I

Journal
VOLUME VIII

AN
Institute
1970

of

the

British

of

Persian

Studies

CONTENTS
. . . . . . Governing Council Professor A. J. Arberry; Dr. S. M. Stern . Obituaries, Director's Report . . . . . . .

Page
.
.

.
.

. .

vi vii xi
I

Excavations at Sir&f:Third Interim Report, by David Whitehouse . . Pyramidal Stamp Seals in the Persian Empire, by John Boardman Persian Lacquer in the Bern Historical Museum, by B. W. Robinson . Charles I and the Antiquities of Persia: The Mission of Nicholas Wilford,
by R. W. Ferrier
. . . . . . . . . . . . . .

19 47
51 57

. .

Iletmish or Iltutmish? A Reconsideration of the Name of the Dehli Saka Studies: The Ancient Kingdom of Khotan, by H. W. Bailey . Dailamis in Central Iran: The Kikfiyids of Jib~l and Yazd, by C. E.
Bosworth . . . . . . . . . . Sultan, by Simon Digby .

65
73. 97
10o5

The Archers of the Middle East: The Turco-Iranian Background, by . . . . . . . . . J. D. Latham

Mohammad Ali Jazayery . . . . . The Genesis of Safawid Religious Paintieg, by J. M. Rogers Charles Burney . . . . . The Khurab Pick-axe-Corrigenda, H. Lechtman by . Survey of Excavations in Iran, 1968-69 .

The Arabic Element in Persian Grammar: A Preliminary Report, by

.

Some Kurdish Proverbs, by D. N. MacKenzie

. .

15
125

Excavations at BdbaJgn, 1968: Third PreliminaryReport, by Clare Goff 141 Excavations at Haftavan Tepe 1968: First Preliminary Report, by
.
.

157
173 175

.
.

.
.

. .

Published annuallyby

THE

BRITISH

INSTITUTE

OF

PERSIAN

STUDIES

c/o The British Academy, Burlington House, Piccadilly, London, WIV ONS
Price: ?5 os. 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 subjectsof 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 refreshthemselves 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 edited by the late ProfessorA. J. Arberry, has already texts, the first of which, the Humay-Nama, 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 ?4 entitles the subscriber to receive the Journal.
Application Forms at back of Journal.

ART

1970 TREASURES TOURS

of

IRAN
visiting

TEHERAN - ABADAN - CHOGA ZANBIL HAFT TEPE - SUSA - SHIRAZ PASARGADAE - NAQSH-I RUSTAM PERSEPOLIS - NAQSH-I RAJAB - ISFAHAN HAMADAN - BISITUN - TAQ-I BOSTAN SAHNEH - QAZVIN - TEHERAN
Departures:

March 25 - April 8 - April 22 - May 6 - Sept 23

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. 14, 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 Churchesof the CrusadingKingdom. Miss Margaret Munn-Rankin, M.A.(Oxon), F.S.A., who will accompany Tour No. 18, 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 EncyclopaediaBritannica.

Miss Clare Goff, Ph.D., who will accompany Tour No. 15, 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. Sylvia A. Matheson, now living in Iran, will accompany Tour No. 16. She has worked on archaeological sites in Afghanistan and Pakistan and is a well-known photographer, travel writer, lecturer and broadcaster. Her books include Time Off to Dig (an account of archaeology in Afghanistan), and The Tigers of Baluchistan(a study of the Bugti Baluch tribe). Among forthcoming publications are Persia: an Archaeological Guide, The Pleasures of Persia (travel book), and Iran: a History of 2,500 Years. Mr David Whitehouse,M.A., Ph.D., who will accompany Tour No. 17, is Wainwright Fellow in Near Eastern Archaeology in the University of Oxford. He read Archaeology and Anthropology at Cambridge, was subsequently a Scholar at the British School at Rome, and conducted various excavations in Britain and Italy before taking up the Directorship of the large-scale excavations at Siraf on the Persian Gulf for the British Institute of Persian Studies. Has published articles in Iran and other learned journals.

340 gns.
8 page illustrated brochure available from:

260-1 Tottenham Court Road, London, W1P OAL, England Tel: 01-6368070 Cables: Swanelnic LondonW1 Telex: 22682

W. F. & R. K. SWAN (Hellenic) Ltd.

AND ITS NIMRUD REMAINS

by M. E. L. MALLOWAN

Royal 4to - 121in. x 81 in.

Two volumes - 360 pp. and 320 pp.

Bound in superfine cloth in slipcase; nine colour plates; nearly 600 monochrome illustrations (including reproductions of about 200 of the ivories).These volumes are important not only to the archaeologistand the historian concerned with Western Asia and the Mediterranean, but also to all who are interested in art, ancient or modern, and for the art historian they have a significant message. Published by William Collins Sons & Co. Sixteen guineas February 1966 Please order as follows: In the United States of America from Dodd, Mