord de la mer, le mot s'applique ici ~ un territoire, selon toute vraisemblance & la partie littorale du FMI situee l'ouest de la principaute de Qais; de sorte qu'il est hautement probable que c'est par Shilau

que FdI regoit l'ambre gris et le bois d'aigle destin6s a Rashid al-Din, et que le nom de FdM est en

6quivalence de celui de Sirif. C'est a la lumibre de cette substitution du nom de FMl & celui de Siraf qu'il faut interpreter Yqfit

lorqu'il ecrit qu'on passe par FMl pour aller de Shiriz a Ormuz: " FdI est un gros village semblable a une ville, aux confins miridionaux du Firs, pris des c6tes. On y passe quand on va a Ormuz et B Qais, sur la route de Huzi."'6 Sur la foi de ce passage, on a 6mis l'hypothbse que FMI mididval 6tait situ&e

9 Zakariyd Qazwini, Athdr al-bildd, 6d. Wiustenfeld, p. 136; Ibn Battfita, cf. infra, p. 29; selon IIamdulldh Mustaufi, Nuzhat al-qulab, (6d. Le Strange, p. I17; trad. p. 116; 6d. Dabir Siyiqi, p. 141), Siraf produit du blW et des dattes.

10 Cf. Jean Aubin, " Elements pour l'6tude des agglomerations urbaines dans l'Iran mfdifval ", dans The Islamic City, 6d. A. H. Hourani et S. M. Stern (Oxford 1969).

11 Istakhri, 6d. BGA, p. 127-128; trad. persane, 6d. T6heran, p. I13: Kfih-i Jam.

12 Ainsi encore Muhammad ibn Najib Bakrdn, Jahan-ndma, 6d. Moscou 1960, pp. 34 et 121.

13 Ibn Majid, dans G. Ferrand, Instructions nautiques et routiers arabes et portugais I (Paris 1923), fol. 38a.

14 Constable et Stiffe, Persian Gulf Pilot, 6d. 1864, p. 31 et p. 194 (Sir Yafal).

15 Mukdtibdt-iRashidi, d. Muhammad Shafi' (Lahore 1947), P. 196. 16 Yqfit, Mu'jam al-bulddn, IV, p. 846.

LA SURVIE DE SHILAU ET LA ROUTE DU KHUNJ-O-FAL 23

beaucoup plus au sud-est que FMl moderne.17 Bien que Yaqfit refire a FMl en temps que bourg, il faut entendre qu'on passe non par FMl, - qui n'est pas une des etapes de la route de Huz5, la liste desquelles est transmise par le Nuzhat al-qulfib,18 - mais par le FMl, district qui, au temps de

Ydqft, s'6tend loin

vers l'est. Le Siraf, nous le savons par Muqaddasi, gdographe attentif a respecter les divisions administratives, allait jusqu'a la frontiere du Kirman, couvrant ainsi toute la region des garmsirdt du Fars qui formera ulterieurement la fagade c8tiere du Ldristdn.19 C'est done dire que la route de Shirtz a HuzUi et a Qais traversait le territoire de l'ancien Sirif, qui a pris le nom de FMl lorsque Ydqfit recueille ses informations.

Dans les textes du XIVe siecle, le nom de FMl est employe pour d6signer en fait l'arriere-pays de la c6te de Qais. Il figure le plus souvent dans une expression composee oii il est associe, non point au nom de SirdffShildu, qu'il a recouvert, mais a celui d'une autre bourgade, ou d'un autre canton, pareillement inconnu des textes d'epoque bouyide, Khunj. C'est le KhunjubMl familier des lecteurs des voyages de Ibn Battfita.20 Cette gemination signifie que Khunj a connu entre le XIIe et le XIVe sikcle, un essor

marque, qui l'a hauss6 au rang de centre urbain. La ruine de SirMf, toute liee qu'elle soit a une profonde degradation de l'6conomie du Firs, n'a pas

frapp6 de mort la region des Garmsirit, jusqu'alors connue sous la dinomination vague de " r'gion des C6tes " (Irahistan, Sif). Elle a rendu possible la naissance, dans des localites qui etaient demeurees anonymes, de nouveaux centres de vie 6conomique et de rayonnement politique. Centres modestes, certes, mais bien vivants. FMl d'abord, puis Khunj, et Ldr, destine a tre le si'ge d'une petite princi- paut6 dans le sud-est du Fars, aux confins du royaume d'Ormuz. Les " terres chaudes " du Fars furent ainsi partagees entre deux formations politico-administratives distinctes, de structure interne diffhrente et de visees exterieures opposees, celles-la meme dont il est fait 6tat dans la relation de Ibn Battfita, le bildd Ldr et le bildd Khunjubdl.

III

D'un coup d'ceil rapide sur les lieux, en 1934, Sir Aurel Stein a conclu que FMl, ofi il vit de nom- breuses pierres tombales de style sir'fi, avait ete prospere vers la meme 6poque que Sir;f, mais que l'occupation en avait probablement durd plus longtemps.21 11 est, en effet, plausible que le gros village de type presque urbain dont parle Yaqilt, situ6 au debouch6 d'une des pistes montant de la c6te, dans un bassin fertile, entourd de cultures,22 ait ete un des estivages des gens de SirMf, qui a la saison chaude allaient en altitude,23 et qu'il leur ait offert une position de repli24 lorsque le grand port fut expos6 aux coups de main des pirates du Golfe et des tribus de l'interieur. Un tel exode expliquerait que Fil ait remplac6 Siraf comme chef-lieu regional. Que Ibn al-Balkhi, dans son tableau gdographique du Fars, compose au debut du XIIe siecle, n'en fasse point mention n'a rien de surprenant, puisque, au moment oii il &crit, Siraf reste aux yeux du gouvernement de Shirdz la metropole du district, nonobstant le fait que la ville et son territoire sont sous la coupe d'un chef de tribu Jat, un certain Abi'l-Qasim.25 C'est dans la liste des possessions des Jat (au XIIe siecle) que le nom de FMl est cit6 pour la premiere fois dans la litterature historique. L'emir AbSi Dulaf, dont Wassaf vante la distinction et la gdnerosite, regnait

17 K. Lindberg, Voyage dans le Sud de l'Iran, carnet de route d'un mddecin a la poursuite du ver de Midine (Lund 1955), p. 252.

18 Cf. infra, note 61. 19 En effet Muqaddasi, 6d. BGA, p. 52 et p. 422, compte parmi

les mudun qui d6pendent de Sirdf les deux ports de Sairi, face au Kirmdn, et de Zirabdd (connu ensuite sous le nom de Lashtdn).

20 Pour les diff6rentes formes du terme chez Ibn Battuta, cf. I. Hrbek, " The chronology of Ibn Battfita's travels ", Archiv Orientdlni XXX (1962), p. 449. Nous employons systamatique- ment ici la forme Khunj-6-FAl de pr6f6rence a la forme locale Hung-6-P~l.

21 Stein, op. cit., p. 221, et p. 2 7. 22 Cf. SakhAwi, Dau al-ldmi' XI, p. 218 (kathirat al-fawdkih) ; Stein,

op. cit., p. 217. Supprimer, dans JA (1953), p. 104, note 2, la phrase sur un marghzdr de Fgl, due A une erreur de lecture.

23 Ibn Rusta, trad. Wiet, p. 179. 24 Stein, op. cit., p. 218.

25 J'ai d6j; signal6 (" La ruine de Sirkf ", p. 300, note 32) la faute de Le Strange, qui transforme cet Aba'l-Qhsim en

khan de Qais (Description of the province of Ears..., Asiatic Society Monographs, XIV, p. 42). Outre que le titre de khdn n'est guere concevable pour un petit chef du Golfe Persique

'

cette 6poque-lA, le texte laisse fort bien entendre qu'il n'est pas un des chefs de Qais. Apris avoir fait 6tat des tractations, A

SirSf, entre Khumartegin et les envoyds de l'6mir de Qais, Ibn al-Balkhi ajoute: " Pour finir il en fut ainsi qu'il y cut quelqu'un d'entre les Jat (az jumla-yi *Jatdn), nommd Abia'l-Qisim qui s'empara aussi de Sirdff" (Fdrs-ndma, p. 136, 11. I6-I7). La

lemon fautive du ms. de Londres est A 6mender non en khdndn,

comme l'a fait Le Strange, mais en jatdn.


sur " tous les Garmsirat, le Irahistan et les C6tes (sawdhil), de la plaine (sahrd) de Kunduran ' Lar, J.ris, Jfiyum, FMl, Kurin, *Zam et Siraf."26 L'dmigration finale des gens de Sirafvers Qais etant

placte vers

I 10o,27 nous aurions la une date acceptable pour le moment oi FMl succeda a Sirdf d6sertd comme

capitale de district. Chez les historiographes shirizis de l'age salghuride et mongol, les nisba " Fali " et " Sirdfi " sont

employdes en complement ou en valeur l'une de l'autre, ce qui confirme la substitution de vocable que nous avons notde, et recdle vraisemblablement l'indice de la migration de familles originaires de Sirif. La plus illustre des familles du FMl &tait a l'origine appelke Sirdfi. Le premier de ses membres connu

pour avoir accede aux charges gouvernementales, Majd al-Din Ismai'll er (mort en 666/1268) est parfois appelk Fali,28 comme le seront ses descendants. Sa fortune venait de biens fonciers qu'il possedait a Siraf.29 Une autre lignee de Flli, qui a pareillement porte

' date ancienne la nisba SirMfi, tirait ressource de consultations juridiques. Qutb al-Din Muhammad Sirafi Fdli30 racontait que " la premiere chose

qu'il avait gouite au monde 6tait le fruit des legons et desfatwd", son pere, Safi al-Din Mas'fid Sirdfi, couramment appel6 F~li,31 ayant pu l'dlever grace aux dons d'un marchand revenant " des iles ".32 En depit d'un apparent flottement dans l'attribution des nisba, on remarque, en suivant les lignees sur plusieurs generations, que les familles du Fil qui, au FMl, portaient une nisba distinctive de leur origine (par exemple Sir~fi) adoptent ou reqoivent, une fois arrivees a Shiraz, la nisba Fll, identificatrice de leur provenance immediate, et que lorsqu'elles sont etablies depuis deux ou trois g6nerations dans la capitale du Fars l'usage tend a leur conf6rer la nisba Shirazi. On inclinera done a considerer que les nisba Fali et Sirdfi ne sont pas indiff6remment identiques, et que la nisba Sirdfi peut d6signer des familles de Sirdf ayant emigre FSl.

Le cousin paternel de Safi al-Din Mas'fd Sirdfi est un de ces Fll d'occasion, - un texte au moins lui donne cette nisba,33 - dont le petit-fils sera un " Shirazi ", mais dont le pays d'origine est rev6lk par sa nisba propre: 'Amid al-Din As'ad Abzari. Originaire de Abzar/Afzar, bourg et canton des Garmsirdt sur la route de Qais,34 il a fait ses premieres &tudes a FMl,35 of voila exister, a la fin du XIIe siecle dj a, une vie intellectuelle organis&e. 'Amid al-Din devint vizir des Salghurides, et sa memoire a survecu jusqu'a nous comme celle d'un lettre eminent.36 Cette culture, dont il requt a FMl les premiers 16ments, est celle de tout un groupe. Formant a Shiraz un groupe coherent des le debut du XIIIe siecle, les Fali entrent dans le cercle 6troit de cette classe dirigeante nourrie de lettres arabes, solide dans ses principes religieux, sunnite, qui s'associe a l'effort de construction d'un Etat men6 par les derniers Salghurides et qui va continuer, a travers les vicissitudes du regime mongol, a maintenir le prestige des

26 Wass~f, 6d. Bombay, p. 174-175. Kundurin (dendronyme, sur kundur, " arbre A encens ", typique de la flore du Laristan; cf. Ange de S. Joseph, Gazophylacium linguae Persarum (Amster- dam 1684), p. 156) est probablement le KundurAn situ6 au NE du port de Charak (FJ VII, p. 191). J'incline A identifier "J.ris ", 6num6r6 A la suite de Ldr, A Jirds (Mufid Mustaufi Yazdi, XVIIe s., dans Farhang-e Irdn Zamin VI/2-3 (1337 s./ 1958), p. 174), aujourd'hui Girdsh (FJ VII, p. 199). Cette localit6 est qualifi6e de village dans un texte du XIVe s. r6f6rant au XIIe s.: " le dih G.rish, proche de Khunj " (cf. infra, note 39). Je corrige en Zam le " R.m " de la litho- graphie de Bombay; c'est 6videmment le nom indigene dont la forme arabis6e est Jamm; il est dans Yaqfit sous la forme Azam (Mu'jam al-bulddn, I, p. 233).

Le fragment d'inscription qui vient d'6tre d6couvert A Siraf (Iran VII, p. 44) nous transmet le nom d'un de ces chefs Jat du XIIe si cle (renseignement queje dois A I'amiti6 de S. M. Stern).

27 Ibn al-Mujawir, Ta'rikh al-mustabsir, 6d. L6fgren, II, p. 287. 28 'Isd b. Junaid Shirazi, dans la traduction persane du Shadd

al-izdr (connue sous le titre de Hazdr-mazdr, 6d. Shiraz 1320 s., p. 143): *Srdffi, al-mashhir bi Fdlf ".

29 Mu'in al-Din Junaid Shirdzi, Shadd al-izdr, 6d. Qazwini- Eghbal (T&h6ran 1327 s.), p. 421 (min Sirdf).

30 Les textes lui donnent la nisba Sirdfi (Shkrdz-ndma, 6d. Karimi,

p. 145), Fili (Mujmal-i Fasi•zi,

sous 712 H.), Fdli Sirdfi (id., mime ann6e, variante), Sirdfi Fdli (HiAjji Khalifa, Kashf al-Zuniin, s.v. " Kashshdf").

31 Cf. Shadd al-izdr, p. 61, ligne 4. Wagssf, p. 170, le compte parmi les lettr6s auxquels (le ?) Sirdf a donn6 naissance.

32 Shadd al-izdr, p. 434-435; Hazdr-mazdr, p. 148 (avec des liberths de traduction).

33 Ibn al-Fuwati, Talkhis majma' al-dddb ft mujam al-alqdb, 6d. Mustafl Jawad, 4e partie, t. II (Damas 1963), p. 900; et cf. t. I,

p. 505. 34 Ajouter Istakhri, p. Io6, i I6, 136 (trad. persane p. Ioo, 104,

119) aux r6f6rences r6unies dans Shadd al-izdr, p. 215.

35 Dans l'6dition de Ibn al-Fuwati, Talkhis, p. 900, est imprim6 "F1 lat ", ce qui designerait une localit6 du Khuzistin, (Ydqfit, Mu'jam al-bulddn, III, p. 846; auj. Fdleh); Qazwini, qui a publid anterieurement la notice, dans ses notes au Shadd al-izdr, p. 522, d'aprbs le ms. de Damas, imprime " FIl".

36 R*f*rences group6es dans le Shadd al-izdr, p. 215, note I,

pp. 517-527 (notes). Dans I'EI, 2e 6d., l'article "'Amid al-Din al-AbzAri " reprend celui, tr s d6pass6, donn6 par Huart A la Iere 6dition (sous " Abarzi "). Plusieurs lettres extraites des Munsha'dt de Abzari ont &t6 publi6es dans 'A. M. ThAbiti, Asndd va ndmahdy-i tdrikhi (Thh6ran I346 s.), pp. 16I sqq.


madrasa du Firs. Foyer de culture arabo-islamique, FMl, Ddr al-safj, " demeure de la puret6 ",,3 envoie ' Shirdz ses theologiens et ses juristes, fournit a la capitale ses grandes families de cadis et de serviteurs de l'Etat. Les Fdli sont d'autant mieux en mesure de soutenir aupres du gouvernement leurs intirets conomiques, ceux de la route du Golfe, que Fdl se jumelle a Khunj, Ddr al-auliyd, " demeure des saints ".38

IV

Khunj devient au XIIe siecle le siege d'un khdnaqdh de derviches qui aura une grande influence sur les destindes du Laristdn. Le fondateur, Shaikh Rukn al-Din Ddnydl, n6 dans une famille paysanne du village de Jarls, etait venu s'installer ' Khunj " dans le coin d'une mosquee ", en qu&te de sa voie spirituelle. I1 y fut sollicite de s'affilier t l'ordre des derviches Kazirfini, la

.tariqa ishdqjya ou .tariqa murshidiya, et il aurait alors reconnu que le directeur de conscience avec lequel il 6tait en contact onirique constant, depuis son enfance, n'etait autre que le saint 6ponyme de l'ordre, Shaikh AbUi Ishlq Kazirini. Quand il eut requ " le tambour et l'ttendard " des membres de l'ordre, les gens de Khunj lui prWtbrent consid ration et il recruta des disciples Ensuite seulement, il aurait sejourne6 a Kdzirfin et visit6 le tombeau de son maitre.39 De ce recit idifiant et suspect on retiendra essentiellement que DanyMl fut affilie a l'ordre Kazirfini (une source le dit disciple de 'Ali Khafri, disciple de Shaikh Abti

Ish.q),40 et qu'un khdnaqdh kaziriini fonctionnait ' Khunj au XIIe siecle.

Un autre recit edifiant rapporte une vision au cours de laquelle DanyMl reconnut que Shaikh Aba Ishliq ttait " le shihna de tous les shaikhs " (un blame est decochi au passage a l'adresse du saint de

Shiraz, Shaikh 'Abdulldh-i Khafif). On est en droit de penser que l'6pisode date des d6buts de sa carriere, du temps oi0 il 6tait attire vers les Ishliqiya sans avoir encore donne une pleine adhesion a Shaikh Abfi Ishdq. La vision, toutefois, avait pour fin de lui reveler non la priseance de son murshid mais le caractere condamnable des menses de l'agitateur religieux Mahmid-i Fadl, sur la valeur duquel il s'5tait interroge (ce qui renforce l'impression que c'Ctait bien dans la periode ofu DanyMl restait incertain de son orientation mystique). A quelquesjours de 1I un voyageur apporta la nouvelle de la mort de cet

individu,41 pour laquelle nous avons d'autre source une date, 609/1213,42 qui permet de situer DinyMl dans le temps. Il est done nd vraisemblablement autour de I190. Selon l'edition imprimee du Tdrikh-i Guzida, il serait mort en 7Io/ 1310-1311,43 ce qui est beaucoup trop tard; certains manuscrits du Tdrikh-i Guzida donnent 6Io/ I213-12I4,

ce qui est beaucoup trop t6t si on accepte que la vision de DanyMl se situe dans sajeunesse plut6t qu'au terme de sa vie. Partant de cette hypothese on ajuge qu'il ne vfcut pas plus tard que la mi-XIIIe siecle,45 ce qui est peut-etre lui mesurer chichement la longevite. D'apres le Gulzdr-i abrdr, recueil de biographies soufies compose en Inde au XVIIe siecle, il aurait envoye un de ses disciples & Delhi & l'apoque de Balban,"4 qui rigna de 1266 ' I287. Cette donnee, qui

37 Laqab entendu sur place par Sir Aurel Stein, op. cit., p. 220.

38 Le laqab de Khunj est donn6 par Nimdihi, Tabaqdt-i Mah.maid- shdhi, sous 713 H., ainsi que dans le firman de Ya'qfib Beg (ligne 14) dont il sera question ci-apr's.

39 Mahmfild b. 'UthmAn, Firdaus al-murshidfya, 6d. F. Meier, p. 468-470; 6d. Afshdr, p. 426-428.

40 Muhammad Ghauthi, Gulzdr-i abrdr, ms. John Rylands Library, fol. 45b.

41 Firdaus al-murshidiya, d. Meier, p. 475-477; 6d. Afshdr, PP- 432-434.

42 Fasih KhwAfi, Mujmal, s.a.; cf. Shadd al-izdr, p. 232-233, Hazdr-mazdr, p. o07.

43 Hamdullih Mustaufi Qazwini, Tdrikh-i Guzfda, 6d. Navi'i (T6h6ran 1339 s.), p. 675. La date ne figure pas dans l'Mdition en fascimile publi6e par E. G. Browne (G.M.S., XIV); la notice sur DAnyAl, p. 793, n'est pas indiqu6e dans la traduction anglaise abr6g6e p. 21 6-217).

44 Ms. B. N. Paris, Suppl6ment persan 1977, fol. 334a.

'5 Muhammad Amin Khunji, apud A. Iqtiddri, Ldristdn-i kuhan.

(T6h6ran 1334 s.), p. I70. A Khunj, bien que les familles notables tiennent A se rattacher A Shaikh DanyAl, les traditions locales ne savent rien de lui. On considare qu'il aurait v6cu dans les premiers siecles de l'Islam. Un Tadkhira-yi Shaikh

Ddnydl aurait fait partie des archives de la famille DAnydli, qui aurait du^ s'enfuir A Fin, pres de Bender-Abbas, sous N~dir- Shah. Je n'ai pu (en 1953) trouver trace d'un tel tadhkira A Fin. Il me fut dit ensuite par un habitant de Girdsh que le tadhkira 6tait conserv6 dans un imamzdda situ6 entre Tirum et

DfrAb. 46 Dans le ms. de Manchester la nisba est 6crite " Khalji ", ce

qu'explique ais6ment l'origine indienne de la copie. Qu'il faille 6mender en " Khunji " est garanti par le contexte (le shaikh envoie Delhi un de ses disciples natif du Ldr, zddbam-i Ldr). Le laqab Rukn al-Din est confirm6 par l'inscription de

Khunj de 849 H. (voir ci-apres, n. 52). Il n'est pas exclu que Ghauthi emprunte son information A la version mise au net des annales de Nimdihi.


parait s^re,47 n'autorise pas a conclure absolument que Ddnydl 6tait en vie apr's I266, puisque son

propre cas enseigne que les derviches 6taient soumis a la direction onirique de guides qu'ils n'avaient jamais rencontres en ce bas monde: son maitre Shaikh

Abfi Ishaq Kizirfni 6tait mort en 1035, environ un siecle et demi avant sa naissance.

Comme dans toutes les loges de derviches, on accueillait au khdnaqdh de Shaikh Danyil les voyageurs et on les nourrissait. Que le khdnaqdh de Khunj soit devenu une des quatre grandes " nappes " (sufra) du Fdrs, au meme titre que les fondations de Shaikh 'Abdulldh-i Khafif a Shiraz, de Tawuis al-haramain a Abarqfih et de Shaikh Abfi Ishdq a Kdzirtin,48 en dit long sur le prestige du lieu, sur la frdquentation de la route et sur les moyens dont disposaient les shaikhs. Ibn Battita dit de Shaikh Muhammad Abil

Dulaf, un successeur de Shaikh Dinyal: " il fait des dons consid6rables, donne des vetements et des chevaux a beaucoup de gens, est genereux & tout voyageur. Je n' ai vu personne qui lui soit com-

parable dans ce pays-l'."49 Les relations de l'ordre kaziriini avec les milieux commergants, avec ceux du negoce maritime en

particulier, sont connues. Ses relations avec les milieux gouvernementaux, qui n'ont pas et6 mises en evidence avec la meme nettete, n'6taient pas moins 6troites. Au carrefour des pistes das Garmsirat, vers Qais d'une part, vers FMI et Shilau d'autre part, vers Ldr et Ormuz enfin, Khunj semble s'etre d6velopp6 rapidement. Peut-etre FIl, beaucoup moins bien place, efit-il ressenti la concurrence de cette ascension si l'unite du Khunj et du FMl n'avait et6 fondde, autant que sur l'axe routier coupant leur territoire, sur l'identit6 de leur classe supdrieure. La classe des notables, a Khunj, parcourt le

mime cursus honorum chirazien que les Fali, bien qu'elle y entre un siicle plus tard. L'action du

khdnaqdh de Shaikh DanyMl fut conforme a la politique poursuivie dans le milieu Fdli, avec l'appui du

gouvernement de Shiraz. Par un paradoxe apparent, elle visait a affaiblir Qais.

V

Sous le regime mongol Qais avait recouvrd l'autonomie de fait que lui avait fait perdre, en I230, l'annexion a l'Etat salghuride. La compagnie des marchands Tibi en avait fait son fief. Son chef, le Malik al-Islam Jamal al-Din Ibrahim, fermier de l'imp6t en Fars, en arrivait a dominer toute la

province. Le seul de ses adversaires qui soit nommdment d6sign6 par les sources se trouve etre 'Izz al-Din Muzaffar Shirazi, le petit-fils de 'Amid al-Din Abzari, c'est-a-dire un homme du milieu fali. 'Izz al-Din suscita des enqu tes contre le prince-marchand de Qais, et obtint sa destitution sous le

ragne de Gaikhatu. Durant l'interri~gne de Baidu, le Malik al-Islam rdussit a faire nommer son frbre au gouvernement du district c8tier. 'Izz al-Din en profita pour le d6noncer a Ghazan, mais un retour de fortune des Tibi causa sa perte. Livr6 a son adversaire, il fut mis a mort en novembre 1296.50

Vers I300 les derviches de Khunj apportbrent au seigneur d'Ormuz un concours de grande port6e en obtenant des Tibi rticents la cession de l'ile de Jarin aux Ormuzis. Le shaikh de Khunj qui procura ses bons offices est appel6 par Joao de Barros " Xeque Doniar "; Teixeira (mieux informd que Barros de l'histoire d'Ormuz) le nomme Shaikh Isma'il, legon prdf"rable, vu qu'il est impensable que Shaikh Danyal ait vicu encore en I3oo. Shaikh Isme'il 6tait g videmment un de ses successeurs a la tete de la communauti murshidie de Khunj. On sait combien la valeur stratgique de Jariin favorisa les ambitions des princes d'Ormuz. En raison du service rendu la dynastie manifesta sa reconnaissance aux shaikhs de Khunj jusqu'a l'epoque de la domination portugaise. La rente institude en faveur du khanaqdhd de Shaikh Dnyl &tait encore vershe au XVIe sitcle. Le roi d'Ormuz Qutb al-Din Tahamtan (I320-1347) fit idifier un mausolde sur la tombe de Shaikh Dtnyil.5' Au sikcle suivant, en 849/1445- 1446, le roi Fakhr al-Din Tiirin-Shah, intronisi depuis peu, fit faire de nouveaux travaux, ainsi qu'en

47 Gulzdr-i abrdr, l.c. 48 .Hamdulldh 1M ustaufi, Tdrikh-i Guzida, d. Browne, p. 793;

6d. Navd'i, p. 675. 49 Ibn Battfita, )d. Defremery, II, p. 242; trad. Gibb, II, p.

406.

50 Wassdf, p. 338. Pour la filiation, Rashid al-Din, Tdrikh-i Ghdzdnt, 6d. Jahn, p. 105 et io6, 6d. Baku, p. 311 et 313, et Ibn al-Fuwati, Talkhis, t. I, p. 362.

51 Barros, Asia, II, ii, 2 (6d. Cidade, II, p. 51-52); Pedro Teixeira, Relaciones (Anvers I6Io), *II, p. I2; cf. JA (I953), p. 95.

LA SURVIE DE SHILAU ET LA ROUTE DU KHUNJ-0-FAL 27

temoigne une inscription du minaret, seul vestige en partie intact d'un ensemble architectural aujourd'hui ruiniforme.52

Le pouvoir des Tibi prit fin dans les ann'es 1320. Rukn al-Din Ahmad, qui etait vdl du Khunj-6- Fdl, dut se replier a Qais vers 1326-1327, of il fut assassine.53 L'occupation de l'ile par les Ormuzis, vers 1330, leur porta le coup final. Qais cessa dorinavant d'avoir une importance economique, tout le courant du commerce international transitant par Ormuz. Quelle allait etre l'incidence de l'abandon de la route de Qais sur l'avenir du Khunj-6-Fil?

VI

La route des c8tes decrite par les geographes arabes du Xe silcle joignait Shiraz a Sirdf par un trace direct, descendant droit du nord vers le sud, via

Firizibtd et Jam, long de 60 farsakh.54 Cette

route fut ddlaiss'e au XIe sikcle, avant meme que les gouverneurs seljuqides du Ftrs aient renonc6 a r'tablir la pr&eminence de Siraf: Ibn al-Balkhi, au d6but du XIIe si*cle, decrit deja un autre itineraire. Au nombre des causes qui motiverent l'abandon du vieux trac6, on peut songer a d'6ventuelles detiriora- tions de la chaussee, telles que, au passage du Mand, fleuve qui est ingueable lors des hautes eaux,55 l'effondrement du pont ancien dont M. Vanden Berghe a vu les traces.56

L'itineraire jalonne par Ibn al-Balkhi, long de 86 farsakh, reprisentait un d6tour d'environ 150 kilometres (si on adopte pour lefarsakh la valeur conventionnelle de 6 km). De Firfizdbdd il bifurquait vers l'est, par la vallie de la riviere de Simkan rejoignait celle du Qara-Aghach/Mand, qu'il longeait sur la section nord-sud de son cours, puis gagnait Siraf en cinq etapes, dont les quatre dernieres ne sont malheureusement pas detaill6es par Ibn al-Balkhi.57 Ce trajet n'itait pas nouveau t proprement parler. De Shirdz a Firtizabtd il coincidait toujours avec la vieille route " sassanide ". Il empruntait ensuite d'autres trongons de routes anciennes, non signalkes par les geographes arabes, car ceux-ci ne decrivent que le r6seau routier rayonnant en 'toile a partir des capitales provinciales de l'empire abbasside. Que la vallke de Simkin, voie naturelle pour la grande rocade preislamique dont on d6cele gt et li des elements, que la vallke du Qara-Aghach aient ete amenagees, des ponts anciens (pont d'Isfal, pont de Zagh, pont de Dasht-i Dal, Pul-i 'Arfis)58 en temoignent sans 6quivoque. Au sud de Laghar la route du XIIe siecle rejoignait

" Kurdn la midiane Ddrdb-Siraf.59 Ainsi, pour contourner des obstacles dont la nature nous echappe, la route " seljuqide " ne taillait pas un trace inedit: ellejoignait dans une combinaison nouvelle des sections de routes preexistantes. La mort ne saisit pas les grandes voies de communication d'un terme h l'autre de leur parcours: la paralysie n'affecte d'abord que tel ou tel segment, tandis que tel autre capte, et canalise sur une duree plus ou moins longue, la nouvelle orientation des courants d'6change.

Le trongon Simkdn-Ldghar de la route " seljuqide " resta un des elements de la route des C6tes de la periode salghuride et ilkhanienne, qui ne se greffait plus a Firiizabad sur l'ancien trace, mais joignait en

12 Inscription in6dite. L'6difice est connu sous le nom de Masjid-i Shaikh DAnyal. Une tombe an6pigraphe qui a &6t d6gag6e passe pour 8tre celle de DanyAl lui-meme.

53 Cf. JA (1953), p. o04- "4 Isakhri, p. -34; Muqaddasi, p. 454. On avait toute raison de

penser que l'exploration de cette route r6serverait d'heureuses surprises A I'arch6ologue (cf. mes remarques, JA (1956), p. 129), ce qu'a confirm6 le voyage fait en 1960 par M. Vanden Berghe, voir aux pp. 163-171 de son article " R6centes dtcouvertes de monuments sassanides ", Iranica Antiqua I

(x96I). 5 A. Wilson, S. W. Persia, Letters and diary of a young political officer

1907-1914 (Londres 1942), p. 179; K. Lindberg, Voyage dans le Sud de l'Iran, p. 129 et p. 253.

56 Vanden Berghe, op. cit., p. 171 (A Gabr, en amont des con- fluents du Rid-i Shfir et du Rfd-i Ddr al-mizqn).

67 Ibn al-Balkhi texte p. 163; trad. p. 82.

58 Sur le pont d'Isfal, cf. FJ VII, p. 67 col. b; le pont "

qui a di^ etre restaur6 r6cemment, entre Kirateh et Zagh " m'est

signal6 par M. Vanden Berghe (lettre du 16.XI.1962). Pour l'emplacement des deux autres ouvrages, voir la carte donnde par M. Vanden Berghe, op. cit., entre les pages 198 et 199. Je dois A l'amabilit6 de M. Vanden Berghe la communication de remarquables photographies, ainsi que les renseignements suivants (lettre cit6e): " Le pont sur le Qara Aghadj au nord de Dasht-i Dal est situ6 en amont du confluent de la riviere qui vient du Kuh-i Maimand et du Qara Aghadj ... Le Pul-i Arus est beaucoup mieux conserve; il est situ6 a 20 km au sud de Qir. 11 y a aujourd'hui une piste qui conduit de Qir

'

Makku-Saifabad et de la' vers Khundj. ... La partie inf6rieure du pont est probablement de l'6poque sassanide, le reste 6videmment est islamique."

59 M6diane mentionn6e par Muqaddasi, p. 454; cf. Le Strange, Lands, p. 296.

4


droiture Shiraz au Simkdn, et qui de Laghar se prolongeait vers le sud-est en passant par Khunj et gagnait la c6te de Qais.60

Qais abandonn6, par quels chemins Shiraz serait-il relie a Ormuz, le nouvel emporium du Golfe Persique ? En d6pit, ou a cause, des efforts des princes du Shabankara pour contrbler les routes du Fars oriental, le trajet classique par Ddrib, Furg et Tarum, patit sans doute des longues guerres qui les opposerent aux Salghurides. La piste de Ddrdb ne devint ni une grande artere au XIIIe siecle, apres que le Shabdnkara fut rdduit a l'etat de protectorat mongol, ni l'artere principale au XIVe siecle, apris l'andantissement de Qais. Elle subsista certes, - Mustaufi Qazwini en 6numere les &tapes,61 mais avec le caractere de voie secondaire. Des voyageurs europeens de la piriode safavide l'ont suivie et decrite. La route de Lar, toutefois, 6tait incomparablement plus friquentee, et I'6tait depuis qu' Ormuz avait mis fin a l'hegimonie de Qais. Mais alors que la route safavide joindra Ldr t' Shiriz en passant par Jahrum, la route des XIVe et XVe siecles passe par Khunj, Karzin et le Simkn.62 La piste d'Ormuz a Ltr se poursuit jusqu'a Khunj, off elle se greffe sur la route pre'existante, au profit de laquelle s'exerce un ph6nomene de " capture " du reseau en formation.

Ainsi, au contraire de la vraisemblance geographique, le deplacement du centre economique du Golfe Persique vers l'est ne provoqua pas l'abandon de la route du Khunj-6-Fdl, ou du moins cet abandon ne fut que partiel, et Khunj n'en fut pas affect6. Quant h FM, il en tira probablement profit, nonobstant sa position de plus en plus excentrique par rapport a l'axe majeur du trafic.

VII

Le credit des shaikhs de Khunj a la cour d'Ormuz, les accointances de la tariqa murshidfya avec les milieux d'affaires ont assurement facilit6 le maintien du passage de la route des C6tes par Khunj. Notons qu'une loge d6pendant du khdnaqdh de Khunj existait " Lar lorsque Ibn Batfita y passa.63 D'autres intirfts 6taient en cause, non moins puissants. Le plus fameux representant du groupe des cadis Fdli, Majd al-Din Ismd'il II, se fit octroyer par l'ilkhan Oljaitu, - en didommagement des persecutions dont il avait ete l'objet comme sunnite, - cent villages de la vallee de Simkdn, un des districts agricoles les plus riches du Fars.64 Or le Simkdn, nous l'avons vu, etait traverse par la route de Khunj. Il est connu par nombre d'examples que les hauts notables effectuaient systematiquement des acquisitions immobilikres le long des grandes voies commerciales, au bord desquelles ils construisaient toutes sortes d'6difices d'usage public. Bien que rien de precis ne puisse, en l'6tat des connaissances, etre rattach6 't ce fait, on notera qu'un des frbres de Majd al-Din Ismd'il II, le cadi Rfh al-Din Ishtaq, 6tait expert dans l'art de creuser des qandt, de construire des edifices, des mosquies, des madrasa, des ponts et des ribdt.65 Des investissements comme ceux des Fdli portaient a defendre la route du Khunj-6- FMd.

La date approximative de la construction de la " nouvelle " mosqu'e de Sirdf correspond bien au moment oi' la littirature historique nous y laisse entrevoir un certain renouveau. Libere de la tutelle de Qais, le FMl jouit d'une autonomie politique moins fragile, et sur le plan iconomique, Shildu se ranima

60 Le trajet au-dela de Khunj n'a pas 6t6 rep6r6. Le Strange avait du renoncer A restituer les noms des 6tapes 6num6r6es dans le Nuzhat al-qulib. Le premier relais apres Khunj 6tait A la " Citerne du Jujubier ", Ab-anbar-i KunAr (plut6t que KinAr, comme le dit Le Strange, trad. G.M.S., p. 176). Pour le second, Le Strange a adopt Ila lecture Hurmuz, avec une identification totalement inad6quate (p. 176, note 2). II avait d'abord lu Haram (cf. JRAS (1902), p. 762), choix qu'il aura modifiC sous l'influence de la variante " Marmaz " (ibid., en note). Une autre variante qu'il indique ," Siram ", i'efit mis, au sens propre, sur la bonne voie. La graphie et l'6vidence cartographique l1vent toute incertitude: c'est Bairam (Fdrs- ndma-i

Nd4siri, II, p. 288; FJ VII, p. 44). Cette identification suffit A montrer que la route de Qais laissait A l'ouest la route de Khunj A Fdl, dont je ne connais pas les 6tapes, que Stein, op. cit., p. 217, mentionne aussi sans fournir de d6tails, et que je

pr6sume 6tre identique A la piste commerciale de Khunj A Tahiri (= Sirdf) signal6e, mais non d6crite, par A. Iqtiddri, Ldristdn-i kuhan, p. 128.

D'apres les renseignements que je recueillis A Khunj en 1953, la piste alors la plus fr6quent6e 6tait celle menant de Khunj A Gdvbandi, via Hanganfi (petit qal'a A 3 km A l'est de Haftavan; port6 sur la carte du shahristdn de Lir, FJ VII), Khalili (FJVII, p. 90), Shirinfi (id., p. 148), Mir Hasani (id., p. 230: Mir Hoseini), LAmar (pron. locale pour Limard, id., p. 209).

61 Nuzhat al-quliAb, 6d. Le Strange, p. 187, trad. p. 177. 62 Ibn Battata, 6d. Defr6mery, IV, p. 311 . 63 Ibn Battita, II, p. 240; trad. Gibb, II, p. 405. 64 Ibn Battfita, II, p. 6I; trad. Gibb, II, p. 304-305. 65 Shadd al-idr, p. 427-428; Hazar-mazdr, p. 146 (ne parle pas des

ponts).


comme debouch6 du Khunj-6-Fdl. Ibn Battiita raconte etre all de " Khunjubal " s'embarquer & " Qais qu'on appelle Siraf" pour se rendre a " Bahrain ".66 L'identification de tous ces termes, dont le dernier ne nous retiendra pas, offre matiere a discussion. Par " Khunjubal ", le voyageur entend Khunj, car il y visita le tombeau de Shaikh Ddnydl. Les commentateurs se sont partag6s sur le point se savoir s'il designe reellement Qais, quoiqu'il passe sous silence sa situation insulaire. Ce qu'il en dit n'est pas tellement loigne de la description de Yaqfit, qui avait vu Qais de ses yeux,67 et il est plausible qu'en 1331 les voyageurs ta destination de la c6te du Bahrain soient encore allis s'embarquer & Qais. Tout le morceau se lit de fagon plus satisfaisante si on suit M. Hrbek dans sa chronologie rigoureuse des peregrinations de Ibn Battifita, et si on admet que celui-ci n'a fait une brive visite 'a Sirdf qu'en 1347.68 Que l'authenticite de cette partie de la narration soit douteuse n'enlkve rien a la qualite de son information. Meme si on le soupgonnait de n'avoir vu personnellement ni Qais ni Shildu, il est bien certain que le nom de Sirdf ne vient pas dans son recit comme une reminiscence littdraire. Shildu 6tait bien le port d'embarquement pour les voyageurs qui de Khunj voulaient se rendre a Qatif. A defaut de Ibn Battita, on connait un autre voyageur qui est pass' par " Shildu ", vers 1430, venant de Qatif et allant a Shiraz.69

Il est vrai Shildu n'est pas mentionni dans les listes des places du Golfe sujettes des princes d'Ormuz. Ces listes fort incomplktes ne citent que le point extreme de la domination ormuzie sur la c6te iranienne: Bandar Mdchfil sous Qutb al-Din Tahamtan (1320-1347), Dairab sous Saif al-Din (I417/I8-I437).7o Dans les annales de Nimdihi, sous 758/1357, Shildu est donne comme limite occidentale du royayme d'Ormuz,71 mais sans doute n'en faisait-il partie que comme seigneurie semi-independante. Non que les princes n'aient veille a tenir tout le littoral. Ils s'imposerent a Dairdb, port de quelque importance a l'ouest de Shiliu, - c'etait l'ancien Najiram, et le debouche du Dashtistan, - o t ils eurent maille a partir avec des chefs locaux. II n'est donc pas surprenant qu'on y signale la nomination d'un gouverneur par le roi d'Ormuz. La situation etait diff6rente a Shilau. Shildu relevait de FMl, et les Fill entretenaient d'&troites relations avec le royaume d'Ormuz.

Gens du Khunj et du Fdil etaient nombreux parmi les cadres dirigeants du royaume d'Ormuz. Quelques-uns sont connus. Shaikh 'Afifal-Din Isrd'il Khunji, fils d'un des " saints " (auliyd) de Khunj, Asil al-Din 'Abd al-Salam (mort en 713 H. selon Nimdihi) devint " cadi des iles "; il convoya a La Mecque un mahmal envoy6 par le roi Qutb al-Din Tahamtan, puis vint en mission diplomatique a Shiraz, ofi mourut vers 137o.72 Un membre d'une des lignies de cadis Fll, JalMl al-Din Yahya b. 'Izz al-Din Ibrahim II (1355-1425) fut cadi d'Ormuz.7 Citons encore la famille de l'historien Nimdihi, originaire du Afzar; son frere devint grand cadi d'Ormuz a la fin du XVe si'cle.74 Notons aussi que parmi les savants qui donnerent l'i'iza a Tij al-Din Kazirini, l'auteur du Bahr al-sa'ddat, figure un certain Qutb al-Din 'Abdulldh

Dmnyali, tres probablement un Khunji, qui est appel6 " gloire des

'ulamd du continent et de la mer ", ce qui semble indiquer qu'il exerga ses talents dans le Golfe Persique.75

VIII

La classe des notables Khunji et Fali, puissante & Ormuz, ne l'6tait pas moins a Shiraz.76 L'influence politique de la capitale du Firs s'est donc fait sentir directement dans le Khunj-6-Fal. Peut-etre est-ce

66 Ibn Battita, II, p. 244; trad. Gibb, II, p. 407-408. 61 Mu'jam al-bulddn, IV, p. 215-216. 68 M. Hrbek, loc. laud., p. 446-450, conclut A l'impossibilit6 d'un

passage & Sirdf en I331, mais A la r6alit6 d'un aller et retour Khunj-Sirdf-Khunj en mai I347.

69 SakhAwi, Dau' al-ldmi', VIII, p. 194, ligne 13. 70 Cf. J. Aubin, " Les princes d'Ormuz du XIIIe au XVe siecle ",

JA (1953), p. 105 note 2, p. II5. 71 Le passage et corrompu. Le ms. de Tashkent donne: mabda'-i

tflash bandar shfrdz; le ms. de Windsor: mabda'-i talash band-i shirdz ast ki dn sild iist.

72 Shadd al-izar, p. 239-240; il n'y a pas de notice correspondante dans le Hazdr-mazdr.

" Dau' al-ldmi', X, p. 215, no. 936.

74 Cf. J. Aubin, " Indo-Islamica I. La vie et l'oeuvre de Nimdihi ", REI XXXIV (1966), p. 62.

7- Bahr al-sa'ddat (fin XIVe-d6but XVe s.), ms. India Office Persan 368, fol. 479b, oui sont nomm6s d'autres maitres de KAzirfmni, dont Majd al-Din FirfizAbadi (l'auteur du Qdmis, natif de Kdrzin), et un certain Qadi 'AlI al-Din Ahlmad b. Isma'il Khunji.

76 I1 va sans dire que la documentation relative A la soci6t6 shirazie est infiniment plus riche que celle que nous poss6dons sur Ormuz. Aussi ne peut-il en etre trait6 dans les limites du pr6sent article.


pour cette raison que nous ignorons & peu pres tout des seigneurs locaux et des conditions de leur reconnaissance par le gouvernement central. Peu fournies en renseignements sur les Garmsirdt, - region lointaine, mal accessible, quasi 6trangere aux Iraniens des villes du haut plateau, - les chroni- ques ne leur accordent mention qu'a l'occasion des dissidences qui s'y manifestent. On a vu qu'un des Tibi etait gouverneur (vdli) du

Khunj-6-Fdl vers 1325. Il dut abandonner ce poste apres qu'un de

ses freres, qui &tait le chef de la famille, fut mis a mort aN Tabriz en 1325. 1 Le fait montre bien que la

situation au Kunj-6-Fdl d6pendait des d6cisions prises au gouvernement central, soit " Shiraz, soit le

cas echeant a la cour mongole. Cela reste vrai meme apres que la chute des Tibi et l'effacement de Qais ont enlev6 a la conjoncture politique locale sa puissance de repercussion jusque dans les plus hautes spheres gouvernementales.

Aucun texte ne permet de discerner les effets dans le Khunj-6-Fal des mouvements de revolte contre l'autorit6 de Shiraz qui secouent les Garmsirat et oi les sources mentionnent essentiellement l'insub- ordination des Laris. Au ddbut de regne de Shah-i Shuja', le souverain muzaffaride du Fars, " le Lar et les autres garmsirdt se rebellerent. (Le souverain) marcha contre les chateaux-forts de ces r6gions avec des troupes ". Cependant l'exp6dition fut bient6t suspendue, et les Laris rentrerent d'eux-memes dans l'obeissance.78 N'anmoins le Lar n'est pas compt6 parmi les districts sur lesquels s'etendait

l'autoriti de Shah-i Shuja', alors que le Khunj-6-Fal y figure.79 En 1393, le Khunj est au nombre des districts oih 'Umar Shaikh, fils de Tamerlan, prince apanag6 du Fars, soumet des chateaux tenus par des garnisons muzaffarides.80 En 1396, la route commerciale du Khunj-5-Fal vers le Golfe est suivie par une des colonnes timourides participant a l'offensive contre le royaume d'Ormuz, - " par le chemin de Karzin et Fal, (elle) alla au bord de la mer (kindr-i daryd) ",81 - mais rien n'est connu des operations locales dans la region c6ti0re du FIl. En 1411-1412, a la suite de la revolte de Tarum, soutenue par le

prince de Lar, une colonne timouride est envoyde dans les montagnes du " Hung-6-Fil ".82 Le terme revet ici un sens gdographique, et rien ne laisse presumer que les autorites du district soient compromises dans l'insurrection.

II convient d'ailleurs de ne pas confondre les autorites des petits centres urbains et les chefs de tribu avec lesquelles elles ont a 6tablir un modus vivendi; particulierement insubordonndes dans les rdgions des Garmsirdt, les tribus sont actives jusque sur le littoral, oh les princes d'Ormuz sont prises avec elles, a Dairab par exemple.83 L'histoire du problkme tribal dans les Garmsirat a l'apoque muzaffaride et timouride meritera une etude particuliere.

IX Le maintien de l'autorite du gouvernement de Shiriz a Khunj au XVe siecle est atteste par deux

documents, l'un de la chancellerie timouride, l'autre de la chancellerie Aq-qoyunlu, 6mis en faveur des shaikhs d'un khdnaqdh. Un dipl8me anonyme et non date, quej'attribue, pour des raisons queje ne puis developper ici, au timouride 'Abdullah, fils d'Ibrahim Sultan, prince du Fars de 1435

% 1447, renou-

velant un dipl6me decerne par son pere au defunt Kaka Siraj al-Din Hasan, confbre au fils du disparu, le Shaikh al-Islam Shams al-Din Muhammad, la direction du khdnaqdh de Kaka Shams al-Din Muhammad Abni Najm et de l'oratoire (khalvat) de Kdkd Fakhr al-Din Abti Bakr, a charge pour lui de

g6rer ces fondations, de guider les disciples et de nourrir les voyageurs musulmans (it'dm-i sddir-i-varid-i ahl-i isldm).84 La tradition locale, encore qu'elle ait conservI le souvenir de l'existence des kdkd

" Cf. JA (1953), P. 104-105.

8 Natanzi, Muntakhab al-Tawdrikh-i Mu'fni, 6d. J. Aubin, T6hfran, 1957, P. 187 (ce passage ne figure que dans la premiere version de la chronique). Cf. J. Aubin " Rf~6rences pour Lar m6di6vale ", JA (1955), P- 495.

79 Selon Faryfimadi, le continuateur de Shabankara'i, Shih-i Shuja' gouverne " le garmsir de Shiriz jusqu'a Ormuz, le vildyat de *Purg-6-Tirum, le vildyat de Hung-6-Pal et le Shabank~ra " (ms. Yeni Cami 909, fol. 278a).

80so Cf. JA (1953), p. II .

81 Sharaf al-Din Yazdi, ;afar-ndma, 6d. Calcutta, I, p. 808-809; cf. JA4 (1953), p. 1 12 (avec erreur quant A l'ann6e).

82 Voir JA (1955), P. 498. 83 Muhammad-Shah d'Ormuz guerroie contre Amir 'Ali

Turkmini qui occupait Dairab et I'en ddloge (Nimdihi, sous 790/1388).

84 Munsha'dt-i afddil-i ndmddr, ms. Istanbul Universitesi F 399, foll. I I7a-123a (ism et kunya du fondateur au fol. I I7a, dans le titre: hukm-i shaikhakhtyat-i buq'a-yi Kaka Muhammad Abt Najm; son laqab au fol. 12ia: khdnaqdh-i hadrat-i muqaddasa-yi Shamsfya).

LA SERVIE DE SHILAU ET LA ROUTE DU KHUNJ-O-FAL 31

(litt6ralement " grand frre ,")85 de Khunj, est d'un faible secours. Elle les rattache a un certain Shaikh Hlaji Muhammad Khunji, inhume dans la belle mosqu6e cath6drale de Khunj, qui serait mort en 786/1384-1385,86 mais dont les annales de Nimdihi placent la mort en 810/1407-1408. Car il ne nous parait pas douteux que ce personnage v6ndrd ne soit identifiable au Shaikh HIaji Shams al-Din

Muhammad Abi Najm, inhume a Khunj, dont Nimdihi rapporte dans l'obituaire de l'an 8io que, bien que n'ayant requ aucune education il avait une grande reputation de saintet6; ses le?ons sur le Coran 6taient celkbres, au point d'avoir incitd Sayyid SharifJurjani, le grand thdologien de Shiraz, a faire le voyage du Garmsir.87 On retrouvera en lui le Kdkd Shams al-Din Muhammad Abii Najm du

dipl6me timouride. C'est aussi en faveur d'un shaikh de la famille Abti Najmi, Shaikh ... b. (?) Shaikh Muhibb al-Din

'Ali Abti Najmi, que le souverain Aq-qoyunlu Ya'qtib Beg, par un firman dat6 de Tabriz, le 3 jumadd II 883/Ier septembre 1478, soit six semaines seulement apres la victoire sur son frere Khalil Sultan qui lui assurait le pouvoir, confirme des privileges accordds aux Abai Najmi par Uzun Hasan: le village de

Bighird, qu'il possede en entier, et d'autres terres situdes dans le buluk de Jizirg-5-Bushkan sont exemptees de taxes.88 Le nom du bendficiaire de l'acte n'est pas lisible, car le document a 6te gratd.89 II est

probable qu'il s'agit de Shams al-Din Muhammad III Abii Najmi.90 Nous savons par Ibn Battita qu'au milieu du XIVe siecle existait '

Khunj, en sus du riche khdnaqdh de Shaikh Abti Dulaf Muhammad, une autre loge de derviches chafdites, qui se livraient 'a des macera- tions intenses tout en se montrant accueillants aux visiteurs.91 S'il ne fait pas de doute que le shaikh

Abfi Dulaf Muhammad 6tait le successeur de Shaikh Dmnyil, puisque le mausolde de celui-ce se trouve

dans le khdnaqdh dont Abti Dulaf 6tait le sup6rieur, on peut se demander a laquelle des deux confrdries du XIVe si'cle se relie celle des kdkd Abii Najmi du XVe. Il se peut que leur loge ait 6td une fondation distincte des deux premieres. S'il n'y a eu que deux grands khdnaqdh, celui des Abfi Najmi n'est vraisemblablement pas celui des Dtnydli. HJji Shams al-Din Muhammad et Kdkt Fakhr al-Din, en effet, passent pour etre ensevelis dans la mosquee cath6drale de Khunj, alors que Shaikh Dinyal l'est au

Masjid-i Ddnyali.92 A en juger d'apres les seuls textes qui soient connus, le couvent le plus c61kbre de

Khunj etait au XVe siecle le couvent des Abti Najmi. Qu'ils aient eu de bonnes relations avec Ormuz est

85 Le mot est usuel en ce sens en lri (not6 en I953; et cf. A. Iqtidari, Farhang-i ldristdnt, Teh6ran, I334 s., p. 154). Un kdkd shirdzi visiteur de Sayyid 'Ali Hamadani, cf. J. K. Teufel, Eine lebensbeschreibung des Scheichs 'Alf-i Hamaddni (Leiden 1962), p. 107.

86 Les notes prises par moi A Khunj au debut de 1953 corres- pondent aux infobrmations publi6es par M. Khunji en appendice au fort d6cevant Ldristdn-i kuhan de A. IqtidAri, aux pp. 174- 175. Outre Shams al-Din Muhammad, la tradition locale connait un KAk~ Fakhr al-Din en qui on est tent6 de retrouver le Fakhr al-Din Abfi Bakr du dipl6me timouride.

87 Nimdihi, Tabaqdt-i Mahmad-shdhi, sous l'ann6e. 88 Document en possession de la famille du cadi de Khunj; je

cite d'apres ma lecture faite sur place. Le cadi poss6dait egalement une transcription faite vers la fin du siecle dernier par Karamati, I'auteur du Tdrikh-i dilgushd-yi Ivaz, sur lequel voir REI XXXIII (1965), p. 155. Bighird est un village voisin de Khunj (FJ VII, p. 44). Le Bfishkan se trouvait beaucoup plus A l'ouest, cf. Le Strange, Lands, p. 260; c'est aujourd'hui un dihistdn du bakhsh Khurmfij, cf. FJ VII, p. 38. Le Jizirg-6- Bfishkin figure dans la liste des divisions administratives du Fdrs timouride, Shams al-siydq, ms. Ayasofya 3986, fol. I27b.

89 Le nom de DAnyal a 6t6 surimpos6 grossierement, suivi d'un wdw. Un autre document ant6rieur A l'6poque safavide me fut montr6 A Khunj (appartenant A Shaikh 'Aziz DAnydli, propri6taire terrien), dont la date avait 6t6 falsifi6e avec une maladresse aussi patente en 543 H./I 148-1149; c'est un acte de vente de terres de Ra'is Nir al-Din b. Ra'is Shafi' de Arad (cf. FJ VII, p. 8) A Shaikh Majd al-Din b. Shaikh Ahmad Danyali.

90 En effet, dans une lettre 6crite vers 1475-1480, Nimdihi exprime son respect A un Kakd Shams al-Din Mulhammad Abfi Najm qui est un shaikh, et qui ne peut 6tre que celui de Khunj (cf. J. Aubin, REI XXXIV (1966), p. 62, note 3).

La g6n6alogie de la famille s'6tablirait done comme suit:

IHji Shams al-Din Muhammad (m. 786/1384-85 ou 810o/I407-08)

Sira-j al-Din Hasan (m. entre 1435 et 1447)

Shams al-Din Muhammad II

Muhibb al-Din 'Ali

Shams al-Din Muhammad III (1478)

91 Ibn Battita, ed. Defr6mery, II, p. 243; trad. Gibb, II, p. 407.

92 I1 est possible que Shams al-Din Muhammad Abia Najmi soit ce

jeune garcon appel6 Muhammad que Ibn Battfita rencontra

(en 1347, selon la chronologie fix6e par M. Hrbek) au couvent des ascktes chaif6ites dont son pere, qui n'est point nomm6, 6tait le sup6rieur. Mais en ce cas on renoncera A en faire le fils de Asil al-Din 'Abd al-Salim Khunji, puisque la tradition locale veut que Haji Shams al-Din Muhammad soit n6 l'ann6e de la mort de 'Abd al-Salam; qu'il ait 6te son neveu est

acceptable, encore que toutes ces traditions (cf. Khunji, op. cit., p. 172, 175) soient singulierement floues ou inexactes

(ainsi, p. 175 le pire du shaikh est appel6 Abia Nasr, alors

qu'on attend Abfi Najm; Asil al-Din est appel6 Amin al-Din).


probable." Le couvent de Shaikh Danyal semblerait au contraire etre tomb6 en decadence, ce qui pourrait s'expliquer par le refus des biens de ce monde fr6quent dans les ordres de derviches d'obser- vance traditionnaliste. Ibn Battfita rapporte que Abi Dulaf " est reputd ne posseder aucun revenu, excepte ce qui lui vient par dons des freres et des associds."94 Une telle attitude exposait les loges de derviches a une existence besogneuse et precaire, comme le montrent bien, malgr6 la notoriet6 de Shaikh Abti Ishbq, les difficultes financieres du khdnaqdh de Kazirfin. Les Shaikh Ab-i Najmi possedaient des

terres, mais a en juger par le firman de Ya'qfib Beg, ils n'6taient que de modestes propridtaires terriens en comparaison des chefs des nouveaux ordres soufis de leur temps. Meme si le khdnaqdh de Shaikh

DdnyMl sombra dans la misere, le jeu d'influences qui avait aide Khunj a surmonter l'extinction du trafic de Qais ne cessait pas de favoriser les derviches de la " demeure des saints ".

X L'essor de Lar, cependant, condamnait le Khunj-6-Fal & vegeter. Tandis que le Khunj-6-Fdtl etait

d6favoris6 par la stagnation de Shildu, vou a n' tre qu'un port secondaire sous la tutelle du monopole ormuzi, Lar tenait la route d'Ormuz et se fortifiait du d6veloppement pris par celle-ci. Vis-a-vis de la principaute d'Ormuz, les deux formations politiques du Garmsir eurent des positions diff6rentes. Le Khunj-5-Fdl entretenait des relations amicales, voire gravitait dans une semi-dipendance. La

principaut6 de Ldr eut avec Ormuz des rapports aggressifs, et dans la seconde moiti6 du XVe siacle tenta de l'annexer. C'est dire que les relations du Lar et du Khunj-5-Fal furent marquies d'un certain

antagonisme. La minceur de la documentation interdit d'en retracer les p6ripities, ou meme de cerner avec nettet6 les grands traits de leur competition.

Un autre el6mentjoua contre le Khunj-6-Fal: le facteur d'unite qu'efit

constitu6 une petite dynastie autochtone semble bien lui avoir fait defaut. C'est a la fin du XVe siecle seulement que nous discernons au FMI la presence d'une famille puissante, celle des ra'is Fali, seigneurs de Shildu (cf. infra), soit a une

poque ofh le Khunj-5-F~l a proprement cesse d'exister comme entit' territoriale. A Lar, au contraire, une famille de malik assura aux XIVe et XVe si'cles la permanence du pouvoir seigneurial et la con- tinuit6 de la politique d'expansion. La classe de notables nombreuse, competente, influente a Shiraz, influente a Ormuz, en quelque sorte cosmopolite, qui existait au Khunj-6-Fal n'a point d'equivalent au

Lar, principaute autonome, sans grands moyens, sans 6Lite, plus tournee vers l'Inde, oii vont ses

marchands, oi umigrent ses mercenaires, que vers l'Iran intirieur. Dans la seconde moitie du XVe

si&cle, cependant, la formation d'une classe de notables cultives y est en cours.

Mahmtid b. 'Uthman Karmustiji, mort en 834/1430, est le premier auteur residant en Lar dont l'existence soit connue.95 Plus tard apparaissent des oeuvres dediees aux princes de Lar, un traite

d'6thique a Jahdnshah,96 son successeur 'Al al-Mulk une risdla compos&e par Shams al-Din Abni Nasr Muhammad, un ra'is local.'9 Une zdwiya chaf6ite 6tait en activit6 l Lar, dont les shaikhs her6di- taires &taient des Khunji d'origine.98 Mais c'est dans une petite cit6 qui n'a guere fait parler d'elle dans la periode qui precide, et qui ne fait pas partie du Lar mais du Fars proprement dit, a Jahrum, que la vie intellectuelle et religieuse se manifeste dans la second moiti6 du XVe siecle avec le plus d'6clat. Le

theologien 'Abd al-Malik Sawaji, issu d'une bonne famille de Sawa et parent des cadis Sawaji, y passe la derniRre partie de sa vie (il meurt en I148I).99 Au terme d'une vie voyageuse, un des membres de la

93 Sans prendre en consid6ration la trop hypoth6tique parent6 des

Abfi Najmi et de Qd.di 'Afif al-Din Isrd'il (supra, p. 29), notons la v6neration de la famille Nimdihi envers les Abfi Najmi (cf. note 90).

14 Trad. Gibb, II, p. 406-407.

95 Notices dans Sakhawi, Dau' al-ladmi', X, no. 561, p. 140, et dans

Abaf Makhrama, Ta'rfkh Thaghr 'Adan, 6d. L6fgren, Arabische Texte zur Kenntnis der Stadt Aden im Mittelalter, 11/2, no. 302, p. 233 (avec la nisba Kurmusti, non identifi6e par Lofgren, II/2, p. 17 de l'introduction). Karmustaj est un village sunnite A l'E-SE de LAr, cf. FJ VII, p. 185.

96 Eth6, Catalogue of Persian manuscripts in the Library of the India Office, I, no. 2216.

97 Sakhiwi, op. cit., VIII, no. 226, p. 108-1o9. Cette notice ne lui donne que la nisba " lari ", mais le nom du village dont il etait ra'is est A trouver dans la nisba de son petit-fils, A qui le Dau al-ldmi' r6serve deux entr6es (I, p. 263 et p. 265) relatives A un seul personnage; il 6tait de H.nSij, " qui est un village de LAr ". Je n'ai pas identifi6 ce village (Anakh?).

98 SakhAwi, op. cit., VI, no. 1013, P-. 306.

99 Cf. J. Aubin, " Etudes safavides I ", Journal of the Econ. and Social History of the Orient II/i (1959), PP 48, 50.

LA SURVIE DE SHILAU ET LA ROUTE DU KHUNJ-O-FXL 33

grande famille des Iji, Sayyid Mu'in al-Din Muhammad b. Safi al-Din s'y fixe pareillement. II 6tait lie avec les shaikhs Khunji de Lar dont il vient d'etre fait mention.100 Un des soufis sunnites les plus influents en Fars a la veille de la conquete safavide, Qutb al-Din Muhammad b. Muhyi al-Din Kush-

knari, dit Qutb-i Muhyi, fonda a proximite de Jahrum une communautd qu'il appela Ikhwdndbdd.101 C'est le Qutbabdd actuel, a trois lieues au nord de Jahrum, au carrefour des routes de Shiraz, de Fasa et de Lar.

Qutb-i Muhyi refusait le titre de shaikh, a plus forte raison ceux de waliou de qu.tb,102

et manifestait de vives pr6ventions contre le soufisme, ffit-ce celui des ordres se rattachant a la sunna, comme l'ordre Ni'matulldhi.103 Il apparait comme un des derniers repr6sentants, avant le deferlement du pathos de la

ddvotion duodecimaine, de ce piktisme sobre qui avait fleuri parmi les 6lites du Fars durant la p6riode mongole et qui avait fait la force morale de la classe riche. Qutb-i Muhyi lui-meme etait a la tete d'une

grande fortune.104 Le motif pour lequel il choisit de s'installer au Jahrum n'est pas evident. Sa nisba nous apprend qu'il etait originaire de la c6te du Golfe Persique, Kushknar tant un village au sud-est du FAl, immediatement derriere la chaine c6tikre.105

XI

On sait que les voyageurs europdens de l'age safavide qui, en grand nombre, venant d'Ormuz puis de Bender-Abbas, ont d6crit la route de Lar ' Shiraz ne font point mention de Khunj, et ne prennent jamais la route de Khunj et de Kdrzin. A compter du XVIe siecle la grand'route passe par le Jahrum. Il est probable qu'on passa par le Jahrum, j'entends de fagon regulikre, des la seconde moitid du XVe sikcle. Les deux voyageurs qui sont allks de Lir Shiraz a cette epoque, Afanasiy Nikitin106 et Girolamo da S. Stefano,107 ne donnent malheureusement pas de details sur leur parcours. Non plus que Gil

SimOes, le secretaire de l'ambassade portugaise qui en 1515 rentre de Shiraz h Ormuz via Lar.1os La relation de Tenreiro, qui a fait le voyage de Lar ' Shiraz en 1523, mais qui n'a publi6 qu'en I560, pr6sente un temoignage interessant sur la route de Khunj et Karzin au moment o i elle devient une

piste secondaire. II 6crit: " Nous partimes de la cite de Lar avec le visage au nord-ouest, et marchames trois journ6es par une

terre apre et venteuse. Durant ces trois jours entiers je ne vis rien qui soit digne d'etre not6. Dans une vall6e pres d'une rivikre, nous dormimes dans un endroit inhabit6, avec grand'peur des lions. Ce

pourquoi nous ffimes toujours en veille, avec les chevaux tenus en bride, et les mules des muletiers que nous avions engag6s

' Lar mises au milieu, avec de nombreux feux alentour. Eux, avec beaucoup d'effroi, nous dirent qu'en ce lieu il y avait beaucoup de lions, et que la nuit ils tuaient les bates des caravanes qui y faisaient etape. Mais de toute la nuit nous ne vimes aucune bete sauvage, a cause des nombreux feux, de la veille que nous fimes. Cette rivikre court de l'est vers l'ouest, et va se mettre dans la mer du Sein Persique. Le lendemain nous partimes de 1l, par la meme terre de cette seigneurie. Pass6 une montagne, nous trouvames une terre peuplke de villages et de gros habitats de cultivateurs

(aldeas e lugares grandes de lavradores), et aupris d'eux des forteresses, des chateaux de pierre et des citernes d'eau de pluie. Ces forteresses et ces chateaux servent de refuge aux habitants quand ils sentent les brigands qui les viennent piller; car il n'en vient jamais moins de cent. Ces brigands sont des

Turkmines (Turquimadis), naturels de la seigneurie du Sofi (...) Pass6 la terre de ce dit seigneur de Lar, nous marchames avec le visage vers le nord-ouest. Nous entrames dans de larges plaines qui sont entre

100 SakhAwi, op. cit., VIII, no. 17, p. 37-38 (ohi il est dit qu'avant d'aller A Jahrum il s6journa A " BAr "; faute d'impression pour Lar ? ou BAr/Bal pour Fil ?), et cf. VI, p. 306.

101 Cf. un extrait de sa correspondance dans Tdd-ddsht-hdy-i Qazwint VI (T6h6ran 1341 s.), p. 156. Voir la notice de LAri, Mir'dt al-adwdr, ms. B. N. Paris, Suppl. persan 169, fol. 245a.

102 rdd-ddsht-hdy-i Qazwint VI, p. 157. 103 Ibid., p. 158. 104 Ibid., p. I57. 105 Cf. Sakhawi, op. cit., VIII, p. 2 8. Appele aussi Qasr-i kinar,

infra, p. 36, et FasT'i, Fdrs-ndmrna-yi Ndsir II, p. 292 (Qasr-i k.nar) et index, p. 23 (Qasr-i kunar). La lecture tautologique Gfisha-kinAri (E12, art. " Dawani ") est infondee.

106 Khozhenie za tri morja, ed. Moscou 1958, p. 89 (" de LAr il alla A ShirAz, douze jours ").

107 Mario Longhena, "

I1 testo originale del viaggio di Girolamo da S. Stefano ", Studi italiani di filologia indo-iranica V (1905), p. 47: de LAr " caminando per giorni xv gionsemo ad unaltra

grande cita chiamata Siraz ". 108 Cartas de Afonso de Albuquerque II (Lisbonne 1898), p. 250.


deux chaines de montagnes, a savoir l'une qui est du c68t du sud, qui va le long de la mer et sein Persique, qui n'est pas tres haute, dans laquelle il y a de grands bois de cypris, et l'autre est du c6td du nord, et court de la fort loin. La plaine est entre elles deux. Elle a en largeur, d'une chaine a l'autre, huit lieues, et en certains endroits six. Cette plaine est longue de six journees de marche. On l'appelle le Coscojarde, ce qui veut dire dans leur langue le sec jaune. Nous cheminames par elle les cinq dites

journ"es, et fimes halte pris des campements des nomades (aduares), oiu nous trouvames tout le n'ces- saire. Par ces plaines il y a beaucoup de fauves et de gibier de toute sorte; ainsi il y a des lions, des onces (qu'ils craignent plus que les lions), des loups, et des ours, qui causent beaucoup de dommage a ces campements, ce pourquoi ils sont toujours clos; ils tiennent les tentes lides les unes aux autres, et laissent au milieu un enclos dans lequel ils enferment le b6tail de nuit; a l'ext6rieur ils ont des matins

qui veillent avec des cloches attachees au cou (.....) Passe ces montagnes, nous arrivames a Chiraz... ."109

L'imprecision topographique, I'ambiguitd de l'orientation appellent le doute: l'ambassade portugaise passa-t-elle par le Jahrum pour joindre la vallee du Qara-Aghach, ou l'atteignit-elle plus en aval? Lorsqu'on remonte de Lar vers Shirdz, en ayant donc n6cessairement " le visage vers le nord- ouest ", le seul cours d'eau coulant d'est en ouest qu'on ait a traverser est le Raid-i Shfir ou Ab-i Fasa, oued fort modeste qui se jette dans le Qara-Aghach. A vol d'oiseau, les points de passage sur le fleuve ou sur son affluent sont sensiblement a la meme distance de Lar, mais la difficultd de progression au sol est plus grande pour atteindre Jahrum. Le Qara-Aghach coule du nord au sud. Cependant les detours de son lit rendent possible l'erreur de Tenreiro, dans la r6gion du Afzar par exemple, oi' il a di l'atteindre apris de fort longues 6tapes s'il n'a rdellement mis que trois jours a traverser les plateaux desolds du Laristdn occidental, " apres et venteux ".110 S'il avait pris par le Jahrum, il aurait vraisemblablement notd le franchissement des cols caractdristique de cet itindraire. Sa description correspond mieux a la route du Karzin. Tenreiro consid're la vallee ou' il campe comme relevant du prince de Lar. On peut tenir cette assertion pour fausse, nde peut-etre d'une confusion entre pays de Lir et garmsirdt, car le Afzar appartient a la zone des " terres chaudes ".111 Bien qu'il ne dise point la direction de la marche le jour suivant, le r6cit laisse a penser que le cheminement vers le nord-ouest est inter-

rompu, avant de reprendre par la suite. En effet, les Portugais d6couvrent un bassin fertile et peuplk qui se laisse aisdment identifier au Qir-6-K~rzin, apres avoir franchi la passe qui s6pare ce canton du Afzar. La vallie que les Portugais remontent ensuite en marchant vers le nord-ouest est celle du Simkdn, qui n'a pas six journdes de long, mais six lieues, les journdes s'entendant du temps mis depuis l'entr6e dans le Simkdn jusqu'a l'arriv&e a Shiraz.112 A l'appui de la description du paysage, probante en elle-m me, le toponyme - le seul donn6 entre Lir et Shiraz - " Coscojarde ". On y reconnait Ktishk-i Zard, " le Pavillon Jaune ", ou " le Pavillon d'Or ", Kfishk-i Zar ou Qasr-i Zar, lieu-dit bien connu, mais sur la route d'dtd de Shiraz a Ispahan. Je ne crois pas, cependant, a une erreur de Tenreiro (sauf dans son 6tymologie, reposant sur la confusion entre kashk et khushk, " sec "). Coscojarde est Kushksdr, au Simkdn.113

Entre Ldr et la vallie du Qara-Aghach, Tenreiro n'a rien eu a noter; aucune agglomeration, aucune zone habitee n'ont frappi son attention. Sur le bord du fleuve, la caravane campe a la belle 6toile, malgr6 le voisinage des bates fauves. Il1 n'est pas dit comment le fleuve est traverse, - a gue sans doute, car c'est a une p6riode de faibles eaux. La region de Kdrzin est habit6e par des paysans, mais les sidentaires y vivent sous la menace des Turkmanes. Au Simkin, ce territoire jadis si prospire, on se ravitaille aupris des nomades. Le passage a l'&conomie pastorale semble total.

109 Itinerdrios da India a Portugal por terra, *d. A. Baido (Coimbra 1923), PP. 12-16.

11no Sir Thomas Herbert, dont les remarques sur les cours d'eau de cette r6gion, posterieures d'un siecle, repondent A propos a la question que soulive le texte de Tenreiro, a oui dire que le fleuve est a cinq petites journ6es de Ldr (Travels in Persia, 1627-1629, d. abr6g6e par Sir William Foster (Londres 1928), p. 61).

111 Le Ktrzin est 6galement compte comme relevant du Lar, ce qui n'a jamais &t6 le cas.

112 Ibn Battfita (supra note 64) dit que la valle de Simk~ n a six lieues de long (le Fdrs-ndma-yi NJVdird, II, p. 225, lui en donne huit de long sur cinq de large). Le Nuzlhat al-qulib, 6d. Le Strange, p. 185-186, trad. p. 176, compte six 6tapes de Shiriz

t la fin du Simkin.

113 FJ VII, p. 193-194-


XII

Un tel tableau n'explique point la desaffection dont la route de Khunj est devenue l'objet. Il est ta parier que la route du Jahrum, si nous en possidions une description aussi vivante, ne se rivilerait guere mieux equipfe en caravanserails et en relais villageois. A la faveur des trois phases de l'expansion turkmene, l'ensemble du Firs a ete atteint par l'extension du nomadisme. Ce n'est 1l qu'un des

elements du problkme de l'evolution du reseau routier. La concomittance entre l'existence d'un foyer intellectuel au Jahrum (foyer que la conqu&te safavide dteint) et la modification du trajet des caravanes qui circulent entre Ldr et Shiraz suggere que le declin de Khunj remonte at plus haut. Imaginer une competition entre les Abti Najmi et les Kushknari serait timtraire et, en l'6tat de notre documentation, gratuit. La creation de

Ikhwdnmbad contribua assurement au developpement de la route, mais

n'en provoqua pas la mise en service. Sans doute, au contraire, est-ce l'existence de la route qui entraina le choix du site de Ikhwn•dbtd. Il conviendra d'attendre de possider une documentation plus etoff6e pour disserter sur les mecanismes qui ont provoque l'abandon de la route de Khunj.

Les bouleversements politiques, la perte d'influence de tel groupe social, la decadence de telle ou telle famille protectrice ont pu jouer un r61e important. On ne n6gligera pas non plus l'effet possible d'une deterioration de l'infrastructure routiere. Nimdihi, que nous pouvons consid6rer comme un chroniqueur local, bien qu'il compose ses annales a la cour du roi de Gudjarate, rapporte que le Karzin fut devast6 au milieu du XVe sikcle par un s6isme en 1440 ou 1441, et par des inondations en 1442.114 Lors de ces inondations le pont ancien construit par le souverain bouyide 'Adud al-Daula, qui franchissait le Qara-Aghach, s'effondra, min6 par le choc des troncs de palmiers et d'autres arbres que charriait le fleuve en crue115 (et, ajouterons-nous, peut-6tre 6branl6 par le r6cent s6isme). Nimdihi ne designe pas d'un nom particulier ce pont, que j'identifie au Pul-i 'Arfis, construction qui fit l'admira- tion de Stack au sikcle dernier,116 et dont les photographies prises par M. Vanden Berghe r6vivlent l'aspect impressionnant. La mise hors d'usage de cet ouvrage d'art, en d6r6glant la circulation, a dui contribuer at faire rechercher un itin6raire qui ne soit pas soumis aux al6as du niveau du fleuve.

Il est vrai qu'un document de la fin du XVIIe sikcle donne a penser que le Pul-i 'ArUs 6tait encore en service. Le 23 shawwal I Io6/6 juin 1695, un certain Amir Abai Talib, mustaufi-i ulkd-i Fdrs, qui s'interessait beaucoup au d6veloppement de l'agriculture et avait acquis de nombreux biens, constitua en waqf" sur le pont dit Pul-i 'Aris " des terres situdes " dans le buluk de Kdrzin, des dipendances du Garmsir Turc du Fdrs (garmsir-i turki-i Fdrs) ", et dans les buluk de Qir, de Afzar et de Khunj, appartenant 6galement au district susdit. Les revenus seront affectis a l'entretien des biens constitues en waqf au paiement des imp6ts, etc.; de ce qui restera, un tiers ira au mutawalli & titre de pension, un tiers a* l'entretien des nattes et du luminaire du Jami'-i 'Atiq de Shiraz, et le dernier tiers aux pauvres de

Shirdz.117 Rien n'est privu pour l'entretien du pont, qui sert manifestement de pr6texte frauduleux a l'rection en main-morte d'un bien qui est essentiellement pour l'habile fonctionnaire un patrimoine.

XIII

Creation du XIVe sikcle, le Khunj-6-Fdl survecut a la mort de la route de Qais, mais se dissocia au XVe sikcle. Le dernier emploi du terme que nous ayons rencontrd est sous la date de 814/1411-1412

(cf. supra p. 30). Les deux firmans en faveur des Abi Najmi sont adresses aux autorites de Khunj, et non t celles du Khunj-6-FSil. Les deux cantons cessbrent de former une unit6 administrative, ce qui n'exclut pas la persistance des liens sociaux. Vers l'est, l'extension territoriale du Fil fut considdrable- ment riduite par la descente des princes de Lar vers le littoral. On aura une idie de l'ampleur de leur

114 Cf. REI XXXIV (1966), p. 62, note 2. Comme Ja'fari place sous 845/mai I44I-mai 1442 les pluies sur le bassin sup6rieur du Qara-Aghach, suivies de graves inondations, et Nimdihi sous 846/mai 1442-mai 1443 la crue de son cours moyen, je pr6sume que celle-ci a lieu au printemps 1442. Cependant Ja'fari semble placer la crue au d6but de 845 H.

115 Ce pont bouyide (?) de Kirzin n'est pas connu des textes g6ographiques, le seul pont qu'ils mentionnent dans la r6gion 6tant celui de SimkSn, sur le Riad-i Simkbn (Ibn al-Balkhi,

Fdrs-ndma, 6d. Le Strange, p. 139, d'oui Nuzhat al-qulizb, d. Le Strange, p. I 17). Ce ne peut etre le pont mentionn6 par Nimdihi, car le Rfid-i Simkdn se jette dans le Qara-Aghach en amont du Karzin, et, d'autre part, le palmier ne poussait pas en amont de Simkin (Ibn al-Balkhi, loc. cit.).

116 E. Stack, Six months in Persia I, p. iI i. 117 Bibliotheque de la Facult6 des Lettres de Teh6ran, ms. 48/J

(cf. le catalogue = Revue de la Faculte des Lettres de TIhe'ran VIII/I (1339 s.), P- 466).


penetration dans les hautes vallkes intermediaires en constatant qu'a la fin du XVe siecle ils avaient etabli leur autorite sur Kushknar. A preuve, deux lettres de Qutb-i Muhyi au prince 'AlI al-Mulk, et & son vizir Nizim al-Din 'All. A 'Ala al-Mulk il demande la restitution au mutawalli du mausolde de Shaikh Fakhr al-Din AbMi Bakr Qasr-i Kinari de la portion (le " sixieme ", sadas) du mazra'a Burdtil, qui est waqfdu dit mausolee, et que les agents du prince viennent de remettre en culture. Au vizir, il notifie que le prince a renouvel6 l'exemption des biens constituds en waqf par la lignde des shaikhs (mashdyikh) de Qasr-i Kinar, et il le prie de faire appliquer la mesure.118

Fil, cependant, restait une seigneurie puissante, dont les ra'is gardaient une large clientele dans l'int6rieur. Un chateau-fort (qal'a) protdgeait la petite citd,119 oh se maintenait la tradition de culture arabo-islamique sunnitex20 qui ne disparaitra qu'avec l'implantation de l'influence safavide. Au contraire du Lar, qui restera un bastion du sunnisme, le FIl, en effet, virera au chiisme.

\ 5HiRZ DARAB FURG TARUM

QAIS

++f+++++++++ roues a ++es++es au x icle - rout-e de Qois, xII xve sitcHUes

- - - - roue d'H-urmuz

i kIhunj, x/ve-xve sikcles route de Lar Shiraz, xvH e si~cle route de Lbr 'a 5hiriz, Xv/e si~clc

Qu'en depit de sa position ecartie, de la digradation de ses liens avec Shirdz et de la pression de plus en plus forte des princes de Lar, FMl se soit maintenu comme une des petites m6tropoles du Fdrs est incontestablement dfi au fait que les ra'is du FMl tenaient Shildu. Les ra'is Fall devinrent vizirs d'Ormuz dans les ultimes d6cennies du XVe sikcle, et ils le resterent 'a titre h6riditaire jusqu'a la fin de la sujetion portugaise. Quel fait pourrait mieux attester l'importance du FMl dans la vie politique et 6conomique du Golfe? Les ra'is Fqll avaient au Bahrain de gros intirets, cons6quence 6vidente de l'activite des relations maritimes entre Shildu et la c6te arabe. Le puissant riseau des alliances fdlies assura aux ra'is une position prepond6rante dans le royaume d'Ormuz, dont ils 6taient t l'apoque portugaise les veritables dirigeants. En 1528, Tenreiro parle de " Chilaao " en ces termes: c'est " un bourg grand et tres noble (villa grande e muy nobre) ", oui les ra'is Fdli ont " beaucoup de gens et artillerie ",

" bien defendue et fort abondamment pourvue de toutes choses ".121 En 1590 Gasparo Balbi mentionne

118 Cf. Tdd-ddsht-hdy-i Qazwint, VI, p. 162-163. Bardfil (FJ VII, p. 26; Le Fdrs-ndma-yi Ndsiri vocalise Birdfl, II, p. 290, et Burdfil, II, index, p. 5) se trouva & quelques kilometres , l'est Qasr-i Kinir. Sakhdwi, op. cit., VIII, p. 218 dit que Kush- k.nar est un des villages du Lar.

119 Sakhdwi, op. cit., I, p. 172, et VI, p. 280: ville (bulda) & dix jours de marche de Shiraz; XI, p. 218: a un qal'a: II faut souligner qu'un document persan de la Torre do Tombo ne connait, en I537, Fdl que comme nom de buluk, la localit6 de ce buluk 6tant Asir (FJ VII, p. I i; Stein, op. cit., p. 217).

120 Cf. les personnages qui passent leur enfance & F&I et y com-

mencent leurs 6tudes auprbs de membres de leur famille avant de venir & Shiraz: Sakhawi, op. cit., I, p. 172 (mort 874/ 1469); VI, no. 938 (836-891/1432-1486); II, no. 658 (n6 856/1452).

121 Frederico Gavazzo Perry Vidal, " Uma nova li&io da ' Viagem por terra ' de Ant6nio Tenreiro ", Primeiro Congresso da Histdria da Expansdo Portuguesa no Mundo. Publicafies da 2a

Secfdo. II. Portugueses no Oriente (Lisbonne 1938), p. I2o; autre edition, sur un autre manuscrit, et ignorant la premiere, par Luciano Ribeiro, " A viagem da india a Portugal por terra, feita por Ant6nio Tenreiro ", Studia III (1959), P. I 17.

LA SURVIE DE SHILXU ET LA ROUTE DU KHUNJ-O-FAL 37

" Silaj " sans le decrire.122 Les auteurs plus recents ne donnent Shildu que comme un simple havre, marquant ainsi qu' un nouveau degre de sa decheance avait 6te franchi. Il le fut sans doute vers la fin du XVIe sikcle. L'exp6dition safavide qui, sous la conduite d'un ra'is FPli partit a la conquete de Balhrain au debut du XVIIe sidcle, embarqua non pas

' Shil~u mais a " Bandar 'Asali ".123

Il serait hautement souhaitable que l'exploration archeologique de Sirifsoit completde par l'examen des sites medievaux du FMl et de sa region, et par la collecte des materiaux 6pigraphiques islamiques qui peuvent y subsister.

sIs Gasparo Balbi, Viaggio dell'Indie Orientali, 6d. Venise, 1590, fol. 41b. ed. D. Pinta, Rome 1962, p. I3.

123 Jall al-Din Munajjim Yazdi, Tdrikh-i 'Abbdsi, ms. British Museum Or. 6263, fol. 98b.

39

EXCAVATIONS AT SIRAF Second Interim Report

By David Whitehouse

The second season of excavations at Siraf took place between October and January I967-68.1 During the season we continued work at the Great Mosque and in addition excavated an impressive house of the tenth or eleventh century and a complex of buildings occupied in the fifteenth century.

We are grateful to H.E. the Minister of Culture, Mr. Mehrdad Pahlbod, and the Director-General of the Archaeological Service, Mr. A. Pourmand, for permission to excavate at Sirdf. The Director of the Archaeological Museum, Prof. E. Negahban, showed continued interest in the site and gave us valuable advice. Mr. Taghi Rahbar accompanied us to Sirif as the representative of the Archaeological Service and we are indebted to him for his assistance.

As in 1966, the excavation was sponsored by the British Institute of Persian Studies. We also received generous contributions from the Calouste Gulbenkian Foundation of Lisbon, the British Academy, the British Museum, the Ashmolean Museum and the Royal Scottish Museum, Edinburgh. We acknowledge their support with gratitude.2

The excavation staff was as follows: David Whitehouse (director), Ruth Whitehouse, Alan Carter, Jan Roberts and Andrew Williamson (site supervisors), Hugh Ainsley (surveyor), Oliver Drerup (assistant surveyor), Sarah Jennings and Jennifer Sibson (finds assistants), Frances Ainger (conser- vator) and Giles Sholl (photographer). We also received help from Susanne Russell, Peter Farries, Sean McMahon and Laurence Smith.

We offer our thanks to the Director and staff of Entreprise de Travaux Petroliers Maritimes for their hospitality at Bushire; the staff of Decca Services, Bushire, for their hospitality and assistance; and the villagers of Taheri, who once more acted as our hosts and provided the major part of the labour force.3

Finally, I am personally grateful to the following for advice on material mentioned in this report: Miss Helen Mitchell of the Heberden Coin Room, the Ashmolean Museum (coins), Prof. K. A. C. Creswell (architecture), Mr. Basil Gray of the Department of Oriental Antiquities, the British Museum (porcelain) and Mr. Ralph Pinder-Wilson, also of the British Museum (epigraphy).

THE EXCAVATION

The excavation lasted twelve weeks, during which we employed up to i6o workmen. Throughout the season work continued at the Great Mosque, with the result that the whole of the main enclosure has now been excavated to the level of the latest floors and most of the extension has been cleared to original floor level. We completely excavated and then demolished the Recent Mosque. We also began to examine the numerous small buildings which abut onto the outer walls of the main enclosure. Meanwhile, at the beginning of the season, we began work a short distance west of the Sounding (Site A), where surface indications suggested the existence of several buildings. These buildings (Site E)

1 For preliminary reports on the excavation, see Iran VI (1968), pp. 163-4; Antiquity XLII, no. I66 (1968), p. 99; and Illustrated London News (June Ist 1968), pp. 26-7.

2 In addition to giving financial support, our sponsors offered a Fellowship for competition among members of the excavation staff and lent us a Land Rover. The Fellowship was awarded jointly to Ruth Whitehouse and Hugh Ainsley. The Director

of the Institute, Mr. David Stronach, gave us unfailing encouragement and advice, for which we are deeply grateful.

3 We also thank the following for lending equipment: the Department of Civil Engineering at Imperial College of Science, London; the Winchester Excavations Committee; and Messrs. Hilger and Watts Ltd.

k c

r a o

14,04t t

s ea

I~ ,j? 1•

.

It 4. ' 40a

40 d ~~I( I bo

LIL

tooY

jr _4L

.jr r

SiR it

L \ jr D 'L

1L

t Jr7<•-i 0jr I LCL

IfctI

r I v . . I L

, jrir

r L .

i

r I L

"L "o . . . . .

__--L td 0 44Lr

j r j r r w m z, jjL

1 0

O " •

.

.,.

.jr

10.V4

IL 11

jr oC rvL.

1L-. jAN I

IL 7 %0Mode ngae n e

s1L h inL0 1$Go

j L % jL 1L I L of j\

I j41 . Jr Jr

Jr-1L•t 0rIL

0IL IL

IL IL j

.

rr It. 9 ILe

"'- -," +j 49

IL Jr

j1' I

" 1L

U L tot 564

jr II

4 t 0 "

jrd r OI

It L s

: r

ILILj r/ ?L

A. IL :

ir to'

d

j 1. tL ItJ, S."

C

•_r 4

T0,4IRI

1st~

1 SC,

AL??~

jr jri

B m ..../CT

j

r -ca o- r j r o r

=

jr•• R oc IL

g o e . . . . . L IL

L.

IL w : iJ r r ro

t Lirit Itcult v4 on--ir

d1. It I LR o A l IRs I

Jr t ir

oer grvsan el

P L 2----~---4` r IL

~i6ILl ~ SCALE

r Soo~~,it~b~r%~; K 200 0 Soo ON so Feen

it 31*'. i I .. ... - T -- #%-@NL IPA w; va A

Fig. I. Stein's plan of Straf, with the positions of Sites A, B, C, D, E and F added. (Reproduced by courtesy of Macmillan & Co. Ltd.)

EXCAVATIONS AT SiRXF 41

proved to belong principally to the fifteenth century. Despite torrential rain we completed our investigation of the structures in late November. Rather than continue digging at Site E, a task which would have involved the demolition of the fifteenth century complex, we moved to a site 18o m. farther west. Here a mound of earth and rubble indicated the existence of a substantial building. Excavation in fact revealed that the mound (Site F) contained the remains of at least three buildings in a good state of preservation. By the end of the season we had excavated one of them, a private house of the tenth or eleventh century.4

The three sites are described in the following order:

I. Site B. The Great Mosque. 2. Site F. The tenth or eleventh century house (p. 48). 3. Site E. The fifteenth century buildings (p. 54).

I. Site B. The Great Mosque The Great Mosque was discovered by Sir Aurel Stein, who in I933 recorded the rubble and pier

bases of a large building approximately 8oo m. west of the modern village. In 1966 we learnt that road works threatened the site and therefore carried out a rescue excavation designed to identify the date and function of the building and elucidate the plan. The excavation revealed a mosque which, in its final form, consisted of a rectangular courtyard flanked by double arcades on three sides, with an arcade five bays deep on the fourth, or qibla, side. Abutting on to the south-east side of the mosque was a large building of uncertain form. Some time after the Great Mosque had fallen into decay, a small structure, known here as the Recent Mosque, was built in the ruined sanctuary. In the First Interim Report I suggested that the Great Mosque was begun in the ninth century, while the Recent Mosque was built shortly before the nineteenth century.5

The Great Mosque Thanks to the intervention of the representative, Mr. Khordvani, the new road was diverted to the

east of the mosque and last season we resumed work with the object of uncovering the entire structure. The main enclosure emerged as a rectangular building consisting of arcades fifteen bays deep and thirteen bays wide, with a narrow entrance in the north-east wall. The overall dimensions were 57 x 44 m. and the major axis was aligned at an angle of 320 south of west. The large building on the south-east side comprised eleven parallel bays extending eastwards from the qibla wall (Fig. 2 and Pl. Ia and b). In 1966 I concluded that the existing qibla wall was an addition and that in its original form the enclosure was only fourteen bays deep, with a length of 52 m. I also suggested that the extension was a secondary feature. During the second season we discovered that bay 5 of the main building was also an addition and we now suggest that the history of the Great Mosque was as follows:

I. The original mosque was a rectangular enclosure fourteen bays deep and thirteen bays wide. The courtyard was surrounded on three sides by a double arcade, with a triple arcade in front of the qibla wall.

2. The mosque was enlarged by the addition of a fifteenth bay at the qibla end. Probably at the same time, a second additional bay was inserted in the courtyard, providing the enlarged mosque with a sanctuary five bays deep.

3. The extension was build on the south-east side. 4. Subsequently several minor alterations were made to both the main building and the extension.

Throughout its history the mosque was built mainly of stone, although at a late date wooden columns may have been placed in the sanctuary. The roof was always of wood. The walls were faced

4 During the excavation the surveyors began to re-map the site at a scale of i : 500 with contours at 2.5 m. intervals. We hope to complete the map, which will cover a slightly larger area than Stein's plan (Fig. I), by the close of the third season.

5 "Excavations at Sirdf. First Interim Report", Iran VI (1968), pp. 1-22, hereafter cited as Strdf I. For the Great Mosque, see pp. 9-11.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 i i I I I I I i1 1

-A

SL SL

E3 11m) 0l E -G MB WELL

CISTERN DR

---

F L

IDR

JIAR

VAULTED CELLAR METR

IHA OD & DW

Period 1 Period 4 DR Drain Period 3 BD Blocked doorway MR Mimbar SL Slot

-- ,0 15

Wku l

Period~~ ~~~~ 3 D Bokddora R Mma

L Slo

Fig. 2. Site B. The Great Mosque. Plan.

EXCAVATIONS AT SIRAF 43

with dressed blocks, usually measuring about o* 25 xo. 15 x o0 15 m., laid in regular courses. In the arcades the circular piers were constructed of wedge-shaped stones radiating from the core. Both walls and piers were finished with white plaster.

The outer walls and the arcades were supported by massive foundations, I m. wide and at least

2.5 m. deep. The foundations of arcades 2, 3, 4 and 5 extended the full width of the mosque, while the foundations of 6 and 14 ran between arcades B and C. The extent of the foundation of arcade 15 is not yet known. In the lateral arcades, the foundations of B and M ran between 5 and either 14 or 15, while C and L ran to 14 or 15 from arcade 6. A comparison of floor levels in the arcades and in the numerous buildings abutting onto the outside of the mosque showed that the foundations created a platform more than 2 m. above the ground."

Above the foundations, only the bases of the piers survived in situ. However, we recovered fragments of several fallen shafts, imposts and arches, which provided useful evidence for the original appearance of the arcades. Each pier normally consisted of a square base, a circular shaft and a square impost supporting a pointed arch. We found no trace of vaults or domes and so presumably the arcades supported a flat wooden roof. Most piers had bases approximately I m. square and up to

o. 6 m. high.

However, in the north-west arcade, piers B6, B 13 and B 14 did not have bases and in the sanctuary the square bases of piers B3, C3, C4, D3 and J4 were built round circular shafts resting directly on the foundations. It is possible, therefore, that originally part, if not all the building had circular piers without bases, as in the Tarik-khine at Damghan. If this were so, the absence of piers resting directly on the foundations in arcades 2 and 6 might suggest that the addition of square bases took place not later than the construction of bays I and 5. We found four broken shafts lying on the floor of the north-west arcades. These varied in diameter from

o'84 to

o.96 m. and the largest fragment was

I -3 m. long. Two pieces retained square imposts I to I 2 m. across and 0o7 m. deep. Masonry from bay 5 showed that the arcades of Period 2 had shallow pointed arches.

Although most of the piers were made of mortared rubble, a group of bases in the sanctuary had as their core a single dressed boulder. The boulders had a flat top which was either square or circular. With one exception, all the stone bases were in arcades 2 and 3: they were bases H, I, J, K and L in 2; E, I, J and K in 3; and base D in arcade 4. The occurrence of five stone bases in arcade 2 which replaced the original qibla wall, showed that the stone bases were secondary features not present in Period I. Clearly the boulders had a special function and they may explain Yaqfit's reference to wooden columns in the mosque at Siraf.7 While most of the arcades contained rubble piers, part of the sanctuary may have had wooden columns resting on bases of solid stone.

In Period I the sanctuary had a mortar floor and when bays I and 5 were added these, too, were given mortar floors. Narrow slots in the floor, found between bases 3J and N, EI and 3 and 5A and C, may mark the positions of wooden rails or screens. The double arcades, unlike the sanctuary, had pavements made of sandstone blocks o0 3 to 0o4 m. square. On the north-west side the paving began in bay 5. There, as on the north-east side of the mosque, the pavement was generally well preserved. It was, however, completely missing from the south-east arcades. The existence of paving in bays 5A and B showed that the double arcades were already paved in Period I, when the sanctuary occupied only bays 2, 3 and 4. In the south-east arcades we found that the blocks had been lifted before the construction of a small room in bay 15M and they may have been removed when the extension was built or re-designed (see below, pp. 44 and 45).

The courtyard (P1. IIa), which measured first 30 X 27 m. and later 26 x 27 m., was also paved. However, recent cultivation has destroyed all but a small patch of pavement between bays IIE and I2D. On the major axis of the mosque, opposite bay 1o, was a stone-lined " well ", o.

6 m. across, surrounded by an irregular mortar spread. It is possible that the well, which we shall investigate next season, fed a pool or fountain. The courtyard also contained a large cistern, 13 m. long, 3 m. wide and 4 m. deep. On the north-west edge a small platform provided access for drawing water. The cistern, which had the same shape and proportions as the numerous cisterns elsewhere in Sir~f, is inappropriate

* The footings of small buildings, probably shops, were found on the north-west, north-east and south-east sides of the mosque. At the time of construction, the ground surface probably

sloped downwards to the qibla end of the mosque, which overlooked the sea.

'Quoted by Sir Arnold T. Wilson, The Persian Gulf (1928), p. 95. 5


to the courtyard of a mosque and, in the absence of secure evidence for its date, we suspect that it was constructed after the main building fell into decay, perhaps to provide ablution facilities for worship- pers at the Recent Mosque (see below, p. 46).

The Great Mosque possessed a small minaret, situated on the north-east wall, immediately east of the entrance (P1. Ic). Only the stump survived and this consisted of a solid base 3-8 m. square, preserved to a height of 2 - 2 m. An external stair, of which five steps remained, provided access from the arcade. The minaret was apparently bonded with the enclosure wall and so was probably original.

Although little ornament survived in the mosque, three bases in the sanctuary retained traces of decoration. These were bases 3J, 3K and 3L, each of which had three short vertical plaster strips applied to the north-east face. 'Two bases, 3J and 3L, had a stone core, suggesting that the strips belonged to a date after the sanctuary was enlarged. Other ornament from the sanctuary included the carved stucco found in 19668 and fragments of plaster bearing incised circles and traces of reddish paint. Elsewhere in the mosque, we recovered two fragmentary crenellations from outside the north-east wall, opposite bay I5K. The crenellations (P1. VIIb and Fig. 8) were made of stucco with moulded decoration and presumably formed part of a series which crowned the outer walls of the main enclosure.

From time to time decay, hastened by earth movements like the tremor recorded in 977,9 made it necessary to strengthen the fabric of the mosque. The semi-circular buttresses in bays 3 and 4 of the sanctuary were probably among the earliest reinforcements, while the rectangular buttresses added to the north-west wall and to arcades 6, 14, C and M appeared to be late. The qibla wall of Period 2 showed clear signs of frequent restoration. Other late features in the main enclosure were two small rooms built in the angles of the north-east wall. Room I, at the north-west end, occupied two bays and was probably a store room. A stone hearth was built against the south-east wall. We found a bronze bowl decorated with punched ring-and-dot motifs (P1. Va) lying on the floor at the base of pier BI5. Just outside the room were two openings cut through the north-east wall. Room 2, at the south-east end of the wall, contained a broken grave cover and so was possibly a shrine. However, the cover was not in situ and the room also contained a fragmentary building inscription. It may, therefore, have been a second store room in which the two objects were deposited for safe keeping. The inscription (P1. VIIa) evidently refers to building or re-building at the mosque. The text is as follows:

~WldL ?r Jl (3)

(i) (In the name of God) the Clement, the Merciful. This is what was ordered by the Amir (2) ... an (or ab) ad-Dawla 'Izz al-Milla Nasir al-Umma Jamal (3) ... (al)-Ma'ali Kamrti b. Hazarasb b. Ka(mrfi).

To the south-east of the mosque was the extension, a large building which measured 43 m. deep by II - 3 m. wide and was divided into eleven parallel bays corresponding with bays I-I I in the main enclosure. The extension projected onto the beach and part has been destroyed by the sea. Indeed, hardly anything survived of bays I and 2, the existence of which was indicated by a mass of tumbled

masonry extending from the qibla wall. However, the surviving bays revealed a standard plan which

probably repeated throughout the building, with the exception of bay I I. In every bay the earliest floor was nearly 2 m. below the floor of the main enclosure. Despite this difference of level, the south- east wall of the main building had been replaced by an arcade with openings overlooking every bay of the extension. We found no trace of stone stairs and, if it was ever possible to move directly from the extension to the main building, wooden steps must have been used.

The bays of the extension were separated by partitions which were solid to at least the level of the main enclosure floor. However, the discovery of fragmentary shafts and arches in bay 9 showed that the

s Sfraf I, pl. VIIIb. 9 Wilson, p. 94.


upper parts consisted of arcades comparable with the arcades on the main enclosure. Initially, each surviving bay had a door at the south-east end and at least six were divided into two unequal parts connected by an opening I *2 to 2 - 4 m. wide. Bay io, which has yet to be excavated, may have had the same plan, but bay I I was originally undivided and had two openings in the north-east wall. All the bays had stone or mortar floors.

The extension was used for ablutions and probably also for storage. The bases of two large unglazed jars survived in the floor of bay 3, while bays 5 and 7 contained plastered basins for ablution and latrines (Pl. Id). Outside bay 7 was a domed clay oven.

At a later date, at least six bays were divided further by the insertion of partition walls, between

2"5 and 5'7 m. from the wall of the main enclosure. The partitions had no openings and so the areas

between them and the main enclosure were sealed off from the rest of the extension. In bays Io and I I they were filled with earth and rubble to the level of the main enclosure.

The masonry of the extension differed from that of the original mosque. The foundations were built of roughly coursed boulders, many of which had been collected on the beach, embedded in reddish mortar. Above floor level, the walls were of carefully prepared sandstone blocks,

0.2 m. long, with

levelling courses made of small stones, set in grey mortar. The jambs of the entrances into the extension were of a different build, for they consisted of masonry similar to the walls but without the levelling courses and with red, not grey mortar. This variety of construction, combined with the change of level between the extension and the main building and the solid character of the partitions, suggested that the extension may have been intended to have floors on the same level as the rest of the mosque. If this were so, the partitions were designed as foundations, precisely the same as the foundations in the main enclosure. The south-east side would have had a solid foundation and the whole structure would have been filled with earth. In practice, however, instead of making up the level, the builders laid floors on the existing surface and replaced the south-east wall with a series of doors.

To summarize, the sequence of building at the Great Mosque was as follows:

I. The original mosque consisted of a courtyard flanked by double arcades on three sides and a

triple arcade in front of the qibla wall. At this date, the arcades possibly had circular piers without bases. The sanctuary had a mortar floor, while the courtyard and the double arcades were paved. The minaret appeared to be original, as perhaps was the courtyard " well ". The date of the cistern is unknown.

2. The mosque was enlarged by the addition of a new bay beyond the qibla wall, which was replaced by an arcade. A second additional bay was inserted in the courtyard, making the

sanctuary five bays deep. Most of the piers now had square bases. During or after the enlargement of the sanctuary, several rubble bases in arcades 2, 3 and 4 were replaced by solid stones, which may have supported the wooden columns mentioned by Yaqfit. Three bases in arcade 3 had fluted decoration. Other ornament in the sanctuary included stucco comparable with the stucco at Na'in. At some stage the enclosure was given stucco crenellations.

3. The extension was added. It may have been designed with floors on the same level as the main building. If so, the scheme was never carried out and the floors were almost 2 m. below the

sanctuary floor. However, part of the extension was later made up to the intended level.

4. At a late date two small rooms were built at the north-east end of the main enclosure.

Unfortunately, we cannot yet date the stages of construction with accuracy. The only securely stratified deposit associated with Period I was a layer of earth and rubble found in a sounding in bay I 5A in 1966. The layer consisted of make-up deposited between the foundation walls. The only glazed sherds from the deposit were Sasanian-Islamic ware, while the coarse pottery did not contain a single fragment of flagon, eggshell ware or bowls of the type made in the kilns at Site D.10 The make-up may, therefore, belong to the early ninth century.

In 1966 we believed that stratified material of Period 2 had been found between the early and later

1o Sfrdf I, p. I6.


qibla walls in bay IL. However, excavation in bays IA-C last season revealed that the later qibla wall had been restored and that consequently many of the deposits between the two qiblas were disturbed. Thus the group of coins found in bay IL in 1966 proved to be valueless as evidence for the date of the second qibla wall (see below, p. 48). The lowest levels in IL, which did appear to be stratified, contained only coarse pottery, but similar deposits in IA-C yielded both Sasanian-Islamic and Dusun sherds.11 It may be significant that this latter group, which was stratified under the earlier of two superimposed floors, contained more than 400 fragments, but no Yiieh ware, painted stoneware or tin-glazed pottery. It is possible, therefore, that the enlargement of the main enclosure took place as early as the ninth century.

Although we cannot yet date the building of the extension, evidence exists for the date of the alterations to bays Io and I I and for the latest use of bays 3 to 9. The evidence for the date of the alterations consisted of a large group of pottery dumped against the outer face of the blocked doorway at the north- east end of bay i i. The doorway was sealed when part of the bay was made up to the level of the main enclosure. Among the types of pottery were Sasanian-Islamic, tin glazed, early sgraffiato and later sgraffiato wares (including Pl. VIe), as well as Dusun, Yiieh and Chinese white wares. True celadon and Seljuq pottery were completely lacking. The latest pottery in the group, later sgraffiato ware, came into use in the eleventh or twelfth century.12 Since the group contained complete or nearly complete vessels, it probably consisted of rubbish rather than residual material and the latest pottery may well be nearly contemporary with the blocking of the door. I suggest, therefore, that the alterations to the extension took place some time after c. Iooo, but before the end of the twelfth century, when Seljuq wares and celadon were in use at Site C.13 The pottery from the latest floors and latrines in bays 3-9 included Seljuq wares, lustre ware of north Persian type and celadon. It apparently belonged to the period described by Yaqilt, after which the mosque fell into decay.

Finally, in the main enclosure, pottery sealed below the floor of room 2 included Seljuq fragments, showing that minor alterations to the Great Mosque were still being made in the late twelfth or early thirteenth century.

On the basis of this material, I suggest the following chronology for the Great Mosque: Period I. Ninth century, perhaps before c. 850. Period 2. Possibly also ninth century. Period 3. The extension was begun some time after the building of Period 2 and it was altered

some time after c. ooo. Period 4. The mosque was still in use when

Yaqfit visited Siraf in 1218, but the pottery from

the extension suggested that it fell into disrepair shortly afterwards.

The Recent Mosque The excavation of the Recent Mosque began in 1966 and was completed last season when, in order

to examine the sanctuary of the Great Mosque, the later building was fully recorded and then demolished. Although the Recent Mosque survived to a maximum height of only I - 6 m., we were able to reconstruct its appearance with confidence, chiefly because the plan closely resembled the mosques of the Tdheri district today.

The Recent Mosque was a small rectangular structure, measuring only 6-4 m. deep by I I .8 m. wide, attached to which was a porch (Fig. 3, and P1. IIb). Outside the building was a minaret and two courtyards. The mihrab of the Recent Mosque occupied the same position as the mihrib of the Great Mosque in Period 2 and it was clear that the qibla wall and pier bases of the earlier building were visible at the time of construction. Indeed, it is possible that part of the sanctuary was used continuously until the Recent Mosque was built.

xx " Dusun " is the name used in 1966 for the coarse stoneware imported to Sirdf in large quantities from the ninth century: cp. SirdfI, p. 18.

12 The date of later sgraffiato ware is at present in dispute. In

Sfrdf I, p. 15, I suggested that the ware was already in use by the middle of the eleventh century. On the basis of his finds in

East Africa, Mr. Neville Chittick has proposed an earlier starting date, nearer to c. 950 than 1ooo (Neville Chittick, " Discoveries in the Lamu Archipelago ", Azania II (1967), pp. 1-31, particularly p. 13). Others, however, prefer a date in the twelfth century.

13 Sfrf I, p. 12.


Like the village mosque at Taheri, the Recent Mosque was two bays deep and three bays across. The mihrab was contained in a rectangular structure projecting from the qibla wall. In the centre of the north-east wall was a porch which diverged some 40 from the axis of the mosque. The lower parts of the mosque were built of mortared stone, with a levelling course

o.44 m. above the ground. The

footings were o.

6 m. thick and rested either on rubble or directly on the floor of the Great Mosque. A section of fallen qibla wall, found in 1966, showed that the upper parts consisted of arcades filled with mortared rubble.

The porch was 2.7

m. deep and 2 m. wide. It was made of two parallel walls, linked at the outer ends by a foundation which served also as a threshold. In each wall, I- I m. above the floor, was a niche o 55 m. across and 0o75 m. deep.

r -1

r' ' r -. r-

I I I

I L.. LJ L-LJ L-J

LI

minaret graves

fr'-'-~"===--'T~

rI I

I I

w3

sI i I

l N 5m-ete

elwlaiw32sions

"YI I i'

SII

I------------------- I -II

early walls additions

5metres

Fig. 3. Site B. Plan of the Recent Mosque.

The facade of the mosque was pierced at regular intervals by five openings. The central opening, which was entered through the porch, was I

.33 m. wide, while the other openings were only

o.74 m.

across. The openings were faced with plaster which retained the impressions of wooden frames, presumably for doors. Openings also existed in the sides of the mosque, each wall possessing a single door or window in the centre of the first bay.

In the interior, the floor was o '4

m. thick and consisted of a deposit of earth and rubble, above which was a layer of beach pebbles. Overlying the pebbles was a plaster skin. The building was divided by an arcade supported on two rectangular piers, with pilasters at either end. The mihrib consisted of an opening in the qibla wall o -9 m. wide, beyond which was a cavity I -42 m. wide and at least

o. 5 m. deep contained in the projecting structure. To the right of the mihrdb was the base of a mortared rubble


mimbar I - 16 m. long and 0 - 75 m. wide. In the mosque at Tdheri the mimbar is a stone structure measuring

o.6xo-65 m.

The remains of the minaret consisted of a long foundation abutting onto the east angle of the mosque. The foundation, which contained at least two periods of building, was 6 5 m. long and between 2 2 and 2 7 m. wide. At the north-east end was the bottom of a flight of steps which began at right angles to the main construction. Analogy with modern minarets suggests that the long foundation supported a

flight of steps leading to the roof of the mosque. The muezzin gives the call to prayer either from the

top of the steps or from a cage-like structure on the roof. Here the position of the foundation suggested that the muezzin stood on the topmost step.

Courtyards existed on the north-east and south-east sides of the mosque. In both cases, they were

badly ruined and their precise extent is unknown. The north-east courtyard, through which one entered the mosque, survived only to the east. It was 6 m. deep and probably extended for the full width of the mosque. On the east side it was bounded by the minaret. Beyond the minaret, the south- east courtyard probably extended from the front of the first courtyard to the qibla wall. Both courtyards were enclosed by thin mortared walls, reinforced at regular intervals by buttresses. In the mosque at

Ttheri the courtyard wall has similar buttresses which support a raised walk. Outside the courtyards, the cistern was either built or cleaned out during the use of the Recent

Mosque, for the earliest filling contained pottery closely similar to material from the courtyard floors. A short distance south-east of the mosque, in bays 3K-L and 4K-L were five Shi'ite burials.

We found little evidence for the date of the Recent Mosque. The latest datable object from the

make-up for the sanctuary floor was a Chinese copper coin of Wan-yen Liang (I149-61), minted between I 156 and I 161, some half century before Yaqfit reported that the Great Mosque was still intact. However, deposits resting on the courtyard floors contained scraps of unglazed painted pottery and a bowl decorated in black under a turquoise glaze,14 while the earliest fill in the cistern yielded more

painted pottery and fragments of a green celadon bowl. All had parallels at Site E, which was occupied mainly in the fifteenth century (see below, p. 56). Finally, a group of Portuguese lead coins, one of which was issued for Malacca probably by Manuel I (1495-1521), were found in a disturbed layer outside the courtyard wall, between the two qiblas of the Great Mosque. Unfortunately, it was im-

possible to determine whether the coins were deposited during restoration or demolition of the Great

Mosque, building of the Recent Mosque or activity on the site at a later date. Thus our evidence for the date of the Recent Mosque is limited to Yaqilt's description of SirMf, which establishes that the Great

Mosque was still standing in 1218, and to the pottery from the courtyard floors and the cistern. The latter suggested, but did not prove that the Recent Mosque was already in use in the fifteenth century.

2. Site F. The Tenth or Eleventh Century House

Before excavation, Site F was a mound, 4 m. high, on the east side of the water course which leads from Shilau valley to the sea. Although normally dry, the water course becomes a torrent after heavy rain and storm water has swept away part of the mound, damaging some of the structures it contained. Examination of the sides of the mound showed that it contained a group of substantial buildings. Last season we excavated one of these, a private house at the west end of the site. During the excavation we found that a narrow street ran along the east side of the house (P1. IIId), separating it from two other buildings divided by a second street which met the first at an angle of 900 (Fig. 4). Clearly, therefore, the mound conceals part of a crowded quarter of SirMf, with large buildings and narrow streets. We intend to continue work at the site in 1968.

The excavated house was a rectangular building, 27 m. long and 18 m. wide, with the major axis

aligned approximately north-south (P1. IIc and d, and P1. IIIc). It was well preserved and the walls stood to a maximum height of nearly 4 m. However, even the highest wall did not survive to the level of the roof or ceiling and our knowledge of the building is limited almost exclusively to the ground floor. This had a symmetrical plan with a courtyard,

I2"7 m. long and

9"5 m. across, surrounded by rooms

14 Sirdf I, pl. VId.

t'j

0

z

o U4 CA

CA

>1

Ot

covered drain

Sstreet

I well D , 3 4 5 6

2 7

B A

14 9

1

13 12 steps I

I

I I

0 ' soak - away

"< " ' x~

edge of water course

S5 10 blocked doorway

metres

RW & AC m DWd

N

t3 Cr.

?1

B 3-? ca iah"

cy. ;=s-

cy F;t

o

"CI h,

;y


on all four sides. If we include its two entrances, the house had fourteen ground floor rooms. On the north side was a narrow yard, with a well-house at the east end and a group of structures to the west.

In the house itself, the main entrance was in the north wall. It was centrally placed and consisted of a recessed doorway, behind which was a passage leading to the courtyard. In its final form, the entrance had a four-centred arch flanked by triple pilasters made of mortared rubble (P1. IIIa). In an earlier phase, the entrance also had a pair of double pilasters in the " porch " in front of the door. The

A X

ROOM I

I0 I

X B

ROOM 8 Section S of I j Room8 is Im

I i I

LIII Agricultural Soil &c. _ Rubble

Soft Loam Sand and Wash

Compact Loam Charcoal

0 10

SMetres

Fig. 5. Site F. Section.


doorway was I - 6 m. wide, with massive jambs and a threshold of stone slabs. On the inner side of the jambs were recesses to receive a draw-bar. The passage (room I) was

4"5 m. long, with a pair of double

pilasters at the inner end. One of the double pilasters retained a decorated stucco capital (Pl. IIIb). The total distance between the entrance and the courtyard was 6-9 m.

On each side of the entrance was a rectangular room measuring about 5"3

x 2 -8 m. internally, with a doorway into the courtyard. In room 2, on the left of the entrance, the doorway was in the centre of the south wall and survived intact. It was 2-5 m. high. In room 14, on the right, the doorway was originally central, but had been blocked and replaced by a new opening at the east end of the wall. The west wall of the room contained a niche. Rooms 3 and 13, which occupied the north-east and north-west angles of the house, were larger than I, 2 and 14, with internal dimensions of about

7" 2 X 2-8 m. The doorway of room 3 was intact, with a height of

2"45 m. However, the floor of the

room was o- 15 m. higher than floors elsewhere in the building and the doorway may have been built for an earlier floor at a slightly lower level. On the south sides of rooms 3 and 13 were two smaller rooms (4 and 12), measuring 4 2 and 4*8 m. by 2 - 8 m. respectively. The doorway of room 4 was again intact, but for reasons of safety we dismantled the lintel in the course of the excavation.

Rooms I to 4 and 12 to 14 made up the front half of the house. The plan of the rear half was almost identical. The rear entrance (room 8) was in the centre of the south wall. It consisted of a simple opening, originally I 7 m. wide but later reduced to

o.75 m., leading into a passage 6-3 m. long and

3" I m. wide. The passage was flanked by rooms 7 and 9, which measured 6 to 6-3 m. by 2 8 m. The south-east and south-west angles were occupied by rooms 6 and I o, with rooms 5 and I I completing the plan of the ground floor accommodation.

Like the front of the house, the rear revealed clear evidence of rebuilding. Blocked entrances existed between rooms 4 and 8 and Io and I I. There was another blocked entrance in the courtyard wall of room I I. In room 9 we found two floor levels, separated by a layer of rubble containing fragments of carved stucco, presumably from an early scheme of decoration.

Two groups of features suggested that the house had an upper storey: flights of steps and pier bases in the courtyard and a row of piers outside the rear wall. The features in the courtyard comprised a flight of steps outside the blocked door of room II (Pl. IId), with traces of a corresponding flight outside room 5. The steps led upwards towards the front of the house. At the front of the courtyard, outside rooms 4 and 12, were the scars left by two L-shaped piers, parts of which were found nearby. The steps and piers are perhaps best explained as the supporting members of a gallery which ran round the northern half of the courtyard, giving access to first floor rooms. In Bushire, modern courtyard houses commonly have a wooden gallery occupying a similar position. Such a gallery is useless in a single- storey building and its apparent presence at Site F suggests that an upper floor existed over at least part of the house during the late phase of occupation. Similarly, the row of piers outside the south wall is most easily explained as the support for a gallery projecting beyond the rear of the house. Rooms in this position would have overlooked the sea and provided the occupants with a well-ventilated loggia.

The house was built of mortared stone. The outer walls were o 065 m. thick, with buttresses every

2 to 3 m. Most of the buttresses were semicircular, but in the south-east corner, where re-building evidently took place, rectangular buttresses were used. While the outer walls were stoutly built, some of the internal partitions were only o0 3 m. thick and could scarcely have supported more than a single upper storey. It is unlikely, therefore, that the house was one of the multi-storey buildings mentioned by Istakhri in the tenth century.15

As at the Great Mosque, the roof was probably flat and made of wood. We found no fragment of domed or vaulted roofs. Furthermore, the roofs were consistently

2.8 to 3 I m. wide, a width also

found in some of the rooms at Site A,1" in the fifteenth century rooms at Site E (cp. Fig. 6) and in the modern buildings of Taheri. The width of the buildings in Taheri is determined by the length of the poles which are laid across the wall tops to support a cover of matting and earth. Timber is unavailable locally and the poles are imported by boat.1 Timber was imported in the medieval period, too:

15 Quoted by Wilson, p. 94.

16 Sifrf I, fig. 3 on p. 6.

17 A. H. J. Prins, " The Persian Gulf Dhows: Two Variations in Maritime Trade ", Persica II (1965-66), pp. 1-18.


Istakhri and Ibn IHauqal both remark on the aridity of Siraf and Istakhri notes that timber had to be imported from East Africa.18 It is probable, therefore, that much of the timber mentioned by Istakhri consisted of poles used for constructing roofs and that the roof at Site F closely resembled the roofs of modern Taheri.19

The roof was protected against storm damage by a series of drains which carried rain water into stone-lined pits at the foot of the outer walls. The drains consisted of columns of earthenware pipes and the soak-aways had corbelled tops. There were four soak-aways on the east side of the house, one survived on the south side and three on the west.

Two additional features of the architecture of Site F call for attention: the buttresses and the pilasters flanking the main entrance. Both have Sasanian parallels and show that the Sirdfi building style contained a number of traditional features. While the pilasters are at present unique at Sirtf, semicircular buttresses are common. They occur, for example, in the fragment of " sea wall " east of Site D20 and among the footings visible on the slopes of the first ridge.

Although only one piece of decorated stucco remained in position, we recovered numerous examples from the filling of the rooms and the courtyard. The majority of the fragments came from the rubble

deposited when the house collapsed and it is reasonable to assume that most of it fell from the walls. Pieces of stucco occurred in all the rooms except 4, 1O and 12. Room Io had been almost completely destroyed by storm water in the gully and little deposit survived. Thus only rooms 4 and 12 undoubtedly did not have stucco ornament.

Unlike the carved stucco from between the floors of room 9, all the plaster in place when the house was abandoned bore geometric motifs, vegetable elements and Kilfic inscriptions moulded in low relief.

Three types of moulded stucco were found:

I. Narrow friezes, o.

I m. wide, perhaps applied to the wall just below the ceiling. 2. Broader friezes, at least

0.2 m. wide.

3. Panels, at least I m. across and 0 - 5 m. high.

The different types of stucco were distributed as follows:

Friezes

Room Narrow Broad Panels Illustrations

I - 2 x - x Pl. VIId 3 x x x Fig. 9, 2 4 - -

5 - x -

6 - x -

7 x 8 - x

9 x - - Pl. VIIc IO

12

13 - x 14- - x Fig. 9, 3

Building to north-west of house x x - Fig. 9, I

18 Quoted by Wilson, p. 94. 19 In 1966 Mr. Neville Chittick discovered abundant Islamic

pottery, much of it identical to material from Sirdf, at Manda, an island site off the coast of Kenya. Manda is surrounded by

mangrove swamps and Chittick has drawn attention to the

probability that the settlement was exporting poles to the Persian Gulf (Chittick, op. cit., p. 19).

20 Sirdf I, p. 5.


The excavation yielded a huge number of finds, mostly pottery (see below) and glass. Among the small finds, objects of bronze and iron were common, as also were fragmentary stone vessels. Although we recovered little anhydrite from the site, coarse grey steatite was abundant. Among the commonest types was a pan-like vessel with a cylindrical body, four legs and a single handle (P1. Vb). Discs with a rebate on the underside of the lip and carved or fretted ornament may be lids for these vessels, which resemble metal incense burners and probably served the same purpose.21 Other grey steatite vessels included straight-sided jars, sometimes bearing carved decoration (Pl. Vc). Among the objects made from other types of stone were three fragmentary vessels of smooth whitish talc: a molar flask, a rilled cylindrical vessel22 and a pen stand. We also found a sandstone mortar and a small three-legged table (P1. Vd).

Other finds from Site F included a group of hemispherical plaster objects, about 25 cm. across and 12 cm. high, with a vertical hole through the centre. They closely resemble the stone bases of wooden stands used today in Gilan and Mazandardn to support skeins of silk. It is possible that the plaster objects from Siraf had a similar function and in this context we should note that, according to Istakhri, linen napkins and veils were made in the city in the tenth century.23

Until we have carried out more extensive excavations under the floors of the house, we cannot determine when it was built or refurbished. In room 9, where we removed half of the visible floor, the only artifacts were pieces of carved stucco which cannot be closely dated. Similarly, the moulded stucco which decorated the house during the latter part of its use cannot be dated on stylistic grounds. However, evidence for an approximate terminus ante quem was provided by the material from a number of deposits overlying the latest floors. The layers of refuse deposited after the house went out of use but before it collapsed (e.g. layers I6o and 181 in the section in Fig. 5) yielded a vast amount of pottery, much of which was comparable with material found in the sounding (Site A) at and the kilns (Site D) in 1966.24 Indeed, most of the coarse pottery, which included eggshell ware, flagons and bowls, was identical with pottery from the kilns.25 Last year I suggested that the kilns belonged to a relatively late stage in the growth of Siraf and were roughly contemporary with stages 2C and 2D in the sequence at Site A. Comparison of pottery from Site F and the sounding support the suggestion and allow us to be more precise about the sequence at all three sites. The refuse layers at Site F yielded the following types of glazed pottery: Sasanian-Islamic, tin-glazed and early sgraffiato ware and Yiieh (P1. VIa), Dusun, relief stoneware and Chinese white ware. There was, in addition, one sherd of painted stoneware. Later sgraffiato ware, Seljuq pottery and green celadon were all completely lacking. It seems clear that painted stoneware, which occurred at Site A in all deposits of stage 2A or later, had already gone out of use by the time the refuse was dumped at Site F. Furthermore, later sgraffiato ware, Seljuq pottery and true celadon (none of which was found at Site D) had not yet appeared. In the sounding, the earliest occurrence of later sgraffiato ware was in stage 2D. We may assume, therefore, that the rubbish layers at Site F were roughly contemporary with the kilns and with stage 2C in the sounding. It follows that painted stoneware was in use in stages 2A and B only, the sherds from later levels being residual.26

The results of the excavation at Site F may be summarized as follows: the building we excavated was, in its final form, an impressive private house with an upper storey over the northern half and a first floor loggia supported on piers projecting beyond the south wall. The house was abandoned before Sirdf fell into decay and by the eleventh century the ruin was used as a dump for refuse.

21 Cp. metal incense burners illustrated in Pope, Survey VI, pl. I278b and c. For a fragmentary steatite vessel from Samarri, see Iraq Department of Antiquities, Excavations at Sdmarrd 1936-9, part 2, pl. CXXIX, lower centre.

22 An identical fragment was found in a late layer at Site A in 1966.

23 A modern stand for skeins of silk is illustrated by Hans E. Wulff, The Traditional Crafts of Persia (i966), fig. 261 on p. 183. Istakhri's reference was quoted in Sirdf I, p. 3.

24 Several deposits may, in fact, be contemporary with the later occupation of the house, but this is not certain. The levels in question include the lowest layers of trampled earth in the

street, which yielded Sasanian-Islamic, tin-glazed, early sgraffiato, Dusun, Yiieh and Chinese white wares. The filling of the well in front of the house also appeared to be contemporary with the occupation. Here the Far Eastern pottery consisted of Dusun (including Pl. VIc), Yiueh, coarse white stoneware, white porcelain and painted stoneware, while the Islamic sherds included Sasanian-Islamic, tin-glazed and eggshell wares.

25 Sirdf I, p. I6. 26 It is noteworthy that anhydrite, which was already being made

in the vicinity of Site A in stage i, was uncommon at Site F, while coarse grey steatite was abundant. Grey steatite appeared first in the sounding in stage 2B.


3. Site E. The Fifteenth Century Buildings Site E was an area roughly 30 m. square, 44 m. north-west of Site A. Before excavation, the site

consisted of two adjacent mounds of earth and rubble. Investigation revealed a group of buildings which differed markedly from the structures found elsewhere at Sirdf (Fig. 6 and P1. IVa). The buildings belonged to two principal phases, in the second of which the walls were made of rubble held together with mud and the floors were usually trampled earth. Mortar was used mainly for doorways, pilasters and arches and stucco was completely missing.

The earliest building in the complex was the square structure which occupied the north-east angle of the site (Pl. IVb). This structure, known as the " square building ", was 13 - 8 m. across and had a symmetrical plan consisting of eight rooms surrounding a central courtyard. In the centre of each side was an iwdn-like room open at the inner end. At the outer end was an opening I 2 to I .3 m. wide. The openings in the south and west rooms were undoubtedly doorways and the latter retained the impression of a wooden frame. The openings in the north and east rooms, on the other hand, were probably windows and at a late date the north opening was blocked with mud and stones. Each iwdn had a pair of niches,

o. 8 m. wide and 0- 35 m. deep, in the lateral walls, o .4 to o 5 m. from the floor.

The niches were carefully plastered and we observed traces of a red stripe under the west niche in the north iwdn.

Each angle of the square building was occupied by a room entered through an opening in the corner of the courtyard. The rooms were completely plain. The courtyard was only 5'4 m. across, with a mortar floor

o. I to o - 15 m. below the floors of the iwdns. At the centre of the courtyard was a

drain. The square building was made of mortared stone. The walls had levelling courses and a " cellular "

structure composed of rows of mortared stones laid at right angles to the faces of the walls. The spaces between these rows were filled with random rubble, producing the appearance of continuous blocked niches. The rooms had plastered walls and floors of mortar or trampled earth.

Some time after the construction of the square building, a group of rooms was added on the south and west sides, making a square complex enclosing an L-shaped yard. The complex was entered from the south through a rectangular room, part of which was paved with eight millstones, making an impressive and durable entrance (P. IVc). The room led into a passage which gave access to the yard. At the west end of the passage was a rubble platform measuring 2-2 x

I. 5 X

o.7 m. high. Comparison

with contemporary buildings at Taheri shows that the unpaved part of the room was probably used for storing visitors' baggage, while the platform was occupied by a watchman.

The other rooms were entered from the yard. At the east end of the south range, room 2 yielded a cache of ornaments and domestic objects. These included a stone pendant; nodules of ironstone, red ochre and natron;27 an illegible coin, two mounts with zoomorphic terminals and a broken feeding vessel of copper or bronze; a knife blade, split pin and a length of chain made of iron; beads with a turquoise glaze; and a glass bottle. Other items included an ivory chess piece and four cylindrical wooden objects, also thought to be gaming pieces (P1. Ve).

The north-western part of the site originally comprised two adjoining rooms, each measuring Io to I I m. by 3 m. internally. One of these, room I12, had two floor levels, the first of which consisted of beach pebbles. At a later date the level was made up by tipping o 75 m. of earth onto the cobbled floor and covering this with a mortar skin. The repeated use of stones and mortar instead of trampled earth suggested that room 12 must have been the principal living room of Site E. The other room was considerably altered at a late stage in the occupation of the site. Probably at the same time, a long wall was built across the northern edge of the site, creating a corridor and a series of rooms. One of these, room 14, was used as a rubbish dump and yielded a valuable group of pottery and other finds (see below, p. 56).

27 Natron is available today in Bushire and some of the villages on the coast, where it is burnt and made into shampoo. An informant in Bushire stated that it is quarried near FirfizAbhd.

Natron could have been carried to Sirdf along the route from FirfizSb•d to Jamm and Tdheri.


rooml

I

- courtyard

mangers CXo

L '

oovens-

---bench--

mi lstone r . .

- hoard

"earliest walls Period 2b 0 5 15

metres Period 1 late walls 0 50

uncerta •fee

t

1~'i ri~~~a;\ - = //LLL /

Perio 1 • late wall 0 10 5

Peiods~n 2a L date fee fluncertain

Fig. 6. Site E. The fourteenth and fifteenth century buildings. Plan.


The buildings enclosed an L-shaped yard with a floor of earth and gravel. On the south side of the yard was a row of eight mangers attached to the outside of the entrance corridor. Each manger consisted of a plastered trough, I m. above the ground, below which was a small recess. The recess was intended to hold a stick to which the animal was tethered. The yard also contained numerous bread ovens, each consisting of a clay-lined pit with an air duct at one side. Ovens of this type, which are widely used today,28 have a long history and examples were found in early levels at Site C in 1966.29 During the use of the complex, a number of structures were built in the yard, including a wall running from the square building to the south range, dividing the area into two. The eastern part of the yard could be entered from outside the complex by an opening in the east wall.

A X

X B

soil etc sand 0 5

rubble mortar chips metres

ILTiI soft loam etc clayey soil

RWm DWd

- - - - - - - - - - -

RWm DWd Fig. 7. Site E. Section.

The evidence of the coins, the Islamic glazed pottery and the Chinese porcelain all points to the conclusion that the site was occupied principally in the fifteenth century, with occupation perhaps beginning some time before c. 14oo00. Among the Islamic copper or bronze coins trampled into the floor of the yard were pieces minted by Ghzdin Mahmiid (c. 1296-1304), Abti Sa'id (c. 1320-30) and Shah

Shujd' (1386-87), while a Chinese copper coin from below the surface of the yard was minted by either Wan-yen Liang (II149-6I1) between 1 156 and 1161 or Ying Tsung (1436-39). The Islamic pottery included a group of glazed bowls from room 14. These have a ring base, a curving side and a slightly everted rim. They have a white friable fabric and are decorated in black under a colourless or turquoise glaze. The decoration consists of a rosette surrounded by a zone of interlaced leaves on the inside of

28 Wulff, p. 292 and fig. 41 I. 29 Sfrdf I, p. 12.


the vessel, with a row of stars or circles on the outside (Pl. VIb and d). Pottery of this type is usually dated to the late fourteenth or fifteenth century and Chittick has published a group of bowls, which includes vessels very similar to the material from Siraf, from a fifteenth century building at Kilwa.30 The Chinese pottery included celadon, sherds of coarse stoneware jars with a green or brown-black glaze and fragments of blue and white porcelain, some of which belongs to the late fifteenth century. Thus the datable finds from Site E are of the fourteenth and fifteenth centuries and we suggest that the main occupation took place after c. 1400.

The excavation of Site E was particularly instructive, for not only did it throw light on a period of occupation apparently unnoticed by contemporary writers, but also provided information on types of pottery and glass which could not be dated in 1966. Among the pottery from Site E were the following wares:

(a) Stoneware with an unglazed ring on the interior. The site yielded numerous sherds of stoneware bowls with a ring base and a curving side. The fabric is pale grey and the glaze, which covers most of the interior and the upper part of the outside, is colourless. A few fragments have painted decoration in grey or greyish green. The most distinctive feature, however, is a concentric unglazed ring on the interior, above the foot (P1. Vf). Although the earliest examples come from layers 40 and 49 and antedate the construction of the square building (see section, Fig. 7), most of the sherds belong to the main occupation and it is probable that the ware belongs mainly to the fifteenth century.

Fig. 8. Site B. The Great Mosque. Stucco crenellations. (Scale I: 4)

30 Neville Chittick, " Kilwa: A Preliminary Report ", Azania I (1966), pp. 1-36, esp. p. 23 and pl. XI and XII.


(b) Unglazed painted pottery. The most distinctive types of coarse pottery found at Sirdf are hand- built vessels decorated with red or purplish slip.31 Two varieties occur: ablution jugs and bowls with a coarse reddish fabric, a cream slip and simple striped decoration, and larger jugs or jars with a soft cream fabric and intricate geometric decoration (P1. VIf). Painted pottery was extremely rare in the earliest deposits at Site E, but common during the occupation of the house complex. Although the cream fabric was considerably more common than the red, both occurred in association with the glazed bowls in room 14.

Fig. 9. Site F. The tenth or eleventh century house. Stucco. (Scale I: 4)

DISCUSSION

According to the documentary evidence, Sir~f was already trading with India and the Far East in the middle of the ninth century. A hundred years later, Istakhri reported that Sirdf was almost as large as Shiraz. It was a wealthy city with fine multi-storey houses. In 977 the city suffered an earthquake and, with the fall of the Biyid dynasty c. 1055, it declined. However, Sirdf still supported wealthy citizens in the early twelfth century,32 but by 1218, when Yaqiit described the site, it was almost deserted. The gradual decline which set in after the earthquake of 977 must have accelerated sharply in the twelfth century and it is noteworthy that the latest elaborate grave cover from the site was made in II32 (see below, p. 61).

31 STfrf l, p. 15.

32 S. M. Stern, " Ramisht of Sirdf, a Merchant Millionaire of the Twelfth Century ", Journal of the Royal Asiatic Society (April 1967), pp. o10-14.

EXCAVATIONS AT SiRAF 59

In 1966 our first task was to cut a trench through the accumulated deposits at Site A. This sounding was intended to examine the sequence of events at Sirdf and to provide a relative chronology for the pottery and other artifacts found in area excavations elsewhere on the site. With its series of clearly- defined levels, the sounding served its purpose well for the period of Siraf's great prosperity, but told us little of the later history of the site. The results obtained in the second season help to repair this omission and compel us to re-interpret the later levels at Site A. We know now that Period 3 in the sounding, like Site E and probably the Recent Mosque at Site B, belongs almost entirely to the fifteenth century. The break in continuity between Periods 2D and 3 in the sounding is partly filled by the latest use of the Great Mosque which, like the final phase at Site C, belongs to the late twelfth and early thirteenth centuries. Thus, we now possess archaeological information for almost all of the period covered by the documents, from the middle of the ninth century to 1218, with additional information on an unrecorded settlement of the fifteenth century.

Our information comes from six sites and the following table summarizes the development they represent. It must be emphasized that the numbering of Periods I to 4 is a provisional scheme, which may be altered in the light of new evidence from other parts of the site.

Site Period

A B C D E F

4 Period 3 Recent - - Main

Mosque occupation

3 - Latest use of Latest Great occupation of

Mosque the house

2c Period 2d Period 3

2b Periods - - Kilns - Occupation 2b and 2c in use and collapse

2a Period 2a Period 2 Construction of the house

I Period I ? Period I Earliest

occupation

Period I. Our knowledge of the earliest occupation at Siraf is based on the excavation of Sites A and C in 1966. Although we found traces of buildings with walls of mortared rubble, no plan could be reconstructed. Site A yielded evidence of making anhydrite bowls and jars and Site C of working iron. The only types of glazed pottery were Sasanian-Islamic and Dusun wares. We concluded in 1966 that the occupation belonged to a period after Sirdf had begun to handle long-distance traffic, but before it became a wealthy port. It is possible, but by no means certain, that the Great Mosque was begun in Period I.

Period 2a. To this period belongs the first phase of the building at Site A, the Great Mosque and the house at Site C. All three buildings had walls of mortared rubble and paved or mortar floors. At Site A, the period was marked by the sudden introduction of a variety of Far Eastern wares, including white ware, Yiieh, painted stoneware and stoneware with decoration in relief. Among the Islamic pottery is tin-glazed ware with decoration in blue. The only coin of this period,

6A


found at Site A and identified after Sirdf I had gone to press, was a badly-preserved dirham, probably of the ninth century. Period 2a clearly belongs to an early stage in the period of Sir~f's great prosperity which, according to the documents, had already begun c. 850. Unless there is a break in continuity at Site A-and we found no suggestion of this-the period falls within the ninth century.

Period 2b. Period 2b represents the maximum prosperity of Sirif. The city possessed a mint and probably reached its greatest size, with suburbs extending westwards from the original nucleus to the west end of TTheri Bay. Periods 2b and c in the building at Site A, the Great Mosque, the house at Site C, the kilns and the impressive house at Site F were all occupied during this period. Decorated stucco was used at both the mosque and the house at Site F. Chinese pottery was abundant and among the sherds from Period 2b in the sounding are the earliest translucent white porcelain and sherds with a bluish ch'ing pai glaze. The early deposits at Site F also contained fragments of translucent porcelain and ch'ing pai, but hardly any painted stoneware, suggesting that the latter type had gone out of use before the end of the period. The earliest steatite belongs to stage 2b in the sounding, and at Site F fragments of steatite lamps, incense burners and jars are common. Before the end of the period, the house at Site F had gone out of use and the ruin served as tip for earth and rubbish.

Period 2c. This was evidently a period of decline. The kilns had gone out of use at the beginning of the period and the site was never subsequently built on; the city was beginning to contract. Nevertheless, the Great Mosque and the houses at Sites A and C remained in use. Indeed, the extension to the mosque probably belongs to this period. Such an apparent contradiction corresponds well with the picture of the eleventh and twelfth centuries conveyed by the documents: on the one hand, the port declined in the face of growing competition from Qais while, on the other, the city continued to support wealthy merchants.

Period 3. For this period we have only the evidence of the Great Mosque and the house at Site C. In both buildings the latest occupation was represented by deposits containing Seljuq white ware and gold and blue lustre ware of north Persian type. The period probably ended shortly after YaqUit reported the lowly status of Siraf in 1218.

Period 4. Before excavations began, the only building at Sirif with walls standing to an appreciable height were two structures on the first ridge: a mosque of the fourteenth or fifteenth century and a small structure known to the villagers as a madrasa and probably of recent origin. In 1966 we discovered a second late mosque at Site B and last season Site E disclosed a complex of domestic buildings associated with coins and pottery of the fourteenth and fifteenth centuries. The finds from this complex closely resembled those from Period 3 in the sounding and were similar to material from the Recent Mosque at Site B. It is clear, therefore, that a modest settlement occupied the site of Siraf in the fifteenth century. Presumably, this settlement had its origin in the wretched community described by Yaqfit and was the forerunner of the modern village of Taheri.

The chronology of the four periods w as probably as follows:

4. Fifteenth century. 3. Twelfth and early thirteenth centuries. 2c. Later eleventh century. 2b. Tenth and early eleventh centuries. 2a. Later ninth century. I. Early ninth century.

A preliminary account of the third season of excavations at Siraf appears on p. 182.


APPENDIX

STONE GRAVE COVER

The most impressive object found last season at Siraf was a stone grave cover with an elaborate floriated Kilfic inscription (P1. VIII and Fig. 10).

0 50 100

cms

Fig. io. Stone grave cover.

The cover was found approximately 500 m. north of the ruined mosque labelled masjid on Fig. i, in a shallow re-entrant just below the crest of the ridge. Although not in position, the cover probably came from the small cemetery which occupies the east side of the valley and contains other, less elaborate monuments.

The cover is carved from a single rectangular block and measures I -76 X 0o46 x 0o47 m. The inside is hollow and the top and sides all bear Kific inscriptions in low relief. The text is as follows:


Top

Z~~ ~3 3TJ s~ cSC 3~ rCS11 1x3~ j~31 s~ ~ 4UI 4~-, 4~4~f;~f~ j! ~LcC ~! C~j.~l ~LL cj! L~~ FJ~*I

" In the name of God the Clement, the Merciful: the blessing of God upon Muhammad and the family of Muhammad and forgive thy servant 'All b. 'Abd ar-Rahman b. Muhammad b. Juwdna- waih. May God have mercy upon him."

Sides

. (4)

.. . (). ( ) "

.,- . . i

, . " In the name of God, the Clement, the Merciful: Say, Lo! Those of old and those of later time will all be brought together to the tryst of an appointed day " (Qur'dn LVI, 49-50).

Grave covers with floriated and, less often, plaited Kilfic inscriptions are common at Sirif.33 One example, now in the British Museum, bears a date equivalent to 991, while a second cover, in the Louvre, is dated I 132. Covers of this type were, therefore, used at Sirdf between the late tenth and the early twelfth centuries. Close parallels for the script used on the Sirdfi covers are difficult to find, although it has affinities with Ghaznavid inscriptions of the eleventh and twelfth centuries and, more distantly, with the floriated Ktific in the Mosque of

al-.H1kim (990-1003). Perhaps the closest analogy

for the grave cover illustrated here is an inscription in the Kisimkazi mosque at Zanzibar, which bears a date equivalent to I lo7. It is probable, therefore, that the cover belongs to the early years of the twelfth century.

33 Sirdf I, p. 20o and pl. VIIIe and f.

Pl. I.

(a) Site B

. G

eneral view looking east, w

ith the qibla w

all on the right.

P1. I.

(b) Site B

. G

eneral view looking east.

P1. I.

(c) Site B

. T

he minaret.

Pl. I.

(d) Site B

. B

ay 7 of the

extension looking south-east.

P1. II.

(a) Site B

. T

he courtyard looking south. P

1. II. (b)

Site B.

The R

ecent Mosque looking w

est.

Pl. II.

(c) Site F

. G

eneral view looking south.

Pl.

II. (d)

Site F.

General view

looking north-west.

P1. IIH. (a) Site F. The main entrance.

Pl. III. (b) Site F. The north-east angle of the courtyard.

Pl. III. (c) Site F. General view looking east.

Pl. III. (d) Site F. The street looking south.

P1. IV

. (a)

Site E.

General view

looking north-east.

P1. IV

. (b)

Site E.

The square building looking north-w

est. P

1. IV.

(c) Site E

. T

he main entrance.

Pl.

V.

(a) Site B

. B

ronze bowl, diam

eter of rim 14 cm

.

P1. V

. (b)

Site F.

Steatite incense burner. O

verall length: c. 21 cm

.

P1. V

. (c)

Site F.

Fragm

entary steatite vessel. D

iameter of base:

20 cm.

P1. V

. (d)

Site F.

Underside of sandstone table.

Diam

eter: 35 -6 cm

.

Pl.

V.

(e) Site E

. Ivory and w

ooden gaming pieces.

Height of ivory piece:

3 - 8 cm.

Height of largest w

ooden piece: 3

7 cm.

P1. V

. (f) Site E

. F

ifteenth century stoneware.

Length of largest piece: io cm

.

P1. VI. (a) Site F. riieh ware. Base diameter: 9 cm.

P1. VI. (b) Site E. Bowl with underglaze decoration. Diameter qf rim: 23 cm.

P1. VI. (c) Site F. Detail of stonewarejar with incised decoration. Height: approx. 50 cm.

P1. VI. (d) Site E. Bowl with underglaze decoration. Diameter of rim: 24 cm.

Pl. VI. (e) Site B. Later sgraffiato ware. Diameter: 24-5 cm.

P1. VI. (f) Site E. Unglazed painted pottery. Height: 39- 6 cm.

P1. V

II. (a)

Site B.

Fragm

ent of dedicatory inscription. P

l. V

II. (b)

Site B.

Stucco crenellations.

P1. V

II. (c)

Site F.

Fragm

ents of inscribed stucco. P

l. V

II. (d)

Site F.

Stucco with m

oulded decoration.

P1. V

III. Stone grave cover, length I -76 m

.

63

THE DARABGIRD RELIEF - ARDASHIR OR SHAHPUR?

A Discussion in the Context of Early Sasanian Sculpture'

By Georgina Herrmann

One of the earliest European travellers to visit the rock relief at Dardbgird was Sir William Ouseley, brother of and Private Secretary to the British Ambassador, H.E. Sir Gore Ouseley. In April I8I I Sir

WVilliam was given a sumptuous luncheon by the pool in front of the carving:2 he has left us a vivid

Fig. i. Ouseley's Sketch of the Ddrdbgird Relief (by courtesy of the Bodleian Library).

1 The award of the Joanna Randall MacIver Junior Research Fellowship by the Principal and Fellows of St. Hilda's College in 1966 enabled me to devote two years' research to the subject of Sasanian rock reliefs. I am most grateful to all my colleagues at St. Hilda's for their friendship and encouragement during my tenure of the fellowship. Thanks to a generous grant from the Gerald Averay Wainwright Near Eastern Archaeological Fund I spent four weeks visiting the relevant sites in Iraq and Iran in the spring of 1967. The loan of a Land Rover by the British Institute of Persian Studies made it possible to reach all the rock reliefs in Fars in the short time at my disposal. For this and much other help I owe a special debt of gratitude to the Director of the Institute, Mr. David Stronach. I have received valuable advice from Dr. A. D. H. Bivar, who kindly read and criticized my text; he is, of course, in no way responsible for any errors or omissions. Dr. C. H. V. Sutherland, Keeper of the Heberden Coin Room, Ashmolean Museum,

selected the coins which are reproduced in this article and gave me much help concerning the problems of Roman portraiture.

The majority of the plates have been made from photographs skilfully printed from my own negatives by Mr. Leonard H. Bell. The line drawings are the work of Mrs. Pat Clarke of the Department of Antiquities of the Ashmolean Museum. All her drawings are based on photographs, and despite this she has produced impressions of the reliefs which would need only slight amendment by a draughtsman able to work on the sites.

It was Sir Max Mallowan who initially suggested the subject of Sasanian rock reliefs as one worthy of further study. I am as always extremely grateful to him for his unfailing support and inspiration in all my work.

2 Sir William Ouseley, Travels in Various Countries of the East,

,1fore Particularly Persia .. ., vol. II (1821), pp. 145-8.


impression of the frustration he suffered throughout the leisurely course of this feast, for he was longing to be free to examine and sketch the relief (Fig. I)! When he visited the Darabgird sculpture Ouseley had already seen the six Sasanian reliefs at Bishapfir; despite the considerable variations in style and subject, he had assumed that Shdhpfir was the king depicted in each of them.3 Thus he was able to state with authority of the Darabgird relief that, " One glance enabled me to recognise in the supposed figure of Rustam, another monument ... of the glory or the vanity of Shapur ";4 and he was further able to

suggest that the old man, on whose head the king's hand rests, " may be a real portrait of the unfortunate Valerian ".5

Fig. 2. Flandin's Sketch of the Ddrdbgird Relief (by courtesy of the Bodleian Library).

Some thirty years later Eugene Flandin, the French painter, also visited and sketched the Ddrdbgird relief (Fig. 2). While Flandin was a sympathetic recorder of contemporary genre scenes, his sketches of bas-reliefs leave much to be desired: Curzon commented that Flandin's engravings " share the inaccuracy not uncommon in the Frenchmen's reproductions of the Sasanian sculptures ".6 Flandin's drawing of the Dtrdbgird relief is particularly inaccurate and the reason for this may be discovered in his comment about the relief, which he considered to be " de tous ceux de la meme epoque, ... l'un des moins habilement executes ".1 While his sketch is poor as a whole, it is correct in one important respect: it shows the king wearing a crown of a skull-cap surmounted by a korymbos-this is the crown of Ardashir; it is not the mural crown of Ardashir's son, Shthpir. At this time, however, Herzfeld's

3 Ibid., vol. I (1819), pp. 279 and 28I-90. 4 Ibid., vol. II, p. 145. - Ibid., vol. II, p. I48. 6 G. N. Curzon, Persia and the Persian Question, vol. II (1892),

p. 2Io, note I.

E. Flandin and P. Coste, Voyage en Perse de MM. Eugine Flandin, Peintre, et Pascal Coste, Architecte, Attachis a l'Ambassade de France en Perse, Pendant les annees i84o and 1841: Perse Ancienne Tome I,

pl. 33; Perse Ancienne--Texte, p. 35.

THE DARABGIRD RELIEF-ARDASHIR OR SHAHPUiR ? 65

ingenious system of identification by crown was still unknown, and as local tradition identified the king as Shahptir, Flandin, like Ouseley, assumed that the scene once again recorded Shahpfir's victory over the unfortunate Valerian.8 Even today, despite Herzfeld's publication of the method of identification by crown in the I930's9-which naturally led to the re-attribution of many Sasanian sculptures-the Darabgird relief continues to be attributed to Shahpfir. It is relatively hard of access-more on account of its geographical position than of the location of the actual sculpture-and most scholars have only been able to consult either the inaccurate drawings of Ouseley and Flandin (Figs. I and 2) or the photographs published by Sir Aurel Stein in 1936.10 Because of the large and deep pool of water in front of the relief, these were not wholly satisfactory: one was too distant to show any detail, while the other was taken from the side and presented a heavily slanted view. Surprisingly, some thirty years later, these two photographs remain the only published ones known to the author; they have of course been frequently re-used. It is probably because of the lack of more reliable photographs and drawings that the attribution of the Ddrabgird relief to Shahpiir has remained unquestioned for so long, for such authorities as Herzfeld,'" Vanden Berghel2 and more recently Ghirshman13 have all followed Ouseley's original identification of the king as Shahptir and the old man as Valerian. The first to query this was B. C. MacDermot, who noticed that the king was not wearing Shihptir's distinctive crown. He pointed out " the possibility that this relief is meant to illustrate a Roman triumph of Ardashir I ... and that similar reliefs of Shahpiir repeat a formula used by his father ",14 although he considered that this was unlikely. As recently as 1965, however, Professor Hinz"s reaffirmed Ouseley's attribution of the relief to Shahpfir, although he claimed that it represented that king's triumph over Philip the Arab rather than over Valerian and he proposed that the old man's head was a " portrait " of Philip.

The cornerstone of Sasanian studies was undoubtedly laid when Herzfeld discovered the method of identifying each king by his distinctive crown: and the Daragbird king is clearly depicted wearing the crown of Ardashir. Unless, therefore, we are to abandon Herzfeld's system in this one instance, we must identify the king as Ardashir. That the relief depicts a triumph over the empire's Western foes is no bar to this attribution, for Ardashir waged a number of campaigns against Roman and allied forces throughout the 230's. Not only is this identification obvious from the evidence of the crown and on historical grounds, it is also, as we shall see, stylistically feasible.

THE RELIEFS OF ARDASHIR (224-241 A.D.)

While the six surviving reliefs of his son and successor Shihpfir (242-272 A.D.) all have in common a " baroque " or florid quality, Ardashir's reliefs are remarkable in that no two are carved in the same manner. It is also noteworthy that his early sculptures were nearly all abandoned before they were finished: this may well have been because he was dissatisfied with their quality. Having begun at Salmisl6 and Firtizabid (the battle scene) by simply " drawing " in stone, his later investiture scenes- at Firizabdd, Naqsh-i Rajab and Naqsh-i Rustam-show his increasing awareness of the values of a balanced composition and of high relief. They illustrate the gradual evolution of a Sasanian style. While relief carving had been of a high standard some 500 years earlier in the Achaemenian period, there is little evidence of Iranian sculpture in the intervening years, and what there is-at Tang-i Sarvak, Bard-i NishZndeh and Sqisa17-is of a uniformly low standard.

8 Ibid., Perse Ancienne--Texte, pp. 34-5. 9 E. Herzfeld, " Khusraus II Krone: Al TAdj al Kabir: die

Kronen der sasanidischen K6nige ", AMI IX (1938), p. Ior ff., Abb. i.

10 Sir Aurel Stein, " An Archaeological Tour in the Ancient

Persis ", Iraq III (1936), pp. 194-6, pl. XVII, figs. 28 and 29.

11 E. Herzfeld, Iran in the Ancient East (1941), p. 313- 12 L. Vanden Berghe, Archdologie de l'Iran Ancien (1966), p. 46. 13 R. Ghirshman, Iran: Parthians and Sassanians (1962), pp. 16o- 1. 14 B. C. MacDermot,

" Roman Emperors in the Sassanian Reliefs ", Journal of Roman Studies XLIV (1954), p. 76.

15 W. Hinz, " Das Sassanidische Felsrelief von Salmis ", Iranica V

(1965), p. 158. 16 For a description of this relief see Hinz, Iranica V, p. 148 ff.

In that article Professor Hinz dates the relief to the years of

joint rule, c. 239-42 A.D. In this paper, see pp. 73-4 below, it is suggested that the Salmis relief is the earliest of Ardashir's sculptures.

'1 Ghirshman, Parthians and Sassanians, figs. 67-9 (Tang-i Sarvak), fig. 70 (Sfisa); and R. Ghirshman, " Bard-ei Nechandeh, Rapport Prdliminaire ", Syria XLI (1964), p. 301 ff. and " Bard-e Nechandeh, Rapport prdliminaire de la seconde

campagne (Mars 1965) ", Syria XLII (1965), p. 289 ff.


Ardashir's investiture at Firizdbdd18 (P1. II) is a crude piece of work, both in actual workmanship and in composition. The scene shows the divine investiture of Ardashir by Ahuramazda: the god on the extreme left is handing Ardashir the circular diadem of kingship, from which are hanging two ribbons or diadem ties. This motif of a circle and ribbons was employed by Ardashir as his device in his other relief at FirfizdbTd, the battle scene (P1. IB and Fig. 5). Beneath the beribboned diadem is an altar, consisting of a column on a stepped plinth; the whitish patch just above this once carried an inscription.19

The god is shown wearing a crenellated or mural crown: a diadem terminating in long diadem ties is fastened round the base of the crown. His dress consists of an overcoat, a knee-length tunic, belted at the waist, and trousers: a coat over tunic and trousers was ordinary Parthian attire, for we see it being worn on the sculptures from Bard-i Nishdndeh.20 Ahuramazda's hair is shoulder-length, and his beard is long and curled. He is holding the barsom in his left hand: with his right he offers the diadem to Ardashir, which the king also grasps with his right hand. The king's left hand is held to his mouth in the familiar token of respect. He wears the same tunic and trousers as the god but carries, in addition, a sword. His beard is long and spade-shaped, as it also is at Naqsh-i Rajab: on his other sculptures it is usually drawn through a ring. His hair is shorn, as it is at Salmis and Naqsh-i Rajab. Ardashir is not wearing the jewel-studded tiara of the Parthian kings, which he wears on his early coins,21 but the skull-cap and korymbos or sphere. The composition of the korymbos has long been a matter of conjecture- and indeed it is unlikely that the problem will ever be finally resolved. One school of thought holds that the korymbos was a bunch of hair pulled to the top of the head and covered with silk.22 That this proposal is not altogether practical is clearly indicated in Ardashir's investiture scene at Naqsh-i Rustam (Fig. 4 and P1. IV). There Ahuramazda is shown with luxuriant curls on his shoulders and with a relatively small knop of hair rising through his crown. This style makes the fullest use of all his hair. Ardashir has similar luxuriant curls on his shoulders but his korymbos is at least three times as high as Ahuramazda's knop! Furthermore, Ahuramazda's hair forms a knop inside the circle of his crown. Ardashir's korymbos rises from a point on a tight-fitting skull-cap. If the korymbos were filled with his own hair, this would have to be pulled through a hole in the skull-cap, puffed into a huge ball, wider at the top than the base, and covered with a cloth which had in some way to be attached to the cap. To suggest that this perilous structure would then remain shapely, firm and in place would be bold indeed!

The custom of arranging relatively small bunches of hair in a plume or knop on top of the head was already known in Parthian times. Examples are seen, in profile, on the coins of Osroes (106/7-130? A.D.),23 and full-face on the coins of Vologases IV (191-207/8 A.D.).24 The same fashion occurs on bone plaques of the first and second centuries A.D. from Olbia on the north shore of the Black Sea.25 In Sasanian iconography, plumes of hair on top of the head are regularly worn by Ahuramazda and Anahita, when they invest the king with the right to rule. We see the style without the concealment of the divine mural crown in the relief of Shdhpfir and his sons at Naqsh-i Rajab (Fig. 9 and P1. VIII). The small figure just behind Shdhptir's back has his hair dragged up to the top of his head, secured by a broad ring and puffed up into curls: he has no side bunches of hair at all. Ardashir II at Tdq-i Bustin26 also has this knop of hair, secured by a ring from which flutter ribbons; his knop is not covered by material and again is considerably lower than the usual korymbos.

18 The relief is carved on the cliff bordering the Tang-i Ab river (right-hand side, direction Shirdz-Firfizabad), immediately adjacent to the ruins of a Sasanian bridge. In the Sasanian period the road passed directly beneath the relief, which is sited some 3 to 4 m. above the water. The attribution of the relief to Ardashir I is by crown. There was once an inscription between the god and king but this has been mutilated and erased. The right hand side of the relief is unfinished.

References to the relief include the following: Stein, Iraq III, p. 123; Herzfeld, Iran in the Ancient East,

p. 311, pl. CVIII; Herzfeld, " La Sculpture Rupestre de la Perse Sassanide ", Revue des Arts Asiatiques V (hereafter Herzfeld, RAA V) (1928), p. 130, pl. XXXV, fig. i; R. Ghirshman, " Firfizbhid ", Bulletin de l'Institut FranGais d'Archdo-

logie Orientale du Caire XLVI (hereafter Ghirshman, BIFAO XLVI) (1947), p. 7, pl. VIb; Vanden Berghe, Archdologie de l'Iran Ancien, p. 50; Ghirshman, Parthians and Sassanians, p. 131, fig. 167.

19 Ghirshman, BIFAO XLVI, p. 7. 20 Ghirshman, Syria XLI, pl. XX I; and Syria XLII, pl. XXI 3. 21 Ghirshman, Parthians and Sassanians, fig. 304- 22 Herzfeld, AMI IX, p. io6; Ghirshman, Parthians and Sassanians,

p. 126; and Hinz, Iranica V, p. 154- 23 Ghirshman, Parthians and Sassanians, fig. 150. 24 Ibid., fig. 153. 25 Ibid., fig. 352. 26 Ibid., fig. 233.

THE DARABGIRD RELIEF-ARDASHIR OR SHAHPUJR? 67

While most authorities have considered the korymbos to be formed of the king's hair, Sarre27 thought that it was made of metal strips covered with cloth, but this suggestion also lacks plausibility, for the rising curves of the globe are so smooth.

But to return to the Firfizabad investiture scene: ribbons are tied at the meeting point of korymbos and cap and a diadem terminating in long diadem ties is secured round the base of Ardashir's cap.

The smallest figure on the relief, that of the king's fan-bearer, stands behind Ardashir. His right arm is fully stretched out to hold the fan above the king's head. He wears the same clothes as the god and king and, like the latter, also carries a sword. His hat inclines forward in a way characteristic of " Phrygian " caps, and has on it a bud-like device.28 The fan-bearer is wearing a similar hat with the same device on Ardashir's second investiture scene at Naqsh-i Rajab (Fig. 3 and Pl. III), though on Ardashir's final investiture scene at Naqsh-i Rustam (Fig. 4 and Pl. IV) he is wearing the familiar high Sasanian hat, again embossed with his bud device. The third Sasanian knight in the Firfizdbid joust (Fig. 7) also wears this high helmet embellished with the bud device. In these three reliefs, at Naqsh-i Rajab, Naqsh-i Rustam and Firiizabid this figure can be identified not only by his bearing of the " bud " device, but also by a definite similarity of facial representation. He is regularly shown as a young, clean-shaven man with ornately curled, shoulder-length hair. In fact we may go so far as to say that a specific man is portrayed, a man who not only carried Ardashir's fan but who also fought valiantly in his decisive battle against their Parthian overlord. It seems likely that the short figure in the Firizibzid investiture scene represents the same man: he may have been carved smaller than the others because of the archaic convention of differentiating the size of the figures according to the importance of the person portrayed.

The three figures behind the page are shorter than the king but taller than the former. They are shown wearing Sasanian hats, tunics and trousers and carrying swords. Their right hands are raised to their mouths in respect, their left hands rest on their swords. They all have neat spade-shaped beards and heavy bunches of hair on their shoulders. The first of the three is differentiated from the other two by his bunch of hair being carefully curled, by the diadem tied round the base of his hat, the long crimped ties of which hang down his back, and by the device he bears on his hat. This device consists of a crescent over a circle on a horizontal base and is the same as the one carried by the middle Sasanian in the jousting scene (Pl. IA and Fig. 6). Like the fan-bearer, he is also represented on Ardashir's investiture relief at Naqsh-i Rajab (Fig. 3 and Pl. II); and he has been identified as Shdhpfir, Ardashir's son and successor.29 The helmets of the men following Shahptir are also embellished with diadems; and, according to Ghirshman, both originally carried devices, though these have now disappeared.30

Ker Porter had little good to say of Ardashir's next investiture scene at Naqsh-i Rajab31 (Fig. 3 and Pl. III): " it is the most inferior of any specimen I had hitherto seen of the ancient Persian sculptor; the proportions being bad and the work extremely rude ".32 However, when compared with Ardashir's Firfizabad relief, both the workmanship and composition are greatly improved. The god and king now form the central element; behind the king stand two figures, and two others stand behind the god, though these latter two are separated from Ahuramazda by a column or standard. The small figures of two children can still just be seen standing under the diadem, and they replace the altar in the Firiizabad investiture.

27 F. Sarre und E. Herzfeld, Iranische Felsreliefs (1910o), p. 68. 28 Herzfeld, RAA V, p. 130. 29 Ghirshman, BIFAO XLVI, p. 7. 30 Loc. cit.

31 The investiture of Ardashir is the earliest of the three sculptures carved on the rocks of this natural unroofed grotto, Ij km south of

Issakhr. The great cliff of Naqsh-i Rustam is clearly

visible from the grotto. The Ardashir relief, measuring 4 - 90 m. in length and 3-06 m. in height, takes the central position and faces West. The depth of carving is still fairly shallow-about 10-I2 cm. The finish is poor-there are heavy chisel markings on both the figures and the background, and once again it is

likely that this carving was abandoned before it was completed. The lower half is heavily damaged.

The relief is attributed to Ardashir I on account of the crown the king is wearing.

References include: Sarre and Herzfeld, Iranische Felsreliefs, pp. 94-6, pl. XII;

Herzfeld, Iran in the Ancient East, p. 311, pls. CVIII and CX, and RAA V, p. 130, pl. XXXV, fig. 2; Vanden Berghe, Archiologie de l'Iran Ancien, p. 23, pl. 25b.

32 Sir Robert Ker Porter, Travels in Georgia, Persia, Armenia, Ancient Babylonia, etc. etc., during the years 1817, z818, 1819 and 1820 (1821) I, p. 573.

7


The two men attending the king are shown wearing coats, with long elaborately pleated sleeves, over tunics and trousers: round the waist is a belt from which a strap supporting the long Sasanian sword is slung. The left-hand figure has his right hand held to his mouth, while the left is hidden in his

sleeve, both gestures familiar from many earlier reliefs. He wears a high hat, which does not appear to

carry a device, though tied round it is the diadem terminating in diadem ties. We would expect him, on the analogy of the Firfzabid reliefs, to carry Shahptir's device and could then identify him with

certainty as the king's son participating in this important ceremony. Herzfeld has so identified him,33 although Sarre believed that he represented the Grand Vizier or Commander-in-Chief.34

Fig. 3. Ardashir's investiture at Naqsh-i Rajab.

The man standing beside " Shthpiir " wears the " Phrygian " hat embellished with a bud device

(see above). His left arm is placed across his chest and we can be sure that his hand is concealed in his

sleeve, like ShThptir's. His right hand holds Ardashir's fan or fly-whisk. The king also wears a tunic over trousers, but this time under a cloak, which is just a little shorter

than the tunic. His sword hangs at his left hip, rather than centrally as do those of his attendants, and this enables the sculptor to carve ribbons hanging from his belt. Round his neck is the familiar heavy necklace of roundels. His beard is spade-shaped and his hair beneath his cap is cut very short. Diadem ties hang from the diadem tied round his cap, and short ribbons flutter from the meeting place of

korymbos and cap. The right to wear diadems35 and ribbons (and to embellish one's horse with them)

33 Herzfeld, Iran in the Ancient East, p. 31 I, and RAA V, p. 130. 34 Sarre and Herzfeld, Iranische Felsreliefs, p. 95. 35 It is likely that the Sasanians actually called this attribute a

diadem-Middle Iranian dydym. The diadem has a long history as an attribute of royalty in the Near East. The Parthian kings wore it and we see it on the stela of Artabanus V from Sfisa, now in the Iran Bastan Museum in Tehran- Ghirshman, Parthians and Sassanians, fig. 70; the central figure in the Bard-i Nishdndeh bas-relief, also in the Tehran Museum, wears a diadem with long fluted ties-Syria XLI, pl. XX i.

It was the invariable attribute of royalty as portrayed on Hellenistic coinages: and there is evidence that Alexander the

Great wore the diadem as the insignia of royalty. According to Xenophon, the diadem was worn by the Achaemenians. In his imaginary work Cyropaedia, Xenophon observed that Cyrus " had also a fillet (diadema) about his tiara, and his kinsmen also had the same mark of distinction, and they retain it even now " (VIII iii I3).

I am indebted to Dr. A. D. H. Bivar for much of the above information.

THE DXRABGIRD RELIEF-ARDASHIR OR SHAHPUOR? 69

was certainly a royal or divine privilege-and the king was, of course, considered to be divine. In his Firtizabdd investiture (P1. II) Ardashir, Shihpir and Ahuramazda wear the diadem: in this relief diadems are worn by the god, the king, his son Shdhptir, the Sasanian child who may be Hormuzd I, and Ardashir's queen: in his Naqsh-i Rustam investiture (Fig. 4 and P1. IV) Ahuramazda, Ardashir and the defeated Parthian, Artabanus, wear the royal diadem-the fan-bearer and " Ahriman " do not: and in the Firiz•bdd battle scene Ardashir, his son and their horses wear diadems and ribbons, as do their opponents. While Ardashir granted his royal opponents the right to wear ribbons, Shihptir did not: in his Naqsh-i Rustam triumph (P1. VII) the supplicant Roman, Philip the Arab, wears only his imperial wreath.

Ardashir's left hand is raised to his mouth, his right firmly grasps the diadem. The tie of the diadem is just off the vertical, gently indicating movement and showing that the diadem is handed from the god to the king. In his Naqsh-i Rustam investiture this element of motion is emphasized (Fig. 4 and P1. IV). The god is handing Ardashir the diadem with his right hand, his left holds the barsom or sacred bundle of twigs. Like the king he wears a cloak over his tunic and trousers.

The figures of the two children are heavily damaged. As far as can be seen the left-hand one is a miniature version of Shthpiir, dressed in the Sasanian tunic, trousers and hat, from which hang fluted diadem ties. He has been identified as Hormuzd I (272-73 A.D.), the grandson of Ardashir.36 The right-hand child is naked, except perhaps for a diadem around his head: he holds the barsom in his right hand. It has been suggested that he represents a Greek deity37-but would he then hold the Zoroastrian bundle of twigs ? He is clearly a satellite of Ahuramazda.

The sculptors of the Firizab•d relief certainly realized the compositional vacuum left behind the

god, and at Naqsh-i Rajab they have filled this gap not with divine attendants-Ahuramazda must have been considered complete in himself-but with two women separated from the god by a sort of column. Ker Porter describes them as " two figures, with visages hideously ugly, yet having the contour of women "!38 Herzfeld identified them as Ardashir's wife and her attendant:39 Sarre, however, considered them to be either Ardashir's wife and mother or his wife and a princess;40 while Curzon thought they were eunuchs.41 Only one of the ladies wears a diadem and tie and therefore Herzfeld's identification seems the most probable. They have their backs to Ahuramazda: this clumsy stratagem, to which the sculptors were forced to resort in an attempt to avoid placing a long retinue behind Ardashir, has resulted in the queen raising her hand to her mouth in respect-but only in respect to a blank rock wall!

The queen wears the high Sasanian hat with neck-guard and long unfluted diadem tie and is dressed, as far as can be seen, in an ankle-length coat. Her hair is arranged in heavy ringlets on her shoulder. Her attendant, of whom only the head and shoulders are visible, wears a " Phrygian " hat, which terminates in a feline head.

It is not only its better state of preservation that makes the investiture relief at .Naqsh-i Rustam42 (Fig. 4 and P1. IV) so much more impressive than the earlier sculptures of the reign of Ardashir. While these

36 Herzfeld, Iran in the Ancient East, p. 3I I. 37 Loc. cit. 38 Ker Porter, Travels I, p. 572.

30 Herzfeld, Iran in the Ancient East, p. 312, and RAA V, pp. 130-1. 40 Sarre and Herzfeld, Iranische Felsreliefs, p. 95.

4x G. N. Curzon, Persia and the Persian Question II, p. 127. 42 Ardashir's investiture scene is the most westerly of the eight

Sasanian reliefs carved on the cliff of Naqsh-i Rustam: it is located at the end of the cliff some 2 m. above present ground level. It measures 4- Io m. in height and 6. 70 m. in width. The carving is very deep and some elements are almost in the round. The finish is fine--on the figures all the marks of the chisel have been polished away and this relief is the first to give the impression of having been completed rather than abandoned before the projected work had come to an end. Both

the riders are identified by inscriptions. On the king's horse the Middle Persian reads: " This image (is) this of the Mazda- worshipping god Ardashir, king of kings of Iran, who (is) a scion of the gods, the son of Pdpak, the king ": and on the god's horse the Middle Persian reads: " The image (is) this of Hormuzd, the god ". E. Herzfeld, Paikuli, Monument and Inscrip- tions of the Early History of the Sasanian Empire (I924), pp. 84-5.

References include: Sarre and Herzfeld, Iranische Felsreliefs, pp. 67-71, pl. V;

Herzfeld, Iran in the Ancient East, pp. 312-13, pl. CX, and RAA V, p. 131, pl. XXXVI, fig. 3; Sarre, Survey, p. 494, pl. I54 a; Vanden Berghe, Arche'ologie de l'Iran Ancien, p. 24, pl. 28 c; Ghirshman, Parthians and Sassanians, p. 133, fig. i68.

Inspiration for Ardashir's novel design may have been provided by works similar to the sacrificial bas-relief from Palmyra illustrated in Parthians and Sassanians, fig. 86, which shows opposed horses and riders.


have the characteristics of friezes on the largest possible scale, the Naqsh-i Rustam investiture achieves a three-dimensional quality hitherto unsurpassed in Sasanian sculpture. Both in its overall composition and in its detail this relief proves that the craftsmen concerned could truly claim to be sculptors, while their predecessors were little more than calligraphers in the tradition of the Tang-i Sarvak and Suisa sculptures. By their placing of the two horses these sculptors have progressed beyond the totally vertical nature of the other two investiture reliefs and have transformed into an effective art form the flattened image of the Firtizdbdd joust.

While in his earlier reliefs Ardashir was shown either being invested with divine power or triumphing over his former suzerain Artabanus V, in this relief he combines the two themes. Both god and king are mounted on sturdy ponies who stand beside (but not on) the fallen enemies of their masters. The arrangement of the figures is strictly symmetrical and heraldic. The slight movement of the diadem ties seen in the Naqsh-i Rajab investiture is emphasized here-they now stream out behind the diadem. This movement is echoed by the god's cloak, which billows out behind him, showing a breeze blowing

Fig. 4. Ardashir's investiture at Naqsh-i Rustam.

from the king. As we shall see, Shahpiir is so attracted by the movement of ribbons and the like in his reliefs that he has them fluttering in opposed directions. The billowing of the god's cloak is balanced by a familiar figure, the page or fan-bearer.

The stocky ponies are small in relation to their riders, whose feet, indeed, hardly clear the ground. Their bridles consist of a plethora of straps-a brow-band, two nose-bands, two throat guards and a curb strap: these are embellished with many discs (almost certainly made of gold), and with ribbons tied to the head-piece. The saddle-cloths, with pommels to support the riders' thighs, are held in place by wide straps round the horses' chests (breast-plates) and hindquarters. Like the bridles, these straps are decorated with plaques-simple discs with rosettes on the god's horse, but embossed with lions' heads on Ardashir's. Balloons of hair are suspended from the saddle-cloths by " chains " or plaited ropes: in shape they resemble an inverted pear, terminating in a point. These balloons or " tassels " must have been regulation equipment on Sasanian horses: they can only have weighed a little for,

THE DARABGIRD RELIEF-ARDASHIR OR SHAHPiR ? 71

when the horses are engaged in combat as in the Firtizabad battle and in the later jousting scenes at Naqsh-i Rustam, the balloons are shown flying out above the horses' hindquarters. Ker Porter described one as a pendant tassel consisting of a " quantity of hair, silk, or probably gold fringe, issuing from an acorn-like head; forming a singular appendage to equestrian trappings .43 They must also have been standard Parthian equipment, for one can be seen flying out behind the horse of the leading Parthian in the Firtizabdd joust (P1. IB).

The horses' tails are tied with a ribbon at the top but otherwise hang free. They are standing on three legs, their fourth (the further forelegs) are raised and nearly touch each other; their ears do touch. The hoofs seem to rest on the necks of the corpses: the fleshy curve of the upper foreleg balances the heavy curve of their necks.

Both riders are sitting with their legs hanging straight down and are shown without either feet or any break beneath their flowing trousers to indicate them: ribbons hang from where their feet would be. For the first time the sculptors have succeeded in indicating the soft material of which Sasanian clothes were made. As well as trousers, they wear tunics with banded sleeves and cloaks. Ahuramazda has a ribbon at his waist, the king's is concealed by his right arm reaching forward. Around the god's neck is a heavy chain of office. As usual, he has a spade-shaped beard and bunches of hair on his shoulder: a small knop of hair also rises above his crenellated crown. The diadem is tied round his crown and the long ties probably reached the horse's back, as do the king's, though in this case the ends are hidden by Ahuramazda's fluttering cloak. The positions of the hands of both the god and the king are the same as in the other investiture scenes at Firtizabdd and Naqsh-i Rajab: but, since they are now mounted, they should, of course, use one hand to control the horse. Shdhpiir makes this change in his investiture scenes and we see both god and king holding the reins with their left hands. As at Salmis and in the

Firlizab~d joust, Ardashir's beard is drawn through a ring. His hair falls on his shoulder: his skull-cap, which now has ear-flaps and a neck-guard, is encircled by the diadem, from which hang long crimped diadem ties, and is surmounted by the korymbos.

The king's fan-bearer stands beside the horse's haunch and is visible only above the waist. He wears the sleeved tunic and high hat of Sasanian fashion. On his hat is the bud device, by which we have recognized this figure in Ardashir's other reliefs.

Ardashir's fallen enemy is readily identifiable as his former Parthian overlord, Artabanus V,~4 for we can see the device which he carries at the Firizbb•d joust on his high hat. His hat has the same neck and cheek-guards as the king's and from the diadem encircling it append the crimped diadem ties of royalty. The edge of his hat is serrated or " toothed ", as it is on his stela at SiIsa.45 Some of the Hatra kings wear this " toothed " hat,46 as does one of the family of Bahrdm II in the adjacent relief.47 Artabanus' beard is forked but smooth and beneath it can be seen his heavy necklace.

Ahuramazda's horse stands beside a figure usually identified as Ahriman, Ahuramazda's eternal enemy.48 Ahriman's curls are arranged formally in two opposed rows, bound by a fillet, the ends of which, knotted on his brow, are in the form of snakes' heads. Ker Porter describes his ear as being formed of " the united ears of a rabbit "!49 His beard is wavy and pointed.

While the three investiture scenes discussed above form a coherent group, which shows a logical sequence of development, Ardashir's other two reliefs are carved in a completely different style. The joust at

Frr~zdbddso (Figs. 5-7 and P1. I) is one of his earliest and is simply a two-dimensional picture in

a3 Ker Porter, Travels I, p. 541. 44 Sarre and Herzfeld, Iranische Felsreliefs, p. 69; Ghirshman,

Parthians and Sassanians, p. 133. Herzfeld, Iran in the Ancient East, p. 312, would, however, recognize him as Ardavazd, the last Parthian to issue coins.

45 Ghirshman, Parthians and Sassanians, fig. 70. 46 Ibid., figs. Ioo and Io2. 47 The third figure on the king's left, ibid., fig. 212. 48 Herzfeld, Iran in the Ancient East, p. 312; Sarre and Herzfeld,

Iranische Felsreliefs, p. 7o; and Ghirshman, ibid., p. 133. '* Ker Porter, Travels I, p. 550.

50 The relief is carved near the top of the mountain on the right- hand side of the Tang-i Ab just as the river gorge opens out into the plain. It measures approximately I8*8 m. in length, although it may once have been a little longer, and 4 m. in height. The right-hand side is heavily damaged. Attribution to Ardashir I is neither by crown nor by inscription. It rests on the event depicted, the individuals' monograms and Ardashir's hair style, which Professor Ghirshman maintains is the " origin of the korymbos or sphere which is placed above all the royal crowns of the dynasty " (ibid., p. 126).

[continued overleaf


stone. There is no attempt at modelling. Like his two early investiture scenes, this relief is also unfinished and some areas are only roughly outlined. The incomplete state of the relief and the heavy damage it has sustained through weathering has obscured many of the details, though the general theme is clear. It depicts the victory of three Sasanian mounted knights over their Parthian opponents.

The three Sasanian chargers are in full gallop, their legs extended in true Sartorius style. As well as a " saddle cloth", the horses are covered with a cloth embossed many times with their riders' devices. Flying from straps attached to the saddle cloths are the strange striated balloons, discussed above. Their tails are plaited and doubled up, presumably to keep them out of the way. The tails of the two front horses have ribbons tied to the top; and they also carry beribboned plumes on their heads.

The Sasanian knights wear a jerkin or corselet over a chain mail shirt, the links of which have been carefully sculpted. This is the earliest known representation of chain mail and it may well be that the Sasanians were the first to use it, for their Parthian opponents are shown wearing the more usual " banded " mail, which is formed of strips of metal backed onto leather or cloth. Armour of this type was already common a millennium earlier and was, for instance, used by the Assyrians. Many hundreds of such metal strips, in some cases still attached to their backing of leather or fabric, were found in Fort Shalmaneser at Nimrfid.51 The Sasanian knights still protected their legs with similar metal strips, to which were attached some stiff frills.

Fig. 5. The first Sasanian knight in the Firizdubdd joust, identified as Ardashir.

Parthian armour also consisted of a stiff corselet over a mail shirt, but this shirt was formed of metal strips on the arms and metal squares round the neck and over the hips-the more jointed squares presumably allowed increased mobility. Like the Sasanians, their legs were protected by metal strips. Sasanian chain mail did not immediately oust " banded " metal shirts, for the upturned " Roman " on one of the later jousting scenes at Naqsh-i Rustam is shown still wearing it.52

The first pair of contestants at Firfizbb•d are Ardashir (Fig. 5), identified both by his device and his head-gear, and his Parthian suzerain, Artabanus V. Ardashir, tilting furiously at Artabanus, strikes him with his lance, the impact of which has the unexpected effect of causing Artabanus and his horse to turn a somersault! Ardashir is wearing a necklace of heavy roundels: two straps cross his chest and support a further large disc in the centre. We see him wearing similar crossed straps on the relief at

51 M. E. L. Mallowan, Nimrud and Its Remains II (1966), pp. 52 Ghirshman, Parthians and Sassanians, figs. 219-20. 409-11.

continued from previous page] References include: Stein, Iraq III, p. 122; Herzfeld, Iran in the Ancient East,

p. 310, pl. CIX; RAA V, p. 131, pl. XXXVI, fig. 4, and

Archaeological History of Iran (1934), P. 78, pl. XI; Ghirshman, BIFAO XLVI, pp. 8-12, pl. VIIa, and Parthians and Sassanians, pp. 125-30, figs. 163-6; Vanden Berghe, Arche'ologie de l'Iran Ancien, pp. 50-51, pl. 71.

THE DARABGIRD RELIEF-ARDASHIR OR SHAHPUR? 73

Salinms.53 His neat pointed beard is tied with a ribbon-at Salmins it is passed through a ring.54 What

appears to be his hair forms a great plume on top of his head and also streams out behind him. This mass of hair is carved quite differently from that of Shdhpiir and the page, with both of whom it is carefully curled and it seems likely therefore that Ardashir is wearing a wig. It is perhaps relevant that in his other early sculptures-at Salmins (see below), Firtizabdd and Naqsh-i Rajab-Ardashir is

regularly depicted with close-cropped hair. It is only in his later Naqsh-i Rustam investiture that he is shown with long hair: on the other hand his coin portraits do show him with long hair.

Round his head is tied a diadem, from which fly back the diadem ties: from further down his back another streamer blows out, perhaps representing his cloak. As we shall see this trait of streamers from both the head and lower down the back is paralleled at Salmins.55

Very little remains today of Ardashir's opponent, Artabanus V. We can see the royal ribbons tied to his belt and ankle, and the device just behind his horse's girth. A " balloon " flies up behind his horse and we can therefore assume that the use both of devices and of balloons was common Parthian

practice.

Fig. 6. The second pair of contestants in the Frurzdbddjoust, identified as Shdhpiir and Darbendan.

The second pair of contestants are the best preserved of the three (Fig. 6 and P1. I): their positions mirror those of Ardashir and Artabanus-there are only a few differences of detail. The Sasanian

knight (P1. IA), identified as Shahptir, wears a " Phrygian " style helmet, which possibly once terminated in an animal's head. His hair is loosely curled, his beard long and spade-shaped. Two round balls are attached to his shoulders. His chest is crossed diagonally by one wide strap. The better

preserved details of his horse clearly show the " pommel " against which riders could brace their thighs to help withstand the sudden thrusts of battle-a great aid to stability prior to the introduction of the

stirrup. Shdhpir's Parthian opponent has been identified as the Parthian Grand Vizier, Darbendan, whom,

according to Tabari,56 Shahpir killed. A diadem tie flutters from his plumed Parthian helmet, but he does not wear the waist ribbon, nor does he seem to carry a device.

The third and final pair are engaged in hand to hand combat (Fig. 7). The Sasanian, Ardashir's

page or fan-bearer, has presumably swept his opponent off his horse. His left arm grasps the Parthian over the shoulder and across the chest. His right arm, to which the Parthian is clinging, still holds the reins. These non-royal combatants wear no ribbons-the page is identified by his device, the bud.

The remaining relief ascribed to Ardashir-the one at Salmds-is again extremely crude. The

roughness of the Firizabaid reliefs is readily explicable by their earliness-but the Salmis relief, recently

5- Hinz, Iranica V, pl. XLVI-XLIX.

64 Loc. cit.

-5 Loc. cit.

56 Chronique de Tabari II, p. 73.


described in detail by Professor Hinz,57 has been attributed by him to Ardashir's final years, for he maintains that it shows father and son in their years of joint rule. On stylistic grounds one is tempted to query this dating, for two features definitely link this sculpture to the period of transition from Parthian to Sasanian rule. Firstly the entirely frontal representation of all the figures is a Parthian and not a Sasanian trait-and indeed it does not occur again until the reign of Bahrdm II.58 Secondly the flat method of carving with its mass of superficial detail is clearly related to crude late Parthian work, for instance the stela of Artabanus V at Sisa,59 which it closely resembles: it appears to precede the more competent flat relief of the Firtizdbdd joust. Similarities of detail with Firtizabad have already been referred to above.

Fig. 7. The third pair of contestants in the Firfzedbddjoust, the Sasanian page and his Parthian opponent.

Hinz's argument that the Salmis relief depicts the period of joint rule is certainly a strong one, for both Ardashir and ShdhpSir are shown wearing Ardashir's crown. However, if we abandon this attractive theory and place the relief as the earliest in Ardashir's series, the correct place stylistically, we may consider it as commemorating the campaign of Ardashir and the king of Adiabene up the Tigris into Armenia.60 This foray was undertaken even before Ardashir finally defeated the Parthians. With the Salmis relief thus placed early in the sequence, Ardashir's sculptures all fall into a logical pattern. Beginning in the Parthian tradition of " drawing " in stone at Salmis and Firtizbad, Ardashir then starts to appreciate the value of modelling and the strength of a balanced composition: by somewhat desperate means he forces the Naqsh-i Rajab sculpture into a symmetrical form. From that he develops a totally convincing and successful style, employing high relief, rigid symmetry and stern repression of unnecessary detail. At Dirdbgird, however, we shall see the weakening of this strong Sasanian style through the insidious influence of Rome.

57 See the recent article by W. Hinz, " Das Sassanidische Felsrelief von Salmis ", Iranica V, pp. 148-60, pls. XLIV-LI.

58 The relief at Naqsh-i Bahrim, at present ascribed to BahrAm II, shows the king, represented frontally, accompanied by four Sasanian nobles, who are shown in the usual stance of profile

heads and frontal bodies: Parthians and Sasanians, fig. 214. In his other reliefs Bahrdm II also adopts the Sasanian stance.

5- Ghirshman, Parthians and Sassanians, fig. 70.

60 D. Oates, Studies in the Ancient History of Northern Iraq (1968), p. 74.


THE RELIEFS OF SHAHPOR (241-272 A.D.)

When Ardashir successfully challenged the Parthians in 224 A.D. Shthpfir had already attained manhood, for he is shown with a long beard on both the Firiztbdd reliefs. He must, therefore, have been at least in his teens. The next seventeen years, which were his most formative, were spent in his father's shadow, although he was appointed co-regent in Ardashir's final years. It is not surprising therefore that his earliest relief, the one commemorating his coronation or investiture,61 is modelled on his father's Naqsh-i Rustam relief. But although they are only separated by as short a period as a decade, a

comparison between Ardashir's relief at Naqsh-i Rustam (Fig. 4 and P1. IV) and that ascribed to

Shihpiir at Naqsh-i Rajab (Fig. 8 and P1l. V) is as striking as a similar comparison would be between a

Romanesque sculptural group and one in the Baroque manner. The simplicity of Ardashir's relief is replaced by an emphasis on flamboyant detail. These two reliefs have in common the opposed mounted figures of the king and the god, though in the later group their positions are reversed. However, despite

Fig. 8. Shahpfr's investiture at Naqsh-i Rajab.

this reversal the god and the king still both stretch out their right arms. In the Naqsh-i Rajab relief

ShThpir reaches for the fluttering ribbons rather than the actual ring: it is indeed the sculptural play with these and other ribbons and with the draperies of the two riders that gives ShThpfir's relief such a " baroque " character. Yet despite this accent on movement-the ribbons are blown in both directions -the overall effect is just as static as that of his father's relief. It is as if the details of these fluttering

61 The relief is sited on the right-hand wall of the Naqsh-i Rajab grotto, facing approximately north-west. The width at the base of the relief is 4-68 m., and the figure of Ahuramazda measures 2 -70 m. in height. The depth of the relief work is considerable, though a little less than that on Shdhpfir's other relief opposite. The figures were carefully carved and polished to remove the marks of the chisel, which can still be seen on the ground.

The relief is attributed to Shdhpfir I only by stylistic com- parisons with Shdhpfir's other reliefs, for there is neither an inscription, nor is the king's crown preserved. The relief has

been heavily damaged and little remains of the right-hand horse and rider, representing the king. The carving of the

draperies, the horses and their trappings compare closely with both Shdhpfir's Triumph at Naqsh-i Rustam, attributed by crown, and his other relief at Naqsh-i Rajab, which is inscribed.

References include: Sarre and Herzfeld, Iranische Felreliefs, pp. 97-9, pl. XIII;

Herzfeld, Iran in the Ancient East, p. 314, pl. CXII, and RAA V, p. 132, pl. XXXVII, fig. 6; Ghirshman, Parthians and

Sassanians, fig. 339; Vanden Berghe, Archdologie de l'Iran Ancien, p. 24, pl. 27.


ribbons and garments had been added as an afterthought, though being largely on the outer surface of the stone, they must, of course, have been in the sculptor's mind from the very beginning.

In Shahpfir's relief the king and the god are totally unaccompanied and the horses stand on the stone ledge without the encumbrance of defeated enemies. The omission of the triumphal element in this investiture is probably not just on account of aesthetic reasons, for in his investiture at Bishpiir (P1. VI) not only are corpses trampled on but also a new element is introduced of a supplicant kneeling between the opposed horses. Ardashir and Shdhpiir, despite their considerable successes in the East, appear to have considered only triumphs against their Western foes, first the Parthians, and then the Romans, worthy of commemoration. Shihpir's Naqsh-i Rajab investiture was, therefore, probably carved immediately after his coronation on March 20oth 242 A.D.62 and before his decisive Roman victories in 244 A.D.

The rippling of the clothes in this relief is exaggerated and formalized. The cloak flying out behind is arranged in stiff stacked folds and the folds of the tunic are carved in the same manner. The horses too share in the overall elaboration of the relief: from their " hogged " manes hang three wavy locks, not seen before. A crest is knotted above their heads and ribbons flutter from their bridles. Their riders grasp the reins firmly with their left hands-a revived feature for Ardashir did so only at Salmas. Shahpir holds the reins in all his reliefs, except the Naqsh-i Rustam triumph, and this often results in a clumsy sculptural pose.

From the saddle cloth hang not only the balloons of hair but also three wavy straps, which first occurred at Salmis and again at Dardbgird. In the present relief, however, they are much more elaborate and terminate in roundels. The horses' tails are bound at the top and plaited at the bottom- the middle is left free.

Shdhpiir's second relief was probably the investiture scene carved at Bishdp;ir63 (P1. VI). Like his father's Naqsh-i Rustam relief it is commemorating both his divine investiture and his military triumphs. The arrangement of the scene is still strictly symmetrical. The two heraldically opposed horses of Shdhpiir and Ahuramazda are shown standing on the bodies of their enemies. The horses are slimmer than heretofore and placed further apart so that there is just room for a kneeling figure between them.

Of the left-hand rider, representing the god, though considerably better preserved than the right- hand figure, little exists above waist-level. He is wearing voluminous trousers, freely carved; his feet are mutilated but hanging from their remains is a pair of ribbons. Most of the trappings of his horse have disappeared. Presumably hanging from the saddle cloth are the three straps and " balloon " of hair, which this time is oval in shape. There are indications that the strap running from the saddle cloth to the tail, usually adorned with discs, was also present. The tail was tied with a ribbon at the top and possibly again at the bottom: it was " plaited " in tight spirals, as also occurs on Shdhpiir's late triumphs. It is probable that any horse standing in the position presented by the sculptor would have fallen over, for all the weight is placed on the two outside legs-the inner foreleg is raised and the inner hindleg is simply rested on the corpse. The dead man's head, which is trodden on by the horse's off-fore, is turned to the right. It is probably bearded and the hair is arranged in the same wiry curls as the " Ahriman " hair on Ardashir's Naqsh-i Rustam relief (Fig. 4 and P1. IV). The mouth is open.

62 A. Christensen, L'Iran sous les Sassanides (1944), pp. I8o-81, and CAH XII, ch. IV, p. 130.

63 From the site of Shdhpfir, a track goes along the left bank of the Rud-i Shpfipr. The first of the six reliefs in the river gorge is sited on the cliff wall, immediately adjacent to and about a metre above the road. The Bishdpfir reliefs have been numbered differently by the authors who have recorded them: Vanden Berghe calls this relief No. I, while Sarre has it as No. V and to Flandin it was Bas Relief B. The width at the base is 9 -20 m. and it measures approximately 5.50 m. in height. It is set within a deeply excised rectangular panel and the depth of the figures is considerable--the prostrate prisoners measure some 60o cm.

Unfortunately the relief has been greatly damaged by weathering and the top half has been eroded. Since neither crown nor inscription have survived, the relief is, like the Naqsh-i Rajab investiture, attributed to ShThplir on stylistic grounds. The better preserved lower section is very worn but may once have been well finished.

References include: Flandin and Coste, Voyage I, p. 51, pl. 48; Sarre and

Herzfeld, Iranische Felsreliefs, p. 222, pl. XLIV, fig. IIo; Herzfeld, Iran in the Ancient East, p. 315, pl. CXIV, and RAA V, p. 133, pl. XXXVII, fig. 8; Vanden Berghe, Arche'ologie de

l'Iran Ancien, p. 55, pl. 77a; Ghirshman, Parthians and Sassanians, p. 159, figs. 201-03.

THE DARABGIRD RELIEF-ARDASHIR OR SHAHPUR .

77

Of the right-hand horse and rider only the legs are preserved. They are in the same position and carry the same ribbons and trappings as the opposed horse and rider. The trampled body is, however, different: it depicts a beardless man whose turned head rests on his arm. He is shown wearing a knee- length tunic and ankle-high boots. Herzfeld does not identify this man64 but Ghirshman, following MacDermot, has suggested that he represents the young Roman emperor, Gordion III.65

The kneeling prisoner is looking towards Shahpir. His hands are held out in supplication, the left hand actually touching the horse's near foreleg. He is wearing a belted tunic and short cloak, which billows out in the wind. This is the only sign of movement preserved on the relief, although the equestrian figures probably wore fluttering cloaks. From the waist up the position of the prisoner tallies nearly exactly with that of the supplicating Roman on Shdhpfr's triumphal relief at Naqsh-i Rustam (P1. VII): he may once have been shown with the same closely clipped beard, bobbed hair and imperial wreath of that figure. Herzfeld66 suggested that this kneeling figure represented Valerian, who was only captured in 260 A.D., an identification recently reaffirmed by Professor Hinz.67 MacDermot, followed by Ghirshman,68 proposed more plausibly that it depicted Philip the Arab, who sued for peace in 244 A.D. If we accept this latter identification, we can date the carving of this relief to shortly after 244 A.D.-a date which would still be sufficiently near his coronation in 242 A.D. for Shthpoir to be celebrating that event: it is unlikely that Shdhpfir would still be commemorating his investiture eighteen years later.

In his BishTipir investiture/triumph Shahpir continues to use the " heraldic " composition devised by his father, but in the second record of his victory over Philip, carved at Naqsh-i Rustam69 (P1. VII), he frees himself from this convention, to which he never returns.70 In this relief ShThptir's sculptors have achieved a completely successful and coherent composition. The " fussiness " so characteristic of all Shdhpiir's reliefs is overcome by the dominant strength of the design. The king sits proudly on his horse, his left hand on his sword, his right grasping the raised arms of a standing " Roman ". Another Roman, half-kneeling beside the standing man, reaches forward in supplication.

The apex of the triangular composition is ShThptir's korymbos. The long diadem ties fly back from his crown and are arranged in two controlled volutes. Heavy bunches of hair frame his face, and his neatly trimmed beard is drawn through a ring. Between hair and beard can be seen the lobe of his ear, embellished with an ear-ring. Round his neck is a chain of large round beads: the curve of his cloak, arranged over his shoulders, follows the curve of the necklace. As at Naqsh-i Rajab, the cloak billows out behind and is arranged in regimented stacked folds. His tunic, similarly folded, splits at the thigh, level with the thigh-pommel, and flows back over the saddle cloth. His toes point down sharply and ribbons flutter from his ankle.

The massive horse is simply and effectively rendered in high relief, though his stance, like at

Bishdpir, is slightly unrealistic. His trappings are much simplified. The breast-plate and haunch strap carry only a single large roundel, decorated with a rosette: the balloon of hair is only indicated between the horse's hindlegs. As in Shahpir's Naqsh-i Rajab investiture, the tail is bound with a ribbon at the top, left free in the middle and " twisted " in spirals secured by another ribbon at the bottom.

ShThptir's two prisoners are both wearing belted tunics. The standing man's clings to his body and is therefore presumably made of some soft material. The skirt of the pleading figure is arranged in

64 Herzfeld, Iran in the Ancient East, p. 315. In RAA V, p. 133 Herzfeld suggests that the kneeling and the fallen Romans represent Valerian alive and dead!

65 MacDermot, Journal of Roman Studies XLIV, pp. 76, 78-80; Ghirshman, Parthians and Sassanians, p. 159, and " Le Triomphe de Chipour I ", Mitteilungen des Instituts fur Orientforschung XI (1965), pp. 3-6.

66 See note 64. 67 Hinz, Iranica V, p. 157- 68 See note 65. 69 Shihpfir's relief at Naqsh-i Rustam is sited in the great cliff

below and a little to the left of the Tomb of Darius the Great. The complete width of the relief, including the later bust of

Kartir, is II - 80 m., while the height of the frame (not of the korymbos) is 6. I8 m. The relief work is fairly deep and the finish fine.

The sculpture is attributed to Shdhpiir on account of his distinctive crown.

References include: Sarre and Herzfeld, Iranische Felsreliefs, pp. 77-80, pl. VII;

Herzfeld, Iran in the Ancient East, pp. 314-15, pl. CXIII-IV; RAA V, p. 133, pl. XXXVIII, fig. 7, and Archaeological History of Iran, p. 83; Sarre, Survey, p. 596, pl. 155 A; Ghirshman, Parthians and Sassanians, p. 16o, figs. 171, 204-05; Vanden

Berghe, Archdologie de l'Iran Ancien, p. 25, pl. 29 d.

70 Bahrdm I returned to the symmetrical format in his one relief, the mounted investiture scene at Bishapfir: and his son, Bahrdm II, also used it on his Naqsh-i Bahram relief.


heavy pleats and clearly indicates the thick cloth of which Roman tunics were made. Balancing these pleats are the carefully arranged billows of his cloak. A cloak is also draped over the shoulders of the standing man. Both men have short, chin-length hair. The pleading Roman's hair is bound by a laurel wreath which indicates his imperial status. As we have noted above some authorities have suggested that this prisoner represents Valerian,7 while others have proposed that he is Philip the Arab.72 Certainly the equation of Philip with the pleading figure is again plausible-and there has been a clear attempt at portraiture, for " Philip " is shown with the short curly beard and rather uncouth appearance associated with him from his coin portraits (P1. XVD).

It is difficult to see whether the standing figure's hair is bound by a wreath or not. It is also not certain whether he is a Roman and his identity is therefore something of an enigma. He may be an important prisoner, taken by the king's own hand in his Roman campaign, for his hands, covered by the sleeves of his tunic,73 are firmly grasped by Shahptir.

While it is comparatively easy to date Shahpir's Naqsh-i Rustam triumph-it must have been carved shortly after 244 A.D., just after the Bishapfir investiture/triumph-his other relief in this area, Shdhpar and his Sons (?) at Naqsh-i Rajab74 (Fig. 9 and P1. VIII), presents more of a problem. It clearly belongs to the second phase when he has broken away from the symmetrical format of his two investiture reliefs-and it precedes his final phase, which we see in the " Valerian " triumphs at Bishdpiir, when a single scene no longer suffices to record his victories but has to be supported by rows of additional scenes.

As in the Naqsh-i Rustam triumph, Shihpiir on his horse dominates the composition and Ker Porter has pleasingly described his person as " not deficient in royal port ".75 The composition is, however, much less successful. As on all the equine reliefs of Ardashir and Shihpfir (except the one at Salmis), the king is shown with legs and head in profile and body turned frontally.76 His right arm crosses his breast and in his hand is what appears to be a short whip or sceptre ?, terminating in a bird's head. He holds the reins in his left hand-his left elbow is clumsily concealed by his body. Much of his crown has disappeared; but from it the long and thick diadem ties ripple out in three formally conceived curves. His cloak, fastened below his necklace, flutters up over his shoulders and is also arranged in artificial folds behind him. His tunic crosses his chest diagonally and folds softly above the belt of roundels. Having split at the thigh, it undulates over the saddle cloth.

The horse is standing in exactly the same position as those on Shdhpnr's other reliefs, though he is more " fussily " represented. Ribbons blow from the bridle; the reins fall in curves on his neck; the three straps also seen on Shihpiir's investiture opposite, wave down his thigh; the three curls from his mane are long and wavy; his tail is, however, only bound at the top and is otherwise loose.

Following Shahpir are nine men, all except one wearing the high Sasanian hat. Of these the principal figure, shown full-length, stands immediately behind Shdhpfir's horse. His high hat carries

71 Sarre and Herzfeld, Iranische Felsreliefs, p. 78; Herzfeld, Iran in the Ancient East, p. 315; Archaeological History qf Iran, p. 83, and RAA V, p. I33; Vanden Berghe, Archeologie de l'Iran Ancien, p. 25; Hinz, Iranica V, p. 157.

72 MacDermot, Journal of Roman Studies XLIV, pp. 76, 78-80; Ghirshman, Parthians and Sassanians, p. 16o, and " Le Triomphe de Chipour I ", Mitteilungen des Instituts fir Orientforschung XI, PP. 3-6.

73 Herzfeld, Iran in the Ancient East, p. 315; Ker Porter, Travels I, p. 54o; Hinz, Iranica V, p. 157.

4 The relief is sited on the left-hand wall of the Naqsh-i Rajab grotto, facing south. Its maximum height is 4.20o m., its maximum width, 6.95 m., while the base width is 5.50 m. The depth of relief is at least 6o cm., and the finish of the figures is fine.

Attribution of the relief to Shihpfir is by an inscription, written in Greek and Middle Persian. The Middle Persian reads: " The image (is) this of the Mazda-worshipping god Shdhpiir, king of kings of Iran and non-Iran, who is a scion of the gods, the son of the Mazda-worshipping god Ardashir,

king of kings of Iran, who is a scion of the gods, the grandson of the god Pdpak ", Herzfeld, Paikuli, p. 86.

References include: Sarre and Herzfeld, Iranische Felreliefs, pp. 92-3, pl. XI;

Herzfeld, Iran in the Ancient East, p. 314, and RAA V, p. 132, pl. XXXVII, fig. 5; Sarre, Survey, pl. 154 b; Vanden Berghe, Archiologie de l'Iran Ancien, p. 24, pl. 26b.

75 Ker Porter, Travels I, p. 573-

76 J. Morier in A Journey through Persia, Armenia and Asia Minor to Constantinople in the years 18o8 and 18o09 (1812), p. 139, suggested that the king's head " presents a full face to the spectator ". The head has been entirely defaced and today gives little indication of its former position. However, without positive evidence to the contrary, it is hard to accept that it was carved frontally for in all respects the remains of the head exactly match other representations of profile heads, for instance the method of carving the bunches of hair and the fact that the

korymbos is slightly more to the front than the back of the crown. Had it been frontal, it would have been the only time Shdhpfr was so represented.

THE DARABGIRD RELIEF-ARDASHIR OR SHAHPOR ? 79

the device formerly worn by Shdhpir himself and Ghirshman has suggested that we see here Shihpair's son and heir, Hormuzd I.77 Diadem ties, the ends of which can be seen below his right elbow, hang from the diadem round his hat; his hair is arranged in a rich bunch of curls on his shoulder. Like the others, he wears a belted tunic under a coat, which is fastened by a rayed and ribboned brooch on his chest. Both his necklace and his belt are adorned with roundels. His hands, covered by his sleeves, rest on the hilt of his sword. The two men standing on his right are shown in exactly the same position, though they are a little shorter: behind them are carved the busts of two others. The neck-flaps of their high Sasanian hats conceal their hair: round each hat is fastened a diadem indicating the wearer's royal status-but no other attribute distinguishes this quartet. They are clearly royal but of less significance than Hormuzd, who is not only larger in size but is also given a fine bunch of hair and his own device.

Fig. 9. Shdhpilr and his Sons at Naqsh-i Rajab.

Only the heads and shoulders can be seen of the four men on Hormuzd's left: they are obscured partly by the horse's hindquarters and partly by the figure of Hormuzd himself. The two most important are standing immediately on Hormuzd's left and directly above him. They are distinguished by their devices: the man next to Hormuzd has in addition a foliated diadem, while the one above has a heavy bunch of curls. His hat tilts gently forward but does not seem to carry a diadem. The remaining two men, carved somewhat smaller in size, both wear diadems: one has it fastened round the usual hat and is not distinguished in any other way; the other wears his diadem tied directly round his head. His hair, drawn to the top of his head, is passed through a ring and arranged in a knop of curls. His necklace is not decorated with roundels, as the others are, but with crossed rectangles.

This relief is usually considered to represent Shahpfir and his nobles or his retinue.78 Since the reliefs of Ardashir and Shihpiir usually served a specific propaganda purpose, it is doubtful whether a

77 Ghirshman, BIFAO XLVI, p. 7. ?7 Herzfeld, Iran in the Ancient East, p. 314; Vanden Berghe Archiologie de l'Iran Ancien, p. 24.


relief simply illustrating Shthpuir and his immediate entourage would fulfil this criterion. More probable would be a scene showing Shahpfir with his sons, who doubtless ruled the various satrapies in his empire, for such a relief would serve to deify his entire family. Eight of the nine figures wear the royal diadem and the ninth is distinguished by his luxuriant hairstyle and his device. If they are not all Shahpir's sons (one is plausibly identified as his heir Hormuzd), they are certainly close relatives. It is perhaps relevant that no less than three of Shdhpiir's sons, Hormuzd I (272-73), Bahrim I (273-76) and Narseh (293-302), eventually ruled the Sasanian empire.

Since no historical incident is depicted in this relief its dating has to be based on purely stylistic criteria. We have already assigned it to Shahpfir's second stage, after he had freed himself from the heraldic convention devised by his father: and we know that this second stage began shortly after 244 A.D. Shahpir's third stage, when a single scene was no longer sufficient, is illustrated in the next two reliefs to be discussed, which it is here suggested were carved to commemorate Shdhpir's capture of a Roman emperor-i.e. after 260 A.D. So, Shahpir and his sons could have been carved any time between 244 and 260 A.D. However, the stiffness and massiveness of his sons suggests a date fairly early in Shahpir's reign: and indeed, this relief, the two investiture scenes at Naqsh-i Rajab and Bishdpir, and the Naqsh-i Rustam triumph were probably all carved in the first decade of Shahpfir's reign.

ShThpir's remaining two reliefs, Bishdpfar IV79 (Pls. IX and X) and Bishdpar VI80 (Pl. XI) are markedly different from the preceding four in a number of ways. Perhaps the most striking change is that the single scene featuring the king is no longer considered to record enough of the event commemorated. Though it continues as the focal point of the relief, it is now framed and supported in the one case by four, and in the other by two superimposed registers of men on horse and foot. Another notable innovation is that we see for the first time the entirely Western feature of a winged cherub, putto or Niki.

The central scene of both these reliefs (cf. Pls. X and XI) shows Shdhpiir triumphing over no less than three " Roman emperors "--one lying between his horse's legs, one, whose hand Shahpfir grasps, standing by the animal's hindquarters, and the third kneeling in front with outstretched arms. MacDermot has identified these three as the murdered Gordion III, the captured Valerian and Philip the Arab.81 If his identifications are correct, these two reliefs would epitomize Shihpiir's major triumphs over Rome, even though these were separated in time by some sixteen years-a theory which the author finds hard to accept.

When Valerian was captured in 260 A.D. and taken to Shdhpfr's city at Bishapir he was accom-

panied by a large number of Roman prisoners of war, who were settled in Iran.82 Bishipir IV is

probably the work of some of these Roman prisoners, for a novel feeling of life and movement can be

9 This is the first of the four reliefs on the left bank (looking from the site of ShApfir) of the Rud-i Shapfsr. It is markedly different from all other Sasanian reliefs, for the scale is half life size instead of " larger than life ". It is also carved on a heavily concave surface, which has given rise to the idea that it was modelled on Roman victory columns. The length over the curve is some 9 - 4 m., while the total height of all four registers is only 5.6 m. The relief is attributed to Shihpfir I on account of his distinctive crown.

Like all the other Bishdpfir reliefs it has been variously numbered-Vanden Berghe No. 3, Sarre No. IV and Flandin Bas Relief F.

References include: Flandin and Coste, Voyage I, pl. 53, PP. 59-64; Sarre and

Herzfeld, Iranische Felsreliefs, pp. 220-I, pl. XLIII, fig. iog; Herzfeld, Iran in the Ancient East, p. 318, pl. CXV; RAA V, p. 134, pl. XXXIX, fig. Io, and Archaeological History of Iran, pp. 84-6; Vanden Berghe, Archdologie de l'Iran, p. 55, pl. 78; Ghirshman, Parthians and Sassanians, p. 158, fig. 2oo00, and " Le Triomphe de Chfpour I ", Mitteilungen des Instituts fir Orient- forschung XI, pp. 3-6; Jean Gag6, " Comment Sapor a-t-il 'Triomph6' de Val6rien ", Syria XLII (1965), PP- 373-88.

80 Carved on the right-hand bank of the Rfid-i Shpfipr, shortly after Shdhpfir's investiture scene. It measures 4. 16 m. in

height and 2 - 20 m. in length and is attributed by the king's crown. This attribution would probably be reinforced by the

inscription a little in front of the horse's head, but this, un-

fortunately, has not yet been translated. References include: Sarre and Herzfeld, Iranische Felsreliefs, p. 223, pl. XLV;

Herzfeld, Iran in the Ancient East, pp. 316-17, pls. CXVI and

CXIX; RAA V, p. 133, pl. XXXIX, fig. 9, and Archaeological History of Iran, p. 84; Vanden Berghe, Archeologie de l'Iran Ancien, p. 55, pl. 77 b; Ghirshman, Parthians and Sassanians, p. 152-7, figs. 196-9, and Mitteilungen des Instituts fir Orient-

forschung XI, pp. 3-6, pl. I.

81 MacDermot, Journal of Roman Studies XLIV, pp. 76, 78-8o. He is followed by Ghirshman in ibid., p. 152, and in Mitteilun-

gen des Instituts fiir Orientforschung XI, pp. 3-6.

82 Herzfeld, Archaeological History of Iran, p. 84; Jean Gag6, Syria XLII, p. 380.

THE DARABGIRD RELIEF-ARDASHIR OR SHAHPUR ? 81

seen in this relief: the horses are not simply standing with a foreleg in mid-air-they are eagerly stepping forward; the kneeling prisoner is not frozen in supplication, he is pleading earnestly. Flandin commented :83

" Ce bas-relief F est sans contredit le plus remarquable de tous ceux que nous ayons 6t6 assez heureux pour retrouver en Perse. Cette superiorit6, il ne la doit pas seulement au sujet qu'il repr6sente, au nombre des figures qu'il contient, 'a la vari6t6 des scenes, il la doit encore a l'execution de toutes ses parties et t l'habilete de ciseau qui distingue cet important ouvrage."

Unfortunately the relief has been heavily damaged and many of the details are lost. What remains is, however, a spirited record of Shdhpiir's triumph.

The mass of the king, his charger, the corpse and the standing Roman are balanced by a group formed by the kneeling Roman, backed by two Persians, one offering a ring, the other standing with crossed arms. Flying above them is the winged cherub. The two groups are linked by the outstretched arms of the supplicant and by the ring offered by the Persian. The whole forms a complex but coherent sculptural design.

Exactly the same dramatis personae occur in the principal scene of Bishdpiir VI (P1. XI). But while the figures are the same there is a marked contrast in style. The " life " so evident in Bishapiir IV has disappeared: we have returned to the Iranian idiom of frozen movement. Indeed, this relief was probably a copy by local workmen of the " Roman " example on the other side of the gorge. These craftsmen have entirely misunderstood the opposed masses of the main scene. With a horror vacui they have spaced out the figures, so that Bishaptir VI is inferior not only in the actual sculpting but also in its composition.

To return to the original (P1. IX and X), the king is shown holding the prisoner with his right hand -their hands are clasped just above his long quiver-his left grasps the reins. He wears the usual silky tunic over voluminous trousers. His diadem ties and cloak billow out behind him in formal curves: ribbons also flutter from the meeting point of korymbos and crown and from waist and ankle. His horse's tail is plaited in spirals and bound with ribbons: only traces of the bridle and breast-plate remain, and these are simply represented. The animal's mane is hogged and the three curls, seen on most of Shdhpoir's reliefs, are omitted.

Only the legs of the standing prisoner survive; the corpse is, however, nearly complete. His head, which is damaged, is pillowed on his left arm, his right lies across his body. He is wearing a short tunic of some heavy material. A similar tunic is worn by the supplicant, all details of whose head are also lost. A sword hangs at his side. As at Naqsh-i Rustam his arms reach out to the king in pleading. It is unfortunate that all three of the Roman heads are damaged, for it would have been interesting to see whether they were all shown wearing the imperial wreath on a relief which was probably carved by Roman prisoners.84

The two Sasanians behind the supplicant both wear the royal diadem and diadem ties: as usual their bodies are represented frontally, while their heads are in profile. The face of the man nearest

Shdhpfir is one of the few that has escaped damage; it is imperious in expression, with a straight nose, curving moustache and neat pointed beard. He wears a high hat from which hang the diadem ties. His arms are crossed on his chest; his hands are discreetly hidden in his sleeves. The identical man in the same position occurs on Bishpflr VI, where, however, he does not seem to have a diadem. The other Persian prince stretches out in front of him and offers Shihpiir a ring. His left hand rests on his sword-hilt. His face and hat have been much damaged, though indications on the rock suggest that he was clean-shaven and wore one of the " Phrygian " hats, which perhaps terminated in an animal's head, as does the second Sasanian on Bishitpir VI. These men were probably Shahpir's two favourite sons, his heir Hormuzd and perhaps Bahritm. On both reliefs the bearded prince is shown nearer the king and we would expect him to be Hormuzd: there must be some hesitation in accepting this identification

83 Flandin and Coste, Vqyage, p. 60. 84 The laurel wreath in Roman iconography could only be worn

by one man, the emperor, or the person to whom he had delegated his authority.


for the figure named as Hormuzd at Naqsh-i Rajab85 looks totally different: there is however a time-lag of some fifteen years between his representation at Naqsh-i Rajab and in Bish~apir IV and VI, during which he may well have changed his appearance and hair style.

The naked winged child flies directly above the two princes: in his hands he is carrying an untied diadem. He is a completely foreign motif, which has its origins much earlier in Greece. There a male child represented Amor and a female Nike. Since Iranian victory symbols were usually male, we may assume that here the child, which is undoubtedly symbolizing victory, is male. This putto occurs only on these two reliefs: the one other related motif can be seen at Taq-i Bustdn,86 some 300 years later, where two clothed winged females fly above the arched entrance of the large grotto.

Following a short distance behind Shahpfir are a row of knights, presumably his finest cavalry (P1. IX). They clearly depict men of rank for they are differentiated from each other. No less than the first five sport the luxuriant curls, which seem to be a royal prerogative: the first three of these wear the simple high hat, the last two the forward-pointing hat, terminating in an animal's head. And behind this curly-haired quintet (whose curls would conceal their diadem ties) are a trio with high hat, short hair and long, crimped diadem ties. So these eight may also be representing Shdhpoir's sons or members of the royal family. Rows of cavalry, undifferentiated this time, fill the two registers above, and the one below the line of" Household Cavalry ".

To the right of the principal scene, behind the two princes, are four registers of vassals bringing tribute. The top two registers of tributaries are today partly obscured by weed. They show rows of men dressed in belted knee-length tunics over baggy trousers: they hold rings, dishes, bundles and bags and one leads a pair of lions. The bottom register also shows men bearing gifts. One leads a pair of horses

drawing a chariot. They are clothed in loose " dresses ", approximately knee-length, with uneven hems. Their legs are bare.

The third register is not separated in any way from the principal scene, of which it is a continuation

(P1. IX). It shows two rows of figures, one above and behind the other. The upper row depicts six men dressed in a short flowing robe: loose folds of cloth curve from one shoulder to the other, suggesting that these garments were made of pieces of cloth knotted on the shoulders. The men are clean-shaven, with short, straight hair and soft moustaches. The lower group consists of two men, preserved only in bare outline, leading a harnessed stallion. Between the horse and the elephant, who brings up the rear, are two more figures, one carrying a burden on his head. Only the front half of the elephant can be seen: he is ridden by a naked mahout, who is clearly of the same nationality as the moustached men in the

upper row, with whom he is level. Somewhat surprisingly the arrangement of Bishaipir IV has been compared with that of Trajan's

column.87 The positioning of registers, which is employed here-and across the gorge on Bishdpiir VI, bears little relation to the continuous spiral of narrative, which runs round the Roman column: it is indeed a return to the archaic Near Eastern convention of superimposed registers. We can assume that the Sasanians drew their inspiration for such an arrangement from the ruins of Persepolis, much of which would still have been visible, rather than from Rome.

Only two registers flank the principal scene in Bishdpir VI and neither forms a continuation of it, as the third register does in the other relief. We have returned to the familiar " larger than life " scale and the execution is clumsy. As in Bishpiir IV, the king is followed by his cavalry, of which the figures in the bottom row are differentiated: all five have thick bunches of hair and the final three wear " Phrygian " hats. The registers on the right again show men on foot. They are probably all soldiers, drawn from different parts of the empire and portrayed with their individual weapons-swords, spears, clubs and pikes.

Though entirely different in the scale and manner of carving, the almost identical principal scenes of

Bishpnpr IV and VI show that the two reliefs commemorated the same event and it is probable that

they were carved within a few years of each other. The innovations occurring in these two clearly separate them in time from Shihpir's earlier reliefs, and we must look for an event which would

81 See pp. 79-80 above. 86 Vanden Berghe, Archiologie de l'Iran Ancien, pl. 128 c.

87 Herzfeld, ibid., p. 84; Ghirshman, Parthians and Sassanians, p. 158.

P1. IA. Detail of IB; the second Sasanian knight, identified as Shahpar (photo: Rustami).

PI. IB. General view of the Fir-zaibad Joust (photo: Rustami).

c~ ? .r r (~i)l~?C

Z '~ I?' r ~z??1! Ii

r: n

t~F~'

I: U

t

Pl. III.

Ardashir's investiture at N

aqsh-i Rajab (photo: R

ustami).

Pl. IV

. A

rdashfr's investiture at Naqsh-i R

ustam.

Pl.

V.

Shdhpur's investiture at Naqsh-i

Rajab (photo: R

ustami).

Pl. VIA. Shdhpur's investiture/triumph at Bishapur.

P1. VIB: Detail qf VIA: Philip the Arab.

P1. VIIA. Shdhplir's triumph over Philip the Arab at Naqsh-i Rustam.

P1. VIIB. Detail of VIIA: the prisoners.

P1. VIIIA. Shahpur and his Sons at Naqsh-i Rajab.

Pl. VIIIB. Detail of VIIIA.

P1. IX

. P

art of Bishapiir IV

, Shahpar's triumph over V

alerian.

P1. X

. T

he principal scene of Bishapar IV

, Shdhpu7r's triumph over V

alerian.

P1. X

I. T

he principal scene of Bishapfir V

I, ShdhpfIr's triumph over V

alerian.

P1. XIIA. Composite view of the Ddrabgird relief.

Pl. XIIB. Distant view of the Darabgird relief.

Pl. XIII. The king and the four principal characters in the Darabgird relief.

Pl. XIV. Detail of the right half of the Darabgird relief showing three of the principal characters.

Pl. XVA. The caricatures of the Romans: the top right section of the Ddrdbgird relief.

P1. XVB. Detail showing the old man's head.

P1. XVC. Coin of the Roman emperor, Valerian (by courtesy of the Ashmolean Museum).

Pl. XVD. Coin of the Roman emperor, Philip the Arab

(by courtesy of the Ashmolean Museum).

Pl. X

VIA

. T

he Persian nobles behind the king in the D

drabgird relief. P

l. XV

IB.

Detail qof the heads of the P

ersians: the top left section of the Dtrabgird relief.

THE DARABGIRD RELIEF-ARDASHIR OR SHAHPJR ? 83

provide sufficient stimulus for a further series of victory reliefs. As he himself recorded, Shthpir was immensely proud of his seizure of a living Roman emperor,88 who was accompanied into captivity at Bishdpiir by many of his men. Here we clearly have ample stimulus for new efforts at relief carving, which are moreover aptly sited at Bishdpair, and we also have to hand foreign workmen to account for the importation of new ideas. It is therefore proposed here that Bishpiipr IV and VI-and only Bishdpfir IV and VI-commemorate Shahptir's final victory over the Romans. We do not know which of the Romans in Bishpfipr IV wore laurel wreaths, because their heads are so damaged: and in Bishdptzr VI they seem to be wearing caps with a thick rolled edge. It is impossible to say whether this rolled edge, which is much more coarsely carved than Shahptir's diadem, is meant to represent the imperial wreath. We cannot, therefore, be certain of the status of the three Romans: one must represent Valerian, but who the other two are is open to conjecture. As already noted, many identifications have been proposed :89 and we can be certain that, unless the inscription on Bishptir VI finally solves the problem, debate on their identity will continue.

Herzfeld claimed that the sculptures of Shdhpiir I showed " no logical development " and that all the victory scenes-the three at Bishpfir, the one at Naqsh-i Rustam and the Dardbgird relief- represented Shahptir and Valerian.90 He therefore wrote that all five works " are strictly contemporary and their subject is the same. Therefore the differences of design best elucidate how strongly not only foreign influence but foreign hands must have been at work during that period of Sasanian art."91 If, however, we can accept the chronological arrangement of Shahpir's sculptures proposed above, we see that they do show an evolution, principally in the changing form of composition but also in style. At the beginning of his reign Shahpiir was still influenced by the symmetry of his father's reliefs, from the domination of which he freed himself within a few years. After a period of considerable sculptural activity in the first decade of his reign, when four reliefs were probably carved, he commissioned no more for the next ten years. His decisive victory over Valerian inspired him to fresh sculptural efforts, and these two final reliefs are no longer bound in any way by the heraldic convention, nor do they rely on a single scene to record the event.

THE DARABGIRD RELIEF92

Having discussed the other reliefs of Ardashir and Shihpfir, we can now return to the vexed problem of the Darmbgird relief (Figs. I, 2 and io and Pls. XII-XVI). Which king does it represent and what event is he commemorating? The Sasanian king on his charger, who dominates the scene, is facing a group of three figures on foot-a young man eagerly stepping forward, an old man on whose head the king's hand rests, and a man leading a donkey. Less prominent, but also belonging to the principal characters, is the corpse lying alongside the horse. The Persians arranged in lines behind the king do not take part in the action-they are simply witnesses-and they are balanced sculpturally by serried ranks of excellently caricatured Roman soldiery.

The relief is strictly symmetrical and this style of composition is in itself an indication of the date of the relief. Ardashir only successfully achieved a balanced composition in his final investiture scene,

88 A long inscription on the walls of the Ka'bah-i Zardusht at Naqsh-i Rustam: see, Sprengling, " A New Pahlavi Inscrip- tion ", AJSL LIII (1936), pp. I26-44, and " Zur Parsik- Inschrift an der ' Kaaba des Zoroaster'", ZDMG 91 (i937), pp. 652-72; Henning, BSOS IX, pp. 823-49; A. Maricq, " Res Gestae Divi Saporis ", Syria XXXV (1958), pp. 295-360.

89 See p. 77-8 above. 90 Herzfeld, Archaeological History of Iran, p. 8o. 91 Herzfeld, Iran in the Ancient East, p. 314- 92 The relief, located some 8 km. over rough tracks from the

modern town of Darib, is carved on a cliff bordering a deep pool, on account of which direct access is difficult. Stein gave approximate measurements of height: 20 ft.; width: 29 ft. (Iraq III, p. I94). The horse and king are in high relief, while the other figures are more flatly carved. The finish is fine.

9

Attribution: Ouseley, Flandin, Stein, Herzfeld, Vanden Berghe, Ghirshman and Hinz have all identified the king as Shdhpfir I, even though the king is not wearing Shdhpfir's crown but Ardashir's. In this paper identification by crown is the method followed and it is assumed that the king represented is Ardashir. Stylistic criteria, see below, would agree with this attribution.

References include:

Ouseley, Travels II, pp. 145 ff., pl. XXXV; Flandin and Coste, Voyage, pp. 34-5, pl. 33; Stein, Iraq III, pp. 194-6, pl. XVII; Herzfeld, Iran in the Ancient East, p. 314; Vanden

Berghe, Archiologie de l'Iran Ancien, p. 46, pl. 64 b and c; Ghirshman, Parthians and Sassanians, pp. I6o-6i, fig. 206; Hinz, Iranica V, pp. 156-8.


which was probably carved c. 230 A.D.:93 and Shihpiir only retained the heraldic manner in his first two sculptures, after which he abandoned this convention. On grounds of composition, therefore, we would expect this relief to be carved between c. 230-46 A.D.

If this is correct, then Ouseley's94 original identification of the old man as Valerian must in any case be wrong, and, as we shall see below, the only similarity between the Ddrabgird head and coin portraits of Valerian is that both portray old men. Professor Hinz's95 recent proposition that this head is a portrait of Philip the Arab also lacks plausibility, though in this case the commemoration of that triumph would fit within our proposed dates.

The Roman heads clearly indicate that the victory scene commemorated is a triumph over the king's Western foes. Ardashir fought several successful battles against the Imperial forces.96 In 230 he attacked Roman Mesopotamia and threatened Syria. The Romans, led by Severus Alexander, retaliated two

Fig. io. The Dardbgird relief.

years later, though Herodian claimed that they failed. The campaign cannot, however, have been an unqualified Sasanian success for Ardashir did not follow it up. He again swept through Mesopotamia in 237, and then, in 241, Shahptir launched a further attack in the period ofjoint rule. This culminated in his success of 244 A.D. We have, therefore, no less than five campaigns which the relief could commemorate: Ardashir's successful raid in 230; his defeat (?) of the emperor Severus Alexander in 232; his 237 foray; Shahpir's battles in 241 carrying his father's standard; and Shdhptir's own triumph three years later over Gordion and Philip the Arab.

It is now time to make a detailed survey of the Dardbgird relief. The king and his horse (P1. XIII) are carved in high relief and they occupy about half the total area of the sculpture. The king is wearing Ardashir's distinctive crown of a neat skull-cap surmounted by the korymbos. Flying up from the diadem tied round his cap are the long heavy diadem ties, which denote royal status. His hair is long and is

93 See p. 70 above.

94 Ouseley, Travels II, pp. 147-8. 95 Hinz, Iranica V, p. 158.

96 Oates, Studies in the Ancient History of Northern Iraq, pp. 74-5.


arranged in luxuriant curls on both sides of his head in a manner that foreshadows Shihpfir's hair-style. Shahpiir is regularly shown with ornately curled hair: Ardashir was first shown with long hair, which then was fairly straight, at Naqsh-i Rustam. Round the Dratbgird king's neck is the familiar heavy chain of office: he is dressed in the usual tunic and trousers, but the material of which these are made has here become really soft and pliable. His tunic is arranged in pleats over the saddle cloth; his loose trousers fall softly around his leg; ribbons float from his feet, the toes of which point sharply to the ground. His cloak, fastened by roundels on his chest, flutters out in rich billowing folds behind him. The flying diadem ties and cloak are arranged in loose ripples which clearly precede the totally formal folds of Shahpfir's first relief, his investiture at Naqsh-i Rajab (cf. Figs. 8 and i o and Pls. V and XIII).

The king's left hand is stretched out to rest on the head of the old man; his right holds a short thonged whip. The reins hang loose on the horse's neck: as we have noted above, Shdhpiir usually holds his horse's reins, while Ardashir does not. The stallion stands on all four legs. Apart from the knight's charger at Taq-i Busttn,97 carved some 300 years later, this is the only relief in which a horse is shown standing with all his legs on the ground, for the further foreleg is usually raised. As in Ardashir's investiture at Naqsh-i Rustam, the corpse is lying alongside the horse: Shahpfir's horses at Bish-piir either stand on the corpse, for instance his investiture/triumph, or straddle it, as in his two late triumphs at that site.

The Daribgird horse's head is held high and in the " flexed " position. The bridle is embellished with many discs, almost certainly made of gold, and the long curb bit has two reins. The saddle cloth is held in place by wide straps round the chest and haunch: like the bridle these straps are decorated with circular plaques. Attached to the cloth is the " pommel " against which riders could brace their thighs. The balloon hanging from the saddle-cloth is again an inverted pear shape, sharply pointed at the base, a characteristic of Ardashir's rather than of Shahpir's balloons. There are three waving straps on the horse's hindquarters, such as were first seen at Salmis. They become regulation equipment on Shdhpiir's early carvings, where they are further embellished by roundels attached to their ends. They do not occur in his Naqsh-i Rustam triumph; and in his late Bishpipir triumphs the horses' hindquarters are obscured by the standing Roman. The Darabgird horse's tail is tied halfway down with one of the distinctive crimped ribbons, below which the tail floats out in three wavy tresses.

While the king on his horse is the focal figure, the next in importance are the young man and the old man (Pls. XIII-XIV). They are probably of equal significance, for while it is the young one who is portrayed full-length, the old man is larger in scale, which compensates for the fact that only his upper body is represented. The young man has short hair, bound by a diadem from which flutter short crimped ties, and a close clipped curly beard. He wears a knee-length belted tunic of flimsy material, and some sort of swathed leggings. A short cloak, also of soft cloth, flutters back over his shoulders and indicates the swiftness of his movement towards the king. A sword can be seen hanging between his legs. His right foot is actually stepping on the arm of the dead man. His left arm is stretched out towards the king and his hand appears to rest on the horse's shoulder between reins and breast-plate. Unfortunately his hand is damaged and we cannot see exactly what gesture he was making with it, though he appears to be making a sort of" V " sign-his two upper fore-fingers are separated from the lower pair. Until we know more about the significance of gesture at this time it is hard to decide what this young man is " saying ". He does not look downcast, as the old man and many of the Romans do; he is not on his knees pleading; he appears neither to be giving nor to be receiving anything. Could he be asking for something, or perhaps handing over his prisoner, the old man ?

It is not only what he is " saying " that is uncertain: his nationality poses another problem. Stein took it for granted that he was a Roman98-his dress, however, though somewhat similar, is not distinctively Roman. Roman tunics were made from a heavy material, as is well illustrated in that of the kneeling figure in Shahptir's trumph at Naqsh-i Rustam (P1. VII), and not of this flimsy cloth. Also the Romans did not wear the swathed leggings which we see on this man's legs. That he was royal is indicated by the diadem and diadem ties-was he perhaps one of the king's allies ?

0' Ghirshman, Parthians and Sassanians, fig. 235. '* Stein, Iraq III, p. 195.


The old man's head (P1. XVB) is a skilful portrait. He has a straight nose, heavy-lidded eyes, fleshy cheeks, thick neck, a strong chin and a large curving mouth. He appears to be clean shaven. His short hair is bound by a diadem ending in fluted diadem ties. The top of his head is covered by the

king's hand, but the angle of its palm suggests a receding brow. Many heavy folds of cloth, perhaps of a cloak, are arranged over his shoulder. Rising above the king's fore-arm is the old man's hand, with the

fingers separated and loosely curved. Once again this gesture is impossible to interpret, though his

expression is certainly one of dejection. It has been suggested that this old man is Valerian99-and in consequence that the king is Shihpiir.

However, when this head is compared with the heads on Valerian's coins (Pl. XVC), it is obvious that the only common factor is that both are the heads of old men, with strong noses and thick necks. The

principal points of difference are the following: Valerian's nose is sharply indented at eye level-the

Darabgird nose is straight; Valerian has a " high " forehead-the Darabgird forehead recedes; the

Dar~bgird mouth is full and curls up at the end-Valerian's does not; and, finally, Valerian has a

relatively long space between his lower lip and the indentation between mouth and chin-at Ddrdbgird this is very short. Dr. Sutherland, Keeper of the Heberden Coin Room at the Ashmolean Museum, who

kindly examined photographs of the Ddrdbgird head, was of the opinion that it did not present the

Imperial view of Valerian. He considered this head to be Eastern in physiognomy rather than Western. Professor Hinz also rejects the identification of this figure as Valerian:100 but his own suggestion

that it is a portrait of Philip the Arab (P1. XVD) is equally untenable, for the Darabgird head clearly represents a man of advanced years. Once again there are many differences of detail between the

Darabgird head and coin portraits of Philip, particularly in the shape of mouth, nose and chin. The third figure of the trio facing the king (P1. XIV) was identified by Stein as a Persian, on account

of his " conical head-dress, beard and ample locks ".101 Once again Stein's identification must be

questioned, for the man has a close cropped beard and neat wavy " page-boy " length hair, bound by a

plaited fillet: he is not accorded the royal diadem and ties. He wears a belted tunic of some stiff material which comes to just below his knees; it has banded sleeves and the belt consists of linked circles. Below this tunic he is wearing a loose skirt, with two groups of pleats, one falling over each leg; this completely conceals his feet. In his left hand, which is partly concealed by the young man's cloak, he is holding a strap or stick which touches the donkey's shoulder.

The corpse's head (P1. XIII) is pillowed on his left arm: he has a pointed beard and short curly hair, bound by a foliated wreath, from which flutter ties. His eyes are closed. Little can be seen of his dress for his body is largely obscured by the horse's four legs and the ankle ribbons of the king. His legs are either bare or in close-fitting smooth trousers, and some sort of strap crosses his chest.

From our examination of the principal characters of the Darabgird relief we can come to the

following conclusions: three of the four figures, the young man, the old man and the corpse, represent kings, but probably only minor kings. Their diadem ties are short while in Ardashir's other reliefs the defeated King of Kings, Artabanus, is given diadem ties the same size as Ardashir's. The author can, unfortunately, shed no light on their probable nationality, except that each appears to originate from a different country: the old man may be wearing a " toga "; the young man has silky clothes and swathed leggings; and only a little can be seen of the corpse with his foliated diadem. The fourth man, who is not royal, is dressed in a tunic somewhat similar to the Sasanian tunic, though his elaborately pleated skirt is entirely foreign. Were it not for the unmistakable caricatures of the Roman soldiery on the left we might, in fact, be tempted to query whether this relief is depicting a Sasanian victory over their Western foes, for none of the four principal figures looks specifically " Roman ".

The main event depicted in this relief is the submission of the old man to the Sasanian king, which is vividly recorded by the placing of the Great King's hand on his head. The young king may be an ally,

99 Ouseley, Travels II, pp. I47-8; Flandin and Coste, Voyage, p. 34; Herzfeld, Iran in the Ancient East, p. 313; Vanden Berghe, Archiologie de l'Iran Ancien, p. 46; Ghirshman, Parthians and Sassanians, pp. 16o-I.

100 Hinz, Iranica V, p. 157. 101 Stein, Iraq III, p. 195.

THE DARABGIRD RELIEF-ARDASHIR OR SHAHPUOR 87

for he looks buoyant, as if expecting a reward. A trampled-on corpse must always have been the

symbol of the total defeat of the enemy: not only is your opponent slain but you possess the battlefield and the bodies. In this case it may represent " defeat " or it may represent the death of a specific member of the enemy force.

Balancing the block of Roman caricatures on the right of the relief is a group of Persians, who are

standing in serried ranks behind their king (P1. XVI). They are arranged in four rows-of the majority only the heads and necks can be seen. The top row consists of four heads, three wearing the " Phrygian " hat. Two of these are clean shaven, the other is bearded, as is the fourth figure, who has a round hat and long curls. The second row also consists of four men-all of whom have neat pointed beards and long moustaches. Their short curled hair appears to be surmounted by a close fitting skull-

cap though it may be bound with a wide fillet. The leading man of this group is holding the hilt of his sword up to his face, thereby doubtless swearing allegiance to his suzerain.

Five men, all wearing the curved " Phrygian " hat, are shown in the third row: three are moustached and bearded, the other two are clean-shaven. The foremost man has a bunch of curls on his shoulder, while the third and fifth men have straight hair. Only two men can be seen full-length in the front row; half can be seen of a third and of the remaining two of the quintet only the heads are visible: all are bearded and moustached and wear the high hat. The two full-length figures are dressed in the familiar belted tunic and loose trousers of the Sasanian court. Their cloaks are fastened a little below the heavy necklace of roundels worn by the king and members of his court: similar necklaces can be seen on all the Persians in this relief. Swords hang by their left sides. They are standing on their toes: their soft boots are secured by straps with a roundel round their ankles.

The right hand of both these figures is held in front of their chests and holds between thumb and

forefinger some emblem or flower. Of the three left hands which are visible that belonging to the last

full-length figure in this row appears to be resting on the right arm of his neighbour, whose own left hand is cupped over his sword. The third left hand is seen nearly touching the horse. The leading figure has a foliate ribbon round his hat; the other full-length figure has a plain diadem round his, from which float the diadem ties of royalty. He has thick bunches of curly hair.

Each of these eighteen figures is clearly meant to portray a specific man, and it is tempting to see in the three back rows representatives of the subject nations: the men in the front row clearly depict the

king's sons or high-ranking members of the Sasanian court.

Since its discovery by European travellers the Dirdbgird relief has been thought to commemorate Shdhptir's victory over Valerian. As both the style and the composition of the relief clearly belong to the period of either late Ardashir or of early Shahpiir, we have dismissed this attribution. Professor Hinz has suggested that the triumph recorded was Shdhpiir's victory over Philip the Arab. This is just possible stylistically but there are, nonetheless, strong arguments against it; for instance the fact that the crown worn by the king is the crown of Ardashir and not that of Shahpir. By 244 A.D. Shdhpir had been crowned for a full two years and even on his early reliefs he is shown wearing his own mural crown. It would not only be surprising if Shdhpir was still wearing his father's crown at DdrTbgird but it would, in addition, be disrespectful to Ahuramazda, who had invested Shihpir with his own crown some two years earlier. Indeed, if this relief does show Shihpfr commemorating his 244 triumph, then all Sasanian attributions by crown will have to be reconsidered. It could, however, show Shihpnr in the period of joint rule before he had been awarded a crown of his own, though this, too, is unlikely for surely Ardashir would not have allowed himself to be so overshadowed in his own lifetime. The most probable event being commemorated at Darabgird is, in fact, one of Ardashir's own victories over the Romans during his final decade.

In the investiture scene at Naqsh-i Rustam, Ardashir had achieved a really successful sculpture, combining a strongly symmetrical composition with an almost abstract simplicity. In the relief at D~irabgird we see the stylistic influence of the Western world, whose very defeat Ardashir is depicting. The composition is still rigidly symmetrical, but, owing to the intervening influence of the Roman manner, the simplicity of Naqsh-i Rustam has given way to a mass of fluttering detail. Clothes become silky and ribbons float up behind the king, instead of hanging down his back. In Shahpir's reliefs this


enrichment, which is often not altogether successful, is carried still further and gives his works their " baroque " character.

The allocation of the Ddrdbgird relief to Ardashir rather than to Shihpir is endorsed by a number of details. Briefly, these are as follows.

(I) The folds of the king's floating cloak and ribbons are carved in loose ripples. One of the characteristics of all Shahpir's reliefs are the hard geometric pleats of the cloak and the doubled back waves of the ribbons.

(2) The king is not holding the reins, which simply rest on the horse's neck as they do in Ardashir's investiture at Naqsh-i Rustam. Shihpoir, on the other hand, usually holds his reins and in his investiture scenes the god does likewise.

(3) Ardashir at Naqsh-i Rustam and Darabgird places the fallen prisoner alongside his horse: the animal is clearly not treading on the corpse. Shihpir, however, either has the horse literally standing on the body or it is between the animal's legs.

(4) Ardashir gives his defeated royal foes diadems: Sh~hpair does not. (5) The " balloon " hanging from the saddle has a sharply pointed end as in Ardashir's relief at

Naqsh-i Rustam: Shahpir's " balloons " are less pointed and at Bishapuir are cut straight. (6) Triple curls on the horse's neck are seen only on Shahpflr's reliefs, and they do not occur at

Ddrdbgird. (7) Ardashir has gradually increased the amount of" movement " on his reliefs: at Naqsh-i Rajab

only the tie of the diadem offered by Ahuramazda moves slightly; at Naqsh-i Rustam, these ties blow out more vigorously and the god's cloak also billows out; at Darabgird the king's ribbons and cloak flutter up boldly to fill the space behind the rider, and the cloak of the young king flutters out to indicate his movement towards the king. In his reliefs, Shjhpfir always exaggerates movement-ribbons and cloaks billow in all directions ignoring the laws of nature, which Ardashir observes.

And finally:

(8) If the young king is pleading for something, he is allowed to do so as an equal, standing on his feet; Shdhpir makes both Philip and Valerian kneel.

Therefore, on all counts, of crown identification, of historical probability, of general style, and of minor points of detail, this relief should be attributed to Ardashir. It must have been carved in his final decade, some years after the Naqsh-i Rustam investiture, and it clearly precedes the reliefs of Shahpfir, both in style and composition. We will have to know much more of the events of the 230's before we can decide which of Ardashir's many forays into Mesopotamia inspired this victory relief: we may then be able to propose convincing identifications for the other characters carved on the relief.

89

THE BELIEFS AND PRACTICES OF THE AHL-I HAQQ OF IRAQ

By C. J. Edmonds

In my book Kurds, Turks and Arabs (London I957), I gave some account of the Kdkais, as the Ahl-i HIaqq are generally called in Iraq. At that time I thought it only fair to my informants to limit myself, in a book intended for the general reader, to the history, organization and geographical distribution of the sect, avoiding as far as possible all mention of their esoteric beliefs and secret ceremonies. But, now that a detailed account of these beliefs and practices (albeit with some playing down of their heterodoxy) has been published in Tehran by a leader of the community in north Persia,' I feel free to make good the omission, recording again, as before, only such information as I have myself obtained at first hand from Kakai sources. These are: (a) a small pamphlet, or Tazkara, written in 1933 especially for me, in old Turkish, by an educated and intelligent Kikai who had formerly served in the Ottoman Civil Service;2 and (b) careful and detailed records preserved in my diaries of many conversations with a dozen or more adepts, supported by several poems in the G6rdni dialect.

The matter already given in Kurds, Turks and Arabs, hereafter referred to as K. T. & A., in so far as it is relevant to what follows, may be summarized thus:

The Kakais are essentially an Islamic dervish brotherhood, sharing much of their esoteric creed with other Silfi orders, but differing from them in the person of the Founder and in certain other beliefs and ceremonies peculiar to themselves and based on a revelation of the Truth or the Divine Reality (HIaqq, IHaqiqat)3 which they claim to be more advanced than that vouchsafed to the others, who are still in the stage of the Tariqat (the mystic Path), or, more elementary still, of the Shari'at, the literal and legalistic interpretation of Holy Writ.

The order was founded, at Pird-i War in the Hawraman district on the Iraqi-Persian border, by a certain Sayyid or Sultan

Ish.q or Suhak, son of Shaykh 'IsE of Barzinja, a village in Iraq fourteen miles

east of the town of Sulaymani. The name Kikai derives from the following legend. The oratory of Shaykh 'IsE at Barzinja was undergoing repairs, but when the new main beam for the roof was hoisted on to the walls it proved to be too short to span the space between them. Seeing his father's distress

Ish.q, the favourite son of his old age, climbed onto one wall, seized one end of the beam, and called to his elder brother, " Brother, pull! " (K. Kake bikeshe). In this way they made it longer and placed it in position. The main beam of the mosque at Barzinja today is claimed to be the identical beam of the miracle,4

Associated with S. Ishiq at Pird-i War were three groups of seven persons generally called the Haft Tan (his Close Companions), the Haftawdna (his sons) and the Haft Khalifa (the Seven Vicars), as well as certain other groups such as the Haftid-u-dii Pir (the Seventy-two Elders). Five of the sons founded families of Sayyids which have survived to the present day; in course of time there came to be

1 See S. C. R. Weightman, " The Significance of Kitdb Burhdn ul-Haqq ", Iran II (1964), pp. 83-Io3. An earlier book by the same author of this work,

I.Hji Nfir 'Ali Shah of the neo-Sayyid

family of Shaykh Hayis, has recently been translated from the manuscript by my friend Dr. Mohammad Mokri, and has been published with a detailed commentary as L'esotirisme kurde (Paris 1966). Most of my authorities on the Iraqi side of the frontier were Murids of the Ibrahimi and Mirasfiri Sayyids, who claim descent from the Founder himself, and probably represent a more conservative tradition.

2 The Tazkara is written in pencil and in a very small hand; I have for the most part used a more legible Kurdish translation made by my friend Colonel Taufiq Wahby, referring however to the Turkish when any point seemed to require checking.

3 Religious and philosophical terms are generally given in their

Persian forms. Kurdish and G6rdni words and passages quoted as such are transliterated according to the system used in A Kurdish Dictionary (Oxford I966) compiled by T. Wahby and myself, and are preceded by the indication K. or G.

4 A similar story is told of the Priory Church of Christchurch near Bournemouth. During the work of building, a beam for the roof, which had proved to be a foot too short of the required span and had been left overnight, was found the next morning drawn out to the required length and securely placed in

position. The builders, recalling that on the previous day an extra hand had joined them but had not appeared at meals or in the evening to draw his pay, concluded that this must have been the Carpenter of Nazareth and called it " Christ's Church ". The miraculous beam is still shown over the arch on the south side of the Lady Chapel.


associated with them five other " neo-Sayyid " families named after, but not necessarily descended from, five personages whose saintly life and miraculous gifts had placed them almost on an equality with the descendants of the Founder himself. The Seven Vicars were chosen by S. Ishlq from among the Seventy-two Elders to be the Dalils or Guides, under the superintendence of Daiid, one of the Haft Tan, of all members of the community, not excepting the Sayyids themselves.

Every Kakai must be affiliated to a Sayyid of one of the ten families as his Pir or spiritual chief, and to a member of one of the Khalifa families as his Dalil. Pirs and Dalils have certain ceremonial func- tions; they also have duties towards, and receive perquisites from, the individuals affiliated to them.

The new material has been set out under nine main headings, each divided into two parts: (a) the Tazkara, and (b) personal information. The style of the Tazkara is prolix and repetitive; by what I hope is a more logical arrangement of the matter and, by ignoring a number of digressions comparing the beliefs and practices of the Kakais with those of other mystical sects and religions or refuting certain popular libellous allegations against the sect, I have reduced the length of (a) as far as possible without prejudice, I trust, to any point of substance; on the other hand, in much of what has been retained I have tried to follow the author's own wording. My authorities for (b) include the author of the Tazkara, but the greater part of the matter, in particular the details of the most secret dogmas and ceremonies, together with the G6rdni verses, were obtained from a venerable Safi5 of Hawar, a village which is situated about five miles as the crow flies north-west of Pird-i War and, with the adjacent hamlet of DdratuwZ or Dtratife, is the last and only Kakai community left in their Holy Land where the Founder walked and talked.

I. FUNDAMENTAL BELIEFS

(a) The Tazkara Before earth and sky were created a secret treasure lay concealed within a pearl in a vast and bound-

less sea. Then in six days God created earth and sky, then the seven degrees of heaven, then the angels. From among the angels (whose number is not known to us) he chose four, Jabra'il, Mika'il, Isrifil and 'Izr'ill, to be his personal attendants.

After the lapse of untold ages he created the spirits (arwdh) of the human race in a fixed and limited number. After the lapse of more ages he created Adam out of clay, wind, water and fire, and the human race is his seed. The story of the expulsion of Adam from paradise for having eaten wheat at the instigation of Satan is an exoterical explanation of the occurrence, because in reality (haqiqat) whatever His creatures may do, it is God that does it in his wisdom.

For the spirits this world is a place of striving and testing. For those who struggle against the evil- prompting soul (nafs-i ammdra) so as to attain to real humanity (insdniyyat) there is the reward of paradise, and for those who tread humanity underfoot there is the punishment of hell. A term has been set for these tests, beginning with the departure of Adam from paradise and his arrival on earth, and ending with the trump of Isrdfil and the resurrection. During this interval the spirits come into this world to be tested one thousand and one times, each time in a different guise (d~n, jdma, libds). The self-manifestations of the essence of God (tajalliydt-i zdt-i Khudd) will also have appeared on earth one thousand and one times in human guise each time with a different personality (sifdt) and in a different place, so that the uncertainty in men's minds is a means of testing them. Every embodied spirit, to whatever community or creed he belongs, is responsible for his actions in this world below. He who avoids evil and serves humanity is obeying God's command, and for him the cycle of dying and coming to life one thousand and one times appears quite short and easy. In this belief the Kikais to some extent resemble the Mu'tazilites.

This religion is the first and the last of the seventy-two sects into which Islam has been divided. Its name is IHaqiqat (Truth or Reality) because, in addition to the outward

(.dhiri) meaning, certain

verses and passages in the Qur'in have an inward (bdtini) meaning; the outward is allegorical (majdzi) and the inward is real. The heavenly books previously revealed were abrogated by the Qur'an. But

5 I use this spelling to describe Kurdish adherents of the mystical orders.

THE BELIEFS AND PRACTICES OF THE AHL-I HAQQ OF IRAQ 91

since Islam has not been preserved in its original state, and since the Qur'an has not yet been correctly interpreted or its obscurities explained, authoritative rulings (i'tihdd) and adaptation to change (tajaddud) are permissible.

The faithful entertain no doubt as to the necessary existence (wujab) of God (jandb-i Haqq), almighty, creator, pre-eternal, everlasting. The heavenly books were appropriate to the conditions prevailing at the time when they were revealed, and to this extent they retain their validity. Ever since the primal compact of Alast, when the spirits were created and recognized God as their Lord, men have been precluded from calling in question the acts of the Creator; they are not free to attack these books, or the followers of other religions, or even those who have no religion; they may only seek to guide others in the right way, but in all humility and without hurting their feelings. (To maltreat an ant, or even a mangy dog, is to humiliate one of God's creatures and is a sin.) God created the world for men and granted them the power of reflection (fikr) and intelligence ('aql) for the understanding of the revealed books, and each individual receives the commands of God in accordance with his intelligence and his circumstances. There is no necessity for anyone to be exactly like anybody else, for sense (hiss) and conscience (wujddn) belong to God Almighty. All men are equal and are brothers without distinction of religion; by their mixing and by their mutual love right (haqq) will triumph over wrong (bdtil), and the purpose of humanity will be achieved. Governments and their offices, crimes and wars, these are the marks of oppression and are sinful (hardm).

Until a man dies he cannot know what fate is reserved for him. He must always remain between fear (khawf) and hope (rajd'). To trust in his own works or wealth or physical health or intellectual knowledge ('ilm) or even his faith (imdn) and to take no heed of the Creator is the road to disaster. By helping his fellows and by repentance he may obtain forgiveness for his sins. Human existence will come to an end with the resurrection, when the records of all will be scrutinized, and those destined for hell and those destined for paradise will be sent to their own places. Paradise is believed to be in the sky and the upper universe (dsman u 'dlam-i 'dli), and hell to be a ball of fire in the middle of the earth; there are no such things as Houris or beautiful boys in paradise. By some hell is called bizil and is likened to a spiky threshing-sledge which is driven round and round a threshing-floor of thorns, each round lasting sixty years. If there were no resurrection and no Mahdi it would not be possible for man to attain perfection.

(b) Personal Information The earliest age is known as the Age of the Pearl, when the name of God Himself was Yt. A spirit to reach perfection must appear on earth one thousand and one times in a different guise:

the more righteous the spirit the quicker does it complete the cycle (dinddon, dawriydt); one that com- pletes the cycle becomes a " seer " (didaddr), and the last time the mortal body is buried it is not

corrupted. The IHaqiqat is essentially a religion of universal brotherhood and love; to harm the lowliest animal is a sin. The resurrection will take place in the plain of Shahrazfir, when those who have failed to complete their cycles will be further punished. With the resurrection the human race will come to an end. It is not known what will happen after that; perhaps God will create a new race of beings, just as He created the sons of Adam after the ending of the race of Giants.

God too will have manifested himself one thousand and one times, each time in the person of a human being and accompanied by dons of the four archangels (muqarrabin) and three other Close Companions to make up the Haft Tan, together with other groups of disciples and adherents, all in a new guise each time."

6 It would seem that only a few of the self-manifestations of the Deity have been complete; most are only partial and difficult to recognize.

The feeling that d6ns, of one spirit are " the same " is so strong that the names are frequently transposed in lists or legends relating to what to the uninitiated student appear to be widely separated eras or dispensations, e.g. the substitution of BRwa Yddgar (sixteenth century) for Zardabdn in a list of the Haft Tan at Pird-i War.

In a rather different order of ideas, there often seems to

subsist a feeling that apparently different persons having the same name are themselves esoterically " the same ". For example, the casting of S. Muhammad, eldest son of S. Ishaiq, as the propagator of religion (Section 3a) seems to link him with the Prophet Muhammad, and I was once assured that Shah Ibrahim (Section 3b) was " the same " as Ibrahim Adham whose tomb is in Baghdad between Bab ash-Shaykh and Bab al-Mu'azzam, Ibrahim Samin (the Fat) whose tomb is in the Kirkfik region, and a certain Ibrahim Khayrallah.


2. THE CALIPHATE AND THE STATUS OF 'ALI

(a) The Tazkara

Although, by inheritance and merit, the Caliphate ought to have devolved upon 'Ali and his descendants, his rights were spurned by Abai Bakr, 'Umar and 'Usman and were then usurped by Mu'awiya. 'Umar refused to accept a redaction of the Qur'dn prepared by 'Ali. The now commonly accepted version was compiled by 'Usman, who left out a number of verses relating to 'Ali and the Prophet's family. But, believing that their fault was due to their inability to see with the inward eye, the Ahl-i Haqq do not abuse the first three Caliphs. They also differ from the Shi'a in their attitude to 'A'isha and may call their women by this name.

'Ali is higher in rank than Muhammad because Muhammad is the Lord (sdhib) of the exoteric Shari'at, but in the spiritual world ('dlam-i ma'nd) 'Ali is the repository of the secrets of the self- manifestations of God and the base of the pillars of the Truth. In his lifetime he was the Lord of both the legal and the spiritual dispensations (sdhib-i shari'at va ma'nawiyydt) and performed many miracles such as bringing the dead to life, turning water into milk, and the like. He takes precedence because, as has been explained in the section on the transmigration of spirits, the divine essence was manifested in the guise of 'Ali and he represents the Truth. The claim of the Nusayris and the 'All Ilthis that 'Ali is himself God is rejected by intelligence and logic.

When Muhammad was making his night journey to the Seventh Heaven to have the honour of speaking with God from behind a screen, he encountered in one of the lower heavens a lion, which took from him the ring of prophethood. The lion was in fact 'Ali, and the hand which the Prophet saw stretched out from behind the screen was 'Ali's hand. On the morrow the turquoise of the ring was returned to Muhammad. 'Ali was not buried at Karbala (sic) but, by his own instructions, his body was entrusted to a bedouin cameleer; this cameleer was in fact the archangel Jabra'il, who took the body to his own spiritual hiding-place (ikhtifdgdh).

(b) Personal Information As between the two main sects into which Islam is divided, Sunni and Shi'a, the Kikais must be

reckoned to belong to the latter, but only in the widest acceptation of the term. The Sunnis having generally been more accommodating than the predominating Ja'fari Shi'as and slower to denounce the mystical aberrations of groups that have not openly broken away from the fold, the Kdkais always seemed to me to feel themselves far more in sympathy with the Naqshbandi dervish order, which is accepted within the orthodox Sunni community, than with the Ja'faris. My Safi informant of Hawair once told me that he had been a Murid of one of the two leading Naqshbandi Shaykhs of the Hawraman region. When I asked how this could be the Safi replied with a smile that the Shaykh, like his father and grandfather before him, had seen the light (K. berchaw-yan roshin bu) and had secretly been a Kakai, whereas the other leading Shaykh was still benighted in orthodoxy (K. berchaw-y tarik-e).

3. SULTAN ISHAQ AND THE THREE HEPTADS

(a) The Tazkara

S. Ishiq was doubly holy because he was both the locus for the self-manifestation of God (mazhar-i tajalliydt-i Khudd) and also a descendant of 'Ali. He ascended the throne over the spiritual world, but always preferred the bare earth to the throne. The miracle of the beam was only the first of the many miracles he wrought on earth. He is the Supreme Authority (mutjahid-i a'zam) on all matters of interpretation and adaptation.

The Haft Tan-i Javidmn,

the Seven Eternal Persons, his Close Companions (muqarrabin), were beings in human form endowed with divine attributes (hd'iz-i malakat). Their names were Binyamin, Dftid, Pir Mtise, MustafA Kamanddar (the Huntsman), Daya Razbar, Ytr Zardaban and Ewat. The first four were identical in the spirit world with the archangels Jabra'il, Miki'il, Israfil and 'IzrB'il.

The Haftawina, the seven children of S. Ishaq, who correspond with the Haft Tan, were allotted spiritual spheres of authority as follows: S. Ahmad Mirastir over Iraq, Syria and Egypt; S. MustafA


Safidp6sh (Whiterobe) over the world of beauty; S. Shihab ad-Din (he is the same as Sh. 'Umar Suhrawardi whose tomb is in Baghdad) over wind and rain; S. 'Abd al-Wafa over intellectual know- edge and intelligence; S. Habib Shah over purity and clean living (?); S. Bawa 'Isa over Hindfistan; and S. Muhammad over the propagation of religion.7

At Pird-i War, in the presence of S. Ishaq, they paired off as Pirs to each other in the manner described under the heading " Affiliation " below. In this context the word " Pir " means " spiritual master " and not just " elder " as in the title of the Seventy-two Pirs, from among whom S. Ishitq chose the Seven Vicars of Daiud.8

(b) Personal Information The major dispensations in which God manifested himself in the course of the progressive revelation

from the Shari'at to the Haqiqat are those of 'All in the seventh century, Mubarak ShTh Bawa Kh6shan in the eleventh, and S. Ishaq at the beginning of the fourteenth. Sultan Ishdq (as it will now be more appropriate to call him) reigned at Pird-i War for 300 years. One day, as he was sitting with his Companions, he told them that he must go away for a short time; he reappeared as I;Iji Baktash Wali in Anatolia where (as the G5rani poem records) he founded the 'Alawi religion, and where he reigned for a hundred years; when he returned to Pird-i War it seemed to his Companions that he had been absent for only one hour. At the end of the 300 years he built his own shrine at Pird-i War and disappeared. The Baktashi teaching belongs to the Tariqat stage of revelation; but, in spite of this and other important differences, the members of the two orders feel that a special relationship of mutual moral support exists between them.

The table on the opposite page shows the names of the Haft Tan associated with each of the theophanies just mentioned. The females listed as No. 5 in columns II, III and IV are of especial interest. It is claimed for 'Ali, Kh6shan and Ishaq that they were only " outwardly " the sons of their putative fathers, that they were all three in reality virgin births, and that they could be distinguished from ordinary mortals by their green nipples, " the colour of a Sayyid's turban ". Fatima bint Asad was 'Ali's mother. The mother of Bawa Kh6shan was Mama Jalala, daughter of his companion Mirzt Haman; one night she dreamt that she was pregnant and woke up to find a child in her arms; she died shortly afterwards and reappeared in the guise of Fatima Lara (the Slim), sister of Bdba Tahir of Hamadan. Khtiitn Dayara Jaf, Ishaq's mother, died in the same way and reappeared in the guise of Razbar. (According to one informant Razbar was really a man but took the form of a woman in order to wait upon the Sultan; that is why she sometimes bears the title Pir and sometimes Daya.)

The Haft Tan of the Pird-i War dispensation were produced by Sultan Ishaq: Binyamin from his breast, Dajid from his forehead, MfIsE from his tongue, Mustaff from his upper arm, Razbar from his head or " from the secret of his own heart ", Zardaban from his blood, and Ewat from his voice and complexion. Among the subsequent dons of Zardaban were 'All Qalandar and Bawa Yadgtr, the eponyms of two of the five families of neo-Sayyids previously mentioned.9 The Haftawana, his seven children, were similarly created by the Sultan, without the intervention of a mother, from his navel or from his liver.

Other groups associated with him at Pird-i War, in addition to the Seven Vicars and the Seventy- two Elders already mentioned, were the Ninety-and-nine Elders of Hawraman, the Forty Personages (Chihil Tan) of Shah6,'o the Sixty-six Henchmen (Ghuldm), and the Countless Thousands (Bawan Hazdr), the rank and file of the order. I have in my possession the manuscript of a

Garmni poem giving the names of the Seventy-two Elders and, for each, the name of the martyr who perished with the Imtim IIusayn at Karbala and of whom he was a dgn. The Forty Personages correspond to the Rijl al-Ghayb of the Shari'at who rule the destinies of men; but vacancies in the group are filled by danddan.

In the original this paragraph is more than usually illegible and obscure. The precise nature of the" correspondence " with the Haft Tan is not specified. S. Habib Shah, who as fifth in the list presumably " corresponds " with the female Razbdr, was one of the two children who left no male progeny to found a Sayyid family and was perhaps also female,

8 For details of the pairing see K. T. & A., pp. 186-7. 9 See p. 89-9o above; for further details of these families see

K.T. & A., p. I87.

10 The name of the sector of the chafne magistrale of Zagros immediately south of

Hawrdmmn,


I II III IV V

Td 'Ali Bdwa Khjshin S. Ishdq idji Baktdsh

I. Jabra'il Salmmn Rida Binyamin Qayghasiz 'Abdal 2. Mika'il Qambar Khuwaida Diid Gulbabt 3. Isrifil Ja'far-i Tayyar Falak ad-Din Pir MiisE Shahinbaba 4. 'Izra'il Sa'd ibn Abi Waqqis BifaqE Mustafft-i Datid~ni Qaftan 5. - Fatima bint Asad Fatima Lara Pir (Daya) Razbar

Warmn 'Abdal

6. - Khalid ibn al-Walid Baba-i Buzurg Ytr Zardaban Qizildada 7. - Bilal-i Habashi Mirzat Hamn Ewat Turabi Orman

The correspondences shown in columns III, IV and V are those given in two Ga6rni poems dictated to me by my Hawari informant and bear out the oral statements of the others except that: (a) the author of the

Tazkara gave Turabi Orman at No. 5 and Balim Sultan at No. 7 in column V with no mention of Wrain 'Abdal; and (b) one informant at Sulaymani gave Bawa Yadgar at No. 6 in column IV in place of Yar Zardabin (so pronounced by all my informants but written in the poems with a final m instead of n). In one or other of the poems No. I is described as " the forerunner " (kewt e p6shxan), No. 2 as the " man of battle " (des we choghan, mird be ruay meydan), No. 3 as " the registrar " (defterxan), and No. 4 as " the archer " (des we keman) .

As an example of the syncretistic tendencies of the religion of the IHaqiqat, the author of the Tazkara gave me the names of the dins of the Haft Tan in the Christian dispensation as Jesus, Peter, Simon, John (Yuhanna), Mary, Mark and John the Baptist (Yahya), and of the Haftawana as Andrew, PNmin (sic), Matthew, Paul, Thomas, James and Luke. This seems to make Jesus a din of the senior

archangel, Binyamin, and not a major self-manifestation of God Himself as might have been expected. Of the minor theophanies my friends on the Iraqi side mentioned only two. The first was Bahltil,

who was " outwardly " the brother of the Caliph Haruin ar-Rashid (eighth century), with his four Close Companions the Shaykhs Hasan al-Basri, Ma'rtifal-Karkhi, H1ajat and Qapl~n. The second was a certain Bawa Na'is, who appeared at some time between the dispensations of Bawa Kh6shan and Sultan Ishlq (but one informant made him a din of the seventh in the list of the Haft Tan, between Mirza Haman and Ewat).

They certainly do not accept as major theophanies the three individuals mentioned by Minorsky"1 as being recognized as such by certain groups in Persia. For them Qirmizi seemed to be not the name of a separate " host " of the Deity but a title of 'All as " Lord of the Age ", Sahib az-Zaman. The name

Muhammad Beg seemed to be quite unknown in any such connection. Khin Atash was a neo-Sayyid of the All line.

One informant stated that Sultan Ishaq was the one-thousand-and-first of the self-manifestations of God upon earth. This would seem to rule out any further theophany, major or minor, after Pird-i War, unless the speaker felt, in saying this, that every genuine " host " of the Deity, before or after Pird-i

War, was " the same " as Sultan Ishliq. The exact status and position in the hierarchy of Shah Ibrahim (son or grandson of S. Muhammad,

eldest son of S. Ishiq), who gave his name to the senior Ibrahimi branch of the Haftawana Sayyids, is a little puzzling to the uninitiated mind. The title Shih suggests that he was regarded as having been a " host " to the Deity in a minor manifestation; in the Turkish formula for the affiliation ceremony quoted in Section 6b he appears as the occupant of a throne but is mentioned after Binyamin, Dalid and

Mtis; and in the invocation for the Major Ablution recorded in Section 9a he is included immediately after what (in the Tazkara version) might be a reference to the Sultan himself and before five members of the Haft Tan (the same three with Mustaff and Razbar added). I never heard him described as a manifestation of God having his own distinctive Close Companions and other associated

groups.

x1 Notes sur la Secte des Ahle-Haqq (Paris 1921), p. i I; and Iranica: Twenty Articles (Tehran 1964), p. 307, the latter a revision of

the article " Ahl-i Hakk " contributed to the Encyclopedia of Islam (Ist edn.).


Much of the information about the Ahl-i IHaqq collected in Persia has come from neo-Sayyid, particularly Atashbegi, sources. The story of the origins of the eponymous ancestor of this family, as told to me by the S6fi of Hawar, may therefore be of some interest.

Atash Beg was the son of Shdwdsquli, who came into the world in the following miraculous circumstances. A brother and sister who were devoted to each other left home for a distant country. Fearing that sooner or later the girl would inevitably be demanded in marriage by some powerful person they decided to give out that they were man and wife. One day

Sult.n Ishlq came to them in the form of a

dervish, and, in answer to his question why they had no children, they could only answer that they had been unlucky. The Sultan offered a prayer on their behalf and promised to come again in a year's time. Of course there was no child, so when the dervish came again the girl took the forelegs of a sheep, wrapped them in a cloth, and put the bundle in a cradle. Ishiaq asked to see the child and, brushing aside the girl's excuse that it was asleep, took the bundle in his arms. He offered up a prayer (du'd) and, the forelegs having turned into a child, exclaimed " Shdwdsquli ".12

4. RELIGIOUS OBSERVANCES: PRAYER, ALMS, FASTING, PILGRIMAGE

(a) The Tazkara In the affairs of this world the Ahl-i

.Haqq go along with the people of the Tradition (sunnat) and the

Muslim Community (iamd'at), but in the matter of religious observances they are different. By virtue of his authority (ijtihdd) S. Ish~q reformed the obligations of prayer, alms and fasting.

Generosity is better than formal practices, and for ritual prayer (namdz) S. Ishiq substituted niydz (M for you, the outsiders, Y for us, the initiated ones), that is to say the distribution of food as an act of devotion. It is incumbent upon every head of household to take fresh or dry fruit or vegetables, or, if his means permit, an ox, a sheep, a lamb, a cock or a chicken, and distribute it in equal portions among his co-religionists present, at certain times and seasons, in the manner explained below. The niydzl3 provides the opportunity for communal devotions. Any room in the house of the host, or any clean place in the open air, may be used as the conventicle (niydzkhdna); silence is obligatory during the proceedings, and worldly conversation is forbidden.

Instead of the obligatory payment of alms (zakdt) the richer give voluntarily to the poorer gifts, of money or in kind, to any amount they wish. The lay folk (murid) must pay not less than three annas14 to the Sayyid and one third of that sum to the Dalil. Sayyids are holy, Dalils are respected. The relationship of Sayyid to Murid is that of father to son.

Since Ramazin is a month of the lunar year and does not fall at a fixed season it was decided by ijtihdd that the times of fasting should be determined by the solar year, and that the duration should be three days and not thirty (sih not si). At first it was laid down that the fast should not be broken for the three days on end. When this was found to be unbearable the period was limited to day-time only; but each day the fast is not broken until ten stars can be seen.15 The month of the fast is ushered in on the night before the first day with an invitation-supper, (da'vat) called " The Welcome " (istiqbdl; K. be pirewe chuin), in the manner described in the next section. The three days of actual fasting must be passed in quietude and in recollection of God (gikr); visits to the public bath and even washing the face compromise the merit of the fast and are forbidden. The fast is broken each evening with a distribution of niydz bought with the cash contributions (one shahi, now one anna, in respect of each person in the household including children however young and servants) collected in lieu of the compulsory alms

(sarfitr) of the Shari'at. During the night poems called kaldm are recited and the wonder-working men

12 In Kurdish qul may mean slave, small bundle or ankle, shin. 13 The word niydz, which in Persian generally means supplication

or need, is used by the Kdkais to connote the offerings of fruit, vegetables and sweetmeats, as opposed to nagr or qurban, which denote meat offerings. The author of the Tazkara uses the word qurbdn sometimes to describe any meat offering and sometimes, apparently, to mean flesh as distinct from fowl.

14 The Indian rupee was the official currency at the time the Tagkara was written; the interval between payments is not specified.

15 From this account read in conjunction with part (b) below, it seems to result that the month of the fast is January O.S., that the days of the actual fast are the 12th, 13th and 14th, and that the Ist, 15th and 31st are observed as feast days.


of old are recalled. The night after the third day, the eve of the festival, is very holy, rather like the Night of Power of the Shari'at; it is the night when the records of men's behaviour are scrutinized, the lengthening or shortening of lives is decreed, and sins are forgiven; many people keep vigil all night. On this night an invitation-supper, in which the nazr must be a cock16, is celebrated in the name of the Padshah, that is to say Sultan Ishaq. The next day is the festival; people go to the bath, put on clean clothes, greet their friends and make up their quarrels. Finally a supper is celebrated at the end of the month in the name of Pir MfisE Goldenpen (qalamzar) who is charged with keeping the records. Some people fast during the first ten days of Muharram; but they do not hold mourning ceremonies, because calamity is a means of grace and brings men nearer to God.

By virtue of a pilgrimage to the tomb of Sultan Ishaq at Pird-i War the adept becomes a HIIaji. Visits to Madina and the holy places ('atabdt) of the Shi'a are meritorious, but there is no particular value in the circumambulation of the Ka'ba at Mecca. Forgiveness of sins is obtained through repentance, sinning no more, public confession before the Sayyid and the community in the niydzkhdna, and the payment of a fine fixed by the Sayyid as appropriate to the gravity of the offence and the means of the penitent, the proceeds being used to buy niydz and nazr for ceremonial distribution.

(b) Personal InJbrmation Prayer. The Qur'dn (LXII, 9) says: " When ye are summoned to prayer on the day of assembly

hasten to the commemoration (zikr) of God ". True prayer is zikr, not the set formulae and physical exercises of the Shari'at.

Alms. The Qur'an says: " Help one another in good works and mutual support ". The Haqiqat, unlike the Shari'at, does not lay down any fixed amount to be given as alms (zakdt); help is to be given to the poor in ways and amounts appropriate to the circumstances of the giver and the receiver. The dues paid by a Murid to his Sayyid are different and are called khayr u khidmat.

Fasting. The Shari'at interprets " fasting " as meaning abstinence from food; but in the IHaqiqat it means abstaining from falsehood, slander and oppression. Nevertheless a fast in the form of abstinence from food during the day is observed once a year for three days from the I2th to the 14th of January (Kanfin II) O.S.; it is followed by the Festival of the Three Nights (K. cejny s6 shew), corresponding in a way to the Qurbin festival of the Shari'at.

Pilgrimage. 'Arafat does not mean the hill famous as a place of pilgrimage for the people of the Shari'at but signifies ta'druf, the mingling of the faithful, getting to know each other, and exchanging information with a view to mutual help.

5. THE COMMUNAL SUPPER

(a) The Tazkara In addition to the three occasions already mentioned there is an obligation to celebrate a supper

for each of the four seasons, making seven in all during the year. A cock or a chicken together with half an oke of rice or wheatmeal porridge (K. sdwer), twenty-five dirhams of clarified butter and seven flaps of bread are accounted an appropriate nazr (? for an average household); the rich may offer a qurbdn (e.g. an ox, sheep or lamb) in place of the fowl; for the very poor it is enough to give bread; each gives according to his ability. Supererogatory invitation-suppers may be offered on the occasion of embarking on an important undertaking, or of a dream, or of an eclipse, or of sickness, or (in the name of Mustaft, who is 'Izra'il) of the outbreak of pestilence or of recent bereavement (superfluous in view of the transmigration of spirits); they may also be offered in honour of one long dead in the name of 'arafdti, or once a year simultaneously on an agreed date by widely separated friends (G. yane weyem).

In the assembly room (jamkhina) or conventicle (niydzkhdna) the Sayyid, who is the President (sarjam) and is to say the graces (du'd), sits at the head of the congregation towards the right. The Khalifa, who is to divide the qurban and the niydz, sits either on the Sayyid's left or facing him. No other

16 It is not clear whether the author means " at least a cock " or " a cock to the exclusion of e.g. a lamb ".


places are reserved, and the people take their places side by side so as to form a ring. Each person as he enters says " Yd 'All ", bows, kisses the ground and sits down. Those already seated do not rise but say " MawlA 'Ali ", and the new arrival replies by repeating the same words. One person (or more if the numbers present require), styled Farrash, remains standing to wait and to distribute the food.

Before the distribution of the qurbdn the Farrish takes a ewer in his right hand and a basin in his left hand, stands on the threshold, exclaims " Allah " and solemnly kisses the ground before the people. Then, beginning with the Sayyid and going from right to left, he washes their hands. After the washing the hand must not be wiped, nor may it touch any part of the body. Two flaps of bread are laid on the supper-cloth (sufra) in front of each person, and the qurbdn is placed in dishes in front of the Khalifa. The Sayyid then says, " Bi'smi'llah ... musht-i ... Ptdshah (in the name of God . .., the hand ... the King ,").17 In the meantime any late-comers remain standing, with the right hand placed over the left hand and the big toe of the right foot placed over the big toe of the left foot, until this grace (du'd-i qurbdn u niydz) has been said; they then kiss the ground and sit down.

After saying this first grace the Sayyid takes a piece of meat, the Khalifa touches it with his hand, and the Farrash distributes the meat to the congregation for them to clean it (i.e. remove the meat from the bones). After the meat has been picked off, care being taken that the bones are not broken, the meat is collected and is placed in the dish in front of the Khalifa, who mixes the pieces of meat and adds the fatty gravy; the bread too is collected and placed before him. Salt is sprinkled on the meat in the dishes, and also over the bones still in front of the members of the congregation.

Saying bi'smi'lldh the Khalifa proceeds to divide the meat into equal portions. The Sayyid hands portions to the Khalifa, the Farrash and those close to him. Then starting with the Sayyid and going from right to left the Farrash distributes the portions to all present in turn. When everybody has received his share the Farrash calls out, " Jam rasid " (" the Assembly is consummated ") and thereupon takes up his own piece, stands at the back of the Congregation, and exclaims, " Allah ".

The Sayyid pronounces the same grace as before, and the people say " Amen, Amen ". The Sayyid now " breaks the seal " by biting his own piece and, saying (G.) " Rehmet che we Yar " (" Mercy is from the Friend "),18 so gives permission to the congregation to begin.

Until the qurbdn has been slaughtered by cutting the throat, boiled and brought into the presence of the congregation, and has had the meat separated from the bones, has been distributed, and has had the graces said over it, nobody touches the meat with his mouth or again with his hand; it is under seal. If such a thing is done on purpose the procedure must be repeated from the beginning; if it happens by accident the person responsible is reckoned to have sinned.

After the sharing out of the qurbdn has been completed in this wise the Farrash gathers up the cloth, returns to the threshold, bends over in the manner of the inclination in prayer (ruk'a), and exclaims " Allah ". The Khalifa takes in both hands the dish in which the skull of the qurbdn or the head of the cock or the chicken has been placed, stands up, bends over as for a ruk'a and recites the " grace over the cloth " (du'd-i sufra). The Sayyid then pronounces the final " grace over meat " (du'd-i qurbdn) and the Khalifa sits down. After these graces have been completed the people kiss the ground.

The Farrash then takes up the cloth and the dish from in front of the Khalifa, and puts them down on one side. Then taking warm water and a towel he proceeds to wash the hands of the congregation. Then he washes his own hands, exclaims " Allah ", kisses the ground in front of the Khalifa, and starting with the Sayyid kisses the hands of all those present.

If it is night a portion is given to the lamp at its place in the line, but this portion may then be given to somebody who has been overlooked or arrives late. If there is no such person the portion is given to the man who lights the lamp. When hands are being kissed the lamp too receives a kiss conveyed by the hand as a mark of respect.

17 In the Turkish MS. two short words of the formula, one before and one after musht are illegible. A well-informed Kirkukli, not himself a Kkai but with Kdkai relations by marriage, once told me that the qurbdn is cooked the day before and is left overnight; in the morning, if all is well, the imprint of a

hand will be found on the pile of meat. It seems possible that the formula refers to this.

18 The wordydr, friend, companion, comrade, sometimes denotes God and sometimes any member of the sect; the abstract noun

Ydri is used as a synonym for Haqiqat, the name of the religion.


In this manner the ceremony is completed and the Sayyid, having said a short prayer, dismisses the congregation. After they have gone a select group (zumra-ikhavdss) stays behind in the jamkhdna to recite hymns and relate stories of their noble forbears.

The distribution of niydz, that is fresh and dry fruit, is performed in the same way except that there is no washing of hands and there is no giving of a portion to the Sayyid with a second grace because

grace over the niydz has already been said by the Sayyid when the Khalifa gives the first portion to the Farrish.

The proper graces and invocations vary for different occasions such as the Communal Supper, a

marriage, a petition for God's blessing on a gift, and the like; they are all couched in the Hawrdmdni dialect.

(b) Personal Information The Communal Supper is called Jam-i Haqiqat. It is presided over by a Sayyid assisted by a

Khalifa and a Farrdsh. Beforehand the host offering the Supper takes the special sacrificial knife

(normally kept by the Sayyid) together with the niydz of fruit and sweetmeats to the Sayyid who

pronounces an invocation (K. du'd edd) over the knife. The host slaughters the meat offering (nazr, qurbdn) over a hole in the ground which receives the blood and the guts. The meat offering which may be an ox, a sheep, a fowl or any lawful animal, is boiled complete with the head. At the time appointed it is taken to thejamkhdna, which is devoid of any kind of ornament or furnishing other than the supper- cloth (sufra) on the floor.

The Sayyid sits at the head of the room with the Khalifa on his right. An invocation is pronounced. The Khalifa, having washed his hands, tears the animal apart. The pieces are distributed. The recipients separate the meat from the bones, leaving the bones on the floor before them. The Farrdsh collects the meat in a large pan (K. tesht), which he places in front of the Khalifa. After a third invocation has been recited the Khalifa takes small pieces of meat, as far as possible of equal size, wraps each piece in bread and hands it to the Farrdsh. The Farrdsh, making a separate journey for each, hands the pieces in turn to the Sayyid, the Khalifa and each member of the congregation.

After the meal the Farrdsh kisses the ground in front of everybody present in the same order, all wash their hands, and the invocation of dismissal (du'd-i murakhkhasi) is pronounced. The owner of the

premises (? the host) collects the bones from the supper-cloth where they were left, puts them into a bag, and then buries them or throws them into a river or running stream.

Any guest who cannot get into the room stands outside with his arms folded across his chest, the

palms of his hands laid on the front of his shoulders. Portions may be sent out to members of the families in their homes. Women are not admitted to the Jam lest their presence should give rise to carnal

thoughts.

6. AFFILIATION

(a) The Tazkara

Nobody may take part in a Communal Supper unless he has been ceremoniously affiliated to a Pir and a Dalil. The candidate, or more often the father of the children who are to be affiliated (lit. are to " go to hand "), arranges a meeting in his own house or in the niydzkhana and slaughters a qurbdn for each one, or perhaps a cock for each boy and a chicken (? hen) for each girl; these are distributed in the manner already described. Then, the Pir being seated, nutmegs are brought, one for each candidate, and are placed before him. He cuts them up with a clean knife and distributes the pieces among the congregation. While the nutmeg is being distributed the candidate, standing at the rear of the assembly, grips with his left hand the left tail of the coat of the person who is to be his Dalil and cries out, " Allah ". When the distribution of the nutmeg has been completed and the invocation (du'd) has been pronounced, the Dalil first and the Murids after him, kiss the hands of all those present in turn starting with the Pir; they then return to their places at the rear of the assembly. Next, in the presence of the congregation, the Murid presents to the Pir a sum of money consonant with his means, and the Pir pronounces the proper invocation. Thereupon all the members of the congregation congratulate the


Murids who have performed the ceremony and have thus been affiliated to Pir and Dalil. In this way Pir, Dalil and Murid are bound together as by interlocking chains; the giving of a daughter by one to the other in marriage is forbidden.

(b) Personal Information The ceremony of affiliation, in which the Murid promises obedience to his Pir, is based on the Pact

between BinyAmin and God, and is called Sarsipurdan, Entrusting the Head. A nutmeg and a coin called Huwayzi are used in the ceremony; in recent times the coin used has been a one-qirin piece of Muhammad Shah Qajar and is often referred to as Muhammadi.19

The following formula in Turkish was given to me by an Ibrahimi Sayyid at Tall 'Afar, where the majority of the population are Turki-speaking and where there is an isolated community of Ahl-i IHaqq: Terin, gogin, emrine: Pir Binyamin artftna; Daudin delilligine; Pir Musanin kalemine; Pir $ah Ibrahim tahti altinda; bu talib bana talib oldu; din-i Hak ve Ehl-i Hak ifinde talib oldu (" By the command of earth and sky, by the Pact of Binyamin, by the leadership of Diid, by the pen of Pir MUisE, below the throne of Pir Shah Ibrahim, this candidate made application to me, he made application within the religion of the Haqq and the community of the Ahl-i Haqq ").

The children of a Murid follow the affiliation of their father. On the death of a Pir his Murids are divided among his sons.

7. THE BROTHERHOOD OF THE HEREAFTER

(a) The Tazkara This institution is called variously the Brotherhood " of the Hereafter " (dkhirat), or " in the Truth "

(Haqiqat), or "in the Religious Comradeship " (Tdri). Since this world below is impermanent (fdni) and life in it is unreal (musta'dr), and since after men's lives have run their course there awaits them the life eternal, two respectable and sober-living persons who know and like each other well and express their brotherly feelings in public, come together to the niydzkhdna and, in the presence of the Sayyid and the congregation, avow and confess (iqrdr va i'tirdf) that they are Brothers in the Truth. They are caused by the Sayyid to repeat this, a qurbdn is distributed, the proper invocations are pronounced, and in this way the Brotherhood is confirmed. The two Brothers then exchange girdles, which they may never give back. They are deemed to be closer to each other than brothers-german; they become confidants of each other and share responsibility for each other's sins. This induces them to vie with each other in righteous living, and if one learns of a transgression by the other he must at once warn him, and the transgressor is bound to take notice. If one dies leaving a family in straitened circumstances the surviving Brother must look after it and bring up the children as if they were his own. In the world to come this relationship will bring the two together again.

The Tazin section of the Baktashis and the S6fiydn section of the Rdfizis have a similar institution.20 But instead of exchanging girdles they mix hot water and cold water which, after being mixed, cannot be separated.

(b) Personal Information An initiate of either sex may take a Sister or a Brother in Religious Comradeship (K. biray Yart,

xushky Tart). In a dream an angel appears to a man or a woman with a message to say that so-and-so is his or her brother or sister. The one sends word to the other to ask if that person too has had such a dream. If the answer is in the affirmative (it is a sin to deny it if it is true) a Jam-i IHaqiqat is arranged and, after the ceremonial distribution of nagr and niydz as already described, two half-glasses of water, one hot one cold, are mixed in one glass. The pair drink from the glass in turn and say: By heaven and

19 I have a specimen of this coin brought from HSiwr. 20 The Turki-speaking QizilbAsh of the Kirkfik region of Iraq are

commonly referred to as Rifizis. By " Tdzin " the author

10

presumably means the people of Thza Khurmtfit which is ordinarily referred to as Tdza. For them and the S6fiyin see K.T. & A., pp. 267-9.


earth and this glass as witness I have taken this girl (man) to be my sister (brother)-K. be shayetfy asman u zewi we shayeti y em came em kiche-m (piyaw-im) qebul kird be xushk (bira) y xo-m.

The two then exchange waistbands as a symbol that their relationship is now closer than that of brothers and sisters german. At the resurrection they will recognize each other by this sign; but others, however near they may have been on earth, will not know each other.

Brothers and Sisters in Religious Comradeship, and their descendants for seven generations, are forbidden to marry.

8. MARRIAGE AND DIVORCE

(a) The Tazkara

The marriage contract is concluded in accordance with the shar' and the Hanafi rite. But it muet be confirmed by the summoning of the Sayyid and the congregation and the distribution of niydz in the following manner. Two nutmegs, one for the woman and oae for the man, are broken in pieces into a dish, grace is said over them, and the pieces are distributed to the congregation. A marriage celebrated in this way in the presence of the Sayyid and the cingregation ought to last until death and never be dissolved. To have two wives at the same time is forbidden and is a sin.

Only for special and valid reasons may a man divorce his wife. To say " I divorce you " even three times has no effect. Divorce can take place only in the presence of the Sayyid who presided over the contract proceedings or of another qualified Sayyid. The Sayyid hears the complaints and statements of the two parties, satisfies himself as to the truth of any allegations of misconduct, and then, according to the circumstances, either admonishes them or consents to their separation. Until these formalities have been completed the divorce is not confirmed.

Nowadays in this region (i.e. Kirkilk) some observe these prescriptions, but there are others who follow the shar' in every particular.

(b) Personal Information

Monogamy is the strict rule, but in recent years some of the richer members of the community have taken more than one wife.

Marriage contracts are concluded outwardly as among the Sunnis, but the really valid part of the proceedings is the form of contract used for the marriage of Adam and Eve, with the Haft Tan, the Haftawana, the Seven Khalifas, the Seventy-two Pirs, the Ninety-and-nine Pirs, the Forty Personages, the Sixty-six Henchmen and the Countless Thousands as witnesses. It had remained secret until it was revealed to Pir MiisE at Pird-i War.21

KSkais do not give their daughters to outsiders. If a Kikai man takes a wife from outside the community she may embrace the religion and be affiliated to her husband's Pir. The Haftawdna Sayyids do not give their daughters to commoners of the community or even to neo-Sayyids.

The only valid ground for divorce is adultery.

9. OTHER BELIEFS AND OBSERVANCES

(a) The Tazkara Dawn is the most propitious time for worship. The sun is an atom of power that has been separated

from the light of the self-manifestation of God, and Sult.n

IshIq derived from this atom. His most intimate spiritual companion and Pir of this religion is Binyamin, the perfect essence of Ilusayn. The moon is the radiant face of Binyamin. The stars of the constellation called the Seven Brothers (Ursa

21 I have a transcript of this contract in the G6rdni dialect (too long to reproduce here) taken down from the dictation of my SWfi friend of Hiwar. It consists of eighty-four decasyllabic lines of verse, the first sixteen rhyming in -er, the next ten in -am, and the rest in shorter sequences except the last, also ten

in -4. The old man recited the whole poem the first time from beginning to end without hesitation; but when I asked him to repeat it at dictation speed he seemed to lose the thread from time to time and must, I think, have omitted a number of lines in several places.


major) correspond to the Seven Pirs who were primarily formed from atoms of the light of God and were

separated materially into seven persons. A Fdtiha in Gardn! is recited at funerals and mourning ceremonies as follows: " Peace be upon

you, O captives of the earth; my hope is in Sultdn Suhdk, the beginning of the spiritual road for men, the creation of the permanence of the Eternal Persons, the bounty of the noble family. First and last, the Friend (Ydr) ".

The Major Ablution is performed as prescribed by the Shari'at. The invocation and statement of

intention are as follows: "

By the breath (nafas) of the Persons who wear the robe (khirqa) of light, the

man of Tashdrdn; by the goblet ( jam) of Ibrdhim that mirrors the world; by the bond (iqrdr) of the

King (Shdh) and the pact (shart) of Binydmin; by the breath of MUse of the Golden Pen; by the consent (rigd) of Dadid the Guide for the road; by the cup (tds) of Razbdr the Mistress of the Resurrec-

tion; may I be cleansed and purified from foot to head. First and last, the Friend ".22 The cursing of Satan for refusing to obey God's command that he should bow down to Adam is an

episode limited to the time of Adam himself. After an interval of unknown duration, but in any case

before the fall of Adam, he was pardoned, was named Gazelle, and returned to rejoin the archangels (muqarrabin). Since man, who is superior to all other animals in intelligence, is entirely responsible for

his own actions, the "

Satan "

who is said to lead men astray is not a personal Satan to be identified with the Satan of the Qur'dn, but that inborn quality of man called the evil-prompting soul (nafs-i ammdra) .23

Blood, pork and wine are all forbidden, but it is permissible to eat just enough of any kind of meat

to ward off death by starvation. It is credibly related that at some time in the past permission was given for the eating of pork and even for the drinking of blood, if boiled, of animals sacrificed in the way of

God. But it was later ruled that, since the object of such sacrifice was to spill the blood, the earlier

ruling was wrong. The drinking of wine and other intoxicants is definitely forbidden and a sin; anyone who has drunk may obtain pardon if he repents, does not relapse, and distributes qurbdn and niydz to

the value of one shdhi (anna) for each glass consumed.

The practice of not cutting the moustache is not based on any verse of the Qur'an or tradition of the

Prophet, but on reasoning, legend and custom: (a) God caused hair to grow on the faces of men to

distinguish them from women and this act of God should be respected; (b) it is inconceivable that

'All or Muhammad would have cut his hair to fall on the ground and be defiled; (c) there is a legend that 'Ali, with his spiritual hand (dast-i batind), pulled out one side of Mu'dwiya's moustache and that

Mu'awiya then shaved off the other side (the legend rests on weak authority, but since it redounds to

the glory of 'Ali there is no harm in believing it); (d) it was the distinguishing mark of the Muslims

who were conspiring to overthrow the 'Umayyads.

(b) Personal Information The Major Ablution is obligatory before the slaughter of an animal as nazr (qurbin), and after

copulation. Both pork and alcoholic intoxicants are forbidden; but it is well known that certain groups known

as pig-eaters (K. berazxor), in particular the Qalkhan! section of the Gbrdn tribe in Persia, indulge in

both.

It is not a sin to cut the moustache. The custom of not cutting it is not based on any religious prohibition but is purely voluntary; it commemorates the distinguishing mark adopted by the partisans of 'All in the struggle with Mu'Aiwiya.

22 The formula in the Ta kara corresponds very closely with that

given by Dr. Mokri (op. cit., p. 197). For my "

the breath "

(in both places) he has "

la parole "; in place of my "

the man

of Tasharan "

he has "

cette eau est l'eau de Tashir et ma

qibla est Perdiwar ".

23 The Ahl-i Haqq share with the Yazidis the belief that Satan

was pardoned soon after the little local difficulty with the

Almighty about bowing down to Adam, but then part com-

pany; for the role ascribed by the Yazidis to Satan after his

rehabilitation see my A Pilgrimage to Lalish (1967), p. 4. Other

points of similarity are the institution of "

Brothers and Sisters

of the Hereafter "

and the existence of two grades of hereditary spiritual superiors to a representative of each of which every member of the community must be affiliated.


major) correspond to the Seven Pirs who were primarily formed from atoms of the light of God and were separated materially into seven persons.

A Fatiha in G6rani is recited at funerals and mourning ceremonies as follows: "Peace be upon you, O captives of the earth; my hope is in Sultan Suhak, the beginning of the spiritual road for men, the creation of the permanence of the Eternal Persons, the bounty of the noble family. First and last, the Friend (Yar) ".

The Major Ablution is performed as prescribed by the Shari'at. The invocation and statement of intention are as follows: " By the breath (nafas) of the Persons who wear the robe (khirqa) of light, the man of Tasharan; by the goblet (jdm) of Ibrahim that mirrors the world; by the bond (iqrdr) of the King (Shdh) and the pact (shart) of Binyamin; by the breath of Mfis of the Golden Pen; by the consent (rizd) of Dafid the Guide for the road; by the cup (tds) of Razbdr the Mistress of the Resurrec- tion; may I be cleansed and purified from foot to head. First and last, the Friend ".22

The cursing of Satan for refusing to obey God's command that he should bow down to Adam is an episode limited to the time of Adam himself. After an interval of unknown duration, but in any case before the fall of Adam, he was pardoned, was named Gazelle, and returned to rejoin the archangels (muqarrabin). Since man, who is superior to all other animals in intelligence, is entirely responsible for his own actions, the " Satan " who is said to lead men astray is not a personal Satan to be identified with the Satan of the Qur'mn, but that inborn quality of man called the evil-prompting soul (nafs-i ammdra).23

Blood, pork and wine are all forbidden, but it is permissible to eat just enough of any kind of meat to ward off death by starvation. It is credibly related that at some time in the past permission was given for the eating of pork and even for the drinking of blood, if boiled, of animals sacrificed in the way of God. But it was later ruled that, since the object of such sacrifice was to spill the blood, the earlier ruling was wrong. The drinking of wine and other intoxicants is definitely forbidden and a sin; anyone who has drunk may obtain pardon if he repents, does not relapse, and distributes qurbdn and niydz to the value of one shahi (anna) for each glass consumed.

The practice of not cutting the moustache is not based on any verse of the Qur'an or tradition of the Prophet, but on reasoning, legend and custom: (a) God caused hair to grow on the faces of men to distinguish them from women and this act of God should be respected; (b) it is inconceivable that 'Ali or Muhammad would have cut his hair to fall on the ground and be defiled; (c) there is a legend that 'Ali, with his spiritual hand (dast-i bdtini), pulled out one side of Mu'awiya's moustache and that Mu'awiya then shaved off the other side (the legend rests on weak authority, but since it redounds to the glory of 'Ali there is no harm in believing it); (d) it was the distinguishing mark of the Muslims who were conspiring to overthrow the 'Umayyads.

(b) Personal Information The Major Ablution is obligatory before the slaughter of an animal as nazr (qurbdn), and after

copulation. Both pork and alcoholic intoxicants are forbidden; but it is well known that certain groups known

as pig-eaters (K. berazxor), in particular the Qalkhani section of the Gbran tribe in Persia, indulge in both.

It is not a sin to cut the moustache. The custom of not cutting it is not based on any religious prohibition but is purely voluntary; it commemorates the distinguishing mark adopted by the partisans of 'All in the struggle with Mu'aiwiya.

22 The formula in the Tazkara corresponds very closely with that given by Dr. Mokri (op. cit., p. 197). For my " the breath" (in both places) he has " la parole "; in place of my" the man of Tasharan " he has " cette eau est l'eau de TashSr et ma qibla est Perdiwar ".

23 The Ahl-i IHaqq share with the Yazidis the belief that Satan was pardoned soon after the little local difficulty with the

Almighty about bowing down to Adam, but then part company; for the role ascribed by the Yazidis to Satan after his rehabilitation see my A Pilgrimage to Lalish (1967), p. 4. Other points of similarity are the institution of" Brothers and Sisters of the Hereafter " and the existence of two grades of hereditary spiritual superiors to a representative of each of which every member of the community must be affiliated.

103

THE TAHIRIDS AND PERSIAN LITERATURE

By C. E. Bosworth

As a pendant to my consideration of the Tdhirids and Arabic literature and music in the article " The Tahirids and Arabic Culture ", Journal of Semitic Studies XIV/I (1969), it is useful to examine briefly what is known of the dynasty's attitude vis-a-vis Persian lore and literature in the third/ninth century. Ethnically, this line of governors in Baghdad and Persia was Persian in origin, and the native tongue of the first of the line, Tahir Dhti 1-Yaminain, was actually Persian (see art. cit.). But as I have endeavoured to show in the article mentioned above, the Tahirids followed so many of their fellow- countrymen during this early 'Abbasid period in assimilating themselves almost totally to the socially- dominant Arab ruling institution and its culture. Tahir was highly articulate as an Arabic prose stylist, as the contemporary renown of his moralizing epistle to his son 'Abdallah shows,' and several of his offspring played leading r6les as patrons of the great Arab poets and musicians and as Arabic scholars and authors themselves. During the civil war between al-Amin and al-Ma'mfin (195-98/ 811-13), and during the years immediately after the latter's assumption of the Caliphate, there was hostility and rivalry between Tahir and the Caliph's minister al-Fadl b. Sahl, whom D. Sourdel has called " the most Iranian of the viziers of the 'Abbisid caliphs ",2 and who was the proponent of Persophile policies. TThir, on the other hand, was regarded in the period after 198/813 as the hero of the pro-Arab party (despite his killing of al-Amin), and the Arab poets of Iraq appeal in their verses to Tahir as the natural leader of their group.3

Certain Persian sources allege that the Tahirid governors showed an outright hostility to any manifestations of Persian culture and literature in their Persian territories; unfortunately, these sources are very late. Muhammad 'Aufi, writing in the early seventh/thirteenth century, states in his literary biographical work, the Lubdb al-albdb:

" The Tthirid dynasty was characterised by patent nobility and abundant generosity, yet though the flood of their bounty and largesse flowed over everyone, they had no concern for (literally, ' gave no credence to ', ishdn-rd i'tiqddi nabfid) Firsi and the Dari language. Hence in that period, poets devoted little effort to this form of expression. However, in the auspicious age of the Tahirids, a poet of sugared words arose, called Hanzala of B~dghis ...".4

The other passage is from the mid-ninth/fifteenth century literary biographer Daulat-Shah's Tadhkirat ash-shu'ard', and occurs in the section dealing with the first tabaqa of Persian poets. He begins by speaking of the eclipse of literature in Persian brought about by the Arab invasions and the subsequent dominance of the Arab language for poetry, prose and official correspondence. Then he continues:

" They relate that Amir 'Abdallah b. Tthir, who was governor of Khurisan in the time of the 'Abbasid Caliphs, was holding court at Niship-ir one day. A man came forward with a book and presented it to the Amir. 'Abdallath asked,' What is this book about? ' The man replied, ' It is the story of Wamiq and 'Adhra, a pleasant tale which wise men put together and presented to King Anfishirvmn '.4a Amir 'Abdallth replied, 'We are men whose reading is the Qur'an, and we have no need of anything except the Qur'an and the

1 See my forthcoming study of this, " An Early Arabic Mirror for Princes: Tdhir Dhfi 1-Yaminain's Epistle to his Son 'Abdalldh (206/82I) ", to appear in JNES (1969).

2 In El2 s.v., see also his Le vizirat 'abbdside de 749 ci 936 (132 d 324 de l'Hegire) (Damascus 1959-60), vol. I, pp. 205-o6.

3 Cf. A. Gh6dira, " Deux pontes contemporains de Baiiar: les freres Ibn Abi 'Uyayna ", Arabica X (1963), pp. I58-6o.

4 Ed. Sa'id Nafisi (Tehran 1335/1957), p. 241.

4a It seems that this romance came into Middle Persian from a Hellenistic source, the original of 'Adhrd' being the daughter of Polycrates, tyrant of Samos; an Arabic version by the Shu'libi translator Sahl b. Hrfirn is mentioned in Ibn an-Nadimn's Fihrist, and a New Persian verse rendering was made by the Ghaznavid poet 'Unsuri. See M. Shafi, " 'Unsuri's W~miq wa 'Adhrd' ", Proceedings of the Twenty- Third International Congress of Orientalists held at Cambridge, ed. D. Sinor (London 1954), pp. i6o-i.


traditions of the Prophet. We do not require this kind of book; it was composed by Magians, and in our view, should be rejected '. Then he added that the book should be hurled into the water, and he issued a command that, if any of the books of the Persian and Magians should come to light in his territories, they were all to be burnt. Because of this, right down to the time of the Samanids, no Persian poetry is known; if such poetry was occasionally composed, it was not written down and preserved."5

Contrary to Daulat-Shdh's concluding remarks, we do in fact have quite a few references to Persian

poets of the third/ninth century who lived in the Tahirid and early Saffdrid periods, with some citations of their verses. It is accordingly certain that at least from the middle years of the century onwards, verse which was quantitative, rhymed and written in metres similar to those of Arabic poetry (in contrast to the Middle Persian, which seems to have been characterized by stress, to have had no regard for

quantity and to have had varying numbers of unstressed syllables),6 was being composed. Admittedly, the material surviving is exiguous,7 and poses some problems. We cannot, for instance, accept 'Aufi's story8 that when the future Caliph al-Ma'miin entered Merv in 193/809 he was greeted with a Persian

qasida by a local scholar and poet, 'Abbis or Abi 1l-'Abbds. The authenticity of the lines quoted by 'Aufi has been denied by most competent scholars on the grounds of their polished and advanced style (they are written in an impeccable Arabic ramal metre); Lazard concludes that " Il faut admettre soit

que cette qasida est une contrefagon delib&ree, fabriquee quelques siecles apres, pour assurer ' la ville

de Mary la gloire d'avoir produit le premier porte persan, soit que, par suite d'une confusion, une oeuvre relativement tardive s'est trouvee rapportee i l'6poque de Ma'mtin ".0

None of these Persian poets of the third/ninth century seems to have had links, so far as we know, with the Tahirids. E. G. Browne wrote that the Hanzala of Bddghis mentioned by 'Aufi and also by Nizami 'Aridi Samarqandi in his Chahdr maqdla " lived more or less under their [sc. the Tthirids'] patronage ",10 but there does not appear to be anything to support this assertion. Hanzala probably wrote in his native Badghis, for according to Nizami 'Aridi, the commander Ahmad b. 'AbdallTh

al-Khujistmni11 read Hanzala's verses there and was inspired by them to abandon the life of a herdsman

and launch out on a career of military adventure.12 If Hanzala's fame was purely local, it is unlikely that he could have had any contact with the Tahirid court in Nishapiir. Tahirid authority hardly extended beyond the walls of Herat, and the mountainous hinterland of Bidghis and Gharchistan was

largely the haunt of Khariji sectaries.13 During the governorship of the last Tahirid in the east, Muhammad b. Tahir b. 'Abdallah (248-59/862-73), Tahirid control over Herat itself was lost to rivals like Ya'qiib b. al-Laith and Ahmad al-Khujistdni. It seems, rather, to have been the SaffArids of Sistan,

Ya'qfib and 'Amr b. al-Laith, who first extended the patronage of a ruling house to Persian poets; on the evidence of the local chronicle, the Ta'rikh-i Sistdn, Muhammad b. Wasif acted as their court poet.14

Thus we are left with no solid evidence that the TThirids encouraged the first stirrings of the revival of New Persian language and literature. Yet we cannot, on the contrary, accept Daulat-Shdh's allega- tion that 'Abdallth b. TThir took positive measures to suppress what remained in his lands of the pre- Islamic and Middle Persian literary heritage, hunting out and destroying Zoroastrian books and manu-

scripts; the unlikelihood of his pursuing such a policy was pointed out by Barthold,15 and the whole

5 Ed. Mu..

'Abbisi (Tehran N.D.), p. 35. 6 Such, at least, seems to be the evidence of the Middle Persian

verse texts found amongst the Manichaean fragments in Eastern Turkestan; see M. Boyce, " The Parthian g6sdn and Iranian Minstrel Tradition ", JRAS (1957), PP- 39 iff

7 It is gathered together by G. Lazard in his Les premiers podtes persans (IXe-Xe siecles) (Tehran-Paris 1342/1964), vol. I, pp. Io-2o; vol. II, pp. 12-23; cf. also E. G. Browne, A Literary History of Persia (London and Cambridge I902-24), vol. I, PP- 452-3; J. Rypka et al., Iranische Literaturgeschichte (Leipzig 1959), pp. I36-7; and A. J. Arberry, Classical Persian Literature (London 1958), pp. 3I-2.

8 In the Lubdb al-albdb, ed. Nafisi, pp. 21-2.

9 Lazard, op. cit., vol. I, p. 3; cf. also Chr. Rempis, "Die altesten Dichtungen in Neupersisch", ZDMG CI (1951), p. 221.

10 A Literary History of Persia, vol. I, p. 346. More judicious is the comment of R. N. Frye, The Heritage of Persia (London I962), p. 253: " . . . the creation of a New Persian poetry coincides [my italics] with the dynasty of the Tahirids ".

11 Khujistdn was itself a district of Bddghis, see Ydqilt, Mu'janz al-bulddn (Beirut 1374-76/1955-57), vol. II, p. 347-

12 Chahdr maqdla, ed. Mirza Muh. Qazwini, Gibb Memorial Series, XI/i (London 19Io), p. 26, revised tr. by Browne, GMS, XI/2 (London 1921), pp. 27-8.

13 See Bosworth, " The Armies of the Saff~rids ", BSOAS

XXXI/3 (1968), pp. 543 and 544.

14 Cf. Lazard, op. cit., vol. I, pp. 12 and 18; vol. II, pp. 13-15

15 In his El' Art. " Tdhirids ",

THE TAHIRIDS AND PERSIAN LITERATURE 105

story is on a par with that of 'Umar's ordering the destruction of the library at Alexandria. The Thhirids controlled not merely Khurasan proper, but virtually all Persia with the exception of the north-west and the Caspian provinces. However, it is well-known that the Zoroastrian communities of Persia, though generally on the defensive against the incoming faith of Islam, were quite flourishing and productive during the third/ninth and early fourth/tenth centuries. It is in this period that a Zoroastrian apologetic and polemical literature arose and that the work of making Middle Persian excerpts and compilations from the Avesta proceeded apace: the latest important redaction of the Bundahishn comes from the end of the third/ninth century, and the third/ninth century element in the encyclopaedic Dinkard is also strong.16 The Letters of Manishchihr, High Priest of Firs and Kirman, show the existence of a Zoroastrian community in Nishapiir during the third/ninth century; a hundred years and more later, the leader of the fundamentalist Muslim Karramiyya sect, Abti Ya'qtib Ishlq b.

MahmashSdh (d. 383/993) and the great Sifi shaikh Abi Sa'id b. Abi 1-Khair al-Maihani (d. 440/1049) were still converting Zoroastrians there.17

The alleged hostility of the Tahirids to Persian culture may well have originated in the efforts of the first Tahirids, and especially of Tahir and 'Abdallih, to stress the family's links with the ancient nobility of the Arabs, through the family's original clientage to the tribe of Khuzi'a; this process can be clearly traced in the compositions of the poets and authors attached to the Tahirids' circle.18 At the same time, the Tahirids did not object when certain of their panegyrists, and especially those with

Shu'fbi sympathies, tried to attach them to the old Iranian past, its emperors and heroes.'9 The result was a certain ambivalence of outlook, reflecting the tensions existing in the Caliphate of the time, when the Persian elements who had already achieved the highest political offices within the state were now claiming social and cultural equality with the proud Arabs.

There seems little reason to doubt that within the entourage of the Tahirid Amirs, the Persian language was generally-known and its use not discouraged. When the great Arabic poet Abil Tammdm at-Td'i came to 'AbdallTh's court at Nishaplir seeking the Amir's patronage, he became enamoured there of a slave girl who used to sing in Persian (kdnat tughanni bi-1-Fdrisiyya).20 A practical knowledge of that tongue might even be advantageous, and it might accordingly be learnt by Arabs and other non- Persians. We have explicit information in a revealing passage of Ibn Abi Tahir Taiffir's History of Baghdad that this was the case with the Syrian scholar and poet Kulthiim b. 'Amr al-'Attabi (d. ?220/835), of Taghlibi lineage and originally from Qinnasrin. Al-'Attdbi had been attached to the circle of the Barmakids, but after their downfall, gravitated to the side of Tahir Dhfi 1-Yaminain.21 Ibn Taiffir records thus:

" Abfi 1-Hasan Ahmad b. Muhammad al-Muhallabi related to me. He said, Yahyd b. al-Hasan b. 'Ali b. Mu'adh b. Muslim related to me. He said, 'I was at ar-Raqqa with Muhammad b. Tahir b. al-Husain,22 sitting beside a pool, and I summoned one of my slave boys, addressing him in Persian. Al-'Attabi came forward, having been listening to our conversation, and spoke with me in Persian too. I said to him, " O Abfi 'Amr, how is it that you come to know this foreign tongue? " He answered me, " I have been in your country three times, and copied out Persian books which were in the library at Merv. These books were brought there and added to the existing stock by Yazdigird,23 and have remained there till this day. I transcribed those of the books which I needed, and then I journeyed on to Nishdpfir. I had travelled ten farsakhs to a village called Dh.w.d.r,24 when I remembered a book that I had not finished with; so I retraced my steps towards Merv and

16 Cf. Frye, The Heritage of Persia, pp. 249 ff.; M. Boyce, " Middle

Persian Literature ", in Handbuch der Orientalistik, vol. IV: Iranistik, part ii. Literatur (Leiden-Cologne 1968), pp. 39 ff. I am grateful to Professor Boyce for guidance in these complex questions of later Middle Persian literature.

17 Cf. Bosworth, The Ghaznavids, Their Empire in Afghanistan and Eastern Iran 994-o1040o (Edinburgh 1963), pp. 186, 201-02; Frye, " Zoroastrier in der friihislamischen Zeit ", Der Islam XL (1965), p. 199.

18 See Bosworth, " The Tahirids and Arabic Culture ". 19 See ibid. (poetry of 'Allan ash-Shu'fibi). 20 As-Siili, Akhbdr Abt Tammdm, ed. Khalil Malmfid 'Asakir et al.

(Beirut N.D.), p. 213.

21 See R. Blachere, EIl s.v.

22 Sc. the brother of 'Abdallah and mamd.h of such poets as al-Buhturi.

23 I.e. Yazdigird III, the last Sisdnid emperor, who fled eastwards to Merv before the advancing Arabs, and was murdered there in 31/651. I follow the Cairo text here (see below, n. 25); Keller has burud j.r.d. and translates " alte Manteln aus gestreiftem Stoff" (?).

24 Probably Dizbdd or Dihbid, Arabic Qasr ar-Rih, to the south- east of Nishdpfir, where the roads for Merv and Herat bifur- cated; see G. Le Strange, The Lands of the Eastern Caliphate (Cambridge 1905), p. 388.


stayed there several more months." I enquired, " O Abfi 'Amr, why did you copy out Persian books? " He retorted, " Are fine meanings (al-ma'dni) to be found anywhere except in Persian books and in eloquent expression? We have the classic language (al-lugha), they have the fine meanings! " After that, he used often to discuss and talk with me in Persian '."25

Our final conclusion-a tentative one, given the paucity of information-must be that the Persian language was at least tolerated in the entourage of the Tahirids, and that the Amirs were not positively anti-Iranian. Yet they were indeed highly Arabized in culture and outlook, and eager to be accepted in the Caliphal world where the cultivation of things Arabic gave social and cultural prestige. For this reason, the Tahirids could not play a part in the renaissance of New Persian language and literature. The leading parts in this cultural process were to be played by such dynasties as the Saffarids of Sistan, plebeian in origin and outlook, ignorant of Arabic and contemptuous of Caliphal pretensions; and by the Sdmdnids of Transoxania, who were faithful to the orthodox Sunni aristocratic tradition and res- pectful of the Arabic learning which was its concomitant, but were geographically far enough away from the heart of the Caliphate not to be dominated by its culture and literary norms.26

25 Ed. H. Keller, Sechster Band des Kitab Bagddd von Alhmad ibn abf Tdhir Taifar, 2 vols. (Leipzig I908), vol. I, pp. i57-8, tr. vol. II, p. 7I; ed. Mulh al-Kauthari (Cairo 1368/1949), p. 87.

26 This r6le of the Sdmdnids is well-known, and is treated at length in the standard works on Persian literature. That of the

Saffirids is less-known (though its importance was briefly referred to by Browne, Literary History of Persia, vol. I, p. 347), but has now been extensively discussed by S. M. Stern in his article "Ya'qfib the Coppersmith and Persian National Sentiment ", in the Minorsky Memorial Volume (forthcoming).

107

KIRMAN AND THE MIDDLE EAST: PAUL WARD ENGLISH'S CITY AND VILLAGE IN IRAN: SETTLEMENT AND ECONOMY

IN THE KIRMAN BASIN

By Brian Spooner and Philip C. Salzman

" Few of us realize that geography is the prince of disciplines, combining the fruits of geology, meteorology, anthropology, sociology, economics, and dozens of other specialities. The good geographer is a philosopher " (Coon, 1964: Io).

INTRODUCTION

It would not, I think, be an exaggeration to say that the appearance of City and Village in Iran by Paul Ward English marks the beginning of a new stage in the steady development of academic writing on Iran. The book is concerned not with a ruler, or a period, nor with a city or a tribe or a village, but with a region. This is a new method of research which has been much discussed in recent years and in a larger sense has been responsible for new departures in the university teaching of oriental studies. This book is the first full scale study according to this method to appear concerning Iran. Though it is still perhaps early (at the time of writing it is still not yet two years since the book's publication), the authors of this review article consider that it has not received sufficient attention. The book is by a geographer. The method implies use and co-ordination of the points of view of other disciplines (in so far as they may bear on the subject). The present writers are social anthropologists, one of us with a strong area bias. As will appear below there is much in this book which we would wish to question. We welcome English's study and the wealth of interesting and detailed material which it makes available, and we consider it a real contribution to the study of Iran, but we disagree with certain points of his application of the method and consequently of his interpretation of his material. The concept of regional studies grew out of a growing consciousness of disciplinary bias and isolation. It is possible that what follows will betray an anthropological bias. However, the pioneer aspect of English's work demands thorough public interdisciplinary discussion. It is in this spirit that the following critiques were written.

BRIAN SPOONER

I

1. " The villager of Iran, whether sharecropper, weaver, or herder, is inextricably involved in an urban-dominated, regional economic organization and probably was so in the past " (p. 88).1 Further- more, the " concept of urban dominance ... explains some of the perplexities produced by village studies elsewhere in the Middle East" (p. xix). These conclusions are based upon Dr. English's analysis of settlement and economy in the Kirmin Basin.

English approaches his data with the heuristic proposition that cities, villages and herding camps are not pristine isolates, each solving the problems of survival on its own. Rather, these different settlement types are inextricably interdependent, and therefore the region, and not the single community, must be the unit of study (p. xviii). Such an approach, English suggests, will not only illuminate new dimensions of understanding, but will correct old confusions about social and economic organization in the Middle East which are the result of scholarly emphasis upon the internal structure of single communities (pp. xvii-xviii).

x References are to English 1966, unless otherwise indicated.


Beginning with regional analysis as a heuristic device, English concludes with the substantive

hypothesis of urban dominance in Iran and the Middle East. What are the intervening steps in the

argument, and to what extent do the conclusions follow from the data presented ?

2. To summarize English's data, there are two factors that result in urban dominance in the Kirman Basin: capital for agriculture and industry, and co-ordination of industrial processes. Because of the scanty rainfall, the Basin depends upon an elaborate and extensive system of qandts (underground irrigation tunnels). On such a scale, the qandts are extremely expensive to build and maintain, and therefore necessitate large amounts of capital and complex regulations about ownership and use. The capital is supplied by wealthy city dwellers (of Kirman City) who then control the water rights and usage and receive a return for water use by agriculturalists. In addition to the water rights, most of the land itself is owned by landowners, who often provide land, seed and draught animals as well as water. For providing labour, the sharecropper receives 30 per cent of the crops. However, the town and village dwellers of the Basin are by no means mainly agriculturalists. Only in the small highland villages do the inhabitants live by agriculture; the majority of the inhabitants of the larger villages, of the regional sub-centres, and of Kirman City itself, are professional people, merchants, weavers and unskilled labourers. The weavers, who make up the largest plurality, are part of a highly organized industry which is directed from Kirman City through agents of the merchants. Raw materials, a great deal of which comes from flocks which are under contract and which are urban owned, flow into Kirmdn City and are then distributed along with contracts to the weavers in the villages, and the finished rugs are returned to the City. Through capital control, land and water rights, and the organization of industry, the economic life of Kirmin Basin is owned and controlled by city dwellers, and thus the region can be characterized as urban dominated.

This analysis of the Kirman Basin undermines " the traditional description of Middle Eastern 'villages' as physically isolated, homogeneous, subsistence settlements occupied primarily by agriculturalists " (p. 12). In fact, Kirman Basin villages (i) have internal socio-economic differentiation, (2) vary in size and socio-economic function, and (3) are tied closely to the regional subcentres and urban centres on which they are dependent.

English suggests that there is good reason to believe that " the patterns of the Kirmdn Basin are representative " of the Middle East as a whole (p. iii), and therefore that " village morphology, economic structure, and territorial organization are products of centralized urban control of rural resources " (p. xviii).

3. To what extent is the Kirmin Basin representative of Iran in general and the entire Middle East? How confident can we be in applying these findings to the area as a whole ?

English does not document the " representativeness " of the Kirmdn Basin beyond saying that " the physical environment presents many of the same problems to permanent settlers as other Middle Eastern areas ", that is, little rainfall, scanty vegetation, and poor soils (p. xix). And yet even in this, Kirman seems to be an extreme in Iran. For example, in precipitation, Kirman received a yearly average of I72.75 mm. (based on data for 1957-61, p. Io), one of the lowest of all populated areas in Iran. Compare the 1957 precipitation for Mashhad (263 mm.), Qum (269 mm.), Tabriz (299 mm.), Birjand (331 mm.), Tehran (37I mm.), Zanjan (479 mm.), Arak (516 mm.), Hamaddn (522 mm.), Shirtz (652 mm.), and the lush Caspian Coast region (up to 2380 mm., cf. Meteorological Yearbook 1957). Of course, precipitation data does not tell a great deal, and we do know that one or another form of irrigation is used throughout most of the country. But it is clear that we cannot assume homogeneity throughout Iran, and thus the extent to which Kirman is representative must be demonstrated rather than assumed. And even where there is irrigation, large amounts of capital and central organization might not be necessary. The complexity and scope of the irrigation is the telling factor here, for as we have learned from Gray's The Sonjo of Tanganyika, small scale irrigation which is essential for production can be organized on a local level with control and labour from the village agriculturalists.

Is the extensive organization of the rug industry of Kirman typical of other regions of Iran, and what takes its place in areas of the Middle East which do not produce rugs ? Are the Turkoman, Baluchi and Qashqai rugs produced in the same manner as the rugs of Kirmn ? Two questions thus arise: Are

KIRMAN AND THE MIDDLE EAST 109

there areas of Iran and the Middle East where agriculturalists are in the majority and where craftsmen and other workers make up a small portion ? And where there is large scale non-agricultural production, is it always organized on a centralized contract basis ? It is far from clear that well-documented answers to these questions would indicate that Kirman is representative of Iran and the Middle East.

Another consideration is the political aspect of" urban control ". Land and water rights and central

organization are political as well as economic. And a political factor very representative of Iran and the Middle East in general is not in evidence in the Kirman data: that is, tribal pastoralists. The tribes

throughout the Zagros mountains, which cover a large part of Iran, and in the north-east and south-east of Iran, have not until just recently been under any kind of " urban domination ". Very often the

villages were owned by or were political clients of the tribes (Barth; Cooper), a situation which is not

atypical of other areas of the Middle East (Musil; Doughty). The cities and the tribes were sometimes

opponents vying for control of the villages, and occasionally for control of the cities themselves. And

throughout, the tribes, with their economic produce of meat, milk products, skins, wool, and wool

products including rugs, were equal trading partners, as well as independent political agents, and could hardly be included under the label " urban dominated ". Whether or not there was any evidence of them in Kirman, surely the tribes must be taken into account in any model of regional organization in the Middle East.

Among the studies that English criticizes for following the " tripartite view of the Middle East ", and thus using the city, village or tribe as the unit of study, is Stirling's Turkish Village (p. xvii, note 4). Now although Stirling does indeed emphasize internal structure, he examines two villages of quite different sizes and characteristics and takes pains to indicate the incoming and outgoing of men and material and to account for their effects on the internal structure. It does not seem fair to criticize a man for having written a different book than one might have liked, as long as he takes into account all that is necessary for an accurate picture of the subject as he chooses it. However, the importance of this

example here is its relevance to the substantive question of urban domination. If the Turkish village is urban dominated, and this does not seem to be indicated from the data presented in the monograph, the forces are not those at work in Kirmdn. There are no landlords; almost all of the land is peasant- held. There is no large scale irrigation and thus no urban capital necessary for water. Almost everyone is a farmer, and although the influence of the market is felt, there is no centralized contract system. And since almost everyone farms, there is no urban-controlled industry. Therefore, even if a regional study is made, it would be unlikely to uncover " urban control of rural resources ". This is not to say that regional analysis would not be illuminating, but rather that we must separate English's pleas to examine regions as wholes from his substantive hypothesis about urban domination. Regional studies

ought to be done; urban domination must be tested.

4. If we cannot accept urban domination in the Middle East as a substantive conclusion, we can

accept it as a variable to be examined. Urban domination must be conceptualized as a dimension

having degrees from high urban domination to low urban domination; and the other elements of social life that vary as urban domination varies must be discovered. As suggested above, the degree of local self-sufficiency in production and the political balance of power are likely to be associated, although not necessarily in a simple fashion, with the degree of urban dominance. In any case, it is

likely that English's model of urban dominance will teach us as much in its inapplicability as in its

applicability. PHILIP C. SALZMAN

II

i. The following critique is empirical rather than theoretical. Though I have driven through the Kirmin Basin (see map, p. 6) many times, I have never stopped to study it firsthand. However, my field experience elsewhere on the Iranian plateau leads me to question the generalization implied in the title: City and Village in Iran: Settlement and Economy in the Kirmdn Basin. English is concerned to correct previous assumptions about villages in the Middle East: I disagree with his appreciation of these assumptions,


However, my criticism aims primarily at redressing the balance in his interpretation. The imbalance may have arisen from over-correction of the imperfections in the existing literature. His own field material is limited to the Kirmdn Basin itself, where he worked for eleven months, in 1961-62 and the emphasis throughout his interpretation of this material is on purely economic factors. By drawing attention to situations in other parts of the Iranian plateau and pointing out the relevance of political and oecological factors both in the Kirmdn Basin and elsewhere, I wish on the one hand to show the importance of English's work, and on the other to restrict his generalization.

I am mainly concerned, therefore, with his Introduction, and Chapters 5 and 7 and the conclusion; I consider that the most valuable chapter-apart from the often excellent factual detail in the appendices -is Chapter 3, closely followed by 4 and 6.

2. The theoretical fulcrum of this book is that " the region, not a single community, is the unit of study " ... " There are strong patterns of inter-relatedness between every village and the city, and weaker social and economic ties among villages " ... " The field area (the Kirman Basin) provides an excellent laboratory for testing this regional approach " (pp. xviii-xix).

Of these three propositions the first two are admirable; the third is misleading. I have myself made a plea for a type of area study in Iran in an article published in 1965, and feel strongly that this approach is sorely needed. But what is the unit of study ? What is a region ? English devotes some space (cf. Introduction) to justifying the " approach " (the importance of which is unlikely to be disputed today), but nowhere suggests how we should in general choose and define our regions. Surely this is an important question. Unfortunately, because of the general validity English claims for the urban domination he finds in the Kirmin Basin, he appears to assume that every region-every unit of study- must contain a city. This is of course absurd. And this is one obvious reason why his thesis of urban domination cannot have as general a validity as he appears to claim.

The simple reason why it is absurd is that a number of villages within one hundred miles of the city of

Kirman are not included by English in his regional study (the Kirman Basin) and could not feasibly be included in one region with any other city: they must therefore either be studied separately from any city, or not at all! In so far as my acquaintance with the Kirman Basin allows me to judge, it would indeed seem to be a very interesting region to treat in a regional study-but not because it contains a city. Its fitness depends on the fact that it contains a group of settlements of varying size which bear closer sociological relationships to each other than to other settlements or communities outside the region.

English's argument is almost purely from economics, and he actually says: " Perhaps this inter- dependence between city and village is most striking in the economic sphere ". Of course economic ties are important, but they are not always primary and they seldom provide the whole explanation to any situation. The primary factors in this situation are oecological: firstly, that the coincidence of cultivable soil and water for irrigation together with the given level of technology allows a population of this size and in this particular settlement pattern; secondly, that the topography endorses a distinct break between this and other neighbouring patterns. Given this oecological situation, other factors-in particular those of economics-result. This is not oecological determinism. The oecological situation does little to shape the sociological relationships. What it does, is to define sites and size of possible agricultural (or pastoral) activity according to any given level of technology. Similarly, the topography itself suggests routes of communication according to the level of technology available. In seeking to choose and define, therefore, suitable regions for " regional study " we should look for what might be called " oecological units ".2 Obviously, there are oecological units which from the point of view of any of the social sciences will not be suitable for regional study-because of lack or surfeit of population. Nevertheless, oecological criteria together with a consideration of patterns of settlement and pastoral activity should allow the definition of a suitable region.

3. " Kirmin's physical environment (Chapter i) presents many of the same problems to permanent settlers as other Middle Eastern areas. It lies in a mountain-ringed basin, has little rainfall, scanty

2 More specifically: areas of settled and/or nomadic communities which, because of a combination of factors of topography, natural conditions and technology, enjoy a network of socio-

logical relationships primarily with each other, and only secondarily with communities outside the area.


vegetation, and poor soils. All crops must be irrigated and protected from blowing sand; grazing resources have deteriorated, and the fuel problem is acute " (p. xix).

All this is certainly true of the Kirman Basin, but it is extreme rather than typical. The most striking feature of the KirmAn Basin is the vital importance of qandts (cf. Appendix D) to absolutely all agricultural activity, the length of these qandts, and the amount of capital and day-to-day expenditure which must be continually devoted to them. In only two other regions-whether on the plateau, within Iran, or anywhere else-do qandts occupy a comparable position of economic and oecological importance: the region which includes the city of Yazd, and the plain of Gundbid (both of which are situated on the plateau). Yazd does, it is true, bear a reasonably close resemblance to the picture English draws of Kirman, but the GunabSd plain is very different (cf. Tdbandeh, 1954). Traditionally, the plain of Gunabdd contains seventeen villages, but no city or major centre.3 There can therefore be no question of urban dominance. Another very obvious reason why the Kirman Basin can in no way be regarded as typical, is that it contains no nomadic pastoralists. In Iran, in general, this makes it exceptional rather than typical. It may of course be argued that the presence even of a largish body of fully nomadic pastoralists need not upset English's interpretation of urban dominance, since we could expect them to depend on the facilities of the city to sell their products and buy grain. However, the long history of the relationship between nomadic and settled on the Iranian plateau, and elsewhere in the Islamic world is more complex than this. The crux of the whole situation in Kirmdn is that there can be no agriculture without qandts (this is not unusual on the Iranian plateau, but it cannot be said to be altogether typical), and the topography and sedimentary fill of the basin is such that the cost of these qandts both in terms of capital investment and maintenance is unusually high. Nevertheless, they do produce enough water to support a city which can in turn afford the financial effort required.

4. This is perhaps enough at this stage to show that there is in the Kirman Basin much that is in no way typical of other regions. Allowing a high degree of urban dominance in the Kirman Basin itself, there are other factors which English has omitted. The most important of these is the extent to which processes of centralization, and Westernization or modernization, may have influenced the traditional patterns. For there is now a large degree of government participation in several sectors of the economy, and this is bound to heighten the effect of any economic-or other-domination from the city.

Secondly, English himself admits that " the economic structure of settlements in Kirmdn varies, but agriculturalists are a majority only in the smallest, most remote villages and hamlets. Most of the people are weavers, merchants, professionals, and unskilled laborers. The organisation of economic activity in all settlements is dominated by city dwellers " (my italics) ... but " one reservation should be noted here. All settlements treated in this study lie within forty miles of Kirman City. It seems probable that isolated peasant villages would be more common at greater distances from a city"! (p. xix and note 18). The type of urban domination English is talking about is bound to be economic, and it would be surprising indeed if the villages within forty miles of a city the size of Kirmdn (60o,ooo) were not economically dominated by it. It is surely only natural that the farther the village is from the city the more distant its sociological relationship will also be. A very large proportion of Persian villages are more than forty miles away from any city. According to the official census figures of 1956 the rural population of Iran was 13,OOI,141 distributed among 49,054 villages of a mean size of only 265o0 inhabitants each (Census, vol. I, p. 21). English defines a small village (i.e. settlements where most of the population practise agriculture) as one with a population of between 100 and 1000 (p. 33).

Finally, although it is noted that " shrines and mosques and their environs are often focal points of social and economic power in their respective settlements " (p. 57), religion in general could be made more use of in an assessment of" settlement and economy ". The most obvious point is that the site of a shrine is often a matter of interest to an area, not just an individual settlement. It causes traffic, and can influence the direction of spread, or even the location of a settlement. Again, sectarian differences (a

* During the last decade one of the villages (Jfiymand) has in fact been made into a regional centre for administrative purposes and given the name Gunlbid, which previously did

not apply to any one settlement. This has naturally affected the pattern of relationships between the villages.


special feature in the case of Kirman) tend to cause spatial segregation in the population, and are therefore also likely to leave a mark on settlement. (This is noted for the city but not the villages.)

5. In a sense, the " diachronic study " is another and a necessary aspect of the " regional study " and there is much to be said for the inclusion of a chapter on the " historical development of settlement " in roughly the position in which we find it. However, in this particular case it does not appear to

support either the general argument or the conclusions of the book. In addition to this, much of the historical synopsis given is either based on obscure authorities, or is simply one possible interpretation of rather scanty data. But the most significant part of the chapter is the account of the foundation of the city of Kirman and the colonization of the Basin, which is shown to account for an unusual lack of communalism. Kirman was founded by the first Sasanian king " as a defensive outpost of the ... Empire. .... The site ... was chosen for defensive reasons. .... The present functionally integrated settlement pattern was initiated by this well-equipped feudal society which, as it grew, extended its

capital, labor, and technology outward to the alluvial fans and later into the mountains ..." " This settlement process explains the paucity of communal traits in the villages of Kirman as

compared with other areas of Iran " (pp. 21-3). This ideal is developed later on. For instance: " every qandt in the Kirmin Basin (for which records exist) was built by a member of the urban 6lite, whether merchant or landlord " (p. 66). This does not help to prove that Kirman is typical.4

6. Another factor which could be used to much better advantage in this study is the role of communications. We are told that Kirmdn was not on a major trade route in early Abbasid times (p. 25; the account of the trade routes existing at that time which he provides in the next sentence is very confused), and that " from the European vantage point, the raison d'Xtre of cities such as Mashhad, Tehran, Tabriz, Isfahan, and Kirman has historically been international trade " and that this is wrong (pp. II 1-12). But we are not told of the effects on local prosperity when one arterial route is for some external reason superseded by another. We are even told that the coming of the motor age merely " intensified and expanded an existing pattern " (p. 66), whereas there certainly are cases where

patterns of communications have been completely changed by the advent of motor traffic (cf. e.g. Spooner, 1965). Perhaps this was not the case in the Kirmdn Basin, but English claims general validity for his statements, and the extreme case of Shahdad (just east of Kirman) is close at hand: by donkey it used to be sixty miles from the city; by motorized transport it is now 280 miles.5 Communications, trade, prosperity-these are always surely complementary parameters.

7. Though we have already shown that English's argument from natural conditions does not help his case, nevertheless it still calls for some comment on detail. Firstly, that any use of Persian statistics should be accompanied by a caution-even if only because we know that in the case of rainfall there is not a long enough run to give a valid mean. Secondly, the most important form of precipitation for the Kirmdn Basin is surely not rain in the Basin at all, but the snow which falls in winter on the surrounding high mountains. For it is this which replenishes the aquifers from which the qandts draw the water for

irrigation. No mention is made of this. Rain in the Basin itself of course also plays its part, but it is not so much its quantity and distribution as its intensity and its irregularity which are important. An inch of rain or more may fall in less than twenty-four hours, but in doing so it may generate such force of run-offthat damage to qandts and other capital property will far outweigh any addition to the water table.

8. The fact that the centre of the Kirman Basin is practically a sand sea may perhaps also be seen

as a factor affecting the unity of the region. Such a feature may often be a factor in making the centres of plains or depressions into borders of regions, and causing settlements built on alluvial fans to form their relationships with their neighbours on the other side of the range (assuming there are convenient dry river beds and passes) rather than with the villages they can see across the plain. However, in this case the magnetic pull of an urban centre the size of Kirman would counteract any such tendency.

* It is worth noting that in a footnote to this passage English admits that he is rejecting the opinion of the most generally accepted authority on this particular subject. It is nevertheless a little quicker by jeep than by donkey. However, Shahdid now exports its citrus fruit harvest to

Birjand which by truck is about the same distance as Kirmdn.

Thus, the Kirmdn market for citrus fruit is supplied by Bam, and Shahdtd supplies the north-east of Iran via Birjand. Without motor transport the two producing areas would be forced to compete for the Kirmdn market.


9. English's whole reason for weighting the interpretation of his material in this way is that he considers that " it is customary for writers to divide Middle Eastern society into three sectors-city, village and tribe-each rooted in a separate social environment, each exploiting a different physical environment " (p. xvii). I contend that the prevalence of this " tripartite " view is greatly over- emphasized. English justifies it by four long bibliographical footnotes. I may be wrong, but I do not think any of the works cited actually states this " tripartite " view in so many words. Most of them (in notes 2, 3 and 4) are simply " unit studies ", rather than " regional studies ". And English himself says that he does not mean to " imply that a geographical approach precludes the study of a single community " (p. 156, note Io). Nevertheless, in several places (cf. e.g. pp. 65, 67, 87, 88) it is somewhat vaguely suggested that a whole " literature " exists to support this trichotomy. One of the authorities cited (p. 156, note 4) as a non-regional village study (Stirling, 1965) is in fact used again later (p. 170, note I) in support of the contention that " the peasantry is an integral, functioning part of a larger society, not a divorced element " (p. 88).

Io. The term feudal is used a great deal, but we are left to guess in what sense it is being used. It is certainly not the sense which may be abstracted from medieval European feudal society because it is used to imply something complementary to centralized urban domination. It seems to be equivalent to Sjoberg's (1960) usage of the term, and perhaps we were meant to guess this from footnote 13 on p. 157. However, a deviant usage of such an emotive and often misused term should be made much more explicit.

I I. Finally, English would like his material, as he interprets it, to solve the problems not just of Iran but of the whole of the Middle East (cf. pp. xx and II 1-14). What is the basis of such all-inclusive generalizations? The Middle East is a political concept designed by the West to cover an area which subscribes to one formal religion and has for brief periods during a long history been mostly subject to one political authority. English is by no means the first to attempt to make such broad generalizations, and there is of course at least one sense in which the whole of the Middle East (even from the Atlantic to the borders of China) constitutes one cultural area. But is this really a justification for asking questions like " What is the nature of settlement in the Middle East ? " " What is the nature of Middle Eastern social and economic organization? " (p. xx), and attempting to find comparatively simple answers for them (pp. xx and I 1-14) ? The questions are, of course, of great interest, but require a much greater fund of more varied field data before they can be convincingly answered-if they turn out to be, in fact, real questions. BRIAN SPOONER

BRIAN SPOONER

BIBLIOGRAPHY

Barth, F., Nomads of South Persia (Oslo I96i). Census, 1956: National and Province Statistics of the First Census of

Iran: November 1956 (Teheran, August i961), 2 vols. Coon, C. S., Caravan: The Story of the Middle East (1964, rev.

edn.). Cooper, M. C., Grass (New York 1925). Doughty, C. M., Arabia Deserta (London 1933). English, Paul Ward, City and Village in Iran: Settlement and Economy

in the Kirman Basin (Madison 1966). Gray, Robert F., The Sonjo of Tanganyika (O.U.P. 1963).

Meteorological Yearbook 1957, Iran Meteorological Department (Teheran I960).

Musil, Alois, The Manners and Customs of the Rwala Bedouins (New York 1928).

Sjoberg, G., The Preindustrial City: Past and Present (1960). Spooner, B. " Arghiydn: the Area of Jijarm in Western Khurd-

smn ", Iran III (1965).

Stirling, Paul, Turkish Village (London 1965). Tibandeh, S. H., Ndbigheh-i 'llm va 'Irfdn dar qarn-i chahdrdahum

(1954) (Tehran 1333 A.H.S.).

6 I have noticed the following factual errors which the author may like to correct in future editions:

(i) Vaziri, Ahmad 'Ali Khan: Tdrikh-i Kirmdn (Saldrt- yeh) [History of Kirman] was edited by Muhammad Ibrdhim Bastani Parizi. (HUfiz Farmanfarmaidn is the general editor of the series in which it was published, cf. p. i90.)

(ii) The city of Kirmin never served an Iranian province of Baluchistan, though the area known as Baluchistan was always administratively dependent on it. In the past Kirmdn and Baluchistan were not separate provinces. At present, both areas are not included within the same administrative subdivision.

They were separated under the Pahlavi regime, and the new city of Zdhidan became first an independent farmdnddrf-i kull, and later (1958) a full province (ustdn, cf. pp. 5 and 157-8, note 5).

(iii) The world mazra'eh is singular and means " field " (literally: " place of cultivation "). The word does not include the idea of residence, and a mazra'eh is often not residential (pp. 16 and 68).

(iv) Kirman is east of Sirjan (p. 25). Lastly, is not the English for Fdrsi " Persian " (cf. p. xviii),

as, for example, deutsch is German ?

115

EXCAVATIONS AT BABA JAN, 1967: Second Preliminary Report

By Clare Goff

The second season's excavations at Bdbd Janl began on July 26th 1967, and continued for ten weeks. As before, the work was sponsored by the University of London, Institute of Archaeology, and directed by the writer. The team consisted of Miss Kay Wright (senior field assistant and photographer), Mr. Robin Dennell, Mr. Hugh Chapman and Mr. Roy Dyckhoff (field assistants), Mr. George Learmouth (surveyor and conservationist), and Miss Jennifer Davison (pottery assistant and registrar). Miss Wright also took charge of the field " first-aid post ", and Mr. Dyckhoff of the accounts. Mr. Mohammed Sarraf, as representative of the Iranian Department of Antiquities, gave invaluable help with the interpretation, drawing and driving, and Agha Heyder from Takht-e-Jamshid once again acted as foreman. My husband, Mr. Price Meade, was able to spend some days helping us at the end of the season. Most of the pottery and small finds were afterwards drawn in Tehran by Mr. Alan Bates and Mr. Richard Jones.

Thanks are also owed to Mr. Pourmand, Director General of the Iranian Archaeological Service, the local authorities of Khorramdbad, and Agha Jazayeri of Khorramabad and Agha Sharifi of Nfirabtd, who gave invaluable local help. The excavations were financed by grants from the British Academy; the Institute of Archaeology, London; the Ashmolean Museum, Oxford; and the City Museums of Bolton and Birmingham. The personal expenses of the writer while on the dig, and on surveys connected with it, were met by a grant from the Wainwright Trust, Oxford.

THE EXCAVATIONS

Work was carried on in five areas (Site Plan, Fig. I). On the Central Mound we completed the excavation of the two " Fortified Manors " discovered the previous season (Trenches H to L); a 10 X IO m. sounding was made to the south-west of these buildings on the ridge connecting the Central Mound to the Islamic grave yard (Trench Z2); and the deep sounding in Trench F was carried down to a depth of just over 6 m. below the summit of the tepe when it became unsafe to continue. On the East Mound, last year's stepped trench was enlarged at its eastern end (Trench A) to obtain more information about Levels v and iv, and a deep sounding was made on the summit of the mound (Trench B) to elucidate the later part of the sequence. Finally three small trenches, P, Q and R, which have not yet been surveyed into the plan, were dug into the saddle separating the two mounds to ascertain the nature of the occupation in this area.

As it is obviously impossible to describe all these operations in a single preliminary report, this year's account will be confined to the Iron II architecture and associated finds.

The Fortified Manors Level 2 (Figs. 2 and 3)

The eastern half of Level 2 was much better preserved than the western sector which we had excavated in 1966. Several of the walls, notably those surrounding the East Long Room, and the South-East and North Towers, were reused in Level I and had their mud brick superstructure preserved for several courses (Section, Fig. 5a and b). The average height of one of these courses was I4 to 15 cm. with about 5 cm. of mortar between the bricks, which were either orange-brown or grey.

1 For the previous season's excavations, see C. Goff Meade, " LiiristSn in the First Half of the First Millennium B.c.: A preliminary report on the first season's excavations at Babd

11A

Jan, and associated surveys in the Eastern Pish-i-Kfih ", Iran VI (1968), pp. 1o5-34; henceforward referred to as Goff Meade, " LfiristAn ", Iran VI.


Fig. I. Plan of Babd Jan showing excavated areas 1966 and 1967.

EXCAVATIONS AT BABA JAN, 1967 117

The completed plan shows a compact, symmetrical building, with a Central Court or Hall,2 flanked to the east and west by two long rectangular rooms, and surrounded on the four sides and corners by eight smaller rooms, which probably continued upwards in the form of towers. The overall dimensions of the building were 33 x 35 m., with the West Long Room, Hall and East Long Room

measuring 18-6o x4'4o m., 186o x 96o m. and 17 4 x 5 20 m. respectively. The building was entered through the Central Tower in the east wall. The entrance was paved

with large flag stones and probably had a mud-brick bench at either side against the wall,3 where

guards, or those seeking audience with the local chief could wait until summoned. From there one entered the East Long Room over a stone threshold. This room, like the entrance, was also partially paved and had benches along the far wall. If the Central Hall was roofed at this period, this room may have served as an ante-chamber, and the main audience have taken place beyond. It seems fairly certain that the West Long Room was primarily domestic in function, since it contained a large quern and several storage vessels.

Level i (Figs. 3 and 4) The excavation of the eastern section of this building has enabled us to gauge more accurately its

relationship to the earlier Manor of Level 2. As the composite plan (Fig. 3) will show, any one of three

procedures could be adopted by the builders: either the Level 2 plan could be completely ignored, the walls stripped to their foundations, and the Level i rooms laid out on a totally new plan, as in the west front; or, the Level 2 foundations could be stripped and be incorporated into the foundations of the Level I walls, as in the area surrounding the West Long Room; or, finally the Level 2 walls could be retained and left standing, either alone, as in the South East Tower, or with Level I walls built up against them, as with much of the centre and east of the building. In this last case, the Level I foundations often start at a slightly higher level than those from Level 2 and also slightly overlap them. This

suggests that, as in contemporary mud-brick structures, the lower sections of the Level 2 walls had been

partially eaten away and must consequently have been left standing for some time before the rebuilding took place. The dual origin of the walls in this area also clears up a point that puzzled us last year-the apparently flimsy character of the northern and eastern walls of the Central Hall.4

The final plan of Level i is essentially that of Level 2 with certain important modifications. The

tripartite division of the central area into a Hall flanked by two Long Rooms is retained, and, as described in the earlier report, the Hall was now definitely roofed in with a double row of columns down the centre.5 The main entrance still lay to the east of the building, but the Central Tower had been abolished in favour of a long rectangular Portico with a small guard room at the northern end. The front of this Portico was almost certainly lined by a row of columns but unfortunately the whole of the south-east end had been destroyed by an intrusive horse burial, and we only found evidence for two stone supports, one definite and one possible (Fig. 4, Nos. XI and XII).

This elaborate arrangement of Portico, Long Room and Hall reinforces the parallels drawn last

year between Level I and Burned Building III at IHasanlfi,6 and leads one to speculate again on the nature and purpose of the pair of buttresses in the west wall of the Hall. This was originally interpreted as a doorway, both because of the presence of a door-socket, and on analogy with a similar arrangement in Level vb on the East Mound (below, p. 126). This left unexplained the hearth directly in front of the

entrance, and the discrepancies in height between this floor level of the Hall and the much higher one of the Long Room beyond, to which the door socket corresponded. In the IHasanlfi IV Burned Buildings, the centre of the wall immediately opposite the main entrance was invariably occupied by a bench and

2 Further excavations were carried out this season on the original Level 2 floor of the Hall, without producing definite evidence for post holes earlier than what is taken to be the earliest phase of Level I. The possibility that this may just conceivably be a late phase of 2 is discussed below on pp. 120--1.

3 This tower was destroyed in Level I, except for the stone paving which was reused. The location of benches, here, as in the East Long Room, is deduced from the sudden gap between

the stone flooring and wall, which, in each case, was filled with a slight bricky fill, suggesting a bricky superstructure.

4 Goff Meade, " Lfiristdn ", Iran VI, p. x 5.

5 Goff Meade, " Lfiristan ", Iran VI, p. I 14-. 6 Goff Meade,

" Liristtn ", Iran Vd Goff, " BabA

Jan Tepe " in " Survey of Excavations in Iran during I965- 66 ", Iran V, pp. 133-4-


BABA JAN LEVEL Z

Fig. 2. Plan of Central Mound, Level 2. To avoid ove;crowding the plan with numbers, 0oo m. has been subtracted from all levels.

LEVEL I LEVEL 2A LEVEL 2

Fig. 3 Plan of Central Mound, Levels I and 2, showing reuse of Level 2 walls in the later rebuilding. To avoid overcrowding the plan with numbers, ioo m. has been subtracted from all levels.


the same arrangement can be observed in the Apadana at Godin Tepe.' If there had been a mud brick platform between the two buttresses, this could have served both as a bench on which the chief could sit while conducting business, and at the same time a step to the high threshold over which he could retire to his private apartments. Alternatively the door socket could belong to a later stage in the life of the

BABA JAN LEVEL I

Fig. 4. Plan of Central Mound, Level i. To avoid overcrowding the plan with numbers, ioo m. has been subtracted from all levels.

7 T. C. Young, Jr., " Godin Tepe", in " Survey of Excavations in Iran during 1966-67 ", Iran VI, p. 161.


SECTION X-X

?~~ " . : " : i : : ".I

ce ____@ptooa

"'T"

&SE CT1ON Y--Y

_c~cdved mosl-11ILCL

•m7 Foa

-W4ALL.~c WL

Irey.-, ~

4e•- \#-•a I r9,

r.1" 77777 • •,':r?

,77 T "•'--" /. . ' ..O P.. . .... . .. 5eoqjc~ry

Fig. 5. Sections through South-East Corner Tower (Level 2) and south wall of Portico (Level z) illustrating the inter-relationships of the

two main building phases.

building and correspond to a high secondary floor level in the Hall itself.8 A further possibility is that

what we are calling the early Level I floor with its hearth is really a late stage of Level 2. Certainly it lies below the level of the top of the Level I stone foundations in contrast to the structures of this

period on the East Mound, where the Level vb wall foundations are invariably sunk into a trench.

However this seems unlikely in view of the fact that the column bases, which undoubtedly go with this

floor, also line up with the Level I buttresses and with stones set along the sides of the Level I walls

which must have supported pilasters. The remaining modifications to the Level 2 plan can be dealt with fairly rapidly. As described last

year, the towers along the west front were replaced by a buttressed Alcove, matching the eastern

Portico. Elsewhere they were either incorporated into the Level I structure as side rooms; replaced by

bigger units, such as the Long Room to the south, or possibly, as on the north east corner, rebuilt on a

more massive scale. This, at any rate, seems to be the most likely explanation for the impossibly small

rooms in this area with their massive stone partitions. These foundations, though lying c. I m. below

the surface of the tepe, had no mud brick superstructure, and several large querns, obviously of later

8 Traces of a high floor were found at the northern end of the Hall associated with a hearth and short secondary stretch of

walling. Many of the better preserved small finds, including a bronze bowl, were also found at this level, just under the

plough soil, but it is impossible to tell whether they were associated with a high, eroded floor, or part of a fallen upper storey or roof.


date, lay directly above them. The mud-brick superstructure of the south wall of the Portico was equally amorphous, whereas that of the adjoining Level 2 wall seems to have been preserved to a higher level (Section, Fig. 5). There is, therefore, just the possibility that the reconstruction at this end of the building was never completed, but one cannot be certain about this when so much erosion has obviously taken place.

Two further problems remain to be settled, both connected with the entrance. One is the date and purpose of the wall (Fig. 3, Level " 2a ") which runs straight across the east front of the manors. This is certainly later than Level 2 since it cuts the foundations of the Entrance Tower. Its relationship to Level I is ambiguous. On the one hand it appears just to underly the eastern end of the Portico's north wall, and its mud-brick superstructure also underlies what is assumed to be the stone base for a column (XII).9 On the other hand its stone foundations do not run under the better established stone base of column XI suggesting that this column may already have been in existence at the time the wall was built. An alternative, and perhaps easier theory, would regard it as a low step or bench, built at the same time as Level I and running along the whole of the front of the building as a narrow plinth. But if this is so, then what are we to make of the walls running off it ?

Level 2a apart, there is the final problem of how the entrance was reached from the surrounding fields. The main doorway overlooks the steepest side of the mound which is here scoured by a slight dent, or gully, clearly visible in the contour map (Fig. I). Was the deliberate angling of the building to face the most difficult route of access a defensive measure ? And does the gully mark the position of an earlier ramp or staircase ?

Apart from the buildings, the Central Mound produced two further very unexpected discoveries: an anthropomorphic vase, and a horse burial with accompanying grave goods.

The Anthropomorphic Vase The painted vessel in the shape of a man (P1. IIIa, b) is the first of its kind to be recovered in situ in

Liiristmn, although at least three other examples are known from private collections. One of these, an example from the collection of Mr. James Bomford, is illustrated here for comparative purposes (P1. IIIc).10

The BTbh Jtn figure stands 20- 7 cm. high and is 13- I cm. wide and 9 -o cm. thick across the hips. It is made of the same buff clay as the rest of the pottery from these levels" with burnished surface and brown painted decoration. The painting makes little attempt to represent clothes, but merely depicts a typical " kite and ring " design around the body. Only on the head is there the suggestion of a cap, or hair, and on the feet, possibly laced boots with curled up toes. The body is squat with exaggeratedly wide hips and legs which contrast with its tiny forearms, and schematic, inhuman head, in which eyes and ears are represented by applied clay buttons. The male sexual organs are suggested in a similar fashion. The jar which it clutches to its chest was broken in antiquity, but was presumably similar in shape to that carried by the example from the Bomford collection.

The Bomford figure is perhaps slightly more naturalistically rendered, although the face has an odd resemblance to a bear (a phenomenon also to be seen in the Marlik figures of the same class)."2 In this as in the other two known examples, the body is oval, without the widely exaggerated hips, the arms have the same squat proportions as the rest of the body, and, as they hold the jar away from the face, must also serve as spouts. There also seems to have been more attempt to represent clothing, though, again, many of the motifs used can be paralleled on the other pottery. The figure appears to wear a tunic or coat, with high collar and long sleeves, below which are striped leggings and the usual pointed shoes. The shape of the head may be intended to suggest some form of peaked cap.

9 The bricks from this wall are normally squarish, and about 13-15 cm. thick. The sizes obtained are as follows: 34 x 34 cm.; 36 X 37 cm.; 38 x 17 cm. (half brick); I8 x 40 cm. (half brick).

10 The writer wishes to express her gratitude to Mr. James Bomford for allowing her to study his collection and for the photograph reproduced here. A better reproduction of the Bomford figure is contained in Antiquities from the Bomford

Collection (Exhibition Catalogue, Ashmolean Museum, Oxford 1966), no. 28o, pl. xxiii. For the other figures cf. R. Ghirshman, Persia from the Origins to Alexander the Great (1964), p. 83, pl. Io8 and p. 320, pl. 391.

11 For a description of this pottery cf. Goff Meade, " Lfiristn ", Iran VI, pp. I5-8.

12 Below, note 18.

P1. I. (a) View of Baba Jan from east.

Pl. I. (b) Central Mound, Levels I and 2. View of Entrance from south.

P1. II. (a) Central Mound, Levels I and 2. View of Guard Room at northern end of entrance showing column base, 2a wall and storerooms beyond.

Pl. II. (b) Central Mound, Levels I and 2. View of entrance from south-east, showing relationship of 2a wall to the main structure.

Pl. II. (c) Central Mound, horse burial.

Pl. III. (a) and (b) Anthropomorphic vase from Level I.

Pl. III. (c) Anthropomorphic vase from the Bomford Collection.

Pl. IV. (a) East Mound, Level vb, "Painted Chamber" from east: (b) -(f) Tiles from " Painted Chamber".


The two other known examples13 both resemble the Bomford figure in the shape of their bodies and stance. In one example however most of the painted decoration-not in any case very extensive- seems to have worn away, while the other is unpainted. This last example carries a typical Liristan vessel with attached beak spout ornamented by large bosses.14

The closest parallels for these anthropomorphic vessels occur at Marlik,15 but here they are executed in a somewhat more vigorous style and in a far greater number of poses. Besides the man carrying a jug, there are women,16 warriors,17 and seated " bears ",18 suggesting that the idea possibly originated at Marlik, and reached the south-west in a conventionalized form, possibly through the medium of

.Hasanli, with whose Iron II pottery the " genre Ltirist~n " wares are undoubtedly connected.19

The Horse Burial The first traces of the horse burial were discovered when, in cleaning down the north face of the

H-L baulk, our foreman uncovered the side of a bronze bowl which was only just visible. A guard was left in charge that evening, and the next morning the baulk was removed, revealing the well-preserved skeleton of a horse, accompanied by a small but interesting collection of grave-goods (P1. IIc and Figs. 6 and 7).

The horse was presumably buried in a pit which appears to have been dug into the eastern edge of the stone paving of the entrance, just south of the main axis. The skeleton was surrounded by a hard bricky deposit, but no actual tomb structure could be recovered, though the shallowness of the pit makes a superstructure likely. The burial had disturbed the Level 2-1 paving and would also have obstructed the main entrance. It may therefore be assumed to be intrusive.

The corpse (P1. IIc and Fig. 6a) had apparently been thrown into the grave and weighted down with stones. It lay on its back with the right hind cannon bone pointing upwards and the left hind doubled up with the cannon bone pointing towards the head. The head and forequarters had been twisted round so that the skull faced south with the neck upright as if it had rested against the side of the grave. The forelegs were stretched out in front.

On the dismantling of the skeleton (Fig. 6b), a bronze nose plate (Fig. 7: 3) was discovered lying under the horse's left shoulder. Beneath it lay a very corroded iron bit (Fig. 7: 5), and five looped harness studs (Fig. 7: 4) were scattered in the immediate vicinity, with traces of the harness straps as impressions in the surrounding soil. Not far from the hind-quarters were the bronze bowl (Fig. 7: I) and a large bronze lamp (Fig. 7: 2).

Horse graves have otherwise been found in Iran at Godin Tepe, where the burial was associated with a second millennium tomb, 20 and at Hasanlil,21 where four horses were buried in the upper part of a tomb on the lower level of which lay a human skeleton. A horse's skull was also buried on the East Mound probably in a sixth-seventh century context, but just possibly intrusive. The attribution of the manors to the ninth and eighth centuries22 gives us a terminus post quem for the Central Mound burial, which must otherwise be dated internally.

Most of the objects found with the skeleton have parallels from other sites of the late ninth to the early seventh centuries, suggesting that though intrusive, the burial may belong to the period directly following the abandonment of the manors, when such occupation as there was on the site had shifted to the East Mound. The nose plate can perhaps be compared to more elaborate ivory examples from

13 Cf. note Io.

14 G. Contenau and R. Ghirshman, Fouilles du Tipd Giyan pris de

NdVhavend, 1931 et 1932 (Paris I935), pl. XVIII, second row down, right.

15 E. Negahban, A Preliminary Report on Marlik Excavation (Tehran 1964), pl. XI. See also idem, " Men and Beasts in Pottery from Marlik Burials ", ILN (December 26th 1964), fig. 5.

16 Negahban, " Men and Beasts ", fig. 2.

17 Ibid., figs. 3 and 4- 8 Ibid., figs. io and 15-.

19 R. Ghirshman, Persia from the Origins, pp. 24-5 and pl. 24. For the connections between the pottery of IHasanlf IV and the "genre Liristan " wares of the Manors, cf. Goff Meade, " Liristdn ", Iran VI, pp. 124-5-

20 T. C. Young, in Iran VI, p. 161. 21 M. Rad and A. Hakemi, " The Description and Results of the

Scientific Excavations of IHasanlfi, ' Solduz ' ", Guzarish-yi bdstdn shinds I (1950). Translation of relevant passage by Parvin Bdrzin in R. H. Dyson, Jr., " Problems of Protohistoric Iran as seen from IHasanlfi ", JNES XXIV (1965), pp. 208-9; cf. also R. Ghirshman, Persia from the Origins, p. 99, fig. 131.

22 Goff Meade, " Lfiristan ", Iran VI, p. 125-


lb.ICb. EDoE PIT? BOWL

"/ 7

L. 4IND.K

LAMP

6 L.HIND.

ANGLE of HEAD FROM 51DE.

2.PAV INGs.

51T

STU DS

12. FEMU1. 05A

.G. SCPULAD.

C.c. i C. L.

Fig. 6. Horse burial in H-L baulk, Central Mound: (a) in process of excavation; (b) with upper bones removed showing position of harness.


Fig. 7. Grave goods from the horse burial on the Central Mound.


Nimrtid dated by Mallowan to the first half of the eighth century,23 and to rather similar early seventh century examples from the chariot burials at Salamis.24 These ivory nose pieces are larger and more nearly triangular than the Bdbd Jan example but the Nimrtid ones have a similar row of holes at the broader end, presumably for attachment to a leather backing. Unpublished bronze examples from the early seventh century were also found at Salamis, together with frontlets of leather with a thin coating of gold from the eighth, and these may eventually give us an even closer parallel. It is noticeable that the Bdbd Jan nose plate was not accompanied by a set of blinkers or cheek pieces such as were found both in the Salamis graves and among the Nimrfid ivories.25 Bridle bosses also occur at Nimrid26 and at HasanlGi IV as well as in the undated horse burial of Rad and Hakemi's excavations27 but appear to be of a different type to the Bdbd Jdn ones. The bit is perhaps more significant. Although too corroded to allow us to see the details of its construction, it appears to have a twisted, jointed snaffle, circular rein rings, and separate cheek pieces. Whether these were pierced like the rather similar examples from Sialk, Necropolis B,28 or had small loops attached to the side corresponding to the blobs of corrosion remains uncertain. Finally, the bronze lamp has been compared to an example from Grave 30 at Assur which also contains a bowl inscribed with the name of Assur-taklaku.29 An official of that name was eponym for the year 805 B.c. in Assur, but, as the grave also contains an elbow fibula of a type which apparently first occurs in the seventh century and continues to the fifth, a slightly later dating for the tomb is indicated.30

The East Mound Our excavations this season have given us a much clearer idea of the buildings in Level vb, which

must be approximately contemporary with the " Manors ", and the slightly later " squatting levels " of va and iv.31 As can be seen from the site plan (Fig. I), the west end of the deep sounding, which in 1966 was excavated only as far as Level iv,32 was continued down to virgin soil, and then the whole 13 m. of this trench (A) extended 4 m. to the north (D, F). Meanwhile the original deep sounding was extended 5 m. to the south (C).

In the lowest level, vb (Fig. 8), the solid mud-brick wall, which in 1966 was discovered running along the northern edge of the deep sounding, can now be seen to form the partition between a large room and an open courtyard, both still imperfectly excavated. The stone structure in the south-east corner of the 1966 deep sounding,33 originally thought to be a phase of Level vb, can now be shown to rest directly on the Level vb floor and to belong to the earliest of the Level va rebuildings. The only original features of the courtyard, which had at least two floor levels of greenish tamped clay, were the stone paving in front of the doorway to the west of the trench, and three post holes, which do not seem to form a sufficiently coherent pattern to imply the existence of a roof.

The room to the north, dubbed the " Painted Chamber " on account of its colourful interior decoration, was entered through an elaborately plastered and buttressed doorway (P1. IVa). The floor level of the Chamber lay some 20 cm. lower than that of the courtyard, and on the inner side of the doorway was a step made by fixing a wooden beam-later burnt-over a layer of pebbles. To the east was a small clay door socket, and the area just beyond the entrance showed traces of charring as if a burnt wooden door had fallen inwards.

23 M. E. L. Mallowan, Nimrud and its Remains (London 2966), vol. II, p. 538 and fig. 458; and p. 582, fig. 549.

24 V. Karageorghis, " Tombs of Warrior Kings ", ILN (Decem- ber 2nd 1967), Arch. Sect. 2279, and pl. 8.

25 Mallowan, Nimrud, vol. I, p. 126 and p. 330, note 8; Kara- georgis, loc. cit., pl. 9.

26 Mallowan, Nimrud, vol. I, pp. 124-5, pl. 65. Shell buttons or studs made in the shape of a figure of eight or double palmette.

27 Dyson, JNES XXIV (1965), p. 209. Rad and Hakemi, loc. cit., fig. 3, opp. p. 95.

28 R. Ghirshman, Fouilles de Sialk pros de KIashan (Paris 1939), vol. II, pl. LVI: S 835, S 841, S 588.

29 A. Haller, Die Graber und Griifte von Assur (Berlin 1954), pp. Iog-0o, and Taf. 22b (lamp), c-e (bowl).

30 A. Haller, loc. cit., Taf. 22 i (ak). The writer is indebted to Miss K. Wright for first calling her attention to the Assur parallel, and to Professor Porada and 0. Muscarella for their help in dating it, and particularly for their comments on the fibula.

31 The sequence is described in Goff Meade, " Liristin ", Iran

VI, pp. 118-9; and Goff Meade, " BabA JYn ", Iran VI, PP. 157-8.

32 Goff Meade, " Lfiristan ", Iran VI, p. I2o and figs. 8 and 9.

33 Ibid., p. I 19, note 22 and p. 120, fig. 8, right.


990O4

97.67 97.45

50CV-ET

97.55

Op / PAINTED

COURTYA Q D. CHAMBER.

.ED PLASTE..

8-M04

M.

Fig. 8. Plan of Level vb East Mound.

Other traces of sudden conflagration and destruction could be seen when we came to excavate the " Painted Chamber " itself. The north face of the wall showed traces of red plaster which extended only a few centimetres up the face of the wall before ending in a line of burning. This line seemed to coincide with a layer of burnt bricky material that may be interpreted as a patch of fallen roofing clay. The most striking discovery in this room, however, was the fragments of up to forty-three complete or nearly complete painted tiles of a type so far unique in the Ancient Orient.34 These tiles were found either standing on edge, or tilting with the north side up, south side down, indicating that they had fallen from a higher level-the top of the walls, ceiling or upper storey. Fragments of the plaster " backing " with which they must have been attached to the building can be seen very clearly in section.

4 Glazed tiles are found in Elam at Chogha Zanbil from the reign of Untash GAL (c. 1265-45) and at Susa from that of Shutruk Nahunte (c. 1207-1171); E. Porada, Ancient Irah (London 1965), p. 116. They also occur in the Iron II period at IHasanlfi IV, ibid., pp. 68-9; R. H. Dyson, " Problems of Proto-historic Iran as seen from IHasanlii ", JNES XXIV (1965), p. 199 and note ii. In this case, however, the tiles follow the Mesopotamian models in having a hollow centre

which fits over a wooden peg, and examples from Elam seem to have been attached in a similar fashion. The Babdt Jan tiles seem to represent an independent local development, though the pottery designs from Sialk, Necropolis B which parallel them closely, suggest that similar methods of decoration could be found elsewhere in the Zagros at this period did one but know where to look.


Otherwise, as with the manors on the Central Mound, there was very little in the way of pottery or any other finds on the floor at all, suggesting that the Chamber and the courtyard had been cleared out and abandoned before the conflagration.

The Tiles The tiles are made of buff clay with large crushed rock inclusions. The underside is left rough, but

the outer surface is self slipped and hard smoothed. The paint, as on the pottery, is usually brick red, but darker overfired examples also occur. Their dimensions are fairly standard. With one or two exceptions the length falls between 46-3 and 47'3 cm., the width between 41 .o and 42-o cm. and the thickness averages 6 - o cm. with a range of 4* 8 to 7

. o cm. As can be seen from the sections (Fig. 9)

they are slightly concave with a hollow face. Most came from the floor of the " Painted Chamber " itself but a few fragments were found scattered throughout our excavations in this area.

The designs so far fall into three main groups. Most common are arrangements in which a central, approximately square panel, is divided by a cross into four quarters which are then filled in with subsidiary motifs.

In the " Standard " design, which occurs by far the most frequently, the central cross is white with a red centre, and the corners are filled by a simple arrangement of concentric red and white squares (P1. IVd). Also fairly common is the " Reverse " where the central cross is red, and the corner design made slightly more complex by the addition of an extra red rectangle (P1. IVc). In both examples the central cross is picked out with rows of dots. In the rarer and elaborate " Chequer Board " designs, the corners are filled in with a small panel of sixteen red and white squares, surrounded by one or more bands of dots (Fig. 9: 3).

In the second sub-group, the " Nine Squares ", the cruciform framework is abandoned in favour of one which divides the central rectangle into nine equal squares. Each of these is then filled by a solid or open rectangle surrounded by a band of dots (P1. IVb; Fig. 9: I); or, if our reconstruction in Fig. 9: 2 is correct, by small chequer boards.

Finally in the third sub-group, the rectangular framework is abandoned. A broad " St. Andrew's Cross " is painted across the centre, and the four triangles between the arms are filled by groups of three small squares with dotted borders. This group of designs, although the most interesting, are the least carefully executed of the three, and in certain cases the painter seems to have deliberately introduced irregularities and asymmetrical effects (P1. IVe, f; Fig. 9: 4)-very much as the modern tribal rug makers do today.

As for the derivation of these designs; the " Nine Squares " and " St. Andrew's Cross " so far have no exact parallels.35 The first group however, seem to have as their basis the second millennium " Kassite Cross " which also occurs on the pottery.36 Moreover, both the " Reverse " and " Chequer Board " arrangements have close parallels on the pottery of Sialk, Necropolis B,37 thus strengthening the connections between the two sites which was already becoming evident after the first season.38

Thus as the result of our second season we have further evidence that Bbab Jan was an important local centre in about the eighth century B.C., and that the inhabitants of first millennium Lfiristdn were not merely nomadic warriors but had a well developed and distinctive culture of which the notorious bronzes are but one aspect. That this was due to artistic inventiveness rather than isolation is shown by the connections with the rich contemporary cultures of Solduz, Marlik and

Kdshmn. It is

hoped that our third season of excavations will uncover more of the unique " Painted Chamber " and establish its relationship to the other largely untouched levels on the summit of the East Mound.

35 c" St. Andrew's Crosses " occur occasionally on both " genre

Lfiristin " and Sialk, Necropolis B pottery, but in neither case

is the parallel very close. A. Stein, Old Routes in Western Iran

(London 194o), pl. XVII: io; R. Ghirshman, Fouilles de

Sialk, vol. II, pls. LXXI, A and LXXX, A.

38 Goff Meade, " Lfristdn ", Iran VI, p. i 18 and figs. 7: 1-2 and 6: i6.

3"Reverse", cf. Ghirshman, Sialk II, pl. LXIV, S 872; " Chequer Board ", idem, pl. LIV, S 811 and S 813; " Stand- ard ", idem, pl. LXVII, S 701.

38 Goff Meade, " Lfiristan ", Iran VI, p. 125-


Fiz. 9. Wall tiles from East Mound.


CATALOGUE Fig. 7

I. Bronze bowl; badly damaged in very fragile condition. Diameter: 21 .o

cm.; Height : 7 o cm. BJ/67/86. 2. Bronze lamp in two parts; saucer with straight sides and hollow, circular support; trefoil oil tray. Octa-

gonal sectioned handle divides at the ends and is attached to the two parts by rivets. Height: 14.0 cm.; Diameter: 20o ocm. BJ/67/87.

3. Bronze nose plate with holes at edge for attachment. Length: 26-0 cm.; Width: 15'7 cm. BJ/67/88. 4. Bronze harness stud with knobbed boss, and loop at back for attachment. Diameter:

4"2 cm. BJ/67/89.

5. Half of iron bit with separate cheek piece. Circular rein ring, and possibly twisted snaffle. Very corroded, fragmentary condition; drawing made from stains in earth. BJ/67/230o.

Fig. 9 I. Tile No. 12 from " Painted Chamber ", East Mound, Level vb. Complete but for small chip off corner.

Coarse buff clay, cream slip, red paint. One corner faded. Length: 47'2 cm.; Width: 41'5 cm.; Breadth: 6o cm. BJ/67/I6o.

2. Tile fragment from East Mound, erosion deposit below Level i in Trench F. Buff with large grits, surface smoother, paint red. Slight trace of burning on " flat " point. Reconstruction tentative on basis of other " nine square " designs. BJ/67/I42.

3. Tile No. 19 from " Painted Chamber ". Made up from nine fragments with one corner missing. Very coarse, light brown clay, buff slip and red painted decoration. Paint very poor, slightly vitrified, and missing on what remains of broken corner. Length: 46.-5 cm.; Width: 41 -2 cm.; Breadth:

5"5 cm.

BJ/67/I66. 4. Tile No. 34 from " Painted Chamber ". Made up from ten fragments. Coarse, light buff clay with black

and buff grits, cream slip on smoothed surface, paint varies from red to dark blackish brown. Drawing after reconstruction in museum had plastered joins. Length: 45 cm.; Width: 42 cm.; Breadth: 5 -8 cm. BJ/67/I84.

131

PREHISTORIC COPPER AND BRONZE METALLURGY IN WESTERN IRAN (WITH SPECIAL REFERENCE TO LURISTAN)

By P. R. S. Moorey

The prime purpose of this paper is to present as concisely as possible the results of a series of spectro- scopic analyses of " Liristdn Bronzes " recently undertaken in OxfordI in preparation for a catalogue raisonne' of the collection of Ancient West Iranian Bronzes in the Ashmolean Museum, and to integrate the information they yielded with other more scattered evidence for the history of metalworking in the

region from its earliest appearance until the seventh century B.C. In the last decade work both in the field2 and in the laboratory3 has contributed significantly to a better understanding of the metal industries of western Iran and their place in the cultural history of the area. Although the bronzes from

Liristan are no longer the isolated and enigmatic phenomenon they used to be, they still constitute one of the most remarkable aspects of Iranian achaeology. Following a pattern recurrent in archaeology information on metal working from settlement sites in the area is still very sparse, despite intensified

exploration in the last ten years. Virtually all the available evidence comes without certain context from plundered cemeteries of cist graves varying considerably in date; exceptionally bronzes seem to have been deposited, again in quantity, in shrines.4 Systematic study of these disjecta membra of Liristan's bronze industry is a salvage operation, at all times subject to the difficulties and uncertainties of such an

undertaking. The following essay attempts to provide a framework round which the available evidence

may be assembled; it will need to be constantly dismantled and re-assembled as fresh evidence adds new dimensions to the structure and corrects false perspectives.5

Danish excavations in Hulailan,6 Vanden Berghe's work in Ilam,7 French excavations at Tepes Giyan and Jamshidis and American excavations at Kamtarlan and Chigha-Sabz9 indicate, as is only to be expected, the existence of a local metal industry throughout Lfiristdn from at least the later fourth millennium B.C. So long as a relatively restricted production of simple personal ornaments, tools, weapons and sheetmetal vessels, primarily undecorated, was the total extent of this industry, Liiristan was no different from many areas of south-west Asia in antiquity within reasonable reach of the neces-

sary raw materials. It is the comparatively sudden appearance of an immensely varied and richly decorated range of bronzework in great quantity which defines a distinct cultural phenomenon, restricted both in time and distribution. This type of bronzework, closely associated with the earliest

production of iron in the region, is distinguished by a particularly elaborate and often highly conventionalized form of cast bronze decoration in which animals play the main part, but sub-human

1I am most grateful to Mr. R. W. Hamilton, Keeper of the Ashmolean Museum, for supporting this project, to Dr. E. T. Hall, Director of the Research Laboratory for Archaeology and the History of Art in Oxford, for accepting the programme, and to Mrs. A. Millett of the Laboratory both for the care and energy with which she executed it and for her interest in the interpretation of the results; for a preliminary report see Archaeometry VII (1964), PP. 72 ff.

2 See particularly R. Dyson, The Archaeological Evidence of the Second Millennium B.C. on the Persian Plateau (I968) (hereafter Persian Plateau); JNES XXIV (1965), pp. 193 ff.; T. Cuyler Young, Jr., Iran III (1965), PP. 53 ff.; Iran V (1967), pp. I1 fif.; C. Goff Meade, Iran VI (1968), pp. 105 ff.

3 J. Birmingham, Iran I (1963), pp. 71 ff.; R. Maxwell Hyslop and H. Hodges, Iraq XXIV (1962), pp. 126 ff.; XXVI (1964), pp. 50 ff.; XXVIII (1966), pp. 164 ff.; J. Birmingham, Iraq XXVI (1964), PP- 44 ff.

12

4 Most recently on the Dum Surkh Shrine see M. van Loon, Bib. Or. XXIV (1967), pp. 21 ff.; for another shrine see H. Thrane, Acta Archaeologica XXXV (1964), PP. 155 ff.

5 A fuller examination of the problems involved will be found in

my doctoral thesis-An Archaeological and Historical Investigation of the Liiristdn Bronzes (Oxford 1966), deposited in the Bodleian Library, Oxford.

6 J. Meldgaard, et al., Acta Archaeologica XXIV (1963), PP- 97 ff.; H. Thrane, Acta Archaeologica XXXV (1964), pp. I53 ff.

7 For interim reports on the work in Ilam see Trisors de l'Ancien Iran (Geneva 1966), pp. 23 ff. Phoenix XII, (1966), pp. 337 if.; Phoenix XIV (1968), pp. Iog ff.

8 G. Contenau and R. Ghirshman, Fouilles de Tpe' Giyan pris de

Ndhavand (Paris 1935)-

9 For a convenient summary, until the final report appears, see R. Dyson, Persian Plateau, pp. 13-4-


creatures-be they gods, demons or heroes-have an important subsidiary r61le in which the natural characteristics of man and beast are often inextricably conflated.10 It may be shown, despite an immense variety of detail arising generally from the fact that each casting by the " lost-wax " process is unique, that there is an essential homogeneity in all this decoration. It is most characteristically represented by the " standard-finials " with their mounts, certain elaborate open-cast pinheads (" wands "), whetstone sockets and decorated cheekpieces for horsebits. Technically distinct from these, but linked to them in many ways by style and iconography, is a sheetmetal tradition related to an industry in western Iran considerably more widespread than the cast bronze tradition of Liristan. Excavations in a shrine at Dum Surkh,al an important single stratified find from Bdbd Jnl'2 and comparative evidence both inside Iran and beyond indicate that this specialized metal industry flourished from the ninth to the later seventh century B.C., virtually contemporary with cemetery " B " at Tepe Sialk away to the east, which provides a valuable paradigm for the widely scattered metal artefacts from Liristan. Although the immediate antecedents and demise of this industry are still far from clearly understood, the main features of these changes may be discerned.

The History of West Iranian Metallurgy to c. Iooo B.C.

Native copper was used in Western Iran by at least the seventh millennium B.C., as also in Anatolia.13 A globular bead from Ali Kosh, dated c. 6750-6000 B.C., was made of cold-hammered native copper, cut with a chisel and rolled into shape.14 Simple hammered copper objects: pins, projectile points and a coil of metal were found at Sialk in Level I, of the early fifth millennium B.C.15 By Level II, c. 4500- 4000 B.C., there is evidence for annealing and the range and quality of surviving copper work is greater.16 The form of the spearheads has developed markedly and awls and personal ornaments appear. At Tall-i Iblis on the Iranian plateau, south-west of Kirmin, in settlement levels contemporary with Sialk II, evidence for copper working in a domestic setting has recently been revealed.17 Many fragments of baked clay crucibles showing copper stain and dross were found there among domestic debris dated by Carbon 14 to c. 4000 B.C.; a copper smelting pit is possibly earlier. The first clear evidence for cast copper is the distinguishing feature of substratum 4 of Level III at Sialk, late in the fifth or early in the fourth millennium B.C. At the same time metal is more commonly used for tools previously made in bone and stone. In the earliest phase of Sialk III were found needles, either with a pierced or folded head, " pins " with both ends slightly swollen and awls.18 In III, phase 2, pins are longer and the first tanged dagger blade appears.19 In III, phase 4, long pins with conical heads and a shaft-hole hoe mark a fundamental stage in the development of local metallurgy.20 In phase 5 the metalsmith's repertory has grown considerably to include more needles, and pins, chisels, flat axes, both hammered and cast in open moulds, and fine dagger blades with long tangs.21 The first sheetmetal vessel, a simple open copper bowl, appears in phase 7b, in the second half of the fourth millennium B.C.22

According to Le Breton Susa A, in the early fourth millennium B.C., yielded many flat axes, large and small chisels, pins and flat disk mirrors.23 In Susa B there are shaft-hole hoes like the one from Sialk III (4).24 At Tepe Giydn in LUristdn cast copper flat axes and awls were also in production by the

10 See E. Porada, Ancient Iran (1965), PP. 75 ff.; R. D. Barnett, Syria XLIII (1966), pp. 259 ff.

11 M. van Loon, Bib. Or. XXIV (1967), pp. 21 ff.; I use the name of the site given by the Danes, see note 6 above.

12 C. Goff Meade, Iran VI, p. 128, fig. 12; also the important grave-group from Tepe Tattfilban (Chindn), Ilam-Vanden Berghe, Phoenix XIV, pp. IO9 ff., which includes a standard- finial of the " master-of-animals " type with three faces on the central tube.

13 J. Mellaart, AS XIV (1964), p. 1 I4; H. Neuninger, Archaeologia Austriaca 35 (1964), PP. 98 ff.

14 F. Hole and K. V. Flannery, Proceedings of the Prehistoric Society XXXIII (1967), p. 177; see p. 191 for later metal.

15 R. Ghirshman, Fouilles de Sialk pris de Kashan, '933, 1934, 1937, (Paris 1938), I, pp. i6-7, pl. LII.49, 53-6, 58 (hereafter Sialk).

16 Sialk I, p. 30, pl. LII.46 48, 50-52, 57, 59.

17 J. R. Caldwell, Iran V (1967), p. 145.

18 Sialk I, pp. 53-4, pl. LXXXIV S.57, S.I643, S.I801, 1805, 1408, 402.

19 Ibid., pl. LXXXIV S.31 I, pl. LXXXV S.312. 20 Ibid., pl. LXXXIV S.168, 251; pl. XXIII.8. 21 Ibid., pl. LXXXIV S.178, 171, 183; pl. LXXXV S.I27;

pl. XXIII.6.

22 Ibid., pl. LXXXV S.I7I18. 23 Iraq XIX (I957), PP. 93-4, fig. 9.

24 J. Deshayes, Les Outils de Bronze de l'Indus au Danube (Paris 1960), I, pp. 231-2; II, nos. I842-3 (hereafter Outils).

PREHISTORIC COPPER AND BRONZE METALLURGY IN WESTERN IRAN 133

second half of the fourth millennium B.C.25 Away to the south-east at Tall-i-Bakun a comparable range of metalwork in the middle and later fourth millennium B.C. suggests that, sparse as the evidence is, the repertory of contemporary smiths is now reasonably well known.26

With the present uneven and erratic distribution of evidence it is extremely difficult to identify the areas which provided the stimulus for the major metallurgical developments of the later fourth and early third millennium B.C.: the increasing use of bivalve moulds for casting and the introduction of alloys, first of arsenical, then of tin, bronze. The appearance of worked copper at Qatal Hiiyiik soon after the middle of the seventh millennium B.C. and the copper shaft-hole mace-head from Can Hasan, dated before c. 4700 B.C.27 may mean that Anatolia's earliest metalworkers were as skilled and precocious as their better documented successors of the third millennium B.C. Recent discoveries in Israel have made clear that the repertory of objects and technical skill of smiths in the later fourth millennium B.C. were

very considerable, radically altering the conclusions previously drawn from the scant existing evidence.28 It must therefore be constantly borne in mind that for the history of metalworking, so ill-represented in most tell excavations, a single hoard or well equipped cemetery may transform understanding of the

subject. It is not yet known where the casting of shaft-hole tools and weapons in bivalve moulds was first

perfected. Deshayes has argued that this method of hafting was first discovered in Iran, citing the shaft- hole hoes from Sialk III and Susa B to support his hypothesis.29 Here again the nature of existing evidence is crucial. The early appearance of copper shaft-hole axes in Mesopotamia has to be deduced still from baked clay models with splayed blades which, as Childe pointed out, can only copy metal

patterns. These primarily belong to the 'Ubaid period, before about 3500 B.C.30 A thorough metal-

lurgical study of the Can Hasan mace-head would probably reveal whether or not it is the earliest known object cast in a bivalve mould. It is, of course, still possible that smiths in Iran were important pioneers in the development, but it is by no means yet certain that they initiated it.31

Analytical results from various areas in the Near East and the eastern Mediterranean region have made clear that an intentional arsenic-copper alloy was often, if not invariably, an important stage in the transition from cast copper to the use of a tin-copper alloy. Alloys of up to 8 per cent arsenic are

easily worked, hot or cold, and can give strength and hardness equivalent to tin-bronze, whilst the

presence of arsenic is very useful in maintaining the workability of the copper.32 This alloy was probably first employed in the Near East sometime in the second half of the fourth millennium B.C. and might well have arisen from an accidental use of an arsenic-enriched copper ore. In Israel arsenical bronzes

appear in the Nahal Mishmar hoard, late in the fourth millennium B.c., and among the objects in the Kafr Monash hoard dated to Early Bronze Age II, in the early third millennium B.C.33 In western

Anatolia, Cyprus, the Cyclades and in Crete arsenical-bronzes were current before tin-bronzes appeared in the middle and later third millennium B.C.34 Among objects from Horoztepe in Anatolia analyses showed that high quality tin-bronzes appeared concurrently with arsenical-bronzes in the third quarter of the third millennium B.C.35 At Ur the few metal objects analysed from the Jamdat Nasr levels, c.

3000 B.C., included objects with percentages of arsenic ranging from I 35 to o.

36. This is in marked

25 Giyan, pp. 43 ff. 26 R. Dyson in R. W. Ehrich, Chronologies in Old World Archaeology

(1965), p. 245; see this chapter also for the comparative chronology of Iranian sites in this period.

27 J. Mellaart, AS XIV (1964), pp. iIi ff.; AS X (I960), p. 103; D. French, AS XII (1962), p. 33-

28 P. Bar-Adon, Israel Exploration Journal XII (1962), pp. 215 ff. 29 Outils, pp. 231-2. 30 C. L. Woolley, Ur Excavations I, p. 205, pl. XLVI.2 T.38; IV,

p. 9, pl. i6; A. Parrot, Tello, p. 36, fig. 7r; S. Lloyd, JNES II (1943), pl. XVIB; E. Mackay, Report on Excavations at Jemdet Nasr, Iraq (1931), p. 266, pl. LXXV.7.

31 The solution of this problem may not be as simple as it appears. It has been shown, by detailed examination of early Danubian copper shaft-hole axes, that shaft-holes were made by drilling a

circular hole through a flat axe-head after casting in an open mould (H. H. Coghlan, Archaeologia Austriaca XXIX (1961), PP. 57 ff.). In other words, a technique appropriate to making shaft-hole axes in stone was initially applied to their earliest metal counterparts. I am grateful to Mr. H. J. Case for this reference.

32 For the metallurgy see J. Charles, AJA LXXI (1967), pp. 21 ff. with refs.

33 C. A. Key, Israel Exploration Journal XIII (1963), p. 289; Science (1964), pp. I578-80; de Vaux, Palestine: Neolithic and Chalcolithic Periods (CAH 1966), p. 32.

34 C. Renfrew, AJA LXXI (1967), pp. 14 ff.; G. Bucholz, Berliner Jahrbuch VII (1967), pp. 240 ff.; K. Branigan, Copper and Bronze Working in Early Bronze Age Crete (1968), pp. 46 ff.

35 T. Ozgi!?, Anatolia VIII (1964), table facing pl. II.


contrast to metal objects analysed from the " Royal Cemetery ", about 500 years later, in which there is little or no evidence of arsenic.36 Two objects of third millennium date from Tepe Gawra in Iraq contained o

. 67 and I 59 per cent of arsenic respectively.37 Evidence for arsenical-bronze in western Iran is still sparse, but certain. In Level K at Geoy Tepe,

west of Lake Urmia, arsenic appears in a fairly high proportion after a period, Level M, when copper was singularly pure.3s Level K is contemporary with Level II at Kjul'-Tepe in Soviet Azerbaijan which has a Carbon 14 date of 2920?90 B.c. Systematic analysis of the metal objects from the superimposed settlement levels at Kjul'-Tepe (I-IV), where moulds and industrial debris indicate local manufacture, has shown graphically the definite stages of transition from copper in Level I to an arsenical-bronze in Levels II-III and then a tin-bronze in Levels III-IV, from late in the fourth to early in the third millennium B.C.39 In the early third millennium B.c. arsenical-bronzes were widely used in Caucasia and Transcaucasia.40 They were probably obtained by melting together the copper and arsenic ores of the region.

Among the metal objects without context reported from Ltiristin are a few arsenical-bronzes. In

every case they are artefacts which may be dated typologically to the middle or later third millennium

B.c. An axe, closely paralleled at Ur and Susa,41 a dagger, both its hilt and blade, with close typological relatives at Kish and Ur,42 a pick-axe, paralleled in Mesopotamia,43 and an axe-head with zoomorphic decoration44 have arsenic percentages ranging from I 9 to

4"4. An axe analysed in Sydney, again of

mid-third millennium date, has 0 -20--0.50 per cent arsenic; another of similar date in Oxford has

2. I per cent tin, 2 -8 per cent arsenic.45 The few analyses of contemporary artefacts from Susa show that they were predominantly made of cast copper46. Evidence so scanty as this may easily distort the overall picture, but it now looks as if arsenical-bronzes were being produced in the second half of the third millennium B.C. in central west Iran and that they may have been produced for some time

simultaneously with cast copper and tin-bronze objects. It is quite possible that where arsenic ores were

easily available (its occurrence is much wider than that of tin)47 and tin was scarce, it survived in use as an alloy much longer or was used intermittently until well into the second millennium B.C. Charles has suggested that the main objection to its persistence is not so much a metallurgical as it is a medical

one; the extreme toxicity of most arsenic salts made the manufacture of arsenical-bronzes a hazardous

process for the smiths.48 Long before a copper-tin alloy came into general use implements appear with definite inclusions

of tin in the copper.49 These are usually described as " accidental " bronzes produced by using ores from a lode in which tin was associated with copper; a not unusual occurrence. Coghlan has defined a deliberate copper-tin alloy as one in which the tin exceeds approximately 3 per cent.50 Voce, on the other hand, believed as little as o 50 per cent tin was a deliberate addition in periods when contem-

porary bronzes had as much as 10 per cent tin in them, as in Level D at Geoy Tepe,51 in the earlier second millennium B.C. Junghans has also argued that a deliberate alloy may be as low as I per cent tin.52

What may certainly be identified as deliberate copper-tin alloys appear first in the Near East in the

36 C. L. Woolley, Ur Excavations II, pp. 284 ff.; IV, p. 164.

37 E. A. Speiser, Tepe Gawra I, pp. oI ff.; see also analyses of slightly later objects, P. Delougaz, The Temple Oval at Khafajah (1940) (OIP LIII), p. 152.

38 B. Burton Brown, Excavations in Azerbaijan, 1948, pp. 179 ff., I88 ff. with a discussion of arsenical bronzes.

39 I. R. Selimkhanov, Germania 44 (1966), pp. 221 ff.

40 I. R. Selimkhanov, Proceedings of the Prehistoric Society XXVIII (I1962), pp. 68 ff.; Proceedings of the Prehistoric Society XXX (1964), pp. 66 ff.

41 Analysis No. 22. 42 Analyses Nos. 23-4. 43 Analysis No. 54. " Analysis No. 38.

45 J. Birmingham, Iran I (1963), p. 75-48.247; Oxford Analysis No. 48.

46 Outils II, nos. 1204, 1843, 1856; also 1407 of slightly later date.

47 R. J. Forbes, Studies in Ancient Technology IX (1964), pp. i66 ff. 48 J. Charles, AJA LXXI (1967), p. 26.

49 E. Schmidt, Excavations at Tepe Hissar Damghan, 1937, p. 360- level Ib; recent analyses of Hissar objects suggest a virtual absence of tin-bronzes: R. Dyson, Persian Plateau, p. 5; see also p. I6 for absence of tin-bronze in IHasanlia VII (c. 2200- 2000 B.C.).

50 H. H. Coghlan, Notes on the Prehistoric Metallurgy of Copper ... (1952), P. 24.

51 Cited by B. Burton Brown, Excavations in Azerbaijan 1948, p. i92.

52 Metallanalysen Kupferzeitliche und Friihbronzeitlicher Bodenfunde aus Europa (1960), pp. 8 ff.


second quarter of the third millennium B.C. and then become steadily commoner. The earliest history of the alloy is still obscured in many areas by the dearth of analyses. An archaic text from Ur, of the

early third millennium B.C., differentiates between copper and bronze.53 Tin-bronzes appear in the " Royal Cemetery " at Ur54 and at Kish and Khafajah55 by about 2600 B.C.; evidence from Syria and Anatolia is so far slightly later.56 There seems little doubt that the alloy was only used in Egypt as a result of contacts with western Asia.57 In western Anatolia tin-bronzes appear in Troy II5s and at about the same time in the Cyclades.59 It is possible that the earliest tin-bronzes yet identified by analysis are those from Caucasia and Russian Azerbaijan.60

A number of good quality tin-bronzes were found in Level D at Geoy Tepe,61 in the earlier second millennium B.C. In LUristdn tin-bronze shaft-hole axes, pick-axes and adzes, were in production by the late third millennium B.C. or very soon afterwards.62 The tin percentages range from 3 -6 to

13" I

and two examples have comparatively high arsenic percentages."3 Evidence from elsewhere has shown that ancient smiths eventually aimed at a more or less standard alloy with about Io per cent tin; erratic variations are a natural feature of early tin-bronzes. By the later second millennium B.C. in

Ltiristdn, as elsewhere in western Iran, it is clear that local smiths had a keen appreciation of the

properties of this alloy and had achieved some skill in regulating it. The higher the percentage of tin in a copper-tin alloy, up to about 15 per cent, the harder and stronger is the resulting metal, without

increasing the difficulties of hot and cold working under the conditions prevalent in antiquity. Signifi- cantly, it is the flange-hilted dirks and their typological relatives, most widely used in western Iran in the later second millennium B.C., which most consistently show high percentages of tin, up to a maximum of about 15 per cent.

The history of metalworking in Lfiristdn from the middle of the third to the end of the second millennium B.C. may not yet be established in any detail, but its predominant characteristics are clear

already. Whereas in the earlier first millennium B.C. the region's metalwork was distinguished both by objects and decoration without known parallel elsewhere, in the previous period both the range of artefacts and their forms had a great deal in common with those produced in Elam, Mesopotamia and North Syria. Links with Elam were particularly close. For instance, with minor variations, all the main

types of shaft-hole axe occur both in Elam and in Liristdn with striking regularity; only Ur outside western Iran has anything like a parallel range of forms and that only includes half the types reported from Iran.64 Vanden Berghe's excavation of eleven cist graves of the third quarter of the third millen-

nium B.C. at Bani Surmah in the region of Ilam has demonstrated the extremely close affinity between the metalwork of this part of Liristdn and its neighbours to the west and south-west.65 Indeed the area seems to form the most easterly province of the material culture so richly illustrated by American excavations on sites in the Diyala valley,66 which also has many cultural links with Elam. At this time a very similar range of metal artefacts was in use in eastern Liiristan to judge by finds from Tepe Giyan and other unknown sites in the area. The quantity of metalwork in the graves at Bani Surmah certainly suggests local manufacture and that parity of form between artefacts from controlled excavations in Elam and Mesopotamia, and unprovenanced finds from Liiristdn, does not necessarily mean that the

53 E. Burrows, Ur Excavations: Texts II, p. II re. no. 373.

54 C. L. Woolley, Ur Excavations II, pp. 284 ff.

55 H. Peake, Antiquity II (1928), p. 452; P. Delougaz, O.I.P. LIII, pp. 15I ff. see also H. Limet, Le Travail du Mital au Pays de Sumer (1960), pp. 58 ff.

56 For example R. J. Braidwood, Journal of Chemical Education XXVIII (1951), pp. 87 ff.; C. Schaeffer, JEA XXXI (1945), pp. 92 ff.; T. Ozguia, Anatolia VIII (1964), table facing pl. II.

57 A. Lucas, Ancient Egyptian Materials and Industries (4th edn., 1962), pp. 217 ff.

58 K. Bittel, Jahrbuch des Deutschen Archiiologischen Instituts LXXIV (1959), pp. I ff.

59 C. Renfrew, AJA LXXI (1967), pp. 12 ff.

s0 I. R. Selimkhanov, Proceedings of the Prehistoric Society XXVIII

(1962), p. 78, nos. 16-18; Germania XLIV (1966), pp. 221 ff. A possible exception is a late fourth millennium object from Teleilat Ghassul in the Jordan valley-A. Mallon, Teleilat Ghassul I (1934), PP. 75 ff.-which may be a very early intentional alloy of tin and copper; there is a similar problem with the Gawra analyses, see note 37.

61 B. Burton Brown, Excavations in Azerbaian 1948, p. 193- 62 Oxford Analyses Nos. 39-41, 53, 55-6. 63 Oxford Analyses Nos. 40o-.

64 Outils II, map X.

s5 Phoenix 14 (1968), pp. Io9 ff., fig. 38. 66 See particularly the comments of P. Delougaz, Pottery from the

Diyala Region (1952), PP. 125 ff.; also Le Breton, Iraq XIX

(I957), PP. 79 ff.


latter were imports from the better known industries of the south and west. Such cultural affinities are the natural complement to the close political contacts between all these regions which are clear from

surviving inscriptions of the later third and early second millennium B.C.67 It is probable that Elam was the original home of the cast zoomorphic decoration on tools and

weapons which was so characteristic of the west Iranian metalworking tradition, but comparatively rare elsewhere.68 The earliest example of certain date from Susa is on a hammer inscribed for Shulgi of the Third Dynasty of Ur (c. 2095-2048 B.C.).69 As Shulgi's activities in Elam are well documented this

object may be taken, as its form suggests, for an Elamite rather than a Sumerian artefact. A number of other objects from Susa70 demonstrate that by the first quarter of the second millennium B.c., if not

earlier, a tradition of zoomorphic decoration on tools and weapons, possibly all votive or parade pieces rather than serviceable objects, was well established in Elam and slowly diffused northwards. In

Liiristan a series of adzes with crudely cast zoomorphic decoration on their butts and a series of " maces " with relief decoration follow Elamite fashions.7 An axe decorated with a simple lion

passant on the butt, reported to be from Lfristan, belongs to this period.72 Analysis showed that it was an arsenical-bronze with an unusually high percentage of lead--3.6-which was almost certainly added intentionally to facilitate casting the animal.

During the earlier second millennium B.c. at a time when links between the Zagros region, Assyria and Syria are also clear from the distribution of the " Khabur " and cognate styles of painted pottery73 identical tool and weapon types indicate a close relationship between the metal industries of the same

regions. Deshayes argued that this might be correlated with the diffusion of Hurrian-speaking metal- smiths.74 Such a comprehensive historical hypothesis is seductive, but, apart from the hazards of

associating a particular set of artefacts with any given people or state, its chronological and topo- graphical basis needs examination before it may be accepted. de Vaux, in an excellent concise review of Hurrian history,75 has made clear the long period of time, and very wide area, over which Hurrian

activity may be recognized in the Near East from at least the Dynasty of Akkad. The cultural settings in which they may now be traced are many and various. Their presence at Susa is virtually certain, but the extent of their influence on Elamite culture is still obscure.76 It would seem too early to estimate whether the occurrence of an identical type of tool or weapon in North Syria and Lfiristtn indicates a flow of craftsmen, perhaps Hurrian, westwards or vice versa; if indeed it involves migrating peoples at all. It is arguable that two other factors, apart from direct exchange of objects through trade, regulate the diffusion of metal artefacts and, when personal contact is involved, of techniques: the activities of armourers in invading armies and the deportation of captured smiths, neither necessarily linked to a

specific linguistic or ethnic group.77 The extent to which trade in copper and tin from Iran, along the northern route westwards into Assyria and Syria, brought closer contacts between the smiths of the regions through which it passed in the second millennium B.C. is virtually impossible to gauge accurately, but it may have been more significant than is yet allowed.

Throughout the second millennium B.C. the political power of Elam and the remarkable vigour of its metal industries stifled the development of large-scale independent production in Lfiristin, though her smiths continued to provide for local needs without displaying much originality in the forms and decoration of their metalwork. Political changes towards the end of the millennium brought fresh

opportunities. Throughout Elamite relations with Mesopotamia, of which she was in fact geographically a part, were always close, if periodically stormy, and both regions drew upon the other's resources

67 W. Hinz, Persia, c. 2400-18oo B.C. (i963), passim (New CAH fascicule).

6s For Sumer, see A. Parrot, Tello, p. o6, fig. 26b; for Syria, see Th. Dunand, Til Barsib, p. io6, pl. XXVIII.5; for Assyria, H. G. Giiterbock, Landsberger Festschrift (1965), pl. XIII-XV; eastern Iran, R. Maxwell Hyslop, Iraq XVII (p955), p. i6I.

69 de Mecquenem, RA XLVII (1953), PP. 79 ff., fig. 2.4a. 70 de Mecquenem, MDP XXV (934), p. 210o, fig. 53, fig. 58.4;

see also MDP VII (1905), pl. XVII.8; RA XLVII (0953), p. 81, fig. 2.5; see also the weapon on a seal impression from Susa, P. Amiet, Elam, p. 258, fig. i57.

71 L. Budde, Pantheon XXI (1963), PP-. 99 ff. 72 Oxford Analysis No. 38-Ashmolean 1951.156, for type cf.

Outils, p. I66-A5c. 73 B. Hrouda, Die Bemalte Keramik, p. 41-

74 Outils I, pp. 426 ff.

75 Revue Biblique (1967), PP- 481 ff. 76 R. Labat, Elam, c. 16oo-z2oo B.C., p. 5 (New CAH fascicule).

77 For deportation of smiths see for instance: D. D. Luckenbill, Ancient Records (1926) II, 234; 2 Ki. 24.14.


through trade and diplomacy. At the end of the twelfth century the political balance was radically altered. In the last decade of this century Nebuchadnezzar I (c. 1124-1 o3 B.C.) of Babylon crippled the political power and influence of Elam with a crushing victory over Khutelutash-in-Shushinak, whereafter he ravaged Elam, including Susa.78 His successors on the throne of Babylon for the next century or more continued the subjection of Elam and its closest dependencies to the north. Susa was greatly impoverished and very poorly equipped tombs overlay the sanctuary of Inshushinak in the early first millennium B.C.79 For about three centuries Elam disappears from Mesopotamian sources and there are no native texts.80

Such a major catastrophe must have had important repercussions among the mountain peoples to the north of Elam who, in varying degrees, had long fallen under her influence. The most tangible evidence from this region for an increased Babylonian presence from the later twelfth to the middle of the tenth century B.C. are the various inscribed bronze weapons reported from

Liiristmn. By far the

greater number of these are inscribed for Babylonian monarchs of the Second Isin and immediately subsequent dynasties or their subjects. The earliest inscribed dagger is for a late Kassite ruler, Adad-

suma-usur (c. 1218-1189 B.C.)81 with at least one for every ruler from Ninurta-nadin-Sumi to Adad- apla-iddina (c. I 130-Io46 B.C.). It is curious that after this period apparently, with the possible exception of Nabu-mukin-apli (c. 977-942 B.c.), daggers are replaced by arrowheads, which again predomi- nate for one short period from Eulmas-sukin-vumi to Ninurta-kudurri-usur II (c. 1003-942 B.C.).82 It will be interesting to see whether further examples confirm and broaden this pattern.

Although the paucity of comparative material from Elam severely restricts the argument there is sufficient evidence to demonstrate a considerable Elamite legacy in the independent cast and sheet bronze traditions of Lfiristan from the tenth century B.C., particularly its iconography.83 By contrast there is virtually no evidence before the early seventh century B.C. for major northern influences in Liristan, either in ceramics or metalwork;84 what intrusions there were, and their source is still not clear, blended with the existing local cultures85 to a large extent. With the collapse of the Elamite metal industry and the destruction of Elamite political power after the Babylonian devastation it is reasonable to assume that Elamite smiths fled northwards for refuge in the mountains, taking the skills and styles of their homeland with them to invigorate the bronze industry of LSiristan, where the vital raw materials: metals, fuel and wax were readily available. By the late tenth century B.C., no longer overshadowed by the great Elamite metal industries and temporarily freed from direct Elamite or

Babylonian incursions, the smiths of Loiristan were free to develop a style congenial to their local

patrons' way of life and religious beliefs, incorporating the skills and iconography of Elam, and to a lesser extent of Mesopotamia, which is so clear in much of their decorated bronzework. If the patrons were newcomers, then they very readily accepted indigenous traditions; in the absence of documentary evidence it is best to accept their anonymity for the moment. The eclipse of this industry in the later seventh century is probably to be associated with the increasing political unity and expansion of the Medes and their associates, who ousted its patrons, cut off the lines of supply for metal from the north and east, and supported other major metal industries to the north of Liiristan.

THE LCRISTAN INDUSTRY, c. 900-650 B.C.

Organization and Patronage It is probably idle to speculate about the organization of an industry which has left no written

records, but certain statements are commonly made about the L-rist^n bronze industry at its most prolific which require qualification even with the evidence to hand. It is sometimes argued that

78 R. Labat, Elam and West Persia, c. r200oo-ooo B.C., pp. I ff.

79 de Mecquenem, RA, XIX (1922), pp. 13I if., 139. 80 R. Labat, Elam and West Persia, c. I2oo-iooo B.C., p. 24. 81 G. Dossin, Iranica II (1962), pp. I49 ff., no. I; for date see W.

Nagel, Berliner Jahrbuch IV (1964), pp. 153-4- 82 For these weapons see G. Dossin, Iranica II (1962), pp. I49 ff.;

W. Nagel, AfO XIX (1959-60), pp. 95 ff. It will be clear from

the suggestion made here that I endorse the criticisms of Ghirshman's " Cimmerian Hypothesis "; see M. van Loon, Bib. Or. XXIV (1967), p. 22; C. G. Meade, Iran VI (1968), pp. 130 ff.

83 E. Porada, Ancient Iran, pp. 70 ff. 84 R. Dyson, JNES XXIV (1965), p. 200 ff.

85 C. G. Meade, Iran VI (1968), pp. 126 ff.


Liiristdn's industry served a people who were wholly nomadic, because of the prevalence of horse-gear and portable objects."6 On the face of it it seems unlikely that a people with such a way of life-the

Scythian repertory in metal is nothing like so rich and complex-would require, let alone support over a

long period of time, a production of bronzework as varied as that of Liiristdn. That some of the patrons of this industry led a nomadic existence is virtually certain, since seasonal transhumance has long been a familiar pattern of life among the region's pastoralists;87 that they all did seems unlikely. Most metal industries in antiquity, particularly those producing very elaborate artefacts, were supported by a small

minority representing the ruling group.88 On the analogy of cemetery " B " at Sialk the owners of the fine bronzework from Liristan are to be seen as members of a warrior aristocracy, distinguished, as so often in history, by the possession of a horse and the wherewithal to equip it and themselves for battle.

They may be seen as the ruling members of the larger, permanent settlements in the lower western

plains or citadels in the higher eastern plains89 dependent on agriculture, horse-breeding, and perhaps control of the north-south trade route, for their prosperity.

The great complexity and quantity of the bronze artefacts made in Liristdn between the late tenth and early seventh centuries B.C. and the pioneering iron work associated with them, indicates high skill and specialization. It is unlikely to have been the creation of nomadic smiths or tinkers, though some

may have found a place in it. Indeed the activity of such people in the ancient Near East is still obscure;90 where documentary evidence is available metal industries were normally centralized and under careful government supervision.91 The activities of modern travelling tinkers in the area suggest that ancient tinkers belonged primarily to areas where truly nomadic communities were most prevalent, not to areas of settlement, and that they specialized in the production and repair of simple everyday equipment rather than in elaborate and sophisticated metalwork.92 For the moment it may be surmised that the " Liristdn " smiths formed distinct groups in the urban centres of Hulailan, Tarhan and southern Kurdistan.

Sources of Ore and Fuel It must be made clear at the outset that the extent to which analyses can contribute to the identifica-

tion of the ore sources exploited in antiquity is very contentious. Apart from the methodological problems involved93 such work depends initially on detailed and scrupulous investigation of the native ores, in this case not only of Iran, but of the Near East as a whole. These are not available. For the moment the ancient documentary sources, the location of known deposits of various ores and their geographical proximity to the main centres of production for any particular industry must be carefully correlated to

provide answers to this vital preliminary question. As Iran is so rich in the ores vital to a thriving bronze industry it is hardly likely that her local industries were dependent on anything but local

supplies. Even more significant for a country in the Near East, where deposits of the ore are scarce, is the occurrence of cassiterite, the only important ore of tin, in close proximity to all the major areas which archaeological evidence has shown were major centres of bronze-working in antiquity: Kara

Dagh, west of the Caspian, on the southern slopes of the Elburz near Astardbad and Shahrfid, and

probably in the Ktih-i-Bendn in central Iran.94 In the absence of physical and political barriers it may reasonably be assumed that the most accessible source was that most commonly exploited. For Liristan this was the area west of the Caspian where both tin and copper occur, not to mention iron. It was an area with a long established and flourishing metal industry of its own. It would be unwise to ignore completely the possibility that a certain amount of tin, perhaps also copper, worked to the east of Elam,

86 Notably H. Frankfort, The Art and Architecture of the Ancient Orient (1958), pp. 207 ff.

87 A. Godard, L'Art de l'Iran (I962), pp. 86 ff.; J. Meldgaard, et al., Acta Archaeologia XXXIV (1963), PP- 97 if.

88 G. Clark, Archaeology and Society (1957), p. 226. 88 C. G. Meade, Iran VI (1968), pp. 126 ff.

90 R. J. Forbes, Studies in Ancient Technology VIII (1964), PP- 91 if. 91 H. Limet, Le Travail du Mital au Pays de Sumer, pp. 165 if.; de

Vaux, Ancient Israel (1965), pp. 77-8.

92 H. R. P. Dickson, The Arab of the Desert (1959), PP. 515 ff.; F.

Barth, Nomads of South Persia (1965), p. 92; see also Cuyler Young, Jr., Iran V (1967), P. 30, note 89.

93 See comments in H. Hodges, Artifacts (1964), pp. 221 ff.; H. W. Catling, Cypriot Bronzework in the Mycenaean World

(1964), PP- 9 if-

94 Compare the maps in R. J. Forbes, Studies in Ancient Technology IX (1964), figs. 3 and 22, pp. Io ff., 130 ff.


was re-exported northwards from Susa. There is a little scattered documentary evidence to suggest that Susa served as a centre for the westward export of tin to Mesopotamia and beyond, regions where cassiterite deposits were virtually unknown;95 but whether the ultimate source of the tin was to the north or east of Susa is obscure. In the earlier second millennium the earliest evidence appears for a trade in tin, probably from the mines east of Tabriz, through the Zagros to Assur and thence south and west.96

Although the distance between Liristdn and the northern mines was considerable, it was not so great as that along which metals travelled to Assur from eastern Anatolia in the early second millennium B.C. or booty, often including very considerable quantities of metal, was transported by later Assyrian monarchs.9" Almost certainly, for ease of transport by baggage animals or in carts, the metal would have travelled as ingots cast to standard shapes and sizes. Documentary evidence from Sumer and the Cappadocian trade"98 indicates that metals were usually smelted as close to the mines as possible, where combustible material was generally available. This is largely confirmed by fieldwork in two great mining areas of antiquity in the Near East: the Arabah of Israel and the Caucasus.99 That this was not invariably the case may be seen at Mari, where smelting of selected ores certainly took place;100 but practical considerations, particularly the availability of fuel, must have reduced such activities to a minimum.

Another source of raw material for the local bronze founders of Lfristan was provided by broken, worn out or unwanted objects, as well as faulty castings, sent back for reuse. A number of objects which may fall in this category were found by the Danish excavators at Tang-i Hamamlan.11o Bronze founders' hoards have rarely been reported from Near Eastern countries, but a number from Cyprus have been fully published.102 It is of some interest to notice that the contents of the Late Bronze Age wreck off Cape Gelidonya on the south coast of Anatolia suggest that scrap metal was, in certain circumstances, transported over considerable distances as readily as new ingots.103

In antiquity the sole means of obtaining the necessary heat for smelting or melting, casting and soldering the ore was the charcoal fire.104 Bitumen, or other inflammable materials, would not have produced a suitable flame, though for quite small works an oil lamp would have sufficed. Although the problem has yet to be scientifically investigated certain comments in Assyrian records of West Iranian campaigns and the pattern of ancient settlement in Liristin suggest that it was once more forested than it is today, allowing ample scope for charcoal production.

The Cast Bronze Tradition: Techniques and Artefacts In the course of excavating the graves in the " Donjon " at Susa de Mecquenem found a variety of

stone moulds, normally bivalve. They were primarily for casting arrowheads, spearheads, dagger- blades and pins.105 Only two were published with illustrations,106 one for casting two arrowheads set side by side with a pointed awl between, the other for two arrowheads cast with a common tang and later separated. Their date will only be established when it is possible to examine in more detail the type of weapon cast, since the context covers much of the second millennium B.c. and later.

95 J. Bottero, Archives Royales de Mari VII (1957), PP- 293 ff.; W. F. Leemans, Foreign Trade in the Old Babylonian Period (1960), pp. 123 ff. 96 For references see B. Landsberger, JNES XXIV (1965), pp. 285 ff.; M. T. Larsen, Old Assyrian Caravan Procedures (1967), p. 4 with note; see tin in Tell Rimah texts, S. Page, Iraq XXX (1968), pp. 90 and 96; P. Garelli, Les Assyriens en Cappadoce (1963), pp. 265 ff.

7 D. Luckenbill, Ancient Records I, nos. 222-3, 412, 443, 447, 454, 456, 460, 475-6, 585, 589, 6o0; II, no. 172.

9s H. Limet, Le Travail du Mital, p. 84; P. Garelli, op. cit., passim.

99 N. Glueck, The Other Side of Jordan (1945), ch. III; Bib. Arch. (1965), PP. 70 ff.; B. Rothenburg, Palestine Exploration

Quarterly (1962), pp. 5 ff.; Jessen, History of Metallurgy in the Caucasus [in Russian] (I934), ch. IV(2).

100 A. Parrot, Syria XX (1939), pp. 14 ff.; H. Limet, Le Travail du Mital, p. 84, notes 5 and 6.

101 H. Thrane, Acta Archaeologia XXXV (1964), p. 159, fig. 5. 102 H. W. Catling, Cypriot Bronzework in the Mycenaean World

(1964), ch. XII-XIII. 103 G. Bass, Cape Gelidonya: A Bronze Age Shipwreck (1967). 104 In exceptional circumstances where coal outcropped naturally

it may have been used. 105 MDP XXV, p. 227. 106 MDP XXIX, p. 135, fig. IOO-I.


.Hasanli is the only site in western Iran from which industrial equipment of the early first millen-

nium B.C. has yet been published. During his soundings on the site Stein'07 found a large pot of partly burnished red ware which contained a variety of bronze objects: a bundle of twenty-nine roughly cast bronze rods, some corroded together, which had been cast in open moulds; eight flat bronze strips, 8 to 12 in. in length; and five small pieces of copper I A in. in length. Close by the vessel were a number of moulds. A large stone-io- x 4 X 21 in.-had moulds for " broad-bladed chisel-like implements " on the wide faces and for ingots on each narrow face.108 Fragments of two stone moulds109 had apertures for ingots on their narrow faces and one may have been a bivalve mould for casting a snaffle-bit, similar to such moulds from south Russia.110 The find is important in illustrating the simultaneous use of open and closed moulds at this time. Analyses of metal objects from Loristan and Dailamn'111 have shown that copper was still in use in the earlier first millennium B.C. for cast objects. Copper is not easy to cast in closed-moulds, but readily dealt with in open moulds and they were clearly retained for casting simple tools and weapons. The method is simple, involves the minimum of labour to make flat axes, sickles, knives, dagger-blades, chisels, etc., and does not demand great skill or experience. It is also relatively fast.'12 The addition of as little as 2 to 3 per cent tin to copper, however, makes it less gaseous and increases the metal's liquidity, which makes casting it in closed moulds easier with a standard tin-bronze of the quality current in Liiristin by the late second millennium B.C.

The flange-hilted dirks and daggers from Liristan and elsewhere in western Iran, all of good quality, were cast in simple bivalve moulds to a variety of standard patterns. When the hilt was cast solid in imitation of an ordinary flanged hilt with inlay plates in place a different technique was often used in which the blade was made separately first and the hilt then cast onto it. North-west Iranian

weapons of this type, sometimes also reported from Ltiristan, have been metallurgically examined in

Sydney,113 with important results. The blade and tang unit were made first, forged to shape by severe

hammering at high temperature. The tang was covered with clayey material, in some cases it survives,114 which when dry formed a core for the casting. In this way a better balanced weapon would be made than those with a solid hilt. A hollow mould, the exact shape of the shoulder, grooved hilt and pommel, was then formed round the clay core covering the tang. Into this the liquid metal was poured. The

complexity of some pommels suggests that the lost-wax process was used to make at least part of the mould. Light hammer blows around the shoulder and the pommel finished off the hilt; at the same time the blade was finished with repeated hammer blows to work-harden and finish the cutting edge. In some cases sheet gold or silver was used to decorate the hilt,115 more commonly linear decoration was traced after casting with hand-held punches or tracers. In the case of the dirk examined in Sydney the hilt and the blade had different compositions; this discrepancy may indicate either a different

origin for the base metals or perhaps independent manufacture of the blade and hilt units. That this is not always the case is shown by the analyses of typologically related weapons tested in Oxford.116

No detailed metallurgical examination of a Liiristan axe-head of the earlier first millennium B.C. is

yet available, but certain points are clear from surface inspection. Most of the standard axe types were cast in bivalve moulds, probably of stone or baked clay.117 For those with elaborate plastic decoration the moulds, if only in part, must have been made by the lost-wax process, rather than from actual

examples or wooden patterns. Outstanding among the axes reported from the region decorated with three-dimensional zoomorphic motifs are some rare examples with the blade cast as an openwork design of animals.s18 Non-utilitarian or votive axes from the region are common; many apparently practical forms have never had their blades ground down for use. Metallurgical examination of some axe-heads

107 Old Routes of Western Iran, pp. 394 f. 10s Ibid., pl. XXVI.3-6. 109 Ibid., pl. XXVI.23. 110 Ibid., pl. XXVI.20; cf. M. A. Sabeyrova, Sov. Arkh. I (1964),

PP. 304 ff., figs. 1-3- 11 Oxford Analyses Nos. 2, 67, 1oo; N. Egami, Dailaman I, p. 63.

112 H. Drescher, Die Kunde VIII (1957), PP- 74-5. 113 A. Godard, Les Bronzes du Luristan, pl. VII; J. Birmingham,

Iraq XXVI (1964), PP- 44 ff.

114 For example, Ashmolean Museum 1965.771.

115 R. Dyson in Dark Ages and Nomads, pl. XII.2. 116 I am grateful to Mrs. R. Maxwell Hyslop for making these

unpublished analyses available to me.

11• See the mould from I;Iasanlfii: H. E. Wulff, The Traditional Crafts of Persia (1966), p. II, fig. 6.

xl1 For example, Survey IV, pl. 52B; 7000 rears of Iranian Art (U.S.A. 1964-65), no. 421; Louvre AO 20.424.


and adzes from Tepe Gawra and Tepe Billa in Iraq, of late third or early second millennium date, are relevant to the study of metalwork from Iran, since exact typological parallels have been reported from Lfiristan.119

The zoomorphic bracelets and many of the simpler pinheads from Liristan, were cast in two-piece moulds and often mass produced. This is made clear by a number of examples which have not been properly cleaned down after casting.120 The more elaborate heads and terminals were made by the lost wax-process and subsequently cast onto a shank or hoop, sometimes made in a different metal. Many of the most elaborate pinheads are on iron shanks.

It is highly skilled three dimensional modelling of zoomorphic cast bronze decoration on whetstone sockets, standard finials, figured tubes, elaborate pinheads and openwork cheekpieces for horsebits which establishes the individuality and great technical skill of Liiristdn's smiths, in a way the other metalwork from the region does not. These were all cored castings by the lost-wax process; a technique of considerable antiquity,121 which had been exploited for centuries with particular skill and ingenuity by the Elamites. This method of casting has two supreme advantages well appreciated by the smiths of

Liristan: very fine detail can be worked on the model for exact reproduction in casting and any shape of model can be reproduced. Even the most superficial examination of the standard finials indicates how the details were either built-up on the original wax model, which may itself have been made over a clay core, with coils of wax, or cut sharply or punched into the wax. The eye-ridges of the lions illustrate very well the coil technique and the elaborate surface decoration of certain anthropomorphic tubes the cutting technique.122 In the latter case, as with Egyptian and Greek bronzes,123 it is certain that the particularly elaborate decoration was done before, not after, casting. The original would have been cut from a block of hard wax, far harder than was normally used for finger modelling. It is self- evident that only one object with re-entrant curves could be made in a single mould by this process. Thus the finials, though often superficially alike, are never identical and subject to many small variations of detail.

Bees-wax in some quantity would inevitably have been required by an industry of this kind. For- tunately evidence for the prevalence of bee-keeping in western Iran in antiquity is not wholly a matter of inference. Strabo describes the bee-hives used in Media and Armenia and comments on the Persian practice of covering their dead with wax; indicative of ample supplies.124 It is also known that bee-

keeping, not indigeneous in Mesopotamia, was introduced on at least one occasion, c. 750 B.C., from mountains to the north and east.125 Pliny discusses in detail the preparation of bees-wax, but the most easterly centre of production, so far as Roman interests went, was in Pontus.126 The bee occasionally appears on west Iranian metalwork of the earlier first millennium B.C.127

Although the great majority of cast bronzes with elaborate modelled zoomorphic and anthropomorphic decoration from Lfiristan were made by the lost-wax process, it is possible that some of the

pinheads with bold, openwork designs were made differently. On these the design is often only on one face and no three-dimensional modelling is involved. These pinheads could have been made in a type of open stone mould widely used in the Near East during the second and earlier first millennium B.c. for the production of jewellery, trinkets and flat deity statuettes of various sizes.128 The pin shanks, whether of bronze or iron, were made separately and the heads subsequently cast on. Open-mould

119 M. Levey, Chemistry and Chemical Technology in Ancient Meso- potamia (1959), PP. 201 ff.; for actual examples from Lfiristan see, Outils II, nos. 1421-7, 1891.

120 For example, Ashmolean Museum: I951.236. 121 The Nahal Mishmar hoard from Israel, late fourth millen-

nium B.C., contains some of the earliest examples, P. Bar-Adon, Israel Exploration Journal XII (1962), pp. 215 ff.

122 A. Godard, Les Bronzes du Luristan, pl. LII.195-6; for the technique see: H. Wulff, Traditional Crafts of Persia (1966), p. II.

123 R. Raven-Hart, J.H.S. LXXVIII (1958), pp. 87 ff. It may be noted that the Egyptians devised a method for large-scale production of standard shapes by the lost-wax process in the

Late Period: G. Roeder, Agyptische Bronzewerke (I937), pp. 188 ff., but I know no evidence for such a procedure in Liristan.

124 Geography II I 14; XV iii 20.

125 B. Meissner, Babylonien und Assyrien I (1920), pp. 223-4; for the date see M. E. L. Mallowan, Iraq XVI (1954), p. 101, note 3.

126 Natural History XXI.xlix. 127 R. Ghirshman, Artibus Asiae XIII (1950), p. 184, fig. 4; A.

Godard, Bronzes du Luristan (E. Graeffe Collection), no. 44, pl. 18.

128s D. Opitz, Festschrift M. F. von Oppenheim (1933), pp. 179 ft.; J. V. Canby, Iraq XXVII (1965), PP- 42 ff.


castings of this kind exhibit a distinctive appearance on the upper or " free " surface. In many objects cast in open moulds this is obliterated by subsequent grinding to produce a smooth finish. This was not

necessary with pins of this type and the remains of the original casting surface sometimes survive on the back.129

The Cast Bronze Tradition: Ores, Alloys and Impurities

Ores

Apart from the significance of native copper in the earliest phase of metal-working some interest attaches to the more complex problem of identifying its use in later periods, when it has been melted and cast. It is generally agreed, and analysis has largely demonstrated, that native coppers are distin-

guished by their purity. But it is also clear that an exceptionally pure copper, closely approaching or even matching the purity of a native metal, may be obtained by smelting an ore such as malachite.130 In these circumstances identification of native copper in these analyses of Liristan Bronzes must be

regarded with due caution. Definite information is available on only one source of native copper in Iran. At Talmessi, to the north-east of Isfahan, there are such rich occurrences of primary native copper that any early settlers in the area would have had little difficulty in finding and using the raw material. It is the most likely source for the native copper used by the earliest settlers at Sialk. The Talmessi

samples had a relatively high arsenic content with traces of lead, zinc and silver, but apparently not

nickel.131 In native coppers from Azerbaijan and Armenia nickel was traced, but not normally arsenic.132 It is possible that native copper, alloyed of course with tin in both cases, was used in the manufacture of the bar cheekpiece for the horsebit of Sialk B type (Analysis No. 3) and the flange- hilted dirk (Analysis No. 37); both objects produced in the ninth or eighth centuries B.C. It is possible, though less likely, that the mouthpiece of the horsebit was also an alloy of native copper and tin (Analysis No. 4).-

There is already evidence that in the later second millennium B.C. unalloyed copper was still used in Dailamdn at least for weapons deposited in graves.133 One of the pins with a large disk-shaped head, from

Ltiristmn, where they are common in contexts of the eighth and seventh centuries B.C. (Analysis

No. 131), is of copper, whilst the other one analysed is a mild tin-bronze (Analysis No. 132). More

analyses of these pins are now required to see if copper or bronze was the more usual; copper would be

perfectly appropriate, perhaps better even, for this type of pinhead, which was hammered from the top of a single circular-sectioned rod and then decorated with repousse or chased patterns. Two pins with swollen or moulded heads, also of copper, could be earlier (Analysis Nos. 109, I 13). A pair of bracelets in Sydney,134 the mouth piece of a horsebit (Analysis No. 2), one of the harness-rings (Analysis No. 67) and a figured tube (Analysis No. I oo) of the earlier first millennium B.C. with less than 3 per cent tin

would, on Coghlan's definition, be classified as accidental bronzes obtained by working an enriched

ore, rather than deliberate alloys.

Alloys: Tin-bronzes In the majority of the bronzes analysed the tin percentages range between 4 and 13 per cent. Only

six objects had higher percentages, to a maximum of 18 - 3 per cent in a bottle-shaped support, whose silver colouring reflects the higher tin content. Witter has shown that the addition of 5 per cent tin to

copper greatly increases its hardness; this can be doubled when work-hardened by hammering. The

addition of tin, between 5 to 18 per cent, does not produce a proportional increase in hardness as cast, but it is very considerably increased by work-hardening. The hardness again increases with the addition

129 For example, Ashmolean Museum: 1965-799 (Analysis No. 125); 1965.807 (Analysis No. 128).

130 H. H. Coghlan, Proceedings of the Prehistoric Society XXVIII (1962), pp. 58 ff.

'3' M. Maczek Archaeologica Austriaca X (1952), pp. 61 ff.with map.

132 I. R. Selimkhanov, Proceedings of the Prehistoric Society XXX

(1964), pp. 66 ff. 133 N. Egami Dailaman I, p. 63. 134 J. Birmingham, Iran I (1963), p. 75.


of between 13 to 27 per cent tin, which gives a maximum hardness, but is very brittle.135 It is interesting to notice here, when dealing with a region in which iron working was being pioneered at a time when these artefacts were in use, that the hardness of a Io per cent tin-bronze is equal to that of a soft iron; work-hardened bronze can reach the hardness of a medium quality steel. It is clear from these analyses that there was little or no attempt at a uniform standard, and pins and finials, for which hardness and strength were not necessarily prime requisites, have equal, if not higher percentages of tin than the tools and weapons. The horsebits are more consistent in having between 5 and Io per cent tin.

Impurities: Nickel

The most striking feature of these analyses is the regular occurrence of small percentages of nickel, less than about I per cent. This is definitely an inadvertent impurity. Only in three of the artefacts analysed was the nickel percentage higher than this: No. 22: 3'5; No. 109: 18; No. 113: I I per cent. Many ancient bronzes from the Near East and Caucasia contain quantities of nickel generally below I per cent, but rising on occasion to 5 per cent.136 In this context Analysis No. 22 is of some interest. The axe is an arsenical-bronze, closely paralleled at Ur and Susa137 in the third quarter of the third millennium B.C. Its composition may indicate that it is of Elamite or Sumerian manufacture. As nickel was so frequently detected in the third millennium artefacts from Susa and south Mesopotamia which they analysed, the Sumer Copper Committee assumed a copper ore containing nickel as the source of the raw material and devoted considerable research to identifying it. Although nickel is not a particularly common impurity in copper, the Committee was unable to reach any firm conclusions.138 They accumulated a wide range of useful data and suggested that the Jebel Ma'dan deposit, inland from Sohar, in the Wadi Ahin, Oman, had been the primary source for Sumerian copper. Copper lumps from Mohenjo-daro also contained nickel, rising to slightly over I per cent.139 Documentary evidence also indicates that before the Old Babylonian period copper from southern sources was normally used in Sumer.140 Except in the case of the axehead (Analysis No. 22), these conclusions have little relevance to the artefacts from the central Zagros. Even allowing that the nickel content identifies the parent ores of Sumerian bronzes, neither the historical evidence nor geographical probability is in favour of a southern source for the ores used in Ltiristan in the earlier first millennium B.c.

Copper ores containing nickel have been found in Armenia and north-east Anatolia. Nickel, normally in comparatively small percentages, has been detected in artefacts from Caucasian sites and in the south-west Caspian area.141 Analyses of weapons from Amlash have made clear that there is a source of nickelferous ore, also containing arsenic, somewhere in the region.142 This is almost certainly the mining area whence Lfiristtn derived its main copper supplies.

Arsenic

Arsenic is a very common impurity in copper and its presence in virtually all these analyses, below

I per cent, is of no great significance. The historical and metallurgical implications of the higher percentages have already been examined and require no further comment here.

135 Cited by H. H. Coghlan in Notes on the Prehistoric Metallurgy of Copper and Bronze in the Old World (1951), p. 44.

136 C. F. Cheng and C. M. Schwitter, AJA LXI (1957), PP. 351 ff.; I. R. Selimkhanov, Proceedings of the Prehistoric Society XXVIII (1962), pp. 77-9; Proceedings of the Prehistoric Society XXX (1964), pp. 66 ff.; Germania XLIV (1966), pp. 230-I.

137 OUtils I, p. 156: Type Aia; the analysis settles Przeworski's doubts about the object's authenticity.

138 H. J. Plenderleith in Ur Excavations II, p. 287 compare C. H. Desch in Ur Excavations IV, p. 165; H. Peake, Antiquity II

(1928), pp. 452 ff.; S. Przeworski, Die Metallindustrie Anatoliens in der Zeit von 1500oo-7oo vor Chr. (1939), pp. 105 if.

139 J. H. Marshall, Mohenjo-Daro and the Indus Civilization (i93I), II, p. 484-

140 H. Limet, Le Travail du Mital au Pays de Sumer (1960), pp. 87 ff.; W. F. Leemans, Foreign Trade in the Old Babylonian Period (I960), pp. 122 ff.

141 See note 136; N. Egami, Dailaman I, p. 63; J. Birmingham, Iraq XXVI (1964), p. 46: blade only.

142 Mrs. J. Birmingham-personal communication about unpublished analyses recently done in Sydney.


Lead In considering the percentages of lead found in these bronzes it is important to bear in mind its

effect on an alloy. As it produces softness, it is undesirable in any quantity in a metal to be used for tools or weapons, but since it increases the fluidity of the alloy it eases the production of sound castings and is particularly valuable in the manufacture of highly decorated castings. A study of the lead percentages does not lend much weight either to the view that richly ornamented objects invariably have a higher lead content or that a lead enriched ore was systematically exploited at some period during the main production of bronzes in Lfiristdn in the earlier first millennium B.C. The highest percentages, over I o, occur in two cheekpieces for horsebits, which are almost certainly modern reproductions (Analyses Nos. 10/32, 12), in a decorated tube (Analysis No. 98) and a bottle-shaped support (Analysis No. 72), both of very unusual composition, which may also indicate objects of dubious authenticity. Here as in Analyses Nos. 17, 32, 38 and 99, the lead percentages, between 2 and Io, almost certainly indicate a deliberate addition to facilitate casting, either by the ancient or modern smith. The axe

(Analysis No. 38) is of some interest in this connection. It is an early type, probably made in the late third or early second millennium B.C., and the extra lead is certainly to ease casting the animal modelled on the butt; one of the earliest examples of cast zoomorphic decoration from the Zagros and indication that the special metallurgical properties of a leaded alloy were already appreciated. There are, how-

ever, many later objects, richly decorated, which do not have significantly higher lead percentages than the weapons.

Antimony Only 16 per cent of the objects analysed contained antimony and in only two cases (Analyses Nos.

48, 130) was the percentage high enough to merit comment. Antimony, like arsenic, has a hardening effect on copper, but it is quite suitable for casting and a good temper may be given to cutting-tools by hot forging. Sumerian texts indicate that a metal, which might be antimony (Sis-gan), was added in small quantities to molten copper or bronze prior to the casting of an object,143 in the later third millennium B.C. Antimony appears in bronzes from Mesopotamia and Caucasia144 and Kuftin reports copper with 5-10o per cent antimony in the lowest level of dolmens in north Ossetia.145 It is possible, that the axe (Analysis No. 48), with I - 2 per cent antimony, is an example of the technique described in the Sumerian texts. Its composition is unusual, with almost equal percentages of tin and antimony. As it can be dated to the third quarter of the third millennium B.C.146 it may mark an experimental stage in the development of tin-bronze. The even higher percentage in the pin (Analysis No. 130), which has a zoomorphic bronze head cast onto an iron shank, probably made in the eighth or seventh century B.C., may just reflect the use of an enriched copper ore.

Zinc Small percentages of zinc in ancient copper or bronze objects, up to about 2 per cent, may be

regarded as natural impurities. In Caucasia for example zinc and copper ores are closely associated and early occurrences of a very low zinc brass were probably due to an accidental inclusion of zinc ore with the copper. The extremely high percentages of zinc in Analyses Nos. 10/32 and 12 are another matter. Seen in the context of the other analyses they are almost certainly to be taken as indicative of a modern forger's brass, despite certain theoretical uncertainties about the significance of zinc in ancient

objects.147 The matter is complicated by the present obscurity surrounding the early history of brass, although the methods of the ancients in manufacturing it are now sufficiently well known.148

It is not until the Roman period that objects and documentary evidence unequivocally indicate the existence of brass. These early brasses contain 70-90 per cent copper, O10-30 per cent zinc; the metal

143 H. Limet, op. cit., pp. 55 ff. with full philological discussion. 144 0. Davies, Man (1935) 91, note 13.

x45 B. A. Kuftin, Archaeological Excavations in Trialeti I (I941): p. I57-

146 See checklist of objects.

147 H. Otto, Wiss. Zeitschr. Univ. Halle (Ges. Sprachw.) VII/I (1957), pp. 203 ff.; cf. H. H. Coghlan in The Application of Quantitative Methods in Archaeology (1960), p. 29.

14s Unless otherwise stated the following evidence for the earliest

history of brass is based on R. J. Forbes, Studies in Ancient

Technology VIII (1964), pp. 260 ff.


becoming yellower as the zinc proportion is increased. The commonest modern brasses contain a much

higher percentage (30-40) of zinc. There is no obvious reason why the process used by the Romans to

produce a copper alloy with the rich yellow colour of a low zinc brass, by heating ground calamine (a zinc ore) with powdered charcoal and small pieces of copper, should not be of greater antiquity. Limet149 has rejected with some justice the old view that the Sumerians were familiar with brass made

by the calamine process in the later third millennium B.C. For similar reasons, and until analyses lend force to the suggestion, it is impossible to say whether the " s1v hlb " of the Ras Shamra tablets is indeed brass as Schaeffer suggested.150 The Pseudo-Aristotle attributes the discovery of brass to the Mossy- noeci, who lived in Turkey south of Trebizond, in an area rich in zinc, in the early first millennium B.c. The same ancient source relates that Darius had a cup which appeared to be of gold, but had an

unpleasant smell; perhaps a low zinc brass. Analysis of Achaemenian metal bowls which had a

golden colouring has shown them to be bronzes with io to 12 per cent tin and no appreciable trace of zinc.151 However, analysis of fibulae from Gordion and other bronze objects from Altintepe in Anatolia has shown that zinc, over o0 per cent in the fibulae, appears not uncommonly in objects, almost

certainly of local manufacture, made at least as early as the seventh century B.c.152 The status of cheekpieces 1951.194 (Analyses Nos. 10/32) and 1951.195 (Analysis No. 12) must for

the moment remain an open question, since the analyses of objects from Anatolia offer positive grounds for thinking that these cheekpieces might be ancient low zinc brasses, perhaps designed to simulate

gold. If the zinc criterion is then equivocal in significance, forgeries will prove as difficult to detect

metallurgically as they are to isolate with confidence on stylistic grounds,153 since ingenious forgers use

fragments of inferior or broken ancient bronzes to make more attractive pieces, rendering their work

virtually immune from detection by analysis. The only published analysis of a definite forgery was

reported to contain 84"7

per cent copper, 14. 69 per cent tin, o*26 per cent arsenic, traces of silver and

iron, but none of antimony, lead, zinc, nickel or cobalt.154 The comparatively small number of these bronzes in which zinc was traced in relatively small

percentages include a pair of fibulae analysed in Sydney,155 which are unlikely to have been manufactured in Lfiristdn, even if actually found there. Fibulae reached western Iran late in the period of maximum production in the Laristdn workshops and are rarely reported with certain provenance from the area. The straight pin was extremely common until the late seventh century B.C. in the region. In the north some of the earliest fibulae appear to have close Caucasian affinities. Those fibulae which have been reported from Lfristdn are generally typical Western Asiatic types with a very wide distribu- tion156 and bear no traces of the distinctive local styles of decoration. They are probably imports from

Assyrian or eastern Anatolian workshops. But, as the other objects include decorated finials and tubes, characteristic of Liristin, as well as the body of a sheetmetal vessel with a cast bronze handle in the same style as the finials, it is certain that an ore with zinc in it was used in the region. Zinc was not found in the Arghana ore examined by the Sumer Committee,157 but zinc occurs in artefacts from

Talish,158 Dailamdn,159 and Amlash,160 as well as from Caucasian sites.161 Thus northern ore was

probably the source of the small quantities of zinc in Liristin bronzes.

The Groups The number of analyses so far available is too small to warrant detailed statistical analysis at this

stage, particularly in view of the uncertain attribution of certain bronzes. The overall picture presented by the analyses now to hand is, however, seen more clearly if they are broadly grouped by impurities.

149 H. Limet, op. cit., pp. 56 ff. t50 Mission de Ras Shamra VII (1957), p. xxxiv.

151 J. Cooney, Journal of the American Research Center in Egypt IV (1965), P 41.

152 I am most grateful to Dr. A. Steinberg for allowing me to cite these unpublished analyses done for him by the Research Laboratory for Archaeology and the History of Art, Oxford.

153 J. A. H. Potratz, Iranica III (1963), pp. 124 ff.

154 ILN, October 29th 1932, p. 667, note 2.

155 J. Birmingham, Iran I (1963), P. 75.

15s D. Stronach, Iraq XXI (1959), pp. 181 ft. 157 Report of the British Association (1928), pp. 437 f. 158 C. Schaeffer, Stratigraphie Comparde (1948), p. 605. 159 N. Egami, Dailaman I, p. 63. 160 J. Birmingham, Iraq XXVI (1964), p. 46. 161 See note 136.


In the following table figures in bold type denote the objects definitely made in the second half of the third or early in the second millennium B.c.; the remainder probably belong to the very end of the second millennium or the first quarter of the first millennium B.C. It will be seen that objects with elaborate cast bronze decoration characteristic of the Lfiristdn workshops, c. 850-650 B.C., tend to be concentrated under A2, A4 and A5. Analyses with an S-prefix were done in Sydney and published by Mrs. J. Birmingham (Iran I (1963), p. 75). COPPER: 2, 67, I09, 113, 131; SI, S2, S3 (perhaps native metal alloyed with tin: 3, 37).

ARSENICAL-BRONZES: 22, 23, 24, 38, 54 (48).

TIN-BRONZES:

A: with nickel I. Ag, Ni: 4, 29, 52, 60, 66, 84, 87; +Bi: 71, 86. 2. Ag, Ni, Pb: 5, 6, 9, 16, 17, 18, 30, 34, 44, 57, 61, 78, 85, 88, 89; +Bi: 35, 64, 79, I07, IO8,

I IO, 128, 129.

3. Ag, Ni, As, Bi: 47, 65, 68. 3a. Ag, Ni, As, Au: 36. 4. Ag, Ni, As, Pb: I, 7, 11 14, 14, 5, 19, 20, 21, 25, 26, 28, 31, 39, 40, 43, 46, 59, 74, 80, 81, 91,

92; +Bi: 41, 45, 49, 51, 55, 58, 62, 69, 70, 72, 73, 75, 77, 82, 83, 94, 95, 96, 97, 101, 105, IO6, II7, II8, 120, 121, 122, 123, 126, 127, 132.

5. Ag, Ni, As, Pb, Sb: 8, 27; S6, S8, S9; +Bi: 53, 56, 63, 63a, 98, 99, 102, 104, I I, 114, 125;

S4, S5, SIo, SII, Si2, 813, SI4, SI6. 6. Ag, Ni, As, Pb, Sb, Zn, Bi: I 9, 130, SI7, SI8. 7. Ag, Ni, As, Pb, Zn, Bi: 90, 93, 00oo, IOa, 124; (+Bi: 33, 42).

B. no nickel, with arsenic I. Ag, Pb, As: 5o; (+Bi: I15, I i6). 2. Ag, Pb, As, Sb: S7.

C: no nickel or arsenic I. Ag, Bi: 76, 103, 112.

D: high zinc-probably a forger's brass I. Ag, Pb, As, Sb, Ni, Fe, Zn: 10/32, I2. 2. Ag, Pb, As, Ni, Fe, Zn: 13.

The Sheetmetal Tradition The creation of the richly decorated cast bronze objects so characteristic of LSiristdn c. 850o-650 B.c.

is a modeller's or sculptor's art, the decoration of sheetmetal is by contrast a draughtsman's art. This need not necessarily mean separate workshops or even different specialist craftsmen. Indeed sheetmetal vessels with cast fittings and the stylistic affinities between cast and sheetmetal decoration indicate that

single workshops, if not single craftsmen, produced both. The main distinction is an historical one. Whereas the cast bronze tradition of Liiristin is very localized,162 the sheetmetal work is part of a much wider West Iranian tradition of metalworking. To the north there is nothing like the range of technical skill and variety of artefacts in decorated cast bronze which there is in Loiristan, but the range of sheetmetal work, often in gold or silver, is much greater. Certain elements in the decorated sheetmetal work associated with Liristin, notably the large group of decorated situlae,163 are distinct and their history and cultural affinities, primarily Babylonian, are now reasonably clear. In contrast the origin of the decorative style used on a series of sheet bronze quiver plaques, belts, disk-headed pins and vessels, many associated with the Dum Surkh shrine, is still obscure. There is some indication that the immediate source of this style in the ninth and eighth centuries B.c. was to the north of Liristin, where

162 So long as the exact relation to Elam in the earlier first millennium B.C. is unknown.

163 P. Calmeyer, Berliner Jahrbuch V (1965), pp. I ff.; VI (1966), pp. 55 ff.; P. Amandry, Antike Kunst 9 (1966), pp. 57 f.-


similar styles of decoration appear on artefacts found at Ziwiyeh and IHasanli,164 and on uncontexted sheetmetal objects in precious metals.165 Indeed, certain decorated objects of gold and silver, reported as from Loiristan, may well have been made to the north, where these metals were more commonly used.

In both areas the actual rendering of the design was done in the same way. The relief designs on flat pieces of metal were worked from behind, normally not to an excessive height. The designs on metal vessels were generally worked from the front. The linear designs were chased on the surface with a slightly blunt instrument. The metal originally filling the groove was displaced, not removed, indicating that the technique of engraving was not used. As the metal is normally thin and a ridge appears on the underside of the completed piece it was clearly worked whilst resting on a yielding surface. The absence of engraving in Liristan is not surprising, for, according to Maryon's experiments, the engraving of bronze with bronze tools would never have been a practical proposition.166 The bronze-worker would not have wasted his time in trying to engrave with inadequate tools when he could so easily have chased the design. On a fine silver goblet, now in Cleveland,167 the artist partially set out his design in a series of pointilli dots, evidently as a guide to fitting the figures into the available space. In the final stages he made significant modifications. There is virtually no trace of mechanical aids in the laying out of designs. On the great majority of disk-headed pins the designs are very roughly laid out and clearly executed freehand without careful preparation.

Although hammering and annealing was the only method by which sheetmetal could be manufactured in antiquity, the metal could be worked down until it was smooth and a thickness of even less than o*5 mm. was sometimes achieved from thick copper or bronze sheets cast in open moulds. Of the three sheetmetal objects from

Lfiristmn, probably all made in the eighth century B.c., analysed in Oxford, one was copper (Analysis No. 131), two were of bronze (Analyses Nos. Iola, 132). Most of the sheetmetal vessels reported from Liristin were made, either by sinking or raising, from a single sheet. They are almost always of uniform thickness and of a relatively simple form. Handles and spouts of beaten metal were secured to the body of the vessel with rivets; usually split rivets with domed sheetmetal head. Certain seams were folded and hammered, sometimes sealed with an adhesive to make them water-tight. Cast attachments68s are not common and when they appear may well reflect northern influences.'69 Inlay is rarer still, but not unknown. White shell or limestone inlays set in bitumen serve as eyes on the female faces worked in high relief set at the centre of sheetmetal disks. This technique was used on the zoomorphic situla from I;asanlfl.o70 Disk-headed pins were occasionally inlaid with carnelian strips."'7

CONCLUSIONS

This paper has dealt briefly and sketchily with but one aspect of a very complex problem. It cannot then be used to draw general conclusions about the cultural history of Lairistdn, only to indicate certain possibilities which emerge when the metalwork of the region is considered in its wider metallurgical context. It has been necessary to assume, without detailed argument, certain basic facts about the typology and iconography of the " Lfiristan Bronzes " which have only become apparent through systematic study in the last decade or so. This would be inexcusable were many of them not already available in print in books and articles by, among others, Calmeyer,172 Deshayes173 and Potratz.174 For thousands of years, with but a brief interlude in the earlier first millennium B.C., the metal workshops of Lfiristin differed very little in the methods they used and the artefacts they produced from the

164 A. Godard, Le Trisor de Ziw'iy, pp. 105 ff., figs. 9 1-2; R. Dyson, Archaeology i6 (1963), p. 132; 17 (1964), pp. 6 and io.

165 D. Shepherd, Bulletin of the Cleveland Museum of Art (1966), PP. 38 ff.

166 AJA LIII (1949), PP. 115 ff. 167 D. Shepherd, op. cit., p. 46, fig. Ii. 16s L. Speleers, Bulletin des Musees Royaux d'Art et d'Histoire

(Brussels 1942), p. 85, fig. 16 (0.1549); R. Ghirshman, Persia from the Origins to Alexander (1964), fig. 122; B. Goldman, JNES XX (1961), p. 242, fig. 5.

13A

169 0. Muscarella, Hesperia XXXI (1962), pp. 317 ff. 170 E. Porada, Ancient Iran, pl. 32. 171 Louvre-AO 20.554. 172 Particularly Moortgat Festschrift (1964), pp. 68 ff.; Berliner

Jahrbuch V (1965), pp. f.; VI (1966), pp. 55 ff.

173 Les Outils de Bronze de l'Indus au Danube, IVe au lie millenaire, I-II, (Paris 1960).

174 Particularly Die Pferdetrensen des Alten Orient (Rome 1966); Luristan Bronzen (Istanbul 1968).


smiths of their more powerful neighbours to the south and west.175 Their work in this period is remarkable only to the extent that fortune has favoured its survival in some quantity and in excellent repair. Throughout the second millennium B.C. relations with the south and west are clear and unequivocal, though the mechanics of them may not always be easily explained with present evidence. Then, apparently rather suddenly, in the very late second or early first millennium B.C., their repertoire was greatly augmented to include artefacts so far unknown elsewhere and others, which though known elsewhere, are never found with the elaborate cast bronze decoration they have in Lfiristan. This

industry flourished for about 200 years and was then totally eclipsed. Nomadic intruders or migrating Iranians are normally invoked to explain this remarkable episode; the true explanation is probably not so simple. There are reasons for thinking that this development was fundamentally a local one; a resurgence of indigenous traditions under particularly favourable economic and political conditions which followed the eclipse of Elamite power. If for no other reason, the very fact that such an industry did not develop during Iron Age II further north strongly suggests that its distinctive character owed little or nothing to the intrusive Iranians. Its anonymous patrons may have been mounted warriors fresh to the region-that remains to be demonstrated-but it is already clear that the smiths they employed were steeped in much older, more local, traditions of iconography and craftsmanship.

175 Ceramic evidence from the region at this time reflects exactly the same cultural affiliations, see R. Dyson, Persian Plateau, passim.


APPENDIX I

A CHECK-LIST OF THE OBJECTS ANALYSED IN OXFORD

It is impossible for reasons of space and expense to document in full detail the objects which were analysed; the following list is intended to contain all the vital information, until such time as the full catalogue raisonne' of the Ashmolean Museum Collection of Ancient West Iranian Bronzes is available.

The following abbreviations are used here:

AR: Ashmolean Museum Annual Report. Bomford: Antiquities from the Bomford Collection (Exhibition Catalogue, Ashmolean Museum,

1966). Gadd: Transactions of the Oriental Ceramic Society 19 (1942-43), PP- 33 if. Przeworski: Archaeologia 88 (1940), pp. 229 ff. Sarre: J. A. H. Potratz, Luristan Bronzen (Istanbul 1968).

Spencer-Churchill: Exhibition of Antiquities and Coins. . . . Collection of the late Capt. E. G. Spencer-Churchill (Ashmolean Museum, 1965).

Analysis No. 1/2: 195I.I89: complete horsebit; mouflon cheekpieces. Cf. Przeworski, no. 15, pl. LXXVa. 3/4: 1931.23: complete horsebit; bar cheekpieces. Cf. Przeworski, no. 14, pl. LXXIVc.

5: 195I.I9 1: cheekpiece in the form of a horse. Cf. Przeworski, no. 17, pl. LXXVIa. 6: 1951.188: complete horsebit; mouflon cheekpiece. Przeworski, no. 15, pl. LXXVa. 7: 195 I.90: complete horsebit; horse cheekpiece. Przeworski, no. I6, pl. LXXVb. 8: Brown Loan: cheekpiece of a horsebit. Openwork plaque with rampant, winged wild goats flanking a

central figure wearing a horned crown. In the upper part, between the animals and the figure, " cocks' heads "; in the lower beasts with open jaws. Design on both faces. Height: o0098 m.; Width: o0 139 m.

9: I95 I. 198: cheekpiece; " master-of-animals ". Gadd, pl. Ioc. I0/32: I95I.I94: cheekpiece; winged mouflon. Gadd, pl. 9c.

11: 1951.197a: cheekpiece; griffin. Gadd, pl. 9a. I2: 195I.I95: cheekpiece; winged sphinx standing on hares. Gadd, pl. 9b. 13: 1951.199: cheekpiece; " master-of-animals ". Gadd, pl. lob. 14: 195'I.92a: cheekpiece. Cf. Iran I (1963), pl. IIb for form. 15: 1931.31: whetstone socket cast in the form of a " panther's " head and neck, but with ibex horns. 16: 1933.513: whetstone socket cast in the form of a " panther's " head and neck with the head and neck of

an ibex in its claws. 17: 1951.204: harness ring. Przeworski, no. 13, pl. LXXIVb. 18: 1931.34: harness ring. 19: 1951.201: harness ring. Przeworski, no. 12, pl. LXXIVa. 20: I951.I53: axe-blade. Przeworski, no. 3, pl. LXXIIc. 21: I95I.I51: club-headed pin driven into an axe-socket. Przeworski, no. 4, pl. LXXIId. 22: 1951.I50: axe-blade. Przeworski, no. 9, pl. LXXIIIc.

23/24: 1951.141: dagger. Gadd, pl. IIb. 25: 1951.223: bracelet with lion terminals. Przeworski, no. 47, pl. LXXIXe. 26: 1951.234: bracelet with sleeping duck terminals. Przeworski, no. 46, pl. LXXIXd. 27: 1951.261: pin with head exactly like the lion terminals of 25. 28: 1951.255: pin with head in form of a duck. Przeworski, no. 35, pl. LXXVIII. 29: 1951.18: " standard finial " in the form of two opposed lions. Cf. Przeworski, no. 27, pl. LXXVIIc. 30: 1931.27: " standard finial " in the form ofa " master-of-animals ". Cf. Przeworski, no. 25, pl. LXXVII. 31: I95I.I8o: " standard finial" in the form of two opposed rampant goats. Przeworski, no. 28, pl.

LXXVIId. 32: see no. I.

33: I95I.I34: flange-hilted dirk. Cf. Maxwell-Hyslop, Iraq VIII (1946), type 36, p. 46. 34: 195I.136: flange-hilted dirk. Cf. Godard, Les Bronzes du Luristan, pl. VIII.I9. 35: I95I.I37: hilt of a flange-hilted dirk. Sarre, pl. I.2. 36: I965.769: flange-hilted dirk. Bomford, no. I I.

37: 1951.135: flange-hilted dirk. Cf. Dyson in Dark Ages and Nomads, pl. IX.5. 38: I95I.156: axe-head. Cf. British Museum-Antiquities of the Bronze Age (1920o), p. 176, fig. 187. 39: I951.I52: axe-head. Przeworski, no. 5, pl. LXXIIe. 40: 1951.154: axe-head. Przeworski, no. I, pl. LXIIa.

13B


Analysis No. 41: 1951.155: axe-head. Przeworski, no. 2, pl. LXXIIb.

42: 1937-739: spike-butted axe-head. Cf. the following axe.

43: 1951.157: spike-butted axe-head. Przeworski, no. 8, pl. LXXIIIb.

44: 195.i60o: spike-butted axe-head. Przeworski, no. 6, pl. LXXIIf. 45: 1951.161: spike-butted axe-head. Przeworski, no. 7, pl. LXXIIIa. 46: 1965.187: axe-head. Spencer-Churchill, no. 27. 47: 1965-759: axe-head. Bomford, no. 2, pl. II. 48: 1965-762: axe-head. Bomford, no. I, pl. II. 49: 1965-761: spike-butted axe-head. Bomford, no. 5. 50: 1965-760: axe-head. Bomford, no. 4, pl. II. 51: 1965-763: axe-head. Bomford, no. 3, pl. II. 52: 1965. 88: spike-butted axe-head. Spencer-Churchill, no. 28. 53: I965. 86: pick-axe. Spencer-Churchill, no. 26. 54: 1965-764: pick-axe. Bomford, no. 7, pl. III. 55: 1965-767: pick-axe. Bomford, no. 8, pl. III. 56: 1933.512: adze-head. Cf. Deshayes, Les Outils de Bronze, no. 1891, vol. II, pl. XXXI.4. 57: 1965-766: adze-head. Bomford, no. Io, pl. III. 58: 1931.30: whetstone socket. Cf. Survey IV, pl. 53G. 59: I963.I57: whetstone socket. Cf. Survey IV, pl. 53G. 6o: 1951.196: horsebit. Gadd, pl. Ioa. 61: 1965.826: horsebit. Bomford, no. 33. 62: 1965.827: horsebit. Bomford, no. 31, pl. V 63: 1965.828: horsebit. Bomford, no. 32. 64: 1965.829: horsebit. Bomford, no. 35, pl. IV. 65: 1965.830: horsebit. Bomnford, no. 34, pl. IV. 66: 1951.201: harness-ring. Przeworski, no. 12, pl. LXXIV.

67: 1951.2ooa: harness-ring. Gadd, pl. Iod, left. 68: 1951.202: harness-ring. Przeworski, no. II, pl. LXXIII.

69: 1965-196: harness-ring. Spencer-Churchill, no. 34, pl. II. 70: 1965.824: harness-ring. Bomford, no. 37.

7I: 1951.167: " bottle-shaped " support. Cf. No. 72 following. 72: 1951.169: " bottle-shaped " support. Przeworski, no. 21, pl. LXXVIe.

73: 195I.I76: " bottle-shaped " support. Przeworski, no. 23, pl. LXXVIg. 74: 1951.177: " bottle-shaped " support. Cf. No. 73 above.

75: I965-782: "bottle-shaped " support. Bomford, no. IoI. 76: 1965-783: " bottle-shaped " support. Bomford, no. 102.

77: 1951.178: " standard finial ". Przeworski, no. 27, pl. LXXVIIc.

78: 1951.182: " standard finial ". Przeworski, no. 25, pl. LXXVIIa. 79: 1965.I94:

" standard finial ". Spencer-Churchill, no. 32, pl. II.

8o: 1963.I58: " standard finial "; compare No. 9I. 81: 1965-191: " standard finial ". Spencer-Churchill, no. 29, pl. III. 82: 1965.192: " standard finial ". Spencer-Churchill, no. 30, pl. III. 83: 1965.193: " standard finial ". Spencer-Churchill, no. 31, pl. II. 84: 1965-787: " standard finial ". Bomford, no. Io6. 85: 1951.179: " standard finial ". Cf. No. 83. 86: 1965-789: " standard finial ". Bomford, no. 103, pl. VI. 87: 1965-790: " standard finial ". Bomford, no. 104, pl. VI. 88: 1965-791: " standard finial ". Bomford, no. III, pl. VII. 89: 1965.792: " standard finial ". Bomford, no. I o, pl. VI.

90: 1965-793: " standard finial ". Bomford, no. Io9, pl. VI.

9i: 1965-794: " standard finial ". Bomford, no. 107, pl. VII. 92: 1950.214: decorated tube-zoomorphic. 93: 1931.28: decorated tube-zoomorphic. 94: 195I.186: decorated tube-zoomorphic. 95: 195 I. 84: decorated tube-anthropomorphic. 96: 195I.I85: decorated tube. Przeworski, no. I8, pl. LXXVIb. 97: I965.I95: decorated tube. Spencer-Churchill, no. 33, pl. II.

98: 1965.784: decorated tube. Bomford, no. x I2. 99: I965.785: decorated tube. Bomford, no. I I3.

xoo: I965.786: decorated tube. Bomford, no. I 14, pl. VIII.

IoI: 1951.332: cast zoomorphic vessel handle. AR (I95I), pl. IIIB, p. 25. Ioia: I951.332: body of the vessel to which No. IoI is attached.


Analysis No. io2: 1951.243: pin; flat head.

I03: 195 .244: pin; nut-shaped head.

I04: 1951.245: pin; nut-shaped head.

105: 1951.246: pin. Cf. Sialk II, pl. XXIX.2, fourth from left. Io6: 1951.247: pin; nut-shaped head.

10o7: 1965.920: pin. Bomford, no. 86.

Io8: 1965-921: pin. Bomford, no. 84. 1o9: 1965.923: pin. Bonmford, no. 85. I1o: 1965-925: pin. Bomford, no. 74. iii: 1965.927: pin. Bomford, no. 75. 112: 1965-936: pin. Bomford, no. 79.

I13: 1965-937: pin. Bomford, no. 8o.

114: I965-938: pin. Bomford, no. 65.

S15: 1965-917: pin. Bomford, no. 69. i116: 1965-919: pin. Bomford, no. 83.

S17: 1951.264: pin; floral head. 118: 1951.265: pin; floral head.

119: 1951.266: pin. P. Jacobsthal, Greek Pins (1956), fig. 163. 120: 1951.267: pin. Przeworski, no. 36, pl. LXXVIIIa. 121: 1951.269: pin; head in form of an openwork cage. 122: 1931.32: pin with large opencast head. Cf. following. 123: 1951.276: opencast pinhead. Przeworski, no. 42, pl. LXXVIIIb. 124: 1951.283: opencast pinhead. Przeworski, no. 19, pl. LXXVIc.

125: 1965-799: opencast pinhead. Bomford, no. 49- 126: 1965.800: opencast pinhead. Bomford, no. 50. 127: 1965.8o6: opencast pinhead. Bomford, no. 46, pl. X. 128: 1965.807: opencast pinhead. Bomford, no. 47, pl. X.

129: 1965.808: opencast pinhead. Bomford, no. 48, pl. X. 130: 1931.33: zoomorphic pinhead. Sarre, pl. XXV.I43. 131: 1951.272: disk pinhead; crudely chased star.

132: 1965-796: disk pinhead. Bomford, no. 53-

APPENDIX II

TABLE OF ANALYSES

No. Cu Sn Pb As Sb Ni Bi Fe Zn Ag Au

I. 94*5 4'7 0o25 o0 18 n.d. o- I8 <0ooI o II n.d. 0o040 n.d. 2. 97'9 1 9 n.d. n.d. n.d. o0o86 <oo01 o 0057 n.d. oo026 n.d. 3.

92" I 7 9 n.d. n.d. n.d. n.d. <0oo01 n.d. n.d. o oo6 n.d.

4. 95" 4'9 n.d. n.d. n.d. 0o041 <o0oI o o69 n.d. o oo88 n.d.

5. 91 7 7-8 0o21 n.d. n.d. 0-19 <o0.o n.d. n.d. o-oI3 n.d.

6. 92- 2 69 o0'51 n.d. n.d. 0o30 <0.01 o 078 n.d. o 046 n.d. 7. 90'

8 6 0*55 0'33 n.d. o- 55 <0o0o n.d. n.d. o00oo99 n.d. 8. 88-5 9- I 027 0o'44 o035 o-o62 <o-o 1 -3 n.d. 0-037 n.d. 9. 91'3 6-5 0o

12 n.d. n.d. o0o86 <o.o0 1"9

n.d. 0ox15 n.d.

1o. 68-I 2-5 Io.I 0-25 0-15 0-15 <o-o0 0-27 I8-4 0o069 n.d.

II. 85-8 12-7 0-74 0*37 n.d. 0.37

<0o-o0 0'021 n.d. 0-027 n.d. 12. 75'5 1I7 7-0 o019 o018 o 11 <oo01 o019 14'9 0-043 n.d. 13.

88"7 7'4 0oI9

0.35 n.d. 0oII <o-o01 19 1-4 0'020 n.d.

14. 91"7 7.4 0-44 0-29 n.d. 0-072 <o0o0 0o098 n.d. 0-oo0048 n.d. 15. 88-2 11-3 0-035 0-29 n.d. o-I6 <o-o0 0o029

n.d. oo011 n.d.

I6. 92'4 7-2 o.19 n.d. n.d. o-15 <0oo1 o-036 n.d. o0o03 n.d. 17. 79'4 I "-4 9o0 n.d. oog92 o-033 <o-o o 0-047 n.d. o011 n.d. 18.

91.7 8 I 0-o69 n.d. n.d. oo017 <oo01

o'o67 n.d. oo10o n.d.

19. 91"9 7-8 0o030 o-18 n.d. 0-ox7 <o0oI 0-o28 n.d. 0-032 n.d. 20.

89"2 9'3

o.o89 1- I n.d. o- 8 <oo 0o-032 n.d. o0I70 n.d.

21. 93'2 5'3 0o 48 0o59 n.d. 0-29 <01 o 0 13 n.d. 0o045 n.d. 22. 90o8

0'o12 059 4'4 0-29 3'5 <oo01 o 019 n.d. 0-o53 n.d.

23- 95"2

n.d. 0o038 4'3 n.d. o-oi6 <o-oI 0o43 n.d. 0-029 n.d.



24. 95-8 <o0 I o*0o6 3'7 n.d. o0o19 <oo0 o039 n.d. 0ooo46

n.d.

25. 93'7 5-6 o014 0.29 n.d. o0*15 <oo01 0o36 n.d. 0o052 n.d.

26. 92'5 6-6 0o24 0'35 n.d. o-18 <oo01 0-075 n.d. o0-45 n.d.

27. 94"4

4'7 o.012 0"45

o12 0o053 <oo01 0o076 n.d. o-oio n.d.

28. 92-3 5'9 1. I 0o40

n.d. 0o.7 <o-o0 0o054 n.d. 0o025

n.d.

29. 90 2 9 4 n.d. nn.d. .d. o-03 <0 o o- 44 n.d. o o 14 n.d.

30. 87 9 1I 2 o 83 n.d. n.d. 0o050 < oo n.d. n.d. o o83 n.d.

31. 95-8 3-6 0o29 o-24 n.d. 0o067 <o0o0 n.d. n.d. o"oi6

n.d.

32. 699 2'2 9"3

0-30 0-26 0"

-3 <o0ox o033 17'5 o-o68 n.d.

33. 89 3 8- 5 1. 2 0o40 n.d. 0o21 n.d. 0o27 o 20 0o022 n.d.

34. 8I-7 1 57 1 9 n.d. n.d. o 39 n.d. o 35 n.d. o0059 n.d.

35. 88- 9 x "o o *026 n.d. n.d. o0 020 < oo007 o0 028 n.d. o-o 12 n.d. 36. 91'7

7"6 n.d. 0o40 n.d. o-090 n.d. o i8 n.d.

o'ox4 o'oI8

37- 900o 9 8 n.d. n.d. n.d. n.d. n.d. o0055 n.d. o 086 n.d. 38. 92 5 o 67 37 6 I9 n.d. o 65 n.d. 0 33 0 31 0o055 n.d.

39. 95.0 3 -6 0 30 o74 n.d. o -19 n.d. o020 n.d. o0020 n.d. 40. 90'9

6"I '4 I. I d. n o I8 n.d. o-I8 n.d. o0 14 n.d.

41. 87"

7 Io0 3 o"

40 I.I

65 n 0 20 o 1g 0 034 n.d. 0- 24 n.d.

42. 87'5 I036 o0Io o*55 n.d. o'073

n.d. o0055 o0079 o0oi7 n.d.

43. 90"8

7'4 o698 o 042 n.d. o.o84 n.d. o 14 n.d. o00oo77 n.d.

44. 94 -8 4 6 036 n.d. n.d. o 090 n.d. o 14 n.d. o oo84 n.d.

45. 94'3 4-6 0o63 o0 15 73 n 0o059

0-023 o-I6 n.d. o0-o3 n.d.

46. 89- 4 8 9 o 80 1oo n.d. o -29 n.d. o030 n.d. 0o 22 n.d.

47. 85"o

13-o n.d. o-68 n.d. o'78 <o.-o o'56 n.d. o-oo00054 n.d.

48. 92'5 2 I 1-I 2 8 I. 2 o034 n.d. 0o 24 n.d. 0o055 n.d.

49. 95'5 3 9 o097 o-29 n.d. o 022 0 o oI3 o020o n.d. o00oo96 n.d.

50. 88-8 8-4 o'48 2*2 n.d. n.d. n.d. o0-3I n.d. 0o047 n.d.

5i. 87 '3 x 7 0024 036 n.d. 0 35 <00 ox 0 19 n.d. 0-042 n.d.

52. 9 I 8 7 n.d. n.d. n.d. 0-052 n.d. o 087 n.d. 0o-o3 n.d.

53. 87'5 99 0o'74 1-2 00o53 0o38

<o-o o0o038 n.d. 0o15 n.d.

54. 96 -o 0 -32 055 2'3 o o94 0oI-2 <o oI 0 64 n.d. 0-036 n.d.

55. 84.8 13.17 o-I9 I0o n.d. o'047 <o00 0o37 n.d. o- I2 n.d.

56. go"9 8-3

0-22 0o28

<o0o04 o0I3 <o-0o 0o090 n.d. 0-038 n.d.

57. 873 12.2

o 7 15 n.d. n.d. o -I6 n.d. o-093 n.d. o-067 n.d.

58. 87"7 io.6

0'057 o0054 n.d. o0-38 o<oi6 0-038 n.d. o0oI5 n.d.

59. 8i 6 076 o- 26 o 33 n.d. 0 038 n.d. o I5 n.d. o0o24 n.d.

60. 89 4 1- o n.d. n.d. n.d. o 29 n.d. o I9 n.d. o. 064 n.d.

61. 94'2 5'4 o'o85

n.d. n.d. 0o23 n.d. o I2 n.d. 0o027

n.d. 62.

93"6 5"6 o-19

o'32 n.d. 0-23 ooi6 oo-029 n.d. o0034 n.d.

63. 93-8 5'3 o019 0"32

oi8 0.060

0o020 o0oI2 n.d. o0o5I

n.d.

63a. 93'9 5'4 o0-5 o'33 o. II o-o65 oo023 0-021 n.d. 0o043 n.d.

64. 87"5

II-8 0o34 n.d. n.d. o-I6 o024 o0-I3 n.d. o0023 n.d.

65- 94'4 4'9 n.d. o'57 n.d. oo055 o*oI7 oi 8 n.d. o000oo78 n.d.

66. 89 -9 o o n.d. n.d. n.d. o-o 7 n.d. o-o19 n.d. 0oo03 n.d.

67. 980o i-8 o I n.d. n.d. 0-042 ox.8 0ooI9 n.d. 0-025 n.d. 68. 8935 Io.0 n.d. 0o37 n.d. 0'027 o-o2o -o605 n.d. o00oo73 n.d.

69. 89-2 Io'4

0o041 0o20 n.d. o-or5 0o029 o-I3 n.d. o-oo87 n.d.

70. 91 3 8"3

o0o41 0-22 n.d. 0.043

o*oi8 0"045

n.d. 0-022 n.d.

71. 89-9 Io.o n.d. n.d. n.d. o0022 o-oI6 o.084 n.d. o0o02 n.d.

72. 72"4

II I 16 I 0"33

n.d. 0o068 o'oI7 0"030

n.d. -oo1I n.d.

73. 8o0.4 183 o0085 o'34 n.d. 0-027 0-021 0o79 n.d. 0-023 n.d.

74. 89-2 10o2 070o o'41 n.d. 0-043 n.d. oI1o n.d. 00-o4 n.d.

75. 88-7 o107 o-o85 0o41 n.d. 0o033

0oo22 o0o57 n.d. o-oi6 n.d.

76. 84"7

14'9 n.d. n.d. n.d. n.d. o0oI9 o41I n.d. 000oo23 n.d.

77. 8702 12"4

o0126 0o23 n.d. o0043 oo012 o0054 n.d. o020o n.d.

78. 92'2 7 '7 o098 n.d. n.d. o0043 n.d. 0o024 n.d. oo020 n.d.

79- 96-I 3'7 o0o8o n.d. n.d. o0o60 0o023 0ooI4 n.d. o-oi6 n.d.

80. 926 6"7 o"

15 o023 n.d. oI*14 n.d. o0 13 n.d. o0039 n.d.

81. 90-3 9'3 o 10 o0-9 n.d. 0o0o5 n.d. 0ooI3 n.d. oo013 n.d. 82.

92"4 7-o o0044 o040 n.d. o0040 oo023 o0038 n.d. o0oo92 n.d.

83. 86-o I3-2 022 0o-36 n.d. o0071 0"024

o0 I1 n.d. o-051 n.d.

84. 83"3 i6"7

n.d. n.d. n.d. o-oi8 n.d. o0039 n.d. o000oo43 n.d.

85. 89"5

I0o2 o209I n.d. n.d. 0oI2 n.d. o*oI3 n.d. o0038 n.d.

86. 9I'5 8-3 n.d. n.d. n.d. o0o84 o0030 o0ox5 n.d. o0025 n.d.

87. 90"6

9'3 n.d. n.d. n.d. o0o6I n.d. 0-028 n.d. 0-0075 n.d. 88. 9gI o 7 8 o 86 n.d. n.d. o o41 n.d. o I n.d. o*037 n.d.



89. 87"6 I2.2

0o060 n.d. n.d. o011 n.d. o0060 n.d. o01o7 n.d. 90o. 86-8

12"2 0-78 0-14 n.d. 0oo0087 0-017 0-022 o012 02onI n.d.

91. 94'3 4*9 0'055 o021 n.d. 0o-043 n.d. 0o49 n.d. o-026 n.d. 92. 85-8 13'5 0o24 o026 n.d. 0o040 n.d. o-082 n.d. o.-095 n.d. 93. 94'3 5- 0o14

o"16 n.d. 0-045

0"022 o-oi6 o-.8 0-050 n.d.

94. 93"0

6-3 o0 12 0"30

n.d. o"I6

0-022 0-033 n.d. 0-028 n.d. 95. 93-8 5'3 0-057 o-63 n.d. o-059 o0026 o-12 n.d. o0020 n d. 96. 90o2 9-2 0o24 o 14 n.d. 0-041 001o5 o069 n.d. 0-023 n.d. 97.

89"9 9'7 0o093 o014 n.d. 0-048

0"022 0-042 n.d. o0o04 n.d.

98. 76.I 12-7 10-5 0-36 o"16

0-043 oo015 0-027 n.d. o-18 n.d. 99.

86"3 9-8

3"4 0-22 0-o85 0o060 o-o16 0-040 n.d. o012 n.d.

Ioo. 95'4 o"18

0-040 I-4 n.d. o0*2 0o020 2-5 0-30 0oo28 n.d.

10o1. 920o 6-9 o065 o- 16 n.d. 0-097 0-o15 0o056 n.d. o0o10 n.d. iola. 92'4 6-9 0-083

o.16 n.d. 00o67 0.022 0.21 o0 1 0-024 n.d.

xo2. 92'9 6-2 o019 0"34 0-078 o-18 0-014 0-036 n.d. 0o066 n.d.

103. 89"3

Io'5 n.d. n.d. n.d. n.d. 0-028 n.d. n.d. o- I8 n.d. 1o4. 92'4 4-8 I18 0"57 o013

0"090 o01oI 0o078 n.d. o-o16 n.d.

105. 89'2 9'9 0o31 0o29 n.d. o-16 o-15 o-037 n.d. o-079 n.d. io6. 92-2 7-0 o0-14 027 n.d. 0-37 oo015 0-023 n.d. 0-026 n.d. 107. 90 7 9 o o o8o n.d. n.d. 0o031 0-024 o0-093 n.d. o-087 n.d. io8.

91"3 8-6 0o040 n.d. n.d. 0-013 o-oi8 0-o014 n.d. 0-078 n.d.

109. 98 o <o01 n.d. n.d. n.d. I 8 <01 o o0o014 n.d. 0o021 n.d.

Ilo. 93'3 6-0 0-26 n.d. n.d. 0-o18 <o-oI 0-13 n.d. 0-27 n.d. III. 914 7'4 0-15 0-23

0"45 0-027 o-oi6 0-23 n.d. 0-o86 n.d.

12. go906 9.0 n.d. n.d. n.d. n.d. oo014 0o20 n.d. o0 I.7, n.d.

113- 97'9 n.d. o0o71 o067 n.d. I 1 o0017 oo020 n.d. o-o18 n.d. 114. 91"3 7'7 0o24 0-29 0-13 0-30 0o019 0-047 n.d. 0-o47 n.d. 115. 91"9 7-2 0o

10 0*34 n.d. n.d. o-oi8 oo056 n.d. 0"44

n.d.

116. 91"8 8-o 0"022

0o*2 n.d. nd. 0-014 0o068 n.d. 0-027 n.d. 117. 96-5 3I o 0'3 o0II n.d. 0o077 <0o01 o-o65 n.d. 0o046 n.d. 118. 94'4 4'9 0o'37 o014 n.d. o014 <o01o o 0034 n.d. 0-025 n.d. IIg19. 908 6-8 0-50 062 0-48 0-033 0-o013 0-24 0-42 0-076 n.d. 120. 92-6 70 00o o83 o-16 n.d. 0-094 0-ox3 0-ox4 n.d. 0-046 n.d. 121. 94-8 4'7 0-059 0-19 n.d. o- I 0-013 0-039 n.d. 001I2 n.d. 122. 83'9 15-6 o0082 0o34 n.d. oo023 o-o16 0-048 n.d. o-oo84 n.d. 123. 87-4 II'7 0'31

0o22 n.d. o-I8 o-oi6 0-12 n.d. o-oIg n.d.

124. 87- I I2.3 o0o6o o030 n.d. oo045 <o01 o0035 01 5 o-027 n.d. 125. 91.3 7-1 0-83 0-17 oo080o 0-069 0012

0o10 n.d.

0"33 n.d.

126. 91g5 8-I o 19 o I8 n.d. 0-039 o-oi8 0-043 n.d. o-oio

n.d. 127. 910o 8-6 0o020 0-28 n.d. 0-032 <oo01 0.062 n.d. o-oo67 n.d. 128. 88-o I1-6 0oI2 n.d. n.d. 0-050 0o012 0'033 o-18 0.OI9 n.d. 129.

93"9 6-o 0-034 n.d. n.d. oo041 o0orI n.d. n.d. 0o0092 n.d.

130o. 89-o 3'4 o065 o075 3-2 0-050 0-028 2-5 0'33 0-23 n.d. 131. 98-8 0-43 o o69 n.d. n.d. 0-049 0-026 0-43 n.d. o I6 n.d. 132. 90'9 8-4 o0o67 0-48 n.d. 0-041 0-012

o* I n.d. o0oo85 n.d.

n.d. - none detected

155

"SOME ANCIENT METAL BELTS "-A RETRACTION AND A CAUTIONARY NOTE

By P. R. S. Moorey

Since writing this paper, which appeared in Iran V (1967), pp. 83 ff., further reading has brought to my attention evidence which indicates that the silver-gilt belt published in section III of the paper, with Pls. II-III, is definitely not an ancient object. In 1961, in a Russian paper previously unknown to me, A. A. lessen1 demonstrated convincingly that the well-known " Maikop " belt in the Hermitage, Leningrad, and a very similar belt, said to be from Bulgaria, now in the British Museum, were both made in Odessa in the first decade of this century, probably for the dealers Sh. and L. Gokhman who were responsible for the creation of the notorious " Tiara of Saitapharnes " sold to the Louvre in 1896.2 In this paper lessen referred to another belt, exactly answering the description of the one I published, which was offered unsuccessfully to various museums in Berlin in 1923-24. It was then condemned as a Gokhman forgery. lessen was unaware of the belt's present whereabouts. It is now clear that sometime soon after 1924 it passed into the de Walden Collection. Significantly it does not appear in Joubert's catalogue of the collection published in 1923,3 although other objects which may now be recognized as Odessa forgeries do. These include a third belt in the style of the " Maikop " and " Bulgarian " belts and two zoomorphic roundels, like one in the British Museum.4 Joubert published the belt and roundels as " Burmese ".

Two points, for which the belt was originally condemned, were clear to me when I wrote the paper, but I then failed to see what they implied, being innocent of the sources, skill and ingenuity of the Odessa forgers.

I. The use of the cicada motif and the mounting of the stones (probably almandines rather than garnets) are on present evidence difficult to date before the fourth century A.D., whereas the bull motif is an Achaemenian one, unlikely to have been used after the late fourth or early third centuries B.C. I erroneously attributed the belt therefore to a transitional phase in " Sarmatian " art.

2. The bull plaque and swan hook derive, as I appreciated, exactly from objects illustrated by N. Kondakof and others in Antiquites de la Russie Miridionale (Paris 1891), figs. 25 1, 334. lessen showed that this, and related publications, were the primary source for many of the motifs used by the Odessa forgers.

In so far as it requires revision of the relevant parts of standard works by Rostovtzeff, Borovka and Salmony5 lessen's paper is indispensable to students of " Sarmatian " art; it is also of considerable significance for anyone at all concerned with the problems of forgery and detection. In this respect it is

particularly interesting to notice, with informed hindsight, that the three belts made in segments incorporate stylistic traits of the " Art Nouveau " style current among Odessa jewellers at the time of their manufacture.

As lessen pointed out, this is but a fraction of the Odessa forgers' output and I trust that in calling attention to them here in an English publication I may at least have prevented others from even stumbling where I so blindly fell.

1" Maikopskii poias " in Arkheologicheskii Sbornik II (Leningrad I96i), pp. 163 if.; I am most grateful to Dr. Jaromir Milek for translating this paper for me; see also K. Jettmar, Art of the Steppes (1967), p. 64.

2 For the Tiara's chequered history, see Vayson de Pradenne, Les Fraudes en archeologie prihistorique (Paris 1932), PP. 519 ff.

' F. Joubert, The Collection of Arms and Armour formed by Lord Howard de Walden (London 1923).

4 F. Joubert, op. cit., no. 31 with plate; Ashmolean Museum Report

qf the Visitors (1963), pl. VIa, b. For the British Museum roundel, see 0. M. Dalton, The Treasure of the Oxus (2nd edn.

1926), fig. 36; this and both the de Walden roundels may be based on one without inlay found at Krasnokutsk: E. Minns, Scythians and Greeks, fig. 57.

6 M. Rostovtzeff, Iranians and Greeks in South Russia (1922),

p. 134, pl. XXV.i ; The Animal Style in South Russia and China

(1929), PP- 46 ff., pl. XIII; G. Borovka, Scythian Art (1928),

pl. 46B, pp. 58-9; A. Salmony, ESA XI (1937), PP- 91 ff.

157

HINWEISE UND ANMERKUNGEN ZU EINIGEN SASANIDISCHEN MONUMENTEN

By Klaus Schippmann

Nachfolgend m6chte ich zuerst auf zwei bisher unbekannte, m.E. sasanidische Monumente hinweisen. AnschlieBend folgen einige Anmerkungen zu zwei schon lange bekannten sasanidischen Denkmilern, die vielleicht dazu beitragen, die Datierung bzw. Identifizierung dieser beiden Anlagen noch einmal zu tiberprtifen.

I

Wihrend eines Besuches von Persepolis Anfang Oktober 1964 entdeckte ich zwischen dem Siidende der Palastterrasse und dem unvollendeten Grab Darius III. am FuBe des Berges drei stelenartige Gebilde (P1. I). Der Abstand zwischen der obersten und der untersten Stele betrigt ca. 12 m. Die oberste ist 1,85 m hoch, die mittlere 1,90 m und die unterste 1,82 m. Sie haben eine viereckige Basis, und die Schafte sind abgekantet. Alle drei stehen auf der Erde und sind nicht eingegraben.

M.E. handelt es sich um Feueraltaire. Ich glaube allerdings nicht, daB sie achimenidisch sind, da die uns bekannten Altire aus dieser Zeit doch wesentlich anders aussehen. Sie k6nnten vielmehr sasanidisch sein; es fehlen allerdings die Deckplatten. Aufgrund der Analogie mit Darstellungen auf sasanidischen Miinzen diirften sie in das 4. oder 5. nachchristliche Jahrhundert zu setzen sein.1

II

Am 25. Sept. 19642 entdeckte ich bei einem Besuch des Dorfes Nimvar, 15 km hinter Dalidjan an der StraBe, die fiber Gulpa-igan nach Isfahan fiihrt, einen grofen Pfeiler, der von den Einheimischen Takht-i Jamshid genannt wird (Pls. II & III). Wenn man von Dalidjan kommt, muB man das rechter Hand liegende Dorf durchqueren; der Pfeiler befindet sich dann am Dorfrand inmitten von Feldern.

Ich besuchte den Ort, nachdem ich vorher das sasanidische Feuerheiligtum von Atashkiih,3 6 km entfernt von Nimvar, aufgesucht hatte. Ich fand es merkwiirdig, daB Godard, der die Anlage in Atashkiih zum erstenmal ausftihrlich beschrieben hatte, sie mit dem Bericht von IHamd-Alldh Mustaufi fiber Nimvar in Verbindung brachte, ohne daB er gleichzeitig Nimvar nach Alterttimern untersucht hitte.4 Mustaufi5 berichtet im 14. Jahrhundert iiber Nimvar von einem Palast, dessen Uberreste noch zu sehen seien und weiter auch von einem Feuertempel. tVber letzteren findet sich auch schon bei Qummi6 ein Hinweis. Godard ist der Ansicht, daB mit den Angaben Mustaufis fiber Palast und Feuertempel die Anlagen in Atashktih gemeint seien. Es stimmt, daB sich dort ein Feuertempel befindet, von dem Vorhandensein eines Palastes kann m.E. aber nicht die Rede sein. Bei dem Bau, den Godard7 als sasanidischen Palast bezeichnet, handelt es sich um einen einfachen Lehmziegelbau mit Innenhof. Die Anlage ist von Hiusern der Einheimischen umgeben. Zum Teil liegen Stallungen unterhalb des Baues. Ich bin in diese hineingegangen, konnte aber keine ailteren Fundamente feststellen. Wenn dieser Bau sasanidisch sein soll, dann k6nnte man fast jeden Lehmziegelbau in Persien den Sasaniden zuschreiben. Siroux,8 der schon 1938 Atashkiih aufgesucht hatte und 1963 nochmals dort

1 Diesen Hinweis verdanke ich einer Mitteilung von Prof. R. G6bl, Wien.

2 Ein weiteres Mal besuchte ich Nimvar am 13.5.1968.

S s. A. Godard, Athdr-e Iran III (i938), S. 32-39; M. Siroux, Syria 44 (1967), S. 53-71.

4Jedenfalls schreibt G. nichts von einem Besuch in Nimvar;

andernfalls hitte er sehr wahrscheinlich den hier besprochenen Pfeiler entdeckt.

5 Nuzhat al-Quliib (Ed. G. Le Strange), S. 73. 6 Ta'rikh-i Qumm, S. 74, I5 if.

SAthdr-d Irdn III (1938), S. 32.

8 Syria 44 (1967), S. 53 ff-


war, druickt sich daher wohl mit Recht sehr vorsichtig aus, indem er sagt, daB man die Anlage als

sasanidisch datieren k6nnte und dann fortfahrt, daB es sich in diesem Falle aber nicht um einen Palast, sondern um eine kleine Kaserne handeln wurde. Siroux, der ansonsten Godards Ansicht folgt und den Bericht von Mustaufi fiber Nimvar auf

Atashkfih bezieht, ohne daB er Nimvar selbst aufgesucht hiitte,9 muB daher den sasanidischen Palast irgendwo anders in Atashkfih suchen. Nach ihm10 lag der Palast ca. 500 m n6rdlich des Feuertempels am FuB eines Berges, der bei den Einheimischen Atash Kiih heiBt. Nach Siroux gibt es dort Reste von Baulichkeiten, sie scheinen aber sehr geringfiigig zu

sein. Denn sonst ist es nicht zu erkliaren, daB er kein Foto oder keinen Plan von diesen Ruinen bringt, widhrend sein Aufsatz von dem Feuertempel dagegen zahlreiche Plane und Fotos aufweist.1n

Zusammenfassend ist festzustellen, daB m.E. auch Siroux kein Nachweis gelungen ist, daB in Atashkiih ein sasanidischer Palast vorhanden war. Selbst aber dann fehlt es an einem Beweis daffir, daB Mustaufis Beschreibung von den Anlagen in Nimvar nicht auf diesen Ort, sondern auf das 6 km entfernte Atashkiih zu beziehen ist.

Ich vermute daher, daB der Pfeiler in Nimvar zu einer der von Mustaufi genannten Anlagen

geh6rt. Fiir einen sasanidischen Feuertempel, einem Chahir Tdql2 also, erscheint mir der Pfeiler zu

groB, so daB er Bestandteil des Palastes gewesen sein diirfte. Die Bauweise des Pfeilers, Steine verschie- dener Gro6fe im M6rtelverband ist sasanidisch. Allerdings sind die Riickspriinge und die Zierleiste

(P1. IIIb) nicht ausgesprochen ein Kennzeichen sasanidischer Bauweise. Dr. Kleiss13 meinte zu diesen Einzelheiten,14 es seien gewisse Ahnlichkeiten mit mongolischen

Pfeilern auf dem Takht-i Sulaiman in Aserbaidschan zu erkennen; er wies aber gleichzeitig darauf hin, daB dort auch sasanidische Pfeiler in mongolischer Zeit wieder verwandt worden seien. Erwdihnt sei

noch, daB Dr. Triimpelmann,15 der Nimvar in Oktober 1964 besucht hat, ebenso wie ich glaubt, daB es sich um einen Pfeiler des von Mustaufi genannten Palastes handelt. Hinzuweisen waire auch gerade im Zusammenhang mit dem Takht-i Sulaimdn auf Mustaufis16 Bericht tiber die dortigen Anlagen. Er

erwdihnt ausdrticklich, daB der ailtere Palast des Konigs Khusrau von dem Mongolenherrscher Abaqa

Khan wiederhergerichtet sei. Sollte etwas Ahnliches in Nimvar geschehen sein, so liige doch zumin- destens die Vermutung nahe, daB Mustaufi auch fuir Nimvar einen solchen Hinweis gegeben hiitte.

Anzeichen ffir das Vorhandensein eines Feuertempels konnte ich in Nimvar nicht feststellen. Inmitten des Ortes gibt es noch einen gr6Beren Lehmziegelbau, der bei den Einheimischen Qal'a-i Jamshid heiBt. Da er aber um- und iiberbaut ist, kann man nicht mehr viel erkennen.

Trotz gewisser Bedenken - Eigentiimlichkeiten des Pfeilers (Riickspriinge und Zierleisten), Nicht- Auffinden eines Feuertempels - komme ich abschlieBend zu dem Ergebnis, daB der Bericht Mustaufis auf Nimvar zu beziehen ist, und daB der dort noch anstehende Pfeiler zu dem von ihm genannten Palast geh6ren diirfte.

III

Hier m6chte ich auf die unter dem Namen Takht-i Nishin bekannte Terrasse in Firfizabad, Fars hinweisen (P1. IVa). Die neuere Forschung vertritt allgemein die Ansicht, daB die Terrasse sasanidisch sei und auf ihr ein groBer Chahdr Taq mit einer lichten Weite von 16, Io m gestanden habe.7 Demge- geniiber wird oft vergessen, daB es zahlreiche Stimmen gibt, wenn auch etwas iiltere, die ffir eine

9 Jedenfalls berichtet auch er nichts von einem solchen Besuch.

10 a.a.O. S. 68/9.

11 Bezeichnend ist auch, daB S. bei einer der Abbildungen (pl. VII) beziiglich der Lage des Palastes vermerkt " En #r 5,

emplacement pr6sum6 du palais ". 12 Dies ist die gewohnliche Form des Feuertempels in sasanidi-

scher Zeit.

13 2. Direktor des Deutschen Archaiologischen Instituts in Teheran.

14 K. hat allerdings Nimvar nicht besucht, sondern nur Fotos

gesehen.

15 Ehemaliger Assistent am Deutschen Archaiologischen Institut in Teheran.

16 Nuzhat al-Quliib (Ed. G. Le Strange), S. 69. 17 S. A. Godard, Athr-i lrdn III (1938), S. 25/6; L'Art de l'Iran,

S. 230; K. Erdmann, Das Iranische Feuerheiligtum, S. 47/8; R. Ghirshman, Bulletin de l'institut franfais d'archiologie orientale

46 (1947), S. 24; L. Vanden Berghe, Archiologie de l'Irdn ancien, S. 47.

HINWEISE UND ANMERKUNGEN ZU EINIGEN SASANIDISCHEN MONUMENTEN 159

achimenidische Entstehung der Terrasse eintreten.s1 M.E. kann man diese Ansicht nicht einfach mit der Behauptung abtun, es handele sich genau wie im Falle des Feuertempels von BishipiIr um das " Nachleben achatmenidischer Formen in friihsasanidischer Zeit "19 oder um das " Weiterleben parthischer Bauweise ".20 Ohne Zweifel weist der Tempel von Bishaptir Ahnlichkeiten in der Bauweise auf, aber, und das wird hatufig tibersehen: bei der Terrasse von Firfizab•d sind im Gegensatz zu Bishaptir Eisenklammern, also ein typisch achimenisches Kennzeichen, zur Verklammerung der Bl6cke benutzt worden. Diese Tatsache wird schon 1813 von Kinneir21 und 1840 von Flandin-Coste22 berichtet und auch noch spaiter von Ghirshman23 bestditigt.

Daneben gibt es einen weiteren Hinweis fiir eine eventuelle achitmenidische Datierung: es handelt sich um die von Flandin-Coste24 in unmittelbarer Niihe der Terrasse gefundene Stiulentrommel achitmenidischen Stils mit 38 Kanneluren. Herzfeld25 spricht 1924 sogar von zwei Saiulentrommeln; die eine habe sich bei dem in der Niihe liegenden Im~mz~da befinden, die andere sollte der Saulat al-daula in sein SchloB in Firfizb~id iiberffihren lassen. Abgesehen von diesen beiden Forschern niemand mehr diese Saiulentrommel(n) gesehen zu haben; ein Foto existiert tiberhaupt nicht. Im September 1967 bei einem Besuch in Firfizbitd habe ich dann die Stulentrommel in unmittelbarer Nthe des Imamzada wiedergefunden (P1. IVb). Ich halte sie fUir achimenidischer Herkunft. Hinzu- kommt schlieBlich die Form der Anlage: eine Terrasse mit mehreren Aufg~tngen ist fuir die sasanidische Zeit zumindest ungew6hnlich, nicht aber fiir die der Achatmeniden. Im Gegensatz zu Bishpilir haben wir schlieBlich zwei Berichte arabischer Geographen, niimlich von Ibn al-Balkhi26 und IHamd-Allih Mustaufi,27 die beide iibereinstimmend angeben, daB sich schon zur Zeit Alexanders des GroBen an dieser Stelle eine Stadt befunden habe.28

Die schon genannten Gegenargumente von Erdmann und Godard sind daher m.E. nicht iiberzeu- gend. Die Ansicht Godards - Weiterleben parthischer Bauweise - ist dabei am wenigsten stich- haltig. Abgesehen davon, daB Godard kein Beispiel nennt, kennen wir in Iran auch kein als sicher parthisch zu datierendes Bauwerk, das in der Art der Terrasse von Firtizdib•d errichtet worden ist.

SchlieBlich sei noch auf das von allen Vertretern einer sasanidischen Datierung verwandte Argument eingegangen: nach dem Bericht Ibn al-Balkhis29 - Anfang des 12. Jahrhunderts - soll sich auf der Terrasse ein miichtiger Kuppelbau erhoben haben. Unter Verwendung der von Flandin- Coste festgestellten MaBe kommt man dann allgemein zu dem Ergebnis, es habe sich bei dem Kuppel- bau um einen Feuertempel in Form eines Chahar Thq mit einer lichten Weite von 16,10o m gehandelt.30

Selbst wenn man unterstellt, daB Ibn al-Balkhis Bericht und die daraus gezogenen SchluBfolger- ungen stimmen, so ist damit noch keineswegs gesagt, daB die Terrasse als solche auch sasanidisch ist. Ich m6chte einen Gedanken Wachtsmuths31 wieder aufgreifen: ist es nicht m6glich, daB in sasanidischer Zeit eine teilweise Erneuerung der Anlage stattgefunden hat, in der Form etwa, daB man auf eine achlimenidische Terrasse nach einigen kleinen Umbauten einen sasanidischen Chahar Tdq errichtet hat ?

Abgesehen davon bestehen gewisse Zweifel an der Richtigkeit des Berichtes von Ibn al-Balkhi.32 Man

18 s. E. Flandin-P. Coste, Voyage en Perse I (Texte), S. 36 ff.; G. Perrot-Ch. Chipiez, Histoire de l'art dans l'antiquiti V, S. 646; F. Oelmann, Arch.Anz. (1921), Sp. 274/5; F. Wachtsmuth, AfO XVI (1952), S. 102; Survey I, S. 311/12; A. Stein, Geographical Journal LXXXVI (1935), S. 491; Iraq III (1936), S. 12o. In dem letzten Aufsatz geht Stein zwar ebenfalls von einer vor-sasanidischen Datierung aus, spricht aber dann davon, daB auf der Terrasse ein Kuppelbau im Sinne eines Chahar TI~q gestanden habe. Dies ist m.E. ein Widerspruch, da wir aus vor-sasanidischer Zeit keinen Chahdr Tdq kennen.

19 So K. Erdmann, Die Kunst Irans zur Zeit der Sasaniden, S. 20.

20 Athdr-ed rdn III (1938), S. 26. 21 A Geographical Memoir of the Persian Empire, S. 68. 21 Voyage en Perse I (Texte), S. 37. 13 Bulletin de l'institutfranfais d'archeologie orientale 46 (I947), S. 14.

s4 a.a.O. S. 37. 14A

25 ZDMG 80 (1926), S. 255. 26 Fdrs-ndma (Ed. G. Le Strange - Nicholson), S. 137, 11 ff.

27 Nuzhat al-Qulfib (Ed. G. Le Strange), S. 116/17. 28 Fir eine durchgehende Besiedlung dieser Gegend sprechen

auch die Keramikfunde von Stein aus chalcolithischer Zeit (Iraq III (1936), S. 128) und von Ghirshman, a.a.O., S. 20, der im Osten der Ebene von Firfizdbdd einen Steinsarkophag entdeckte, der aus parthischer Zeit stammen soll.

29 a.a.O. S. 138, 11 ff. 30 s. Anm. 17.

31 AfO XVI (1952), S. 102/3.

32 Ich verweise auf die von mir demnichst erscheinende Arbeit " Die Iranischen Feuerheiligtumer ", wo ich u.a. ausfuihrlich zu den arabischen Berichten iiber die verschiedenen Ruinen in FirfizAbdd Stellung nehmen werde.


darf davon ausgehen, daB er den von ihm geschilderten Bau nicht selbst gesehen hat. Denn keiner der islamischen Geographen oder Historiker, die von diesem Ort berichten, sei es vor der Zeit33 Ibn al- Balkhis oder danach,34 erwihnen einen solchen Kuppelbau auf einer Terrasse, zumal noch mit genauen MaBangaben. Bei griindlichem Studium der arabischen Berichte ergibt sich weiter, daB diese sehr unklar sind. Keiner, auch nicht der von Ibn al-Balkhi, spricht klar und eindeutig von zwei Anlagen, namlich von dem heute noch ca. 33 m hohen Turm und der etwa 0oo m entfernten hier behandelten Terrasse. Es fillt weiter auf, etwa in dem Bericht von al-Istakhri35 und die auf diesen zuriickgehenden Angaben von Ibn

.Hauqal36 und Ydqfit,37 daB vielmehr die Eigenschaften beider Anlagen miteinander

vermengt werden. So wird in diesen Berichten von einer Estrade (Tirbdl) gesprochen, was auf die Terrasse Takht-i Nishin hindeutet. Die dann aber sofort anschlieBende Bemerkung, daB diese Estrade so hoch war, daB man von dort die ganze Stadt und ihre Gaue iiberblicken konnte, diirfte sich doch aber wohl auf den turmartigen Bau beziehen. Zu dem Bericht Ibn al-Balkhis ist schlieBlich noch zu

erwihnen, daB er im Zusammenhang mit dem Kuppelbau von einer Leitung spricht, die Wasser auf die Spitze dieses Baues heraufleitet. Diese Anlage wird auch von al-Istakhri und anderen arabischen Autoren erwahnt, dort aber spricht vieles dafuir, daB damit eine Leitung auf die Spitze des turmartigen Baues gemeint, ist, was m.E. auch plausibler ist. Es muB auch noch festgestellt werden, daB Ibn al- Balkhi nicht von einem Feuertempel im Zusammenhang mit dem Kuppelbau spricht.

Ghirshman38 fiihrt zur weiteren Begriindung an, daB das von Ibn al-Balkhi genannte MaB von 75 Gaiz (=78 m) nur ca. 4 m von dem von Flandin-Coste angegebenen LangenmaB von 82,10 m

foir die

Terrasse abweiche. Dem ist entgegenzuhalten, daB Ghirshmans Umrechnung I Gaz = I,04 m39 umstritten ist. Hinz40 ist der Ansicht, daB diese Umrechnung erst neueren Datums sei; er teilte mir ferner mit, daB im I I. Jahrundert I Gaz = 0,62 m betragen habe, d.h. das MaB for den Kuppelbau wiirde dann nicht 82,IO m, sondern nur 46,50 m sein. Im iibrigen sagt Ibn al-Balkhi, daB " die

Liinge der vier Wande dieses Domes bis unter die Kuppel 75 Giiz betrigt ", danach handelt es sich also um das MaB der Hohe des Baues. Ich gebe allerdings zu, daB das Wort "Lange " in diesem

Zusammenhang st6rt, es haitte besser " H6he " geheiBen. Ich m6chte aber bezweifeln, ob diese Ungenauigkeit im Ausdruck AnlaB genug ist, die Angaben Ibn al-Balkhis einfach im Sinne Ghirshmans umzudeuten. Dem steht zumindest auch der zweite Teil des Satzes " bis unter die Kuppel " entgegen.

Ich bin daher der Ansicht, daB die Annahme, die heute unter dem Namen Takht-i Nishin bekannte Terrasse sei sasanidisch und aufihr habe ein Feuertempel in Form eines riesigen Chahar

.Taq gestanden,

nicht tiberzeugend ist; manche Einzelheit spricht vielmehr daffir, daB zumindest die Terrasse achi- menidischen Datums sein k6nnte. Es bleibt zu hoffen, daB in absehbarer Zeit eine Grabung eine

endgiiltige Kliarung herbeiftihren wird.

IV

Unter diesem Punkt m6chte ich die sasanidische Anlage von Qal'a-i Yazdigird in der Niihe von Sar-i Pul Zuhab dem alten IHulwin behandeln. Wir verdanken es den Untersuchungen von E. J. Keall41 in den Jahren 1964/65, daB wir endlich ein umfassenderes Bild von dieser Anlage besitzen.

Nachfolgend m6chte ich zu der von Keall142 vermuteten Identitiit von Qal'a-i Yazdigird mit Madharii- stan Stellung nehmen. Meines Erachtens ist diese Annahme falsch, weil Keall zwei der wichtigsten Quellen, ndimlich Abti-Dulaf und Ibn Rusta tibersehen hat.

33 at-Tabari (Ed. H. Zotenberg) II, S. 71; Istakhri, BGA I, S. I23, 19 ff.; Ibn IHauqal, BGA II, S. 278, I6 ff.

34 Idrisi, Giographie (Ed. P. A. Jaubert) I, S. 394; Yiqfit, Dictionnaire giographique, historique et litteraire de la Perse (Ed. C. A. Barbier de Meynard), S. I75; Qazwini, Kosmographie (Ed. F. Wiistenfeld), 2. Theil, S. I21, I7 ff.; HIamd-AllTh Mustaufi, Nuzhat al-Qulfb (Ed. Le Strange), S. I16/17.

35 s. Anm. 33. 36 s. Anm. 33.

37 s. Anm. 34- 38 a.a.O. S. 24. 39 a.a.O. S. 21, Anm. 9. 40 Islamische Mafle und Gewichte, Erglinzungsband I, Heft I,

Handbuch d. Orientalistik (0955), S. 62.

41 Iran V (1967), S. 99-121.

42 S. 99/10; I20/21.

HINWEISE UND ANMERKUNGEN ZU EINIGEN SASANIDISCHEN MONUMENTEN 161

Abai-Dulaf43 (um 950) macht von .Hulwan

kommend, die nachfolgenden Angaben uiber Mddharii- stfn: " Thence to the Cupper's Arch which is a stone vault at the cross-roads in the gorge between two mountains, a marvellous, tall building. Thence to Madharfistan which is a huge portico standing alone with a raised platform. ... Snow falls only on its eastern side looking towards Khorasan, but not on its western side looking towards Iraq. Thence to Marj al-qal'a .. "

Diesem Bericht ist zu entnehmen, daB Mdidhartistzn hinter dem " Cupper's Arch " lag. Damit diirfte der Taq-i-Girrd gemeint sein.44 Gesttitzt wird diese Auffassung auch durch Ibn Rustas (um 903/13) Bericht,45 der anfiihrt, daB, wenn man aus Richtung HIulwdn kommt, auf der Mitte des Passes

" ein Bogen ist, der aus Felssteinen erbaut und mit Marmor belegt ist .... Danach gelangt man zum

Gipfel des Passes und steigt von da hinab nach Maij Darwistdin ", welches nach Schwarz46 Midhartistdn ist.

Diese beiden Berichte widerlegen m.E. die Ansicht Rawlinsons47 und auch Herzfelds,48 daB der

Tai.q-i Girrd mit Midhariistdn identisch sei.49

Sie lassen sich aber auch nicht mit Kealls Lokalisierung von Madhartistin bei Qal'a-i Yazdi- gird vereinbaren. Ersteres muB an der groBen StraBe Baghdad-Khurasan gelegen haben. Daftir sprechen neben diesen beiden Berichten auch die weiteren von Ibn Khurdadhbih,50 Quddma51 und

Muqaddasi52 tiber Mddhartistin. Qal'a-i Yazdigird aber liegt nach Kealls vom Angabenjeweils 4 Para-

sangen53 von Sar-i Pul Zuhdb und von Tdq-i Girrdt weit entfernt von der groBen StraBe in den Bergen. Eine andere M6glichkeit findet sich bei Minorsky.54 Aus der von Abti-Dulaf, Ydqtit55 und auch

Qazwini56 erwaihnten Tatsache, daB Schnee nur an der 6stlichen Seite des Gebiiudes fiel, schlieBt er, daB Mddhartistdn hinter dem PaB von

Pit.aq lag. Er vermutet es bei Sar-mil, 13 km57 vor Kirind.

Ich habe jedoch Zweifel, ob diese Lokalisierung richtig ist. Minorsky geht nimlich davon aus, daB auf der Strecke Sar-i Pul Zuhlb

(.Hulwdn)

- Kirmdnshdh, die er mit 142 km angibt,58 die Parasange mit 4 km zu rechnen sei, weil diese Distanz in den arabischen Quellen mit etwa 35 Parasangen angegeben wiirde. Dies ist aber unrichtig. Einmal betrigt die heutige Entfernung 152 km59 und vor allem aber rechnen die arabischen Geographen nicht mit 35 Parasangen, sondern mit wesentlich weniger. So geben Ibn Khurdddhbih60 (um 899) und Quddma61 (um 930) 30 Parasangen; Istakhri62 (i. Hiilfte des io. Jahrhunderts) und Ibn IHauqal63 (nach 977/78) 27 Parasangen.

Aber selbst jedoch, wenn man Minorskys MaR von 4 km zugrundelegt, so muB man fragen, wie er den Unterschied erkliren will, der sich daraus ergibt, daB einerseits die Distanz Hulwdn-Sar-mil

(Maidhartistmn) ca. 35 km betriigt, andererseits aber Ibn Khurdddhbih - auf den er sich selbst beruft -

43 Aba-Dulaf Mis'ar Ibn Muhalhil's Travels in Iran; Ubersetzung V. Minorsky (I955), S. 43.

44 So auch Minorsky, a.a.O., S. 43, Anm. 2; S. 87. 45 BGA VII, S. I65, Io. 46 Iran im Mittelalter, S. 902; nach J. Markwart, Untersuchungen

zur Geschichte von Eran, 2. Heft, S. 23 Anm. 4 ist der Ort nach einem Perser Mdh-Druwasp, Dihqdn von Bdbil Mahrfid 76/77 Hedschra benannt.

47 Journal Royal Geographical Soc. IX (1839), S. 34- 48 Urspriinglich hatte H., Petermanns Geographische Mitteilungen

53 (1907), S. 53 vermutet, Mddharfistn lige in der Nahe des PIitiq; spiter in Sarre-Herzfeld, Iranische Felsreliefs, S. 233; Geschichte der Stadt Samarra (1948), S. 52, hat er diese Ansicht aufgegeben und nahm eine Identitit von MrLdharfistan mit dem

.TIq-i Girri an.

9 Dies wird auch schon von Keall hervorgehoben. 50 BGA VI, S. 19, 8.

51 BGA VI, S. 198, 4-

5. BGA III, S. i35, 6.

53 Keall, S. Ioi rechnet, wie allgemein tiblich, die Parasange mit 6 km; s. auch W. Hinz, Islamische Marfe und Gewichte, S. 63. Dennoch diirfte das MaB der Parasange nicht zu allen Zeiten

und ftir alle Gebiete Persiens einheitlich gewesen sein; so sagt Rawlinson, JRGS IX (1839), S. 31, Anm.+, dab die Parasange ein sehr unbestimmtes MaB sei. Fiir Kurdistan rechnet er allerdings auch mit 3 3/4 Meilen, was etwa 6 km entspricht; s. im iibrigen nachfolgend auch Minorskys Berechnungen mit 4 km und den Hinweis des Verfassers auf 5 km als MaB fiir die Parasange.

54 Abti-Dulaf, S. 87.

55 Geographisches Worterbuch (Ed. Wiistenfeld), C 7, 354, 7- 56 Kosmographie (Ed. Wiistenfeld) 2. Theil, S. 302, 23 ff. 57 Nach Farhang-i Joghrdfiyd'f-ye Irdn, io Bande (Ed. Razmdra),

Teheran (1950-1954), Bd. V, S. 238. 58 Aba-Dulaf, S. 87.

59s. Guide-Bleu, Moyen-Orient (1965), S. 697/99; Karte des " Ministry of Roads ", Teheran, 5. Aufl. 1967.

60o BGA VI, S. 19, 8 ff.; diese Angabe sowie die von Quddma diirften am zuverlissigsten sein, da sie als Staatsbeamte Zugang zu den Archiven hatten.

61 BGA VI, S. 198, 4 ff. 62 BGA I, S. 196, 2 ff.

63 BGA II, S. 359, 8 if.


Qudama und Ibn Rusta64 4 Parasangen65 fur diese Strecke geben. Nach Minorskys Berechnung wiren dies nur 16 km.

Herzfeld wiederum hatte, bevor er sich for eine Identitit des Taq-i Girra mit Mddharfistdn entschied, letzteres bei dem Dorf Pititq vermutet. Dort, also noch vor dem Taq-i Girra und der PaBhohe, hatte er im Oktober 190566 sasanidische Ruinen67 entdeckt. Es ftillt jedoch schwer, Herzfeld zu folgen, da diese Lokalisierung den Berichten von Abti-Dulafund Ibn Rusta widerspricht. Denn beide gebenja an, daB M~dharistin hinter dem Taq-i Girra lag.

Ich selbst schliel3e mich insoweit Minorsky an, als dieser Mddhartistdn hinter dem PaB von Paitaq sucht. Ich vermute esjedoch niiher zum PaB hin. Meines Erachtens sollte bei einer Lokalisierung von

Mtdhartistan insbesondere das erste, 4 km nach dem PaB liegende Dorf68 Surkha diza in Betracht gezogen werden. Im August 1844 sahen J. F. Jones9" und Rawlinson dort noch ein Gebaude, das Jones wie folgt beschreibt: " The remains of a fire temple are seen close to the Khan, supported on massive arches ". Heute allerdings sind keine Spuren mehr von diesem Bau erhalten; dies berichtet schon Herzfeld,70 der den Ort 1914 und 1917 passierte. Jones spricht zwar von einem Feuertempel in Form eines Kuppelbaues, aber die Beschreibung der Anlage von Midhartistn durch die arabischen Geo- graphen als Plattform mit einem Garten ist so vage,71 daB eine Identitait beider Anlagen nicht von vornherein auszuschlieBen ist.

Die Lokalisierung von MidhartistAn bei Surkha-diza wiirde der Entfernungsangabe fUr IHulwdn- Madhartistdn von 4 Parasangen entsprechen, wenn man eine Parasange gleich 6 km rechnet. Denn Surkha-diza liegt 25 km hinter IHulw~n. Etwas schwierig wird es allerdings fUr den anschlieBen- den Teil der Strecke: Mddhartist~n - Marj al-Qal'a (Kirind),72 die mit 6 Parasangen, also 36 km, angegeben wird. Die Entfernung Surkha-diza - Kirind betrdgt jedoch nur 23 km.73

Das Verhiltnis von I Parasange gleich 5 km74 wuirde daher im ganzen gesehen etwas besser zutreffen; die Streckenangabe JHulwdn - Madhartistn (Surkha diza) wdire zwar um 5 km tibersch- ritten; die Distanz IHulwan - Marj al-Qal'a (Kirind) - Kirmanshah aber die Io, bzw. 30 Parasan- gen nach Ibn Khurdadhbih und Qudama betriigt, wiirde der heutigen Entfernung von 48 bzw. 152 km entsprechen. Man muB natiurlich immer in Betracht ziehen, daB die Parasangenangaben der arabischen Geographen nur Schaitzungen waren, die an sich schon unzuverliissig sind,75 und daB die StraBe Baghdad - Khurasan friiher vielleicht etwas anders verlief. Allerdings glaube ich nicht, daB dies auf die uns hier interessierende Strecke zutrifft; die natiirlichen Verhdiltnisse, besonders die Berge, schrieben den Verlauf mehr oder minder vor.

AbschlieBend komme ich zu dem Ergebnis, daB m.E. die Berichte der arabischen Geographen ziemlich klar auf eine Lage von Midhartistin unmittelbar an der groBen StraBe Baghdad-Khurasan hinweisen, und zwar nach dem TZq-i Girrt; wo allerdings genau, ist aufgrund der schwierigen Parasangenberechnung nicht mehr einwandfrei festzustellen. Eine Identifizierung von Madhartistan mit den heutigen Ruinen von Qal'a-i Yazdigird dagegen halte ich nicht fur m6glich.

64 BGA VII, S. 165, 7. 65 Minorsky, Abiz-Dulaf, S. 87 gibt zwar 6 Parasangen, aber das

ist falsch. 66 s. Petermanns Geographische Mitteilungen 53 (1907), S. 53. 67 Auf der Karte I : 2ooooo Herzfeld, Paikuli I, Sheet 2, verzeich-

net er allerdings " Sasanian Ruins " hinter dem Tdq-i Girrd auf der Pa83hhe. Es fragt sich, ob es sich hierbei nur um einen Fehler oder um eine weitere Ruinenstitte handelt.

6s Dies wiirde insoweit mit Ibn Rustas Beschreibung iiberein- stimmen.

69 Transactions of the Bombay Geographical Soc. VIII (1849), S. 267/ 68.

70 Gewhichte der Stadt Samarra, S. 51, Anm. 2.

17 So spricht Le Strange, The Lands of the Eastern Caliphate, S. 191, von einem Palast, wihrend K. Erdmann, Das Iranische Feuer- heiligtum (194i), Anm. 39 (S. 71), an eine kultische Hochstitte zu denken scheint.

72 Diese Identitit wird allgemein angenommen; s. z.B W. Tomaschek, " Zur historischen Topographie von Persien ", I, Sitzungsberichte d. philosoph.-hist. Cl. d. Kaiserl. Ak. d. Wiss. Wien, Bd. CII (1883), S. 8 (S. I50); Herzfeld, Geschichte der Stadt Samarra, S. 41; Schwarz, Iran, S. 492 vermutet es in der

Nihe von Kirind.

73 C. Ritter, Die Erdkunde im Verhdltnis zur Natur und zur Geschichte des Menschen, 9. Theil, 3. Buch. West-Asien, 2. Aufl., 1840, S. 391/93, gibt anhand der verschiedenen Berichte von

ilteren Reisenden (G. Keppel, Ker Porter u.a.) ffir die Strecke Sar-i Pul Zuhab - Kirind nur 9 Parasangen. In diesem Fall wuirde sich die Strecke auf 54 km ermil3igen; der Unterschied wiirde also nur noch 6 km (zu 48 km) betragen.

74 Auch J. Hansman, Iranica Antiqua VII (1967), S. 51, legt dieses Mal3verhiltnis fuir Khuzistan zugrunde.

75 s. auch Schwarz, Iran, S. 900, Anm. I8; beachte weiter auch den Hinweis unter Anm. 73 auf Ritters Angabe von 9 anstatt I o Parasangen ffir die Strecke Sar-i Pul Zuhab-Kirind.

P1. L Feueraltdre bei Persepolis.

Pl. Ila. Blick von Siiden vom Gipfel eines Berges am Dorfrand von Nimvar.

Pl. IIb. Blick von Siiden.

Pl. IVa. Terrasse Takht-i Nishin bei Firfzabad.

P1. IVb. Sdulentrommel im Garten des Imdmzdda bei Firuizdbdd.

P1. IIIa. Blick von Westen.

163

FURTHER NOTES ON THE SHAFT-HOLE PICK-AXE FROM KHURAB, MAKRAN

By C. C. Lamberg-Karlovsky

Metallurgical analyses and new dating evidence, resulting from recent archaeological excavations in south-eastern Iran,' place a new perspective on the important shaft-hole pick-axe recovered from a burial at Khurab by Sir Aurel Stein.2 This object has been the subject of separate articles by K. R. Maxwell-Hyslop3 and F. E. Zeuner.4 The pick-axe was found in a grave illustrated by Stein containing thirty-two complete vessels, mostly unpainted, two metal bowls similar in shape to those of ceramic, a complete alabaster jar and an agate bead tipped with gold ferrules at both ends. One small bowl was filled with human(?) bone fragments.5

Mrs. Maxwell-Hyslop has compared this shaft-hole pick-axe to her types 9B and 15,6 while also suggesting parallels for it with the metal rods of Hissar III which have human figures modelled at one end. Compared with the axe-heads with animals on the socket, as the well known example from the hypogeum at Til Barsip or later

Lfiristin examples, it is suggested that the Khurtb blade is crudely fashioned and evidences little technical skill. Metallurgical analysis refutes the absence of a high standard of craftsmanship in the manufacture of this object. No convincing parallels have been found for this object to date.

Following the lead of Professor Piggott who dated the Khurtb burials on the basis of far reaching and generalized parallels to around 2000 B.C.,7 various authors have suggested dates between 2000- 8oo B.C.8 On the basis of recent work in this area their date (and the pick-axe) can be more securely suggested. At KhurTb, Grave E, from which this object was recovered, contained cups almost identical to those recently excavated at Bampir in Period V context.9 The appearance of a distinctive incised grey ware, in late Bamptir IV and throughout V, of a type found also on Oman across the Persian Gulfxo establishes the first firm link between the two sides of the Persian Gulf and sets a terminal date to the Bampair sequence; allowing us to suggest a date for Bampfir V and thus by its ceramic parallels to Khurdb Grave E.

Recent excavations on the island of Bahrain have uncovered a seal impression similar to a stamped seal tablet in the Yale Babylonian Collection.11 This Yale impression is dated to the tenth year of Gungunum, King of Larsa, in southern Babylonia-that is 1923 B.c. The Bahrain seal was found in a Barbar culture level, partially contemporary with the Umm an-Nar culture of Oman, which can in

1 B. de Cardi, " The Bampur Sequence in the Third Millennium B.C.", Antiquity XLI, no. 16I (1967), pp. 33-42; " Excavations at Bampfir S.E. Iran: A Brief Report ", Iran VI (1968). A re-examination of the collections from Bampiar, Khurab and

ChTh IHuseini in the Peabody Museum, Harvard University, has been completed (in press) by the author.

2A. Stein, Archaeological Reconnaissance in Southeastern Iran and Northwestern India (i937), pl. XVIII, E, i, 258, p. 121.

3 K. R. Maxwell-Hyslop, " Note on a Shaft-Hole Axe-Pick from Khurab ", Iraq XVII (0955), p. i61.

SF. E. Zeuner, " The Identity of the Camel on the Khurab Pick ", Iraq XVII (I955), pp. 162-3.

' A. Stein, op. cit., p. 121, fig. 41, pl. XVI, E, i, 253, 254; pl. XVIII, E, i, 251 ; pl. VI, E, i, 252; pl. X, E, 259.

6 K. R. Maxwell-Hyslop, " Western Asiatic Shaft-Hole Axes ",

Iraq XI (i947), pl. XXXV, 15; pl. XXXVIII, 5, 10o, pp. 97 and 1 o4.

7 S. Piggott, Prehistoric India (1950), p. 2 8. 8 K. R. Maxwell-Hyslop, op. cit., p. I6I. See also D. H. Gordon,

" The Pottery Industries of the Indo-Iranian Border: A Restatement and Tentative Chronology ", Ancient India Io/ II (1954-55), pp. 189 if. B. de Cardi, Iran VI, pp. 144-7 and the author's own study of the Bampfir and Khurab collections in the Peabody Museum. Similar cups are also evident in the Peabody collection from Sutkagen Dor.

10 B. de Cardi, op. cit., pl. IVb: I-15 compared with G. Bibby, " Arabian Gulf Archaeology ", Kuml (1967), figs. ii and 12; and K. Thorvildsen, " Burial Cairns on Umm an-Nar ", Kuml (1963), fig. 20o, pp. 191-219.

11 For Yale Babylonian impression see B. Buchanan, " A Dated Seal Impression Connecting Babylonia and Ancient India ", Archaeology 20 (1967), pp. 104-7. For Bahrain seal see G. Bibby, " Arabian Gulf Archaeology ", Kuml (1966), pp. 147 and 152, note I.


turn be paralleled at Bampfir V with the incised grey ware (" hut-pot ") motifs. The general evidence, thus, points to a date c. 1900oo B.C. for the terminus of the Bamptir sequence,12 and for the date of the

Khurtb shaft-hole pick-axe.

APPENDIX

The report and technical analysis of the Khurab pick-axe was undertaken by Heather N. Lechtman of the Laboratory for Research on Archaeological Materials, Massachusetts Institute of Technology, Cambridge, Mass.

Composition of the Metal

Although the radiograph on P1. IIaa3 indicates that the object is solid, the metal of which the camel and the socket were originally constituted has almost entirely corroded. Presently the object measures 18- 4 cm. from the head of the camel to the base of the pick; the width of the object at the base of the camel is 6'2 cm., while the blade is approximately I -6 cm. wide. The shaft-hole is

3"4 x I -5 cm.

Four drillings were made into the object, one in the body of the camel, two in the socket, and one in the blade to obtain metal samples for analysis. The positions of two of these sampling sites, the single drill hole made within the body of the camel to a depth of 7 mm. and one of the holes drilled completely through one side of the socket and into the socket cavity, are indicated by the two white arrows on P1. Ia. The second sample from the socket was removed from the opposite side of the object. None of these samples contained metal. The material removed was composed entirely of the mineral products of corrosion, primarily cuprite (cuprous oxide).

The blade of the pick, although extensively corroded, is still largely metallic, however. Samples of metal obtained from a depth of almost I cm. within the blade were analysed by emission spectrography after all observable mineral had been removed from the drillings. The position of this drill hole, indicated by a white arrow in P1. Ib, is approximately 3"5

cm. below the point at which the blade meets the socket. The results of a qualitative analysis of two samples taken from this site are given in the following table.

Spectrographic Analysis of Metal Drillings Removed from the Blade

Element Percentage, by weight, of total sample

Ag ooI-o- I As I -0-Io Bi o*ooi-o-oi Cu Major Component Fe I 'o-10 Mg o0ooI-o.OI Ni ooo0I-o-'O Pb o0 I-Io Sb o.I-I.o Si 000ooo Sn Not Detectable (detected to 20 p.p.m.)

12 The date of c. 1900 B.C., or after, is further supported if the Bampfir V parallels with Hissar III suggested by de Cardi, op. cit., be accepted. It might be pointed out that the representation of a spoked chariot in Hissar III context suggests, as Childe long ago noted, a low date, after 2000 B.C., for the later Hissar III settlement, see " The First Waggons and Carts--

from the Tigris to the Severn ", Proceedings of the Prehistoric

Society 8 (1951), p. I84.

13 The radiograph reproduced here as P1. IIa was made at the Eastman Kodak Company, Rochester, New York, through the

courtesy of Mr. Charles F. Bridgman,

FURTHER NOTES ON THE SHAFT-HOLE PICK-AXE FROM KHURAB, MAKRAN 165

The metal is therefore a copper-arsenic alloy with antimony and lead as the major impurities. This composition is consistent with an early second millennium date for the object.14

Method of Fabrication The radiograph gives every indication that the pick was cast in a solid single piece. The cavity in

the socket was created by a core placed inside the mould before the metal was poured. From the position of the shrinkage zone shown on the radiograph, the mould could have been in either of two positions at the time that the molten metal solidified. These are given as positions A and B in Fig. I. With the mould in position A, the metal would have been introduced through some sort of gate whose location was at the extreme end of the blade, as shown in the diagram. The second, alternative position is that in which the entire mould was held vertically and the metal poured in from the top. In this case the gate was probably offset to one side of the blade's tip, an unusual displacement.

ga te

4 moutd,

centre line porosity

Fig. I. Two alternative reconstructions of the position of the mould during casting.

The evidence for both of these reconstructions is contained in the radiograph. Examination of the blade in P1. IIa reveals a discontinuous line of high film density running almost precisely along the centre line of the blade from the position of the drill hole to the extreme tip of the blade. This line, which corresponds to a region of low density within the metal and which is often referred to as a centre line porosity, is characteristic of the final stages in the solidification of a cast metal ingot and represents inadequately fed shrinkage of the last portion of the molten metal to solidify. Porosity can also be caused by the evolution of small gas bubbles in the last of the molten metal. In this particular case it is difficult to estimate the relative extent to which each of these processes has contributed to the porosity indicated in the radiograph. On P1. IIa the line of porosity follows the geometric midline of the blade almost along its entire length, but it becomes much broader and more diffuse at the blade's tip and tends to bend toward one narrow side of the blade as indicated in drawings A and B of Fig. I. The

14 See comments by I. R. Selimkhanov in C. F. C. Hawkes, " Analysis by Optical Spectrometry ", Atti del VI Congresso

Internazionale delle Scienze preistoriche e protostoriche 1 (1962), pp. 46-8.


preponderance of porosity at the tip of the blade is due both to the increased amount of metal shrinkage at the place where the metal cooled last and possibly also to the presence of dross in the liquid which may have accumulated and floated to the top of the casting. The large, rectangularly shaped densities near the blade's tip which is particularly clear in the radiograph on P1. IIa are near the surfaces of the blade and are not associated with the centre line shrinkage. The fact that the line of porosity tends to bend towards one side of the blade at its tip may be indicative of the fact that a hot sink of metal existed at this location. The last metal to solidify would do so at a place where the metal was hottest, presumably at the position of the gate. Thus the shrinkage would tend to shift over in the direction of this heated area. Either of the reconstructions given in Fig. I is consistent with the picture of the internal structure of the casting presented in the radiograph.

Small, almost spherical bodies of high film density evident on the radiograph within the body of the camel also indicate porosities in the metal. These represent bubbles formed as the result of gas evolution within the metal itself or from the mould when it was struck by the hot metal. It should be noted that these bubbles tend to form a line along the horizontal where the body of the camel meets the socket of the pick. The socket contained a core made of clay or a sand clay mixture at the time of casting. Had the mould been in a vertical position, as in detail B of Fig. I, it is not unlikely that, during the casting operation, the bubbles formed within the liquid metal of the camel's body were prevented, by the presence of this core, from continuing to move upward through the socket and finally, from escaping by rising up along the length of the blade. The formation of the bubbles along this horizontal may, therefore, indicate the evidence for position B of Fig. I as the actual casting position employed.

The shaft or blade of the pick has broken away from the socket at some time subsequent to excavation. The complete continuity of the corrosion products on the surfaces of the object on either side of the break indicates that the object was intact while buried. The direction of the lines of fracture, which run almost horizontally across both sides of the object at the position of springing of the blade from the socket can also be seen on P1l. Ia. The radiograph also reveals quite clearly the position and direction of the break which appears as a series of dark, horizontal lines on P1. IIa. It is not extraordinary that the pick should have broken where it did. The radiograph indicates that, when the object was fabri- cated, inherent weaknesses in the form of hot tears may have existed in the metal precisely where the socket meets the much thicker blade. The formation of a hot tear in a casting, characterized by an actual parting of the metal, is not unusual in those portions of the casting where the mould changes abruptly from a relatively thick size. Such regions are zones of potential weakness which, in the case of this object, may have facilitated the failure of the mineral precisely along such lines of weakness.

Reconstruction of the mould used in the fabrication of the pick is much more difficult. All that can be said with certainty is that it consisted of at least two parts, the outer material that formed the mould proper and the inner core which was necessary to create the socket cavity. Furthermore, there is some evidence that a single metal chaplet, in the shape of a cylindrical pin, was used to attach the core to the mould in order to hold it in place during the pouring of the metal. The position of this pin is outlined by a white circle on P1. Ia. The distance of the pin as measured along a horizontal from the rear of the pick is approximately 2-4 cm. on one side of the object (P1. Ia) and approximately 2"3

cm. on the opposite side of the socket indicating that the circular depression appearing on P1l. Ia bears a close relationship to the corroded remnants of the pin which are still visible on the other side of the object. The roughly circular outline of the pin can also be seen in the radiograph on P1. IIa. The core was also probably attached directly to the mould at both ends to give it increased rigidity during casting. There is, of course, the possibility that the pin described here as a chaplet may actually have been used to hold the pick to the wooden handle. Given the relatively small diameter of the pin, such a mechanical device would not have been adequate to secure the handle to the pick in a tool that was functional, but it might have served sufficiently well for an object only in ceremonial use.

The question remains as to whether the mould was a piece mould or whether it was a single structure built up over a wax model which was subsequently melted out of the interior in the usual manner of lost wax casting. The advantage of piece moulding is that many castings can be made from the original model; that of lost wax casting is the freedom it allows in the shaping of the wax model which may contain undercuts and intricacies of form easily reproduced in the casting. Careful examination of the

Pl. Ib. Detail of the rear, narrow side of the blade showing the site of the metal samples removed for analysis and the linear corrosion protuberances, possibly related to the structure of the mould.

Pl. Ia. Detail of one side of the socket showing the positions of the pin, the sites of two sample drill holes, and the break between the socket and the blade.

Pl. IIa. X-radiograph, side view (courtesy of the Eastman Kodak Co.). Note particularly the porosity along the centre line. The roughly circular outline of the pin at the centre of

the socket can be seen.

groove

protuberance

head

hump

P1. JIb. View of the camel from above showing the relationships of protuberance, hump and groove. (Photograph by Arsen Charles, Peabody Museum,

Harvard University.)

FURTHER NOTES ON THE SHAFT-HOLE PICK-AXE FROM KHURAB, MAKRAN 167

pick has revealed no obvious flash lines in those places where it would seem most natural for mould

joints to have existed. Since such lines are easily filed or scraped away, however, their absence on the finished object does not preclude their presence in the rough, as cast condition. The extent and uneven- ness of the corrosion of the object, as well as the fact that it has been mechanically cleaned in the past, make it difficult to detect any traces of such lines that may in fact still remain. If the mould was a piece mould, it could have easily been of the bi-valve type with the parting line in the plane that bisects the narrow sides of the pick. Such a mould would draw easily from both sides of the model or, ultimately, from the casting, since the design of the object is such that there are no undercuts, and a simple two

piece mould plus core would be adequate. The strongest argument in favour of lost wax casting is the absence of flash lines, however, as noted, their absence is not conclusive evidence for the use of the

cire-perdue process. There are several roughly linear protruberances on the object that appear on the blade or just

beneath the point at which the blade springs from the socket. Several of these are shown on P1. Ib. They may be due entirely to local differences in corrosion rates associated with cracks in the corrosion

product. There is also the possibility, however, that they represent " fins " of metal having formed when the molten metal ran into cracks in the mould, for they appear in no way associated with the

parting lines of a piece mould. One of these protruberances indicated by a white line arrow on P1. Ib, can also be seen, in silhouette on P1. Ia extending from the bottom of the socket down along the rear, narrow side of the blade.

Although no traces of wood remain in the socket cavity, the corrosion of the inner surfaces of the

cavity, especially on its upper surface, displays the fibrous structure associated with wood. It is evident that this corrosion formed while in contact with the original handle that must have been in place at the time of burial but which has since completely deteriorated. The method of assembly of the pick-axe and its handle is a matter of conjecture. The probability of the use of a metal pin to hold the two

together has already been mentioned. Examination of the outer surfaces of the socket revealed no obvious traces in the corrosion of an organic material, such as leather or a fibre cord, which might have been used to bind the axe to the handle. On the other hand, there is a very deliberate, wide, and deep groove with sharply delineated edges lying posterior to the hump of the camel and continuing diagonally along its side. This groove is part of the original casting and can be seen clearly on P1. Ia. The corrosion with the groove is slightly different from that of the surrounding areas. It is less coarse, lighter in colour, and tends to give the appearance of having been formed in contact with some organic material which

may have once lain within the groove, although no traces of such a material remain. It may be worth- while considering the possibility either that it played some part in securing an organic lacing that was wound around both the pick and the handle or that the ridge was originally intended as the rendering of some portion of a trapping (see below). In the latter case it may or may not have been inlaid with an

organic material representing such an element. In F. E. Zeuner's discussion of the anatomy of this camel he does not mention this groove, although

it is distinctly noticeable on the object.15 This may be due to the fact that he did not inspect the object itself but only photographs, all of which showed the opposite side of the camel reproduced on P1. Ia.16 On the opposite side of the animal, in fact, the groove is decidedly less pronounced, although it does

appear to continue in a shallower form. The speculations here as to the interpretation of this feature are simply several which are suggested by the technical analysis.

Zeuner puzzled over the Bactrian or dromedary identity of the camel, " The Khurab beast has only one hump and thus might perhaps be regarded as a dromedary. But every other feature speaks against such interpretation ".17 It is his conclusion after balancing the contradictory evidence that it is a Bactrian camel. One might note that an earlier, in fact the earliest, representation of the Bactrian camel can be seen on a sherd from Sialk III4*18

16 F. E. Zeuner, op. cit., pp. 162-3. 16 For previously published photographs see D. H. Gordon, op.

cit., as well as F. E. Zeuner, op. cit., K. R. Maxwell Hyslop, op. cit., and V. G. Childe, New Light on the Most Ancient Near East (i954), pl. XXIXb.

15A

17 F. E. Zeuner, op. cit., p. 162.

18 R. Ghirshman, Fouilles de Sialk I (Paris 1938), pl. LXXIX, A2.


P1. IIb reproduces a view looking straight down onto the top of the animal from above. The head, hump and groove are labelled accordingly. The accentuation of the groove on the proper left side of the animal is quite evident here. Of interest is a heavily corroded protruberance which appears only on the left side of the body and extends anteriorly from the hump stopping just behind the neck. This feature can also be seen in P1. Ia, although it is not as obvious in the side view. It is as yet an unexplained element but it seems clear that it is an intentional element that existed on the object originally. The right edge of this protruberance, clearly visualized on P1. IIb, runs diagonally down across the back of the animal and is quite sharply delineated. It is not impossible that this protruberance represents the fragmentary remains of an anterior hump, now eroded away but for this left upper region. The limited possibility of the groove representing some type of trapping would suggest the domestication of camel several centuries prior to the later part of the second millennium where evidence becomes abundant.19

The metallurgical analyses and cultural context of this object, recovered from an archaeologically poorly defined area of Iran, rather confirms Professor Mallowan's recent comment " every part of Iran was in due course affected by the pace of technological development in other parts of Asia ".20

ADDENDUM

Since this article was written, the author has completed a season of excavation at Tepe Yahyd (see p. 184). A single, nearly complete cup from Tepe Yahyd, Period V, can be identically paralleled at KhurTb, Grave E, where this pick-axe was recovered. Radiocarbon dates for this period are presently being run. In the excavations at Tepe Yahya, Period V and VI were completely void of any metal remains, however, the earliest occurrence of metals so far at Tepe Yahya appear in the earlier Period VII, see note 20.

19 W. Beebe, " Camel Domestication ", Bulletin of the New York Zoological Society 48 (1940), p. 117; and A. E. Robinson, " Camel in Antiquity ", Sudan Notes and Records I9 (1936), p. 47.

20 M. E. L. Mallowan, " The Development of Cities ", Cambridge Ancient History, fascicule no. 58 (1967), p. 56.

The earliest documented metal tools in the area of south-

eastern Iran is now evidenced in the recent excavations at Tepe Yahyd (see p. 189). In Period VII square sectioned awls, spatula shaped chisels and long and thick pins were uncovered. Period VII has a single radiocarbon date of 3860o? 140 B.c.

(5730 half-life). The tools appear to be manufactured by simple hammering.

169

SURVEY OF EXCAVATIONS IN IRAN, 1967-8

I. EXCAVATION REPORTS

Bdbd Jdn Work was resumed at Bibd Jan on July Ist and continued for ten weeks. In this, our third season,

we concentrated on the smaller East Mound in the hope of completing the excavation of the " Painted Chamber " and ascertaining its relationship to the Fort discovered on the summit of the tepe at the end of the previous season (Iran VI, pp. 157-8). In this we were entirely successful though a further summer's digging will be needed to finish the excavation of the Fort.

It now appears that there were three main phases of occupation. The earliest building on the site (old Level vb, now iiib) is still the " Painted Chamber " which seems to have been constructed at the same time as, or shortly after, the last Fortified Manor on the Central Mound. It is now revealed as a large " T "-shaped hall with a central rectangular section 12* 50 X 10o45 m., extended 5 m. at either

-4M Wku ,."-' ., U. S. S. R.

'-4. oft-*{ 0 m CA 5 P / AN

... .

0 9 oTABRIZ EA, \ HoFbavrn Tepe ". -.

"'a) .

nu *Tureng Tepe * ',--

Dikh H Kaluraz MESHED

Te p Takhbt-i Sulirrmn

o (R TEHRAN

'Tongavar *HAMADAN ( Sjarb Jan eTepe Sialk

" 1

Bi Suo ahe '., S $

. .SFAHAN hShahr-i Sokhte

Ho? Tee /

*KERMAN .

.Penrspolis ZAHEDAN 'o * Darvazeh Tepe ?

' SH 1I4IAZ --

" 'oTepe Yah Iya

SAUDI ARABIA t r

oO 500

Kilomcetres


end by small rectangular alcoves 3"75

m. across. The walls were covered with red painted plaster and broken by elaborate recessed doorways and blind windows. A large " A " shaped niche in the back wall directly opposite the main entrance was probably a fireplace. There were two large stone column bases in the centre of the room and around them were scattered 133 more painted tiles.

Shortly after the construction of the " Painted Chamber " the Fort (Level iiia) was built directly adjoining it to the west. This building was unusually well preserved, for, in the centre of the tepe, its walls were still standing to a height of

4" 50 m. and stretched from just below the thin Achaemenid deposit of Level i down to virgin soil. In plan the Fort resembled the Fortified Manors on the Central Tepe, built on a rather more massive scale, with a large central hall or court c. I - 15 X II 80 m., flanked to east and west by long rooms

4"90 m. across, with a range of smaller domestic rooms to the

south. Only the western long room was completely excavated. Its long walls were lined with decorative mud brick buttresses or pilasters and at its northern end was a second fireplace identical to that of the " Painted Chamber ". Its lowest floor produced a great mass of broken pithoi together with some fine " genre Liiristdn " painted wares. It was presumably domestic in function.

The central court or hall was reached from the long room through a high corbelled doorway. Only a small part of this area has so far been excavated and it is still uncertain whether it was roofed over or formed an open area in the centre of the building, but red painted plaster found at the foot of one of its walls rather suggests the former.

This complex of buildings shows remarkable parallels with Tepe Nfish-i Jan, c. I6o km. away across the Kfih-i Garin range, and easily reached via the Nihivand pass. For at both sites a fortress so strongly built as to be preserved almost intact until the present day was constructed directly up against a large " N "-shaped chamber elaborately decorated with blind windows and presumably with some specialized function. The plans of our fort are, however, quite different from the Niish-i Jan example and hark back to the Hasanlii IV Burned Buildings.

At the end of Phase iii both " Painted Chamber " and Fortress appear to have been burnt and in the second period of occupation, ii, the former went out of use and was replaced by a series of small private houses and stables (old Levels va and iv). The Fortress however, was reoccupied for all the rooms have one or more high secondary floors associated with minor structural changes such as blocked doorways and heavy replastering. In the pottery from this phase, a degenerate form of " genre Lfristdn " continues side by side with new wheel-made wares with mica tempering in shapes related to the pottery from the fortress of Ntish-i Jan.

Finally the Fort and associated houses were burnt a second time and after being deserted the site was finally reoccupied for a short period in the sixth or early fifth centuries (Levels ia and b). It was then abandoned until the present village was built in the time of Reza Shah.

CLARE GOFF

Bdni Surmah The first season of excavations at Bdni Surmah lasted from October 15th to December 12th 1967.

Most interestingly, the work revealed a new and previously unknown stage in the Bronze Civilization of Lfiristdn: a phase dating from the middle of the third millennium B.C. The funeral vaults of the necropolis at Bdni Surmah are unusually impressive because of their size. With a few exceptions, they measure from 8 to 16 m. in length with a width that varies from I - 70 to 3 m. Their depth varies from I -50 to

2" IO m. The burial chambers are closed by enormous stone slabs, some of which weigh as

much as 6oo kilos. The interior structure with its perfect finish is also striking: the side walls are made of thick limestone blocks, disposed in regular horizontal parallel courses.

These enormous burial chambers represent family or collective sepulchres, containing several bodies. The dead were buried fully clothed and with all their personal belongings. The men were inhumed with their weapons and tools: the women were buried with elegant necklaces, engraved pins, hair-spirals, rings and bracelets. Earthenware vessels of two classes constitute the greater part of the

SURVEY OF EXCAVATIONS 171

funerary furnishings; on the one hand, coarse unpainted wares and on the other, painted pottery, either polychrome or monochrome. One significant feature of the implements and weapons at Bani Surmah is the fact that they are made of copper and bronze and that, in contrast to those found at the neighbouring necropolis at War Kabfid, they include very few iron objects.

Many socketed axes were found, almost all of a type that is known from mid-third millennium Mesopotamia. Flat axes, socketed spearheads, and knives and daggers of a simple, archaic type were also recovered. From close parallels with the Early Elamite Period at S-isa (Sfisa Dc-Dd) and from others with the Early Sumerian Period in Mesopotamia therefore, we may place the family vaults at Bani Surmah in the last phase of the Early Sumerian Period, that is c. 2600-2500 B.C.

L. VANDEN BERGHE

Choga Sefid The Rice University archaeological team, headed by Professor Frank Hole, has resumed work on

prehistoric sites on the Deh Liiran plain, which lies about ioo miles east of Dizful, with the excavation of Chogha Sefid and further surface reconnaissance for sites of all ages and their relation to irrigation systems.

After six weeks of excavation, we can report only a sketchy outline of what we have found. The dig at Chogha Sefid is being made for the purpose of clarifying the relationship between the earliest ceramic phase we have yet found (the Mohammad Jaffar) and the Sabz Phase which begins the well-known tradition of making buff-ware pottery. A transition between the two phases was not found in our previous work at Ali Kosh and Tepe Sabz. According to radiocarbon dates already published the interval falls somewhere between 5ooo-6000 B.C. It was during this time that such innovations as irrigation agriculture and domestic cattle and dogs were introduced locally.

We had expected to find an intervening phase during which changes in ceramic techniques and agricultural innovations would appear. As nearly as we can tell now, we have three distinct phases with clearly marked ceramic differences, although we have not yet had analyses of animal bones or of carbonized seeds. The unexpected variety in the local ceramic tradition is impressive and may prove to show connections with Chogha Mish in Khiizistdn and with sites across the border in Iraq that are currently under investigation by Joan Oates.

We have apparently established a development out of the Mohammad Jaffar Phase but find the direct connection with the Sabz Phase is elusive. Although some Sabz Phase sherds appear on Chogha Sefid we have not yet tested an area where they are superimposed on our new phases.

As yet, we have not assigned names to the new phases. The uppermost seems to be characterized by red, unpainted wares while the lower two include varieties of painted pottery previously unreported and as yet unstudied. The uppermost phase has well-preserved houses founded with large stones as well as areas of pavement. These architectural remains are now being extensively exposed. The lower phases have been found only in midden deposits in the areas tested.

Another five weeks of digging lie ahead during which we will concentrate on recovering architectural features and in linking the Sabz Phase to those we have recently discovered.

While the digging proceeds, Dr. James Neely is making an intensive survey of the Deh Liiran area with the aid of Elizabeth Carter, a student at the Oriental Institute. So far the principal findings have been the discovery of four very large Islamic sites with their irrigation systems and an impressively large SUisa A occupation throughout the area. It is also noteworthy that Tepe Miisiyan, the largest site on the Deh Lfiran plain is covered over nearly two-thirds its area by Mehmeh and Bayat Phase occupations. The major irrigation networks appear to be Sasanian and Islamic in age although as work proceeds it should be possible to relate some canals to much earlier sites.

After the first of the year we will move our operations to Khfizistdn proper and make an intensive surface survey of the area between Dizful and Sfisa. This will be a continuation of work begun by Robert Adams and will be carried out in collaboration with the French Mission and other American archaeologists. This survey should result in the discovery of many unreported sites and be of great


value in planning future excavations in this important region. We will pay particular attention to patterns of settlement and attempt to gain an impression of the changes in population size and the differences in kinds of settlements. In this we are most interested in the development of nucleated cities that are characteristic of the urban civilizations.

FRANK HOLE

Darvdzeh Tepe On September 18th 1968, Harvard University began excavation at Darvazeh Tepe in Fars Ostan.

The mound lies about 38 km. due south-east of Persepolis, and was first sounded by Professor L. Vanden Berghe (" Archaeologische Opzoekingen in de Mary Dasht Vlakte ", Jaarbericht Ex Oriente Lux XIII, p. 398 f.; and Archeologie de l'Iran ancien, p. 43). The mound is one of the largest ancient sites in Fars, containing approximately 63,500 sq. m. During the first season some 1447 sq. m. of the mound were partially excavated. A stratigraphic clarification and recovery of structural remains within the highest part of the mound were the major purposes of the initial three months expedition.

At least eight occupation phases have been discovered in the excavated area. The three topmost phases contain transient occupation and are recent. The five lower phases contain buildings oriented along a north-south axis, and contain at least one principal industry. Clay deposits, wells and reservoirs, ovens, clay slag, ash heaps, ceramic seconds, unfinished figurines and pots seem to indicate a centre for artisans. From an artistic standpoint, the decorated pottery captures the natural beauty of the Mary Dasht plain. Water birds, mountain goats, running water and textiles are recurrent motifs on a distinctively native ceramic.

Most of the pottery and other artifacts appear to be indigenous to ancient cultures in Fars. Only a few pieces of pottery may have been imported into the area. These imports most likely come from Assyria where they were first manufactured about 1245 B.c.; of course the imports arrived at Darvazeh Tepe later. Although an exact chronology for the occupation phases and cultural sequence(s) must await further excavation and laboratory analyses, present typological parallels suggest a date range from c. 750 B.c. to c. 1750 B.C. in the non-transient phases. If such a date range is proved correct through C14 and thermoluminescence dating, the site may provide an important historical link between Persepolis/Pasargadae and the unknown Elamite cities of Fars. Darvazeh Tepe may indeed live up to its name and become a door to better understanding the area.

MURRAY B. NICOL

Tepe Farukhdbdd

Tepe Farukhab•d is located on the banks of the Memeh River about 12 km. south of the village of

Deh Liiran and I I0 km. west-north-west of Andimeshk. A group from the University of Michigan Museum of Anthropology conducted excavations in the tepe during March and April 1968. The work was supported by the United States National Science Foundation and conducted with much valuable assistance from the Iran Archaeological Service and the British Institute of Persian Studies.

The site is 30 m. in height and at present about 250 x 150 m. at its base. Approximately 60o per cent of the tepe has been cut away from the south and west sides by the Memeh River. The site was probably founded about 5000 B.c. and occupied more or less continuously until 3000 B.C. Subsequently it was

reoccupied in the second millennium B.C. and in the first millennium A.D. The primary object of the project was to excavate several deep stratigraphic soundings in order to provide evidence of the economic development of the town during the fourth millennium. The steeply cut south face permitted us to avoid later deposits not of interest and to dispose of excavated debris with a minimum of effort.

Three excavations were undertaken. Two 5 m. wide trenches were located on the south face in what was once the central area of the town. A I m. wide trench was located on the north face in hopes of locating evidences of town walls. Roughly 50 square metres of each fourth millennium level was


excavated. In Excavation A, to the south-east, we defined thirty-six layers. Layers 33 to 36 are of the Bayat Phase, roughly equivalent to Sisiana d; 22 to 32 are of the tentatively defined Farukh Phase, roughly equivalent to Stisa A; 18 to 21 are Uruk, equivalent to Suisa B; I I to 17 are Early Jemdat Nasr, equivalent to Sisa C; and I to Io are Late Jemdat Nasr, equivalent to Sflsa Da. In Excavation B we defined forty-seven layers. Layer 47 is Bayat Phase; Layers 36 to 46 are Farukh Phase; 30 to 35 are Uruk; 23 to 29 are Early Jemdat Nasr; 20 to 22 are Late Jemdat Nasr; and I to 19 are of the second millennium B.C. In Excavation C, layers of Uruk, Jemdat Nasr and second millennium age were sectioned. Now let us briefly consider the outstanding features of each phase.

Only a small area of Bayat Phase layers was excavated. The ceramics are similar to those from the uppermost layers of the Rice University excavations at nearby Tepe Sabz. Small structures were revealed in both excavations A and B. There were few other features.

The tentatively defined Farukh Phase is distinguished by bowls and jars shaped in the SUsian tradition with bold painted designs commonly including motifs composed of dots, vertical zig-zag elements, and diagonally arranged triangles. In Excavation A, a massive building containing almost no artifacts, a possible granary, and a badly eroded platform were partially cleared. In Excavation B there were a series of small structures. Massive bricks of up to 50 X 25 X 12 cm. in size were used in Farukh Phase structures. Exotic flints are the primary surviving traded commodity. Obsidian occurs, though rarely.

The transition from Farukh to Uruk is far from clear. Further study of the samples is required. Uruk ceramics differ from other reported assembleges primarily in the relatively common occurrence of a small jar with painted black designs on a red slip. In Excavation A, small structures and two rectangular, clay-lined pits filled with carbonized grain were found. In Excavation B a group of inhuma- tion burials and a series of small structures were found. In the latest Uruk layer one massive wall of a large structure of very small riemchen bricks was cleared. Exotic flints drop in frequency, but a few scraps of carnelian, marine shell and copper occur.

The Jemdat Nasr Phases are distinguished by a series of polychrome painted jars. In the earlier phase bevelled rim bowls were still in common use. In Excavation A, a massive structure of large riemchen bricks developed through many rebuildings. In the later stages large plaster vats and grinding slabs are frequent features. In Excavation B, the massive Uruk structure was carried through many rebuildings, reaching its height in Early Jemdat Nasr times and decreasing in size in Late Jemdat Nasr times as the structure in Excavation A grew to its maximum. Exotic flints become more common through the Jemdat Nasr development, and copper, carnelian, and other rare materials become remarkably common.

The second millennium occupation, as revealed in Excavation B, begins as a series of small structures. The ceramics in these layers may be of the period of the Third Dynasty of Ur. Later a massive wall complex was built and much refuse was dumped over its face. This is visible in both Excavations B and C. After the deterioration of this wall, more small structures were built. The ceramics and figurines of these layers may be of the Isin-Larsa Period.

Finally, a large building on stone foundations was constructed on the summit of the tepe. This may have been an Early Islamic fort. It was not investigated.

Some preliminary analysis of the collections from Tepe Farukhabdd was conducted in the month of May. Though much more work is required, it seems probable that the desired data on economic

development were obtained. The archaeological section of the final report should be ready for publication in about one year.

HENRY T. WRIGHT

Haft Tepe For three seasons the Archaeological Service of Iran in co-operation with the Institute of Archaeology

of the University of Tehran has been excavating an archaeological complex in the vicinity of the Haft

Tepe sugar cane plantation, in the south-western part of the Khtizistan plain. At this site, which was


Fig. I. General plan of temple and tombs.

P1. IA

. H

aft T

epe: L

arger vaulted tomb.

Pl. IB

. H

aft T

epe: Smaller tom

b with tw

enty-three skeletons.

Pl. IIA. Haft Tepe: One of the temple halls.

Pl. IIB. Haft Tepe: Stone inscription.

Pl. Il4A. Front nord de la terrasse, digagi en 1965.

Pl. IIIB. COramique peinte du secteur stratifie; niveau des 22. 75 m. Vers 2400 av. J.C.

Pl. IIIC. Au premier plan, couche de Djemdet-Nasr; au-deli, face du socle de la phase finale de Suse I.

Pl. IIFE. Ciramique de la premiere phase de la terrasse de Suse L.

P1. IIID. Paroi nord du massif de Suse L De haut en bas; clous dicoratifs; niveau des

21"34 m. avec tessons peints; sol infdrieur

a 21 I2 m. et couche de galets a 21 04 m.

Pl. IIIF. Empreinte sur argile de lapremierephase de la terrasse de Suse L

P1. IVA.: Tepe Traya

Pl. IVB.: Female idol from Tepe Trahyi, Period VIII


selected in order to increase our knowledge of the history of the Elamite people, we have, during past seasons, uncovered a tomb and temple belonging to the middle of the second millenium B.C.

Here I shall make a brief report of the third season of excavation at Haft Tepe which began in early January and ended on April 5th 1968. During this season we continued to clear the temple area with the hope of ascertaining its limits; we began clearing the western edge of the incline where it had been levelled for the construction of a road by the sugar cane plantation; and finally we continued eastward on one grid line only toward one of the larger mounds in order to learn something of the contents of these major mounds, of which eleven exist, all still untouched.

The temple area which we finished uncovering in this last season is shown on Fig. I. It contains (from the north) two tombs with vaulted roofs, one more carefully constructed and larger than the other.

The larger tomb is o- 30 x 3 -25 m. in size and contained twenty-one skeletons (Pl. IA and Fig. 2). A terrace 6o cm. high covers about four-fifths of the floor area, leaving a canal along one side of the length of the tomb and an empty space at the back. This terrace was divided into three sections by very narrow low walls. In the first section seven skeletons were found, lying parallel to one another, together with a very few funerary objects, including some small jars. To the height of the partition the skeletons had been packed in red ochre, traces of which still remained. The middle section, which was the smallest, was empty. The back section contained two very incomplete skeletons. Behind these terraces was a doorway leading into the temple up a flight of stairs. Near this doorway were twelve roughly heaped incomplete skeletons. The low partitions of the terrace have many holes allowing drainage into the canal which runs alongside it (Fig. 2).

The second tomb is much smaller than the first, 6 x 2-5 m., and not so well constructed (P1. IB). The entire floor was packed with fourteen skeletons, buried in an orderly fashion, parallel to each other with straight legs. Above these carefully arranged skeletons, nine more skeletons, carelessly dumped, were uncovered. The first tomb is connected by a passage to one of the main halls, but the second tomb has no such connection although it is located parallel to the first at the north end of the temple. In addition to the tombs, simple burials without construction and jar burials were found in rooms around the temple.

Behind the tombs are two halls, each 9 x 7 m. with floors of baked brick coated with gypsum (P1. IIA). Below the floor of one of these halls there seems to be another room which has not yet been excavated, except for a passage which we have cut through. It is not as yet clear whether or not this floor actually is a roof for a lower room. Unfortunately part of the other hall was shaved off in levelling the ground by the sugar cane factory and only a small part of its paved floor remains.

The halls lead to a rather large portico, 20 m. long and 4 m. wide. This portico opens onto a courtyard about 30 m. long and 20 m. wide. In the centre of the courtyard is a platform, possibly an altar, 50 cm. high and dug I m. below floor level, built of nine layers of baked brick. Next to the platform we found many fragments of inscribed stone, part of two large inscriptions which were apparently installed here. At a corner of the courtyard were a heap of ash and charcoal, a hearth and a little terrace.

All the tomb structures, the floors of the halls and the floors of the portico and courtyard are built of large baked bricks. The rest of the construction is of very large mud bricks. The entire temple area, consisting of the tombs, halls, and courtyard, is surrounded by a wall 4 m. thick with only two openings, on the southern side of the courtyard. One of these openings is a passage leading to a rather small closed room with a floor of gypsum-coated baked brick. The other opening leads to a rather larger room with more passages which are not yet excavated. Possibly they lead to the other large constructions existing on the major mounds which we are just beginning to investigate.

In continuing our excavation on one grid line toward the east where the major mounds are located, we uncovered a wall 5 m. thick extending in a north easterly direction (Fig. 3). The length already opened is more than 50 m. At the point where we had reached one of the major mounds, this wall joined at right angles with a huge wall, possibly for fortification, 12 m. thick with 9 m. of height still standing. At the angle where the two walls joined, we found several large halls, each of which, still only partly excavated, may possibly be Io x 6 m.

Beside the long wall we came across a great deal of ash, charcoal and rubbish in which we found


Fig. 2. Plan of larger vaulted tomb.

HAFT TAPEH

Fig. 3. Plan of excavation extended on grid line to major mound.


many broken pieces of seal impressions, some with designs and some with vertical lines of cuneiform inscriptions. These impressions are being studied by Prof. de Meyer of Belgium, and we hope to gain more information about Haft Tepe from their translation.

A variety of pottery shapes and thousands of sherds were found along the wall in the rubbish

deposits. This pottery, mostly buff without designs, is on the whole comparable to pottery of the middle of the second millenium B.C. in Mesopotamia and Khiizistan. Other objects found in the excavation of this season include broken figurines, clay tablets, a bronze dagger, bronze ingots and a few other bronze objects.

In one of the grids, on this grid line, near the major mound, we dug a deep test trench, 2 m. square. In this test trench we discovered two main constructional levels below the level (middle of second millenium B.C.) which we have been excavating for the past three seasons. Although we have no idea of the plan and date of these two lower constructional levels, we could see that the walls were very thick and built mainly of mud brick. We hope that in the future, after clearing the top constructional level, we may have the opportunity in one area of continuing down to these lower levels and learning more about these older constructions.

As we continue to work at Haft Tepe, we have begun to realize the immense size of the archaeological remains in these mounds, which will require many seasons to excavate. The extremely large walls which we uncovered at the end of the last season as we reached the first of the eleven major mounds to the east of the temple compound, lead us to believe that very large constructions must be buried under these mounds. Therefore I believe that at this point excavation on a much larger scale should be undertaken.

During each season we have excavated written documents, either clay tablets or stone inscriptions. The stone inscriptions discovered in the first season, bearing the name of the king Tepti-ahar, indicate some rather strange religious ideas according to epigraphers and will, I hope, soon be published (P1. IIB). The second season stone inscription is still awaiting translation. During the last season a large clay tablet was found which may be an astrological omen, together with other tablets and seal

inscriptions which also await translation. I am very happy to announce that the project of building a museum and excavation quarters at

Haft Tepe is being realized through the efforts of H.I.M. Queen Farah. The plans are ready, money has been budgeted, and we hope that in a few months construction will begin on the museum which will keep all the objects found at Haft Tepe permanently at the site.

This is only a brief general report of the last season of work, so that my colleagues may be kept informed of the progress of our excavation. I hope soon to publish a systematic preliminary report of the Haft Tepe excavation.

EZAT O. NEGAHBAN

Haftavdn Tepe For nine weeks from July 9th 1968 excavations were carried out at the large mound of Haftavan

Tepe, situated three miles south of the town of Shahpir, in Western Azerbaijan, and about one and a half miles from the village of Haftavan. These excavations were under the direction of the writer and were sponsored by the University of Manchester, with financial support from that University, from the British Academy and from the Calouste Gulbenkian Foundation. A staff averaging ten in number was at work with the expedition, and a labour force starting with small numbers and rising to about I20 men was employed on the site. This is a major site, the area of the mound being over 6oo x 500 m., making it one of the three or four largest prehistoric sites round Lake Reza'iyyeh. Both previous expectations and the results of this season justify plans for a campaign of several seasons' excavations at Haftavdn.

The mound of Haftavdn Tepe is situated in the middle of the small plain at the north-west corner of Lake Reza'iyyeh, and is now accessible only by tracks through the irrigated fields; it has no village


next to it. A spring at the eastern end of the site must explain its situation. The size of the site makes it clearly that of a town of some importance, and attracted an excavator who, about thirty-five years ago, cut a trench right across the highest part of the site (the citadel) and also a second trench along the south side of the citadel. Only the large area of the site has made these trenches less destructive than they might otherwise have been: no record has been published. Quarrying for soil by villagers has also damaged the mound, though the north side has remained almost unaffected. The distribution of these pits and trenches naturally determined the location of the expedition's areas of excavation.

The first area to be excavated revealed a building level immediately beneath the surface of the mound, where the pottery and finds from a burial indicated a date in the eighth or possibly the seventh century B.C. Then work was begun on a long trench down the west side of the site, from the slope of the citadel across the lower skirt which surrounds it. Soon it became clear that a deep fill of decayed mud brick and silty wash had to be removed before any building remains could be reached; and here eventually was uncovered an impressive building constructed of undressed stones set in mud-a method still used in some neighbouring villages-and with its walls preserved up to 2 i70 m., the floor being at a depth of 5 - 60 m. from the surface. At first this building appeared to form a defensive tower, perhaps part of a gateway, but later extensions of the excavations either side of the main trench have shown that it more probably represents the house of a leading member of the community, contrasting with the humbler dwellings found in the same level further along the trench to the west. Overlying this building level were found burials with distinctive grave goods, two of them having bronze chains worn round the neck and numerous bronze pendants. A date for these burials towards the very beginning of the first millennium B.C. seems likely. In occupation overlying the main building level was found pottery of HasanlGi V type, so that a date for the stone-built houses of c. 1500 B.C. would be perfectly possible, though any precise chronology is still unobtainable. The excavations reached part of an earlier level immediately beneath, in which pottery derived from the Early Bronze Age forms was found: this pottery could date within the period c. 1800-1500 B.C.

Meanwhile excavations were being carried out on top of the highest part of the mound, the citadel, in two adjacent areas eventually joined together. Here a building level with stone foundations and a stone-paved street was found just below the surface in part of the excavations, though elsewhere it had been completely robbed out. Stone-robbing and a large number of pits make the detailed interpretation of the results of the excavations of the citadel a difficult task, though the general cultural sequence is now clear. The earliest level reached this season probably dates to the seventh century B.C., with foundations of a massive building and pottery including recognizably Urartian wares, though a date in the sixth century B.c. is not impossible; such a date would make this level attributable to the Medes. The levels overlying this, pock-marked as they are by pits, have yielded pottery providing dating evidence through parallels with Hasanlti IIIA (c. 600-400 B.C.) and with Achaemenian forms found elsewhere. Round the summit of the citadel during this period were built defences of mud brick walls, of the thin square bricks associated with the period. The uncovering of the original defences forming the citadel awaits future work. Then, at some time towards the end of the Achaemenian empire, the settlement was evidently abandoned. In the early Sasanian period the top of the citadel was used as a cemetery, the one secure piece of dating evidence being a silver coin of Shahpir II (309-379 A.D.): these burials have yielded miniature glass bottles, fragments of glass vessels, many colourful beads, signet rings and miniature glass mirrors. The absence of Islamic finds makes it seem probable that the stone street and walls sealing some of these burials-and so presumably the whole cemetery-date to the late Sasanian period rather than any later.

In an area opened up at the very edge of the mound near its east end two building levels with remains of houses and large enclosures were found, together with two phases of burials. Most of these date to the eighth or seventh century B.c., but one grave stratified in an earlier level, in the top of the fill of a large pit, contained a spouted jar of Hasanli V type, attributable to the period c. I200-1000 B.c. The contents were not the same as those of the burials above-mentioned with the bronze chains.

The results of this season show that the centre of this town site seems to have shifted eastward between the mid-second millennium B.c. and the early first millennium, while in the latter period the citadel must have been established. Eventually, by the Achaemenian period, the settlement may have


been restricted to the citadel. These changes should make it possible for areas of different periods to be exposed without always having to clear down through later levels, except of course on the citadel.

C. A. BURNEY

Hasanla Project 1968: Hajji Firuz, Dinkha Tepe, Se Girdan, Qalatgah Hajji Firuz had previously been excavated by Charles Burney (1958) and T. Cuyler Young, Jr.

(1961) as part of the IHasanloi Project. Their work established a sequence for the site running from the Chalcolithic Pisdeli period (c. 3500 B.C.) back into the Late Neolithic Hajji Firuz period (c. 500ooo-6ooo B.c.). The latter material is the earliest known in the Solduz Valley, and equivalent to Period X at Hasanlti Tepe.

The aim of this season's work was to clear as wide an area of a single Hajji Firuz level as possible in order to obtain information on the architecture of the period, and to relate the pottery and small finds to any domestic structures found. Special care was also taken to recover plant and animal remains in an effort to learn about the economy of the Hajji Firuz people. The main area of excavation lay on the east side of the mound: ten 5 x 5 m. squares were opened, eight of these yielding architectural remains, a total of 200 sq. m. The latest Hajji Firuz level in this area had been almost completely destroyed by pitting from Chalcolithic to recent times (two of these contained burials: (I) a HasanlGi V (early Iron Age) burial with a fine black burnished spouted vessel; and (2) a burial dating to the Bronze Age " Painted Orange Ware " period-HasanlGi VII). Excavation was concentrated on the next two Hajji Firuz building levels. The upper building level consisted of two houses separated by a narrow alley. One of these houses is almost complete, measuring 6-6 x 6-2 m. It had a large living area (heated by a horseshoe-shaped hearth), separated by a partition wall from a cooking and storage area which contained two more hearths, a series of bins and large storage jars sunk into the floor. The house is constructed of mud brick and has clay floors which were periodically renewed; the uppermost floor of the building was covered with a red colouring matter, perhaps ochre. Fragments of roofing found in the fill indicate that it was made of mud resting on a bed of reeds, undoubtedly supported by wooden poles.

At the back of the main or living room of this house was a bin, which had been covered by a layer of hard-packed clay and contained the burial of at least thirteen individuals, five children and eight adults. The bodies were presumably put into this tomb over a period of time, and only the uppermost skeletons were fully articulated. The bones were covered with red ochre. Grave goods included fourteen biconical spindle whorls of clay, four small pots, flint and ground stone tools, and a small rectangular stone which had been carved with linear geometric designs.

The second (lower) building phase consisted of a complete house, and parts of at least six other buildings. The house, again of mud brick, was smaller than the building described above, measuring 4'40 x 4'40 m. It consisted of a single room partially divided by a wall, which extended only halfway across the building; this partition wall probably served to help support the roof, enabling the builders to use shorter beams. The doorway to the house had a raised sill, and a single step down to the inside floor level. Interior fittings included a horseshoe-shaped hearth and a low curbing, setting off the area on one side of the partition wall from the main room. Beneath the floor of this " set off " area was a quantity of animal bones, and the scattered remains of at least three human skeletons.

A second partially preserved building is of interest as it has a narrow curving wall attached to it, providing an outside enclosed area adjacent to the main structure.

Small finds from the site were, with the exception of clay spindle whorls, scarce. The most interesting were small ground stone celts and fragments of small female figurines made of clay. The figurines are in a style presently known only from Hajji Firuz, and have narrow skirts decorated with punctuate or fingernail impressions. Only a small quantity of chipped stone was found, consisting of approximately equal amounts of flint and obsidian. The pottery is a poorly fired buff ware, burnished on all but the largest vessels. Included in the ceramic industry is a red slipped ware, and pottery decorated with red or brown paint in geometric patterns, usually chevrons or triangles. The closest parallels in both shape


and design are with the Hassuna culture of Mesopotamia, but the Hajji Firuz ceramics must still be considered as a distinct stylistic unit which seems to extend north along the shores of Lake Reza'iyyeh, perhaps as far as the Araxes River.

Further excavations were carried out in the second millennium B.c. levels at Dinkha Tepe following up problems started in 1966. One area next to the river was taken down to virgin soil with no radical

change apparent in the material.

The sondage at Se Girdan, south of the village of Cheshme Gol, was made in order to establish whether the suggestion made in 194o by Sir Aurel Stein that the tumuli mounds were natural was correct. It was noted during the preliminary survey that there are in fact eleven tumuli, not five as was

thought by Stein. Three tumuli were selected for preliminary investigation: a large one, a medium one, and a small

one. The large tumulus, called Tumulus I, is 8-22 m. in height and close to 60 m. in diameter. Several

test trenches were dug around the perimeter and it was established that the tumulus is constructed of a

very hard clay. Moreover, there is a ring of small stones completely surrounding the tumulus along its

slopes. This rubble ring apparently served as a sealing over the clay and was only 50 to 60 cm. below the surface. A pie-shaped cutting was also made to see if any other features were to be found: none were. Work was then stopped.

The second tumulus examined is called Tumulus II and was only 47 cm. in height above the

surrounding fields. A massive circular stone pile of rubble stones was found and cleared. In two areas of this rubble pile, which was found near the centre of the tumulus, gaps in the filling were noted

indicating that plundering had occurred. Nevertheless, we continued to clear the area in order to get a coherent plan of the structure. A rectangular stone tomb was cleared below one of the plundered areas. There was no evidence of a roof and the tomb was completely empty but for a small shell (?) ring.

The third tumulus, called Tumulus III, was a medium sized one rising only 3 m. above the sur-

rounding plain. Here we removed one quadrant, i.e. one-fourth of the tumulus, and found a rubble

ring surrounding the tumulus in a manner very similar to that noted in Tumulus I. Near the centre of the tumulus we found also a pile of rubble stones similar to that noted in Tumulus II: near the centre of this rubble we found a depression indicating that the pile of stones had collapsed into a hollow area below. Upon removing these stones we came upon a pit cut into the virgin soil: the pit is

3.50 x 2 m.

and is I - 50 m. deep. There was some evidence that the pit had been covered with a wooden roof as traces of wood were found among the stones. The bottom of the pit was covered with small pebbles for a

depth of about 7 cm. and on this floor was found a skeleton of an adult male lying on his side with his feet drawn up. To the south-west of its head were found six objects, mostly badly crushed, one of which was a stone object ending in a stylized lion's head; another was a badly shattered silver vessel; also two shattered silver tubes; a bronze knife; a bronze adze; and beads of gold, silver, paste, and stone, apparently coming from a necklace. No pottery was found and it is not at present possible to suggest a date for this or the other tumuli in the area. More work is needed in the library in order to seek out

parallels for the objects recovered; and it is hoped that excavation of some of the tumuli will aid in

finding out their date and the culture responsible for their erection.

The expedition spent several days working in the area of Cheshme Gol, especially at the site of

Qalatgah immediately above the spring. Sherds were recovered which indicate a time range parallel to that of Hasanlti IIIA and B; namely from c. 750-400 B.C. Sherds of triangle ware and red slipped bowls are characteristic Hasanli types found at Qalatgah. A small alabaster stamp seal of Urartian shape but bearing an Achaemenian design and sherds of late III type indicate that the site continued to be occupied into the Achaemenian period, although continuing local pottery traditions originating under earlier Urartian influence. The architecture of the site is clearly Urartian as may be seen from the remaining cut stone blocks of the retaining walls.

At the spring Cheshme Gol we discovered a reused building block bearing six lines of a cuneiform


inscription. The inscription is Urartian and includes the name of the king Ishpuini in the patronymic form and thus belongs to Menua (although his name is broken away at the edge of the block). The inscription thus belongs to the same time range as the Kel-i-Shin and Tash Tepe inscriptions and helps to establish the line of the Uratian frontier south of Lake Urmia in the early eighth century. The Project hopes to do further work on the Qalatgah site when the current Se Girdan and Dinkha phase is concluded.

ROBERT H. DYSON, Jr., OSCAR WHITE MUSCARELLA and MARY M. VOIGT

Kaluraz Following the excavations of 1967 (cf. Ali Hakimi, " Kaluraz et la civilisation des Mardes ",

Archdologie Vivante no. I, 1968, pp. 63-5), the Archaeological Service of Iran again resumed work at Kaluraz in June 1968. The most recent discoveries from this rich cemetery situated 55 km. south of Resht are reported to come from graves of either the eighth-seventh centuries B.C. or the third century B.C. Of particular interest are more than a dozen handsome bronze animal figurines from the eighth- seventh century graves, each with a uniform, highly stylized design not exactly matched elsewhere.

Kangavar The Archaeological Service of Iran-which also plans to open important excavations at Bishapiir,

Hamadin and Persepolis-has recently begun new work at Kangavar, the site of the renowned temple of Anahita. Under the direction of Mr. Kambaksh Fard, the western limits of the temple have already been exposed in preparation for a more extensive campaign in 1969.

Shahr-i Sokhte The Italian Archaeological Mission has established the existence of one of the main proto-urban

centres of the Indo-Iranian borderland at Shahr-i Sokhte in Iranian Sistin. After two successive campaigns of excavation, this site, which was discovered by Sir Aurel Stein in 1916, is yielding substantial new information about the process of urbanization and civilization east of Sumer.

Salt sedimentation has permitted recognition, without excavation, of an extensive network of walls, lying below ground level. This network of walls extends over of a large clay mesa for some 2000 sq. km. and overlooks the dry portion of Lake Hamun.

The excavations have been conducted in a flat area of some 18oo sq. m. in the centre of the mound. A group of rectangular buildings was discovered, four of them concentrated in an insula isolated by winding alley ways averaging 3 m. in width. The buildings have six or eight symmetrical rooms divided by minor rectangular walls.

Stein thought that violent winds had completely eroded the entire earth deposit, leaving only the potsherds. Contrary to this theory, our work has shown that this deposit has been exceptionally well preserved. The uncovered walls project to a height of over 3 m. Doors, windows, fireplaces, staircases, furnaces and other architectural features are finely preserved. Also, the very dry climate of Sistdn has preserved in good condition most of the wood remains: burnt roof beams, straw mats and tools.

The complete excavation of a single dwelling house shows that these well-preserved walls were not constructed at one time. Periodically, perhaps as a consequence of the Hilmand river floods, the plan of these buildings was raised, thereby filling their rooms with an enormous number of potsherds and bricks. Walls and fireplaces were rebuilt directly on those of the preceding phase, the older doors and windows were blocked with brickwork and plaster. At present four different phases are distinguishable: in the last, most of the rooms were subdivided with some evidence of an extensive destruction by fire. The situation reminds one of the last period of Mohenjo Daro.

The pottery found in these fillings is closely related in shapes and patterns to the Bamptir valley,


Mundigak IV and Northern Baluchistan. There is evidence that Shahr-i Sokhte was a production centre for alabaster vessels and generally for hard stone manufacturing, espcially for lapislazuli, cornelian and turquoise.

MAURIZIO TosI

Sirdf The third season of excavations at Sirdf began in October 1968 and at the time of writing was still

in progress. The excavation, which is sponsored by the Institute, received generous support from the Calouste Gulbenkian Foundation, the British Academy, the British Museum, the Royal Ontario Museum, the Monro Trust, the Pilkington Glass Museum, the Russell Trust and the Corning Museum of Glass. The British Museum and the Royal Scottish Museum kindly released members of staff to join the expedition. During the excavation we began to build a permanent headquarters and museum, the cost of which was borne in part by a generous grant from the Iranian Archaeological Service.

We excavated in three areas: at the Great Mosque (Site B), the group of tenth- or eleventh-century houses (Site F) and at the ruined mosque described by Stein (Site G).

Site B: During the second season, we excavated most of the Great Mosque to the level of the latest floors (Sirdf II, pp. 39-62 of this journal, especially pp. 4 I ff.). This year we completed our work in the extension, made a series of soundings in the main enclosure and excavated a group of buildings on the south-east side of the mosque. The most important result was the discovery of stratified deposits under- neath the Great Mosque. Associated with the footings of rectilinear stone buildings was a vast quantity of pottery, much of which differed markedly from the wares found previously at Siraf. Many of the forms have Sasanian parallels and clearly the structures are Sasanian or very early Islamic. Among the pottery are two sherds bearing Pahlavi graffiti. In the extension, we confirmed that the paved floors found in most of the bays were consistently overlain by trampled earth floors containing Seljuq pottery and later sgraffiato ware. An earth floor in bay 7 yielded a dinar struck in I146-47. Outside the mosque we excavated the area bounded by the outer wall of the main building and bay i i of the extension. Most of this area was occupied by a rectangular enclosure containing a cistern. The enclosure was accessible from the extension and was in use during Period 3 in the development of the mosque. On the south-east and north-east sides of the enclosure were two groups of small structures, identified as shops, beyond which were winding streets. The most distinctive pottery from these buildings was later sgraffiato ware and it is probable that the area was in ruins before the end of the fourteenth century.

Site F: In 1967-68, excavation at Site F disclosed an impressive courtyard house, at the side of which was a narrow street. This year we investigated three buildings to the east of the house and a fourth building to the north. The work revealed a residential quarter with a regular grid of streets, contrasting sharply with the confused plan of the commercial area outside the mosque. On the east side of the courtyard house were two smaller buildings, separated by a narrow alley. The first building (Sirdf II, Fig. 4, vertical hatching) measured approximately 15 X 16 m. and had nine ground floor rooms, with a courtyard in the centre. Two of the rooms had vaulted roofs and throughout the house decorated stucco was abundant. The stucco resembles the plaster found in 1967-68 and includes triangular panels closely similar to SirdfII, P1. VII(d) and Fig. 9, 3. The second building (Sirdfll, Fig. 4, transverse hatching) measured approximately 15 m. square and had a complex history extending until at least the later thirteenth century. Among the finds from the overlying levels was a hoard of sixty Chinese copper coins associated with two Islamic coins of the fourteenth century. Fifty-five of the Chinese coins were identified on the site and, with one exception, all were minted between 976 and 1252. To the east of the two houses was a narrow street, beyond which was another large building. The long wall in front of the house excavated in 1967-68 marks the edge of a broader street, on the north side of which were further impressive buildings. One of these, a house approximately 16 m. wide and 21.5 m. deep, is being excavated as I write.

Site G: Finally, excavation at Site G (Sirdf II, Fig. I, masjid) revealed a mosque with a mihrab decorated in Safavid style with panels of cast plaster.

DAVID WHITEHOUSE


Suse Les travaux sur le site de Suse se sont cantonnes, cette annie, au tell de l'Acropole, et ont durd du

8 janvier au 6 mars. En attendant la publication prochaine d'un rapport exhaustif sur les fouilles

effectudes depuis 1965 dans ce secteur nous en pr6sentons, ici, a grands traits, les principaux r6sultats. C'est dans une aire d'environ 150 m. sur Ioo m., situde dans la partie centre ouest du tell, et qui

avait ~ t~ pargnde par les grands sondages de J. de Morgan et R. de Mecquenem, que nous avons

entrepris de nouvelles fouilles. Les premiers travaux - en 1965 - ont mis en evidence l'existence d'un socle en briques crues dont

nous avons d6gage alors une section de la paroi nord qui s'enfongait a pres de 3 m. sous le niveau de

depart. Ce niveau est situd a la cote des 22 m. au-dessus de la plaine. (P1. IIIA). Le front de ce socle avait etd modifid posterieurement ta plusieurs reprises. Le dernier &tat de refection etait representd par un mur massif, a saillants et rentrants, bati en briques crues de 25 x i xI i I cm., de type " riemchen ". La c6ramique permet de dater cet ouvrage de la periode finale de Djemdet-Nasr ou des debuts du

Proto-dynastique. L'ann6e suivante la fouille progressait en surface vers l'ouest et mettait au jour une longue muraille de plus de 30 m. dans le sens nord-sud, sans parement interieur. La face exterieure etait revetue d'un enduit d'argile d'une epaisseur moyenne de Io cm. Ce mur se pr6sentait comme une extension plus recente, du c6td ouest, des vestiges decouverts dans le sondage de 1965. Trois etats de construction etaient nettement perceptibles; le dernier pouvait etre date par une petite tablette de

l'dpoque d'Agadd mentionnant une livraison d'orge et de farine. En avant de ce mur une bande de terrain, conservee sur une largeur de 5

' 6 metres, livrait un lot de ceramique peinte de type " Suse II ". En 1967 le chantier s'installait au nord-est du premier sondage, dans un secteur oii subsistait une

importante levee des 23 m. A quelques centimetres ' peine sous la surface actuelle apparaissait un

niveau d'occupation qui pouvait tre mis en relation avec la phase la plus r6cente du long mur decouvert en 1966 sur le front ouest. La plupart des structures ne conservaient guere qu'une assise ou deux de

briques crues. Malgrd leur etat lacunaire les dimensions importantes de ces constructions, leur

regularite, 6voquent des edifices publics plut6t que des habitations particulieres. Un clou inscrit, au nom du souverain dlamite Puzur (ou Kutik)-Inshushinak (ca. 2200 av. J.C.), fait 6tat de la construction d'un temple au dieu Shugu. La ceramique du niveau est bien typiquement agaddenne: jarres a 6paule fortement carinde et ornde de c6telures, ta fond arrondi sur anneau, coupes a flancs evases et bords

1egerement evers6s, bols 't rebords retournds, fioles 'a panse piriforme, etc. Quelques rares tessons

peints portaient un decor purement g6omitrique. La date proposde est corroborde par la pr6sence dans ces ruines de deux grands puits construits en briques cuites qui coupaient les murs de notre niveau. Le materiel le plus ancien retire de ces puits - c6ramique et figurines - appartient ' la periode d'Ur III.

La couche archeologique sous-jacente a ces constructions de surface a livrd une forte proportion de la

ceramique peinte dite de "Suse II" (P1. IIIB). Celle-ci est accompagnee de grandes jarres '

decor applique: bandes rubannees '

impressions digitales, cols a contreforts, ouvertures '

capuchon semi-

spherique. De nombreux tessons portaient en relief l'image du serpent; celui-ci se retrouve en entier sur l'epaule d'une jarre intacte placee sous le point de depart d'une canalisation en grandes briques cuites (42 X 21 x 6 cm.), qui traverse une bonne partie du chantier. Cette couche - niveau moyen des 22-75 m. - est a situer au debut de la pariode d'Agade ou a la periode Proto-impiriale (vers 2400 av. J.C.).

Au sud de ce secteur, qui parait avoir conserv6 une stratification normale, et au-del d'une rue de

5 m. de largeur, est apparu un mur de briques crues, sans parement interieur, analogue a celui qui avait 6td decouvert en 1967 sur le front ouest. Ici il s'agissait de la face nord de la meme structure; les sols indiquaient trois remaniements et ils Ctaient en connexion avec les niveaux identiques du secteur stratifiC dtcrit ci-dessus: fin du Proto-dynastique, Agad6 ancien et Agad6 ricent. En retrait de ce front recent, i 25 cm. environ de la surface, le sol est jonchC de poteries de l'6poque de Djemdet-Nasr mel6es t de grands clous d~coratifs en forme de cornets, t des briques cuites, & des Cl6ments de gouttiRres.

L'hypothise d'une terrasse occupant le centre du tell de l'Acropole se prdcisait done. La campagne de 1 968 devait permettre de retrouver le noyau primitif de cette plate-forme dont nous n'avions touch6 jusqu'ici que les remaniements tardifs. Les travaux entrepris sur le front nord, a quelque 5 metres au


sud des murs degages l'annee precedente, ont mis au jour une paroi de briques crues qui representait la face d'un vdritable massif (P1. IIIC). En degageant cette paroi on rencontrait d'abord des clous d6coratifs - cylindre termind par un disque - tres diff6rents des " cornets " de Djemdet-Nasr; au niveau des 21.34 m. le sol, de terre battue, 6tait recouvert de nombreux tessons de caramique peinte du type dit de " Suse I ". Le massif se prologeait jusqu'a un autre sol situ & 21 12 m. et reposait sur une mince couche de galets, parfois melds

' des cendres, au niveau des 21 -04 m. (P1. IIID). Aussi surprenant que cela puisse paraitre il devenait 6vident que des structures de la p6riode de Suse

I, reperdes ailleurs a une dizaine de metres plus bas dans la masse du tell, 6mergeaient ici tout pres de la surface actuelle, jusqu'au niveau des

23"73 m. atteint par le sommet du massif. Ces constructions de

Suse I ont etd recouvertes par un vaste remplissage de briques crues t l'apoque de Djemdet-Nasr. Mais la phase finale de Suse I, reprdsentee par le sol des 21 *34 m., constituait ddji un remaniement assez radical de structures plus anciennes. Ces dernieres 6taient fondees sur le sol des 21 m.; ici aussi les murs ont et6 noyds dans un remplissage de briques crues qui a transformd les batiments primitifs, el6vis au centre de I'Acropole, en une sorte de socle encore conserve sur pris de I - 75 m. de haut. C'est sur ce socle que se dressaient les constructions de la phase finale de Suse I, arasdes a leur tour et niveldes durant la p6riode de Djemdet-Nasr. Ces remaniements sont apparus 6galement en divers points du chantier.

Que reprdsentait le niveau des 21 m. ? Un puits de sondage nous a fourni provisoirement des indications fort intdressantes. Le remplissage de la phase finale de Suse I se termine, ici encore, sur une couche de galets et de terre charbonnde, t 21 .03 m. Au-dessous, les lits de briques crues descendent, sans solution de continuitd, jusqu'au niveau des io* 80 m. Entre le dernier lit et le sol vierge probable, %

950 m., on rencontre encore des tessons peints de Suse I, de la ceramique fine rouge et lustrde, des nuclei et lames de silex.

Nous pensons que les vestiges rvdl16s par ces quatre breves campagnes representent le haut-lieu

primitifde Suse, dleve sur la butte d'environ 9"

50 m. qui devait devenir I'Acropole. La premiere occupation, dont temoignent I - 30 m. de decombres, a ete suivie par une pdriode au cours de laquelle fut drigde cette terrasse, d'une hauteur de IO

- 20m., qui servait de plate-forme gigantesque au sanctuaire de la cit6

et a ses dependances. Plusieurs indices, en effet, portent 'a croire que les edifices qui s'6levaient au centre de la plate-forme, avaient un caractere sacrd: nombreux clous d6coratifs recueillis a la base des murs, petites cornes votives en argile cuite et a decor rouge et noir, fragments de vases peints portant l'image du serpent. Sur le pourtour de la terrasse, au-delh d'un espace libre d'une dizaine de metres, de petites pikces, regulierement aligndes, representent probablement les annexes du temple: reserves et magasins. On les retrove aux deux periodes de Suse I.

II faudra attendre que l'etude du materiel, les analyses de carbone 14 peut-etre, nous fournissent des donnees precises, avant de tenter de situer ces vestiges dans un cadre chronologique. On peut ddj~ tenir pour assure que le dernier remaniement de la periode de Suse I coincide avec la fin de la ceramique peinte de ce style. Il parait bien, a maints indices, que cette ultime phase se prolonge assez loin dans la

periode d'Uruk. C'est par-dessus ces vestiges, apres un temps d'abandon, semble-t-il, et souvent en contact avec le materiel de Suse I, que l'on rencontre les briques et la cdramique de Djemdet-Nasr.

M.-J. STEVE

Tepe ra.hy

d The first full season of excavation at the site of Tepe Yahyd, supported by the National Science

Foundation, Ford Foundation and Harvard University, was carried out between June 28th-September 3rd 1968. The site, discovered the previous summer on survey, is located in the fertile Soghtin Valley, 2oo km. directly south of KirmTn. Tepe Yahyd measures 19 m. in height, forming a near perfect circle at its base, with a diameter of I90 m. From the base of the mound sherds extend along both banks of the Kish-e-Shfir River (directly adjacent to the mound) a distance of I km. to the north and 2? km. to the south. Tepe Yahyd represents the largest and most imposing pre-Islamic mound discovered in south-eastern Iran (see P1. IVA).


Seven Io m. square trenches were excavated, providing a step trench from the top to the foot of the mound. At the base of the mound water was reached at 6.35 m. beneath the present surface. The sequence of occupation on the mound can be tentatively put forward. Further clarification of this sequence, together with greater horizontal exposure of certain periods must await our second season of excavation, scheduled for the summer of 1969. The sequence of occupation is here outlined as a series of VIII Periods-stratigraphically defined from top to bottom.

Period L The top of the mound is capped by a mud-brick construction (brick size: 40 x 40 x 15 cm.), forming a large citadel structure, incompletely exposed. An extensive mud-brick platform has been uncovered which is built up along the edge of the mound to form a fortification wall. Considerable evidence exists for multiple re-facing of both the wall and extending platform. Ceramic evidence consists largely of storage jars with incised and raised motifs, coarse, grit tempered, plain buff wares, predominantly bowls and jars with some spouted jugs. Fine wares, in the minority, are either plain buff slip or orange-red slip bowls. A light greenish-blue glazed ware is also rare. A variety of beads, glass cosmetic jars and bronze and iron implements were recovered. Radiocarbon samples from this period, and all other periods, are presently being run. Period I is identifiable as Parthian.

Period II. No clear architecture has been recovered from the exposed levels of this period. However, the floors here, defined outside of the Period I wall, suggest the possibility that much of the Period I architecture was in use during this period. Many shapes and wares of Period I also characterize Period II; however, markedly different types do appear: a plain fine buff ware with a rich red slip, often burnished, high neck jars with twisted rope handles, trefoil mouth jars and small cosmetic jars and a rare ware with black painted designs over an orange-red slip. This material is clearly distinguishable from the Period I (Parthian) and the earlier Period III (Achaemenian) wares. It is difficult resisting the temptation of identifying this as the site mentioned in classical texts as Carmania, the Achaemenian citadel surrendered to Alexander the Great. Its location fits all too well the classical references while no other known site in the area evidences such an impressive Achaemenian occupation as seen in Period III. It is possible that we have here an Hellenistic occupation.

Period III. This period is characterized architecturally by large-scale constructions in mud-brick (36 x 36 x 13 cm.). No single complete complex was uncovered; however portions of two large rooms were excavated. Dr. B. Rothenburg of the Tel Aviv Museum, Israel, uncovered a large pyrotechnical installation, identified as a bakery. Many shapes and ceramic wares compare well with those from Pasargadae. (My thanks to Mr. David Stronach for making available his drawings and materials from Pasargadae, thus allowing for this comparison.) Glazed ware and rare painted wares: red-brown, or dull plum bands on jars or along rims, and an equally rare incised ware were also recovered from this period. Small finds included iron and bronze implements, i.e. iron trilobate arrow heads and a bronze macehead. The extensive occupation during this period for hundreds of metres in each direction from the mound, adds further support to the identification of Tepe Yahyf as a regional Achaemenian centre (Carmania ?).

Period IV. Directly underlying the Achaemenian structures, without evidence of a stratigraphic hiatus, architecture was exposed, utilizing bricks of the same size as in Period III. Plain coarse ware, buff to orange-red were most abundant. The presence of a buff, grey and red burnished ware, including a burnished grey ware bridge-spouted vessel, was noted. Subdivisions for this period await stratigraphic unravelling; however, there are clear indications that these wares are paralleled in shape and fabric in the Iron Age II and III periods ofnorth-western Iran. Further work on this first millennium sequence, which begins with this period, promises an exciting period of discussion.

Period V. Only the slightest evidence attests to the presence of a late second millennium occupation. The pottery dominantly plain buff and light orange, with rarely a black paint on buff cannot be readily paralleled on the Iranian Plateau but indicates some distant relations with Giyan II. Considerably more excavation is necessary before an understanding of this period becomes evident.

Period VI. Painted pottery, predominantly black on red, find designs and shapes readily paralleled on the one hand at Bampiir V and VI, and on the other hand at Susa D. Coarse plain wares, dominantly large bowls, some with spouts, as well as a reduced grey ware painted with a lustrous black paint below the rim are also found. Elaborately incised steatite bowls were also found together with a steatite

16


stamp seal. Only partial room plans were recovered. For the first time there is no evidence of a major fortification wall about the settlement.

Period VII. This period represented by several building levels, which sub-divide the period into at least three phases, is referred to as the " Yahyd Period ". Its characteristic pottery, red-slipped with black painted geometric designs finds no ready parallel, with the single exception of an illustrated sherd from Chah Husseini excavated by A. Stein. Other wares can be paralleled at Tal-i Iblis, Iblis Period and Bakun A. The most characteristic shape is a beaker, often with potter's marks on the base, which in shape are identical to those of Susa A. The quality of pottery in shape, method of decoration, and manufacture is outstanding. Several copper-base implements, chisels, awls, and pins, stamp seals, polished stone celts, beads, mortars and pestles, door-posts of stone, marble bowls and well preserved architecture, which in one case was associated with an interesting drainage system point to a substantial settlement of long duration.

Period VIII. The neolithic levels are of outstanding importance. Four different levels of architectural construction have been exposed for this period. Pottery is predominantly a plain chaff- tempered coarse ware, sometimes burnished over a red slip or wash. From the uppermost level of construction a few sherds of painted fine ware, black designs on a buff pottery can be identically paralleled at Tall-i Bakun BII. Of great surprise was the exposure of a massive wall 6-8 m. wide, constructed in 6o-8o cm. wide sections paralleling the direction of the wall. Rooms of the lowest exposed level of architecture were built up against this wall. The wall, as well as the rooms, were faced with a fine coat of plaster and washed over with a red paint. Virgin soil was reached outside the wall; however, within the area of the settlement we have not reached virgin soil. In fact, the floor level from which this wall was constructed has not been reached within the settlement. The pottery in this lowest level contains shapes which are sharply carinated, identical to the " milk-jars " of Hassuna- Matarrah. The flint industry is entirely microlithic. Stone bowls as well as mortars and pestles were also in abundance. Up-turned bases of coarse ware bowls, more rarely with haphazardly applied red wash, suggest parallels with Guran and more distant Jarmo. Clearly, this impressive settlement reaches back into the sixth millennium and suggests that this area shared in the early developments not previously suspected for this area of Iran. From the lowest excavated level a female idol was recovered resting on a bed of approximately thirty flints and associated with a bone spatula and three stone " arrow straighteners ". This figurine measures 28 cm. in height and is of a green speckled soapstone. It is unique for both its style and excellence of craftsmanship for this period in the Near East (see P1. IVB).

We now await a series of radiocarbon dates from each of the above periods before we return for our second season in the summer of 1969.

C. C. LAMBERG-KARLOVSKY

II SURVEY REPORTS

Archaeological Service of Iran: the discovery of an Assyrian relief During the summer of 1968 the Archaeological Service of Iran carried out extensive surveys in

eastern and western Azerbaijan as also in the regions of Qazvin, Kermanshah and Isfahin. Unquestionably the outstanding discovery has been that of a new Assyrian relief reported by Mr. Ali

Sarfaraz from the district of Kermanshih. Known as the " Jrdmandt relief ", the carving in question was found near the village of Tang-i Var, on the flanks of Ktih-i Zinaneh, over I00oo km. north-west of Kermdnshdh. Cut into the face of a vertical cliff at a point 40 m. above ground level, the relief depicts an Assyrian monarch standing in a standard pose with one hand raised and the other grasping a staff (cf. Iran V, pl. I, opposite p. 148). A worn and still unstudied inscription runs across the king's body.

Survey in Azerbaijan As an extension of the interest of the University Museum of Philadelphia in north-western Iran, a

surface survey is being undertaken in the province of Western Azerbaijan. Field work began under the


Hasanli! Project on a modest scale in August I966, continued for three months in 1967, and will continue through most of 1968. The survey area includes the Ushnai and Sold-iz complex of valleys south-west of Lake Reza'iyyeh (traditionally Lake Urmia), the plains up the west side of the lake depression (Urmia, Salmis and Khoy), the Marand Plain across the north end of the lake, extreme north- western Iran between the Turkish frontier and the Aras River, and the eastern slopes of the Zagros mountains to their frontiers with Iraq and Turkey.

The purpose of this project is both to extend our knowledge of archaeological sites and ceramic traditions to Iran's modern frontiers in the north-west and to treat a large geographical area as the problem of man adapting to and changing an otherwise " natural " environment. Over 250 archaeo-

logical sites of all periods from the late Palaeolithic to the recent past have been located. On the ground site discovery has been supplemented by aerial photographs and previous survey data. Data on the size and shape of each site and its location relative to geographical features is being collected for quantitative analyses. The location of each site for each major occupation period will later be compared with the distribution of modern settlement.

A basic chronological scheme for Western Azerbaijan has been developed by comparing surface sherds with ceramic materials from the excavated sites in and around the survey area. Most of the

samples collected fit into the occupation phases of Geoy Tepe or Hasanlfi Tepe (both in Western

Azerbaijan); Yanik Tepe (justeastof Lake Rezd'iyyeh in Eastern Azerbaijan) or Ki'ul Tepe, Nakhiche- van (an Autonomous Republic of the U.S.S.R. administered by the Azerbaidzhan S.S.R.).

Palaeolithic occupation seems to have been very sparse in spite of investigation of many rock shelters in the research area. No non-ceramic open sites have been discovered. A generalized soft-ware " Neolithic " with ceramic forms similar to those already discovered at Hiajji Firiiz but with significant local variation of surface treatment and design are found throughout Western Azerbaijan. The

succeeding Early Chalcolithic period, of which a finger impressed ware is especially characteristic, is found west of Lake Rezd'iyyeh and north as far as the Soviet Union. Similar wares have recently been discovered in south-eastern Europe (northern Georgia) and these ceramics may belong to a common tradition going as far west as Eastern Europe. The Late Chalcolithic period, as yet imperfectly understood, appears to be dominated by local ceramic traditions. Painted wares similar to those from Pisdeli Tepe just south of Lake Rezd'iyyeh have not been found north of the lake. From the Urmia Plain north to the Soviet Union and west to Turkey several phases of a widely spread Bronze Age " culture " have been defined. Incised decoration is characteristic of the earlier phase and modelled lugs either

unpierced or pierced horizontally or vertically are common in the middle phase. Modelled lugs of the

type described above are common in eastern Turkey and the Transcaucasus, although they are known

by several different terms, no one of which is entirely satisfactory. Neither the monochrome " Khabur " wares south-west of the lake nor the polychrome " Cappadocian " wares west of the lake have been found in the extreme north-west. Whatever occupation occurred between the end of the Middle Bronze Age and the beginning of the Iron Age must await excavation or adequate sounding at appropriate sites. Dark grey wares analogous to those of Hasanlfi Period V (now called Iron Age I), while

rare, are found north-west of Lake Rezd'iyyeh in the Early Iron Age. Pottery of the succeeding " Late Western Grey Ware Period " (Hasanlfi Period IV, now called Iron Age II) was not found north of the Salmis Plain. It must be admitted however that it is difficult to distinguish Iron I from Iron II wares

by surface survey alone. From the middle of the eighth century B.c. down to the beginning of the " Achaemenid " period a plain brown or buff ware sometimes with simple painted designs or a burnished (and infrequently polished) red ware is especially characteristic. The red wares possibly coincide with " Urartian " political control of at least the mountains and smaller valleys of Western Azerbaijan.

An important supplement to the archaeological survey is the study of geography in the project area as it relates to human use of the landscape in the ethnographic present and the archaeological past. Data are being collected on both natural and " man-made " resources. The geomorphology of the agricultural plains, fresh water resources, and changes in the level of Lake Reza'iyyeh are also being studied.

R. R. B. KEARTON


Survey in Azerbaidjan Das Deutsche Archiologische Institut in Teheran setzte in der Zeit vom 25. April bis zum 26. Mai

1968 den im Vorjahr begonnenen Survey in den Provinzen West- und Ostazerbaidjan des Iran fort. Es wurden wiederum bisher unbekannte urartiische Anlagen festgestellt und zwar:

I. eine urartaische Befestigung auf dem Ausfluss eines Lavastromes des Araratgebietes, unmittelbar

6stlich des Fahrweges von der asphaltierten Hauptstrasse Bazerghan - Tabriz nach Pol-i- Dasht an der sowjetisch-iranischen Grenze. Die Festung liegt etwa drei Kilometer von der Asphaltstrasse entfernt und ist aus schwarzem Lavastein errichtet. In Ermangelung eines anderen Namens ist die Anlage von uns " Siah-Tepe " genannt worden. Die rechteckige Anlage zeigt die typische Bastionen- und Courtinenarchitektur der urartaischen Zeit. Auch einige Felsabarbeitungen und einwandfrei urartaische Keramik wurden gefunden.

2. eine ausgedehnte und in ihrem Mauerbestand noch gut erhaltene urartaiische Festung, Kale Oglu genannt, westlich einer 1968 noch im Bau befindlichen Asphaltstrasse, die von der Einmiindung der alten Khoy-Strasse in die Hauptstrasse Bazerghan-Tabriz zu einem Araxess tbergang zu dem sowjetischen Nakhitchevan fiihrt. Die hier gefundene Keramik ist ebenfall- eindeutig urartiisch. Die Mauern sind bis zu vier Meter hoch erhalten, 4,5-5 m stark und zeigen wiederum typische Bastionen- Courtineneinteilung.

3. am Nordwestrande der Ebene von Shahpur wurde eine zum Teil noch in ihren Mauern erhaltene Befestigungsanlage festgestellt, die sowohl auf Grund der Architektur, als auch der Felsabtrep- pungen und einzelner Keramikreste als urartiische Festung anzusprechen ist. Sie liegt etwa 20 km nordwestlich von Tazeshahr.

4. wenige hundert Meter n6rdlich der von uns 1967 gefundenen Terrassenanlage oberhalb des sasanidischen Felsreliefs siid6stlich Shahpur ist eine weitere, durch die Keramik ebenfalls als

urartiisch zu bezeichnende Anlage festgestellt worden. Sie besteht aus raumartigen Felsabar- beitungen, sowie Felsabtreppungen und Befestigungs-beziehungsweise Terrassenmauern.

5. in der Nihe des Ortes Gauharchin-Kale liegen zwei Felskl6tze in Verbindung zum Ufer im Urmiasee. Auf dem einen, Kazimbashi genannt, befinden sich wahrscheinlich sasanidische und mittelalterliche Baureste, aber auch einige Mauerteile und Felsabtreppungen, die die Annahme einer urartiischen Anlage auch an diesem Platz rechtfertigen. (vergl. Lehmann-Haupt, Armenien Einst und Jetzt I (1910o), pp. 370-3.)

Von den 1967 festgestellten urartiischen Plitzen sind zwei, naimlich Zengar, acht Kilometer westlich von Maku (in Iran Bd. VI, 1968, S.166 muss es heissen: " acht Kilometer westlich von Maku bei Zengar ") und Bastam im Mai 1968 in jeweils sechs Tagen ndiher untersucht worden. Diese Kurzgrabungen wurden im Hinblick auf gr6ssere Grabungen durchgefiihrt, die ffir 1969 geplant sind. In Zengar ist die Grabkammer und ihr Felstreppen-Zugang freigelegt worden. Ausserdem wurden einige Schnitte auf dem Plateau tiber der Grabkammer angelegt.

In Bastam ist eine Toranlage freigelegt worden und die Hinterfiillung der starken Hangmauer mit dem antiken Aufweg an zwei Stellen geschnitten worden.

Uber die Grabungsergebnisse in Zengar und Bastam, sowie fiber saimtliche von uns in den Jahren 1967 und 1968 festgestellten urartaiischen Plaitze und einige weitere, die wahrscheinlich urartaiisch sind, ist ein Bericht in den Istanbuler Mitteilungen 18, 1968, im Druck.

Ausser den urartiischen Objekten sind 1968 wiederum eine Reihe von vorgeschichtlichen, mittel- alterlichen, armenischen und islamischen Objekten festgestellt worden.

Unabhaingig vom Survey in Azerbaidjan sind durch den Assistenten des Deutschen Archiiologischen Instituts, Herrn Dr. Gerd Gropp, eine Anzahl parthischer und sasanidischer Inschriften, hauptsichlich in der Provinz Fars, festgestellt und bearbeitet worden.

WOLFRAM KLEISS


Survey in Western Azerbaijan

A survey for prehistoric sites was made by Columbia University's Department of Anthropology team during July and August 1968 in Western Azerbaijan. This mission was financed by the National Science Foundation with additional support for student participation by the Ford Foundation.

The specific nature of the investigations was an exploratory one, namely a search for suitable Neolithic (aceramic) and palaeolithic sites. It was hoped to find a site, or sites, which offered promise for a future full scale season of excavation. Although many archaeological sites were visited, none fitted the requirements of the mission. However, sites were found which will be of interest to archaeologists specializing in later Neolithic periods.

The area of the search as outlined in the permit granted to us by the Archaeological Service of Iran

ranged from Khoy in the north to Mihabid and Sardasht in the south, on the west side of the Lake Urmia basin. Brief reconnaissances were also made in the areas of Maku and Mardgheh districts in order to investigate caves reported there. Soundings were made in the Chuldbhd caves between Bukan and Meyandowab, in the Kerboota Cave near Taze Kande village in the Maragheh district and in the rock shelters near Malhan south of Shahptir. Tests were also made in the tepe site near Kopi Bib Ali on the southern edge of the Daryicheh-ye-Kopt Lake north-east of Mihabad, and in the Seavan Tepe, near Gerevan in the Ziveh valley. The cave and rock shelter tests were disappointing, in that they yielded Iron Age and later ceramic materials only. Kerboota Cave, which looked the most promising, contained Iron Age ceramics down to the bottom of the sondage, which reached

3"5 m. At that depth,

digging was halted because it was observed that the soil was filtering down among a number of large boulders to an unknown depth below.

In all, some 38 tepes were recorded by the survey, as well as some 18 caves, 7 shelters, 4 tumuli and

3 surface sites of random character. It is not understood why the caves and rock shelters, which looked so promising in aspects, did not yield the expected stone age Palaeolithic remains. No surface indications were found in the stream beds, valley terrace cuts, and other places investigated by the team. Surface collections and the two sondages made of the tepes indicated that none of them contained

pre-ceramic archaeological remains, so far as we could determine. It is possible that we had missed the

palaeolithic sites, but it seemed unlikely that not even a trace could be discovered. Of interest, in our estimation, to students of the later Neolithic, is Tepe Seavan in the Ziveh valley

area. This area lies west of Reza'iyyeh at the head of the Baranduz River. In the two small sondages (one measuring I - 5 x I25 m., the other c. I x 2-25 m.) were found ceramics which were tentatively identified as belonging to the H-jji Firfz and Dalma periods. The deeper of the two sondages, which was made in the flat apron next to the tepe, struck water at c. I 5 m., halting the work. Ceramics, obsidian and chert fragments were found in this sondage as far as we had reached. This mound, measuring about 6 m. high by about 0oo m. in diameter (longest dimension), appeared to be surrounded on about three-quarters of its perimeter by a marsh and ditch. Sondages on the top of the tepe revealed Muslim and Iron Age ceramics. Tepe Seavan is one of several prominent habitation mounds in the southern end of this lush mountain-ringed valley.

It is not clear why this area of Azerbaijan, which is the best watered and greenest and most fertile area in Iran, did not yield any palaeolithic sites. This is in contrast with the Kerminshah area to the

south, which has proven to be extremely rich in palaeolithic materials and occupation sites. One of the team members, the zoologist Dr. Dexter Perkins, Jr., observed that this Azerbaijan region is

presently a very good habitat for all kinds of game animals. From a zoologist's point of view, the lack of palaeolithic man occupational evidence seemed incomprehensible. He suggested that there may be a scarcity of raw chipping material, such as flint, for the flaking of stone implements which may have been an inhibiting factor towards occupation by palaeolithic man. This theory, however, does not seem too likely, since chert materials were observed in outcrops in the mountains, and some flint and chert materials were picked up on the later period Neolithic tepes. It is not proven, but it is possible that an ecological factor may have inhibited occupation by palaeolithic man in west Azerbaijan. The idea is tentatively advanced that this part of Iran may have been inhospitable climatically during the Last Glaciation, or before some 7,00ooo-8,00ooo years B.c. It may be well worth the effort to make pollen core


samplings from the marsh areas in the Ziveh valley to learn something about the climatology of this part of Azerbaijan.

The data are being assembled at Columbia University, and a full report of our season's investigations will be submitted for publication.

RALPH S. SOLECKI

Survey oJf man's fire-using industries in Afghanistan, Iran and Turkey

During the months of August and September Theodore Wertime, research associate of the Smithsonian Institution, led a group variously numbering as many as ten persons through the great mountain backbone of south-western Asia to reconnoitre mines, slag heaps, settlements and contemporary technological sites. The expedition was intended to follow up on the Wertime expedition of 1967 in search of tin and the Smith-Pleiner-Wertime explorations of the great Persian desert during 1966 in search of ancient metallurgy.

The reconnaissance may be summarized as follows. We:

i. Conducted inspections of five zones of ancient mining-Mukfir, Ozbekuh, Anarak, Ahar and Ergani. Mukfir was a zone of gold and iron mining; Ozbekuh of lead and zinc; Anarak of native copper, chalcocite, cobalt, nickel and zinc; Ahar of copper; Ergani of copper.

2. Collected slags at Farinjal, Askar Kot, Karistu, Ozbekuh, Dehuk, Naiband, Sechah (Baft), Qatru, Anarak, Ahar, Tire Bolu and Gedak. We studied smelting and roasting furnaces in some detail at Anarak and Gedak.

3. Examined a range of ancient settlements-Mundigak, Tall-i Iblis, Sialk, Yengi-imam, Altintepe, Kiultepe and ?atal Hiiytik.

4. Observed pottery, brick or glass making at Istdlef, Qara Murdd Bdgh, Ghazni, Herat, Anarak and Kdshin.

5. Carried out a planned conjuncture with the Harvard expedition excavating at Tepe Yahyt, which enabled the team to observe the resource setting of a site of ancient pottery and metallurgy. It had been hoped to do likewise with the site of Cay6nii Tepesi in Turkey; but excavations had not begun there when we arrived.

THEODORE A. WERTIME

The Persepolis Plain and Shirdz: Field Survey 2

The map illustrating my 1966 survey (Iran VI, pp. 169-7o) revealed a lack of sites in the northern and southern extremes, regions which it had not been possible to examine thoroughly. In September 1968, with the aid of a much-appreciated grant from the I.O.O.C., this was achieved and, as expected, the northern sector demonstrated that the high level of occupation extended up from the dense central zone, while the southern sector, also confirming expectations, showed that its salinity and vulnerability to flooding made it apparently as inhospitable to early settlers as it is today. However, the definitive map showing the full range of the various occupational periods of the entire plain, which is to be

published soon by Mr. William Sumner, may show a somewhat different picture between the edge of this map and the shores of Lake Neyriz, which I was not able to explore completely. This, therefore, is an interim report and map of a more restricted area, in which, nevertheless, it was possible to locate an additional forty-one prehistoric and over thirty-one additional Islamic sites.

What may emerge from the survey as significant is the establishment of a distinct plain red-ware culture, using exclusively Bakun A5 pottery (with very occasionally Btkun painted buff continuing) at a hitherto unsuspected frequency of sites in the central zone.

Upon re-examination of the six Shirdz sites, sufficient painted software of the Jari type was found at one of the Bdkun mounds, and at another mound hitherto believed to be Islamic only, to establish a far earlier occupation on a total of three settlements.


In three caves along the northern slopes of the plain, unusual designs painted in bright crimson were discovered. Streaks of lime deposit over some indicated a certain antiquity, but hindered decipher- ment; yet their similarity and proximity to animals, trees and human figures " pecked " on the nearby rocks is perhaps significant.

THE F ERSEPOLIS PLAIN AND SHIRAZ

46 9 00 U~~Va 033 10 2 6 13

1 31 98 7 8 ~13 2

a 14 . 31o % % (4k 19 0 *% % 15-

o\ sqanz t hrc "

27o

Isai Ac Ceeon

16:0 % 11 12 391 % 10 % 0

% %a 24J& 1111%___ ,ara

Bandemir 366 A 2 % 17, % 018 0~

% /707 dNow i;.4 % % % I f 'A OJari *Bakn % AA 2 W Bakun A5 @Shags %% 21r-~~Ly/Y~ar ~LLT~~~V ~ ~-o1 (I)Kftari (]Flint SHIRA

OTaimurano Sterile 41' Islamic Celado

DCave Painted 43

pAltar onknown Y

*Ffort 5 *"'%%10r Mod@ r

M r

' k I m s 7"0.~TO

Prehistoric

i. Ja'fardbad I 2. Ja'fardbad II 3. Ja'fardbd III 4. Mashtekdn* 5. Sa'datibid 6. Haftkian A, B

8. B.bd

SAdegar 9. Tall-Akhodi Bdnesh

xo. Sahlibdd II x i. Puzeh Malami 12. Sahlabad I 13.

Bdzdr.ibdd Cave

14. MAdeh Bdnou* 15. Negaristin I Cemetery I6. Negaristdn II 17. Negaristin III

18. Sahlibhd A, B* 19.1 20. Kfishk 21. 'Alibitd Chdman I, A, B*

22. 'Alidb~d Chdman II 23. 'Aliabdd Chaman III 24. Asdf 25. Gondashlfi I 26. Gondashlf II 27. " Tolombe" 28. Esfirdfin 29. Tall-i-Chirdeh 30. Kushkak 31. Zsrdgariin 32. Aktape I* 33. Aktape II* 34. Ayfib 35. Tall-i-Jalalabad 36. Husaindbad 37. Kafr 38. Ja'fardbdd B*

39. Abgarm Piru 40. Kalfeh 41. Shahgal-Baghi B 42. Askari 43- Malamjar (Askari B)

Islamic

I. B~nesh 2. SahlibTd (Celadon) 3. Koroni I Village 4. Koroni II 5. Bazardbid 6. Madeh Binu* 7. Negaristan 8. Sa'datibhd 9. Karag~li

oo. Hazdr B, A, B I I. " Tolombe " 12. Amrabad

I3- Jaldlabb d 14. Esfidfin A, B I5. Pol-i-No 16. Gashak* I7. Rajdbdd Dehchdst i8. 19. Kamjfin 20. 21. Gavkfin

22. SultanTbad 23. Alivali6la 24. JamalDbdd A, B, C 25. Tang Dast-i-chap 26. Pdru 27. Tall-i-QdsimAbid ('Alidbdd) 28.

I.Hji*bid A, B

29. Tall-i-Asyfin 30. Pol-i-Khdn 3'1. Chastekor

Prehistoric: known

(a) T/Siydh (b) T/Sabz (c) T/Darwdzeh II (d) T/Darwdzeh I (e) T/Taimurdn (f) T/Ja'faribdd (g) T/Qal'eh (h) T/Shurei A, B (i) Naqsh-i Rustam (j) T/Bakun A, B

* Indicated by Mr. William Sumner.


At the extreme eastern corners of both Koih-i Sdqiz and Kfih-i Husain, two platforms of the exposure- altar type were found, fashioned from the huge fallen rocks. The former, slightly to the west of the two well-known ossuaries, has a rectangular-lipped trough, while that at the directly opposite side from the

Naqsh-i Rustam altars has a circular-lipped trough; both have runaway channels.

PAUL GOTCH

Takht-i Suleiman*

Nach einer Unterbrechung im Jahre 1967 wurden die Grabungsarbeiten des Deutschen Archio- logischen Instituts auf dem Takht-i Suleiman (Azerbeidjan) wiederum mit Unterstiitzung der Deutschen Forschungsgemeinschaft fortgesetzt. Innerhalb der Ringmauer wurde neben dem Nordtor, im Hof E, im Areal westlich vom Feuertempel und am grossen Westiwan gegraben.

Die bereits 1966 auf der Innenseite des Nordtores angeschnittene Lehmziegelanlage erwies sich als identisch mit dem in frtiheren Grabungen gefundenen unmittelbar hinter der Steinmauer liegenden Lehmziegelmassiv. Es liess sich nachweisen, dass dieses Lehmziegelmassiv eine der Steinmauer zeitlich vorausgehende Befestigungsmauer darstellt. Hinter dem westlichen Turm des Nordtores wurde der Rest einer zu dieser Lehmziegelmauer geh6renden Rundbastion freigelegt, die gr6sstenteils abgetragen worden ist, als man die Steinmauer vor die Lehmziegelmauer setzte. Die Rundbastion besteht in ihrem unteren Teil aus massivem Bruchstein-M6rtelmauerwerk, im oberen Teil, der eine

iiberwblbte Turmkammer hatte, aus Ziegelmauerwerk. Es ist zu vermuten, dass die Rundbastion zur

Toranlage der Lehmziegelmauer gehorte. Der Hof E nord6stlich vom Feuertempel und die ihn umgehenden Korridore wurden freigelegt.

Uber seiner 6stlichen Aussenwand fanden sich zwei Keramikbrenn6fen und weitere Einrichtungen einer Werkstatt aus il-khanischer Zeit, in welcher Wand- und Fussbodenkacheln fur den Palast Abaqa Khans hergestellt wurden.

Westlich vom Feuertempel wurde die Untersuchung der sasanidischen Pfeilerhallenanlage fortgesetzt. Vor ihrer Sidfassade lag eine zum See hin offene Arkadenhalle mit Rechteckpfeilern. Der Rundpfeilersaal PB wurde freigelegt, in seinen mehrfach aufgehohten Fussb6den wurden kultischen Zwecken dienende Einrichtungen gefunden, Feueraltare, spindelf6rmige Stinder, Wasser- becken und ein von Wasserrinnen umgebenes Podium. M6glicherweise diente der Saal wihrend seiner letzten zoroastrischen Benutzungsperiode als Yesischngah, als Raum far verschiedene

Opferzeremonien. Unter den sasanidischen Kleinfunden sind mehrere Kupfermtinzen des 5. und 6.

Jhdts.n.Chr., zwei Stempelsiegel, ein mit einer stehenden Figur dekoriertes Goldplittchen und einige Fragmente von Bauornamentik aus Gipsstuck zu erwihnen.

Unter den Pfeilerhallen und den sie umgebenden Raumen und Hofen wurden weitere Fundamente der den Steingebauden vorausgehenden Lehmziegelanlage aufgedeckt. Siidlich vor dem schon bekannten Viersaulenraum und seinen Seitenkammern lag ein groBer Hof, an dessen Westseite sich drei hintereinanderliegende Fluchten von rechteckigen Riumen entlangzogen. Die siidliche Begren- zung des Lehmziegelkomplexes stimmt mit der Pfeilerhalle uberein, im Westen reicht der altere Bau iiber den spiteren hinaus, ebenso im Norden, wo sein AbschluB noch nicht festgestellt wurde. Die Ausdehnung nach Osten lad3t sich wegen starker Versinterung des Gelandes nicht klaren. Die

Lehmziegelanlage ist vor allem in ihrem n6rdlichen Teil hiufig umgebaut worden, ehe sie durch die

Steingebaude ersetzt wurde. Ihre Mauern und FuBb6den wurden dabei bis auf die Feldstein- fundamente abgetragen.

Nach Ausweis einiger Keramiktypen ist es wahrscheinlich, daB die Lehmziegelanlage in der Zeit des ersten vorchristlichen bis ersten nachchristlichen Jhdts. errichtet worden ist.

Die Fundamente des Lehmziegelgebiudes stehen unmittelbar auf einer durch ubergetretenes Seewasser gebildete Kalkablagerungsschicht, die auf eine lingere siedlungslose Periode des Takht- Plateaus hinweist. Unter dieser Kalksinterkruste liegt eine Kulturschicht aus achdimenidischer Zeit. In einem Areal von geringer Ausdehnung wurden Griber und sehr zerst6rte und eingeschwemmte

*Editor's note: Report received after the journal had gone to press.


Mauerreste gefunden. Die Mauern nehmen offenbar Bezug auf die Graber. Diese bestehen aus ovalen, in den Kalksteinuntergrund gehackten Gruben, in denen die Skelette ohne erkennbare Regel oder Ausrichtung liegen. Als Grabbeigaben fanden sich Pfeilspitzen, Glasperlen, Bronze- und

Eisenringe. Bei den Arbeiten am Westiwankomplex wurden zunachst die den sasanidischen Bestand iiber-

lagernden Teile des il-khanidischen Palastes suidlich der groBen Halle untersucht. Der dem nord- westlichen Oktogon entsprechende Achteck-Kuppelbau an der Stidwestecke der Anlage und die ihn

umgebenden Nischenkammern wurden freigelegt, ebenso die hochgelegenen durch ein Treppenhaus von den Seearkaden her zuginglichen Raume zwischen Achteckbau und See. Es fanden sich Reste von Baukeramik und Wandstuck in situ. Die aus dem Palast des ausgehenden 13. Jhdts. hervorge- gangene Baukeramik zeigt eine auBergew6hnliche Vielfalt von Formen und Techniken und geh6rt zu den wichtigsten Funden der Grabung.

Die Eingangszone der sasanidischen Iwanhalle wurde untersucht. Im Gegensatz zum Wieder- aufbau des 13. Jhdt. waren ihre Wandendungen unter dem Eingangsbogen wie im Iwan des

Feuertempels durch Risalite verstirkt. Siidlich neben der Halle wurde eine Folge von Raumen und Korridoren der sasanidischen Periode gefunden, von denen die meisten bis zum Gewolbescheitel erhalten sind. Die Raume wurden teilweise als Substruktions-und Kellerraume in den mongolischen Palast einbezogen, in einem fand sich eine Wasserbeckenanlage mit Stufen. Der AbschluB der sich nach Suiden erstreckenden sasanidischen Anlage wurde noch nicht erreicht. Eine aus der Westseite des Traktes vorspringende mit einer Rundbastion bewehrte Mauerecke weist darauf hin, daB der See ebenso durch eine mit Bastionen besetzte Mauer umschlossen war, wie das Tempelgebiude im Norden des Sees.

DIETRICH HUFF

ABBREVIATIONS

AASOR Annual of American Schools of Oriental Research AfO Archiv ffir Orientforschung AJA American Journal of Archaeology AJSL American Journal of Semitic Languages and Literatures AJ Antiquaries' Journal AK Antike Kunst AMI E. E. Herzfeld, Archaeologische Mitteilungen aus Iran ANET Pritchard, Ancient Near Eastern Texts AOr Archiv Orientalny Arch Anz Archaologischer Anzeiger AS Anatolian Studies BA Besch Bulletin van de Vereeniging ... de Antieke Beschaving, Hague BASOR Bulletin of American Schools of Oriental Research Belleten Tiurk Tarih Kurumu: Belleten BGA Bibliotheca Geographorum Arabicorum Bib Or Bibliotheca Orientalis BSA Annual of the British School at Athens BSOAS Bulletin of the School of Oriental and African Studies CAH Cambridge Ancient History CIA Corpus Inscriptionum Arabicorum DAFA D6l6gation Arch6ologique frangaise en Afghanistan, memoires EI Encyclopaedia of Islam ESA Eurasia Septentrionalis Antiqua IAE E. E. Herzfeld, Iran in the Ancient East

(i94i) ILN Illustrated London News Iranica Iranica Antiqua JA Journal Asiatique JAOS Journal of American Oriental Society JEA Journal of Egyptian Archaeology JHS Journal of Hellenic Studies JNES Journal of Near Eastern Studies JRAI Journal of the Royal Anthropological Institute JRAS Journal of the Royal Asiatic Society KF Kleinasiatische Forschungen LAAA Annals of Archaeology and Anthropology, Liverpool MAOG Mitteilungen der altorientalischen Gesellschaft MDOG Mitteilungen der deutschen Orientgesellschaft MDP M6moires de la D6l6gation en Perse MJ Museum Journal, Philadelphia OIC Oriental Institute, Chicago, Communications OIP Oriental Institute, Publications OS Orientalia Suecana PZ Praehistorische Zeitschrift RA Revue d'Assyriologie RCAS Royal Central Asian Journal REI Revue des lRtudes Islamiques SAA Soviet Anthropology and Archaeology SAOC Oriental Institute, Studies in Ancient Oriental Civilisation Sov Arkh Sovetskaya Arkheologiya SS Schmidt, H., Heinrich Schliemanns Sammlung trojanischer Altertumer Survey A Survey of Persian Art from Prehistoric Times to the Present, ed. A. U. Pope,

Oxford, I938 TT Tuirk Tarih, Arkeologya ve Etnografya Dergisi WO Die Welt des Orients WVDOG Wissenschaftliche Ver6ffentlichungen der Deutschen Orientgesellschaft ZA Zeitschrift fir Assyriologie ZDMG Zeitschrift der Deutschen Morgenlaindischen Gesellschaft

NOTES ON TRANSLITERATION FOR CONTRIBUTORS TO IRAN I. OLD AND MIDDLE PERSIAN

It is recognized that no rigid lines can be laid down here, but it is suggested that the Old Persian syllabary should be transliterated according to the table in Kent, Old Persian. Grammar, Texts, Lexicon, p. 12; that for Manichaean Middle Persian and Parthian, the transliteration system given in Andreas- Henning, Mitteliranische Manichaica, vol. III, p. 66, should be used; whilst for Pahlavi, the table of alphabets given in Nyberg, A Manual of Pahlavi, new edition, p. 129, may be used as a reference for transcription. II. ISLAMIC AND MODERN PERSIAN

The system used for the Cambridge History of Islam should be used here as far as possible. Consonants

(a) Arabic z q

Sb , s k t ) sh J 1

s th ,

s ( m t j

, d j n

h .it h - kh w

d ' y 3 dh gh -a (in construct state:

r , f -at)

(b) Persian additional and variant forms. The variant forms should generally be used for Iranian names and for Arabic words used in Persian.

v P 5 z J g

-, s 3 zh j v ch z

(c) The Persian " silent h " should be transliterated a, e.g. ndma. Vowels

Arabic and Persian. Short: a Long: I orDoubled - iyy (final form: i)

u c: i Diphthongs " au

i ai

Notes I. The izdfa should be represented by -i, or after long vowels, by -yi, e.g. umard-yijdnki. 2. The Arabic definite article should be written as al- or 1-, even before the so-called " sun letters ",

e.g. 'Abd al-Malik, Abu 'l-Nasr. 3. The macrons of Abi and Dhi (Zi) should be omitted before the definite article, e.g. Abu

'1- Abbds (but Abil 'Ubaida). It is obvious that for the rendering of linguistic and dialectical material, and possibly also for

contemporary literary and spoken Persian, this rigorous system of transliteration is inappropriate; contributors should use their discretion here. III. GENERAL POINTS

I. Names of persons should be rigorously transliterated. 2. Conventional English equivalents (without macrons or diacritics) should be used for the names

of countries, provinces or large towns, e.g. Khurasan, Shiraz. Otherwise, all place-names should be rigorously transliterated. Archaeologists are asked to be especially careful in representing the names of little-known places at or near sites.

3. Modern Turkish names and words should be written in the current romanized Turkish ortho- graphy.

4. Where classical Greek and Latin renderings of Old and Middle Persian names exist, these familiar forms should be used for preference.

BRITISH INSTITUTE OF PERSIAN STUDIES

APPLICATION FOR INDIVIDUAL MEMBERSHIP

N A M E ............................................................................................................................................................................... DESCRIPTION ........... ...................... . ................. .................. (Title, Decorations, Degrees)

A D D R ESS ........................................................................................................................................

O C C U P A T IO N ............................................................................................................... Interest in o r co n nectio n w ith Iran

.........................................................................................

I wish to apply for Membership of the British Institute of Persian Studies as an individual subscribing member paying *?1/?2 10s. per annum and *enclose cheque/ Banker's order for this sum.

*I wish to make a covenant for a period of seven years. S ig n atu re

................ ............................... D ate........................................................ . . .

.... ....

..... . . * Delete whichever is inapplicable.

Payment may be made in Dollars or Rials at current rates of exchange, e.g. ?2 10s.= $6.25.

APPLICATION FOR MEMBERSHIP BY A CORPORATE BODY

NAM E. ..... ...................................... ...... ......................................

ADDRESS

Application is hereby made for Membership of the British Institute of Persian Studies as a corporate subscribing member paying the sum of ?50 per annum. A *cheque/banker's order for this sum is enclosed.

S ig n a t u re ..........................................................................................................................

D a te .................................... ...........

......... ....................... * Delete whichever is inapplicable.

MEMBERSHIP-SUBSCRIPTIONS

Under paragraph i i of the Articles of Association of the Institute, the Council has prescribed the following scale of subscription for members.

I. Individuals

(a) Those subscribing ?2 Ios. per annum or more. They will have the right to receive an invitation to the Annual General Meeting and other meetings of the Institute and will receive the Journal of the Institute, which it is hoped to publish annually.

(b) Those subscribing ?x per annum. They will have rights as in (a) above, except that they will not receive the Journal; they will, however, be able to purchase the Journal, if they wish, at a price of ?I 15s. for each issue.

2. Corporate bodies

The subscription will be a minimum of ?50 per annum. They will have the right to nominate one representative to attend meetings of the Institute and vote thereat (paragraph 9 of the Articles). They will receive one copy of the Journal.

3. Unincorporated bodies

The subscription will be a minimum of ?50 per annum and the privileges will be as for corporate members subject to the provision of paragraphs 7 and 8 of the Articles of Association.

4. Forms should be sent to the Hon. Secretary, J. E. F. Gueritz, Esq., M.A., 85 Queen's Road, Richmond, Surrey.

FORM OF COVENANT

o f .....................................................................

....... .......... .. . .............. ....................................................................

hereby covenant with the British Institute of Persian Studies that for a period of seven years from the date of this Deed or during my lifetime (whichever shall be the shorter period) I will pay to the said Society from my general fund of taxed income such an amount annually as after deduction of income tax at the Standard Rate yields the net s u m o f ..

.................................................................................................................. The first paym ent is to be m ade on the........................................................

Dated this ........................................ day of ......................................................... 19 ..........

Signed, sealed and delivered by the said:

In the presence of:

NOTICE

It is hoped that as many as possible of those applying for Membership of the Institute will undertake a seven-year covenant, thus making a further substantial contribution to the resources of the Institute.

BANKER'S ORDER

T o ............................................................................................................................B

A N K

ADDRESS

Please pay to the British Bank of the Middle East, 7 King William Street, London E.C.4, for the credit of the British Institute of Persian Studies, the sum of ? : now and a similar sum on Ist January of each year until further notice/for the next six years.*

Signed........................................................................... D ate............................. .................................. ..

* Delete whichever does not apply.

S... 250 Mites

C . 400Xilometres

A U. S. S. R. bLiM z .

? A~RBAIJAN Q c

.......... .i L.'Re zayek ~. ..., 44, G"".Gorgan

,., ' "

Qazvi. 4NDA"RAN .Damg

has

S"TEHRAN '. *

,- l-HHad RKH URAS A N

& ( .Kermanshak *Qu- m

'Kashan..

/s I R A N I

AA AhA

,,Thsargadae *KrmaiL P. ersepolis .Koo.l' RMAO

S-hiraZ • ...,

BandarA'AbbasH

ARABIA A

1,7. A. "

PRINTED IN GREAT BRITAIN BY HEADLEY BROTHERS LTD 109 KINGSWAY LONDON WC2 AND ASHFORD KENT

Iran 07 (1969)

Documents

Transcript of Iran 07 (1969)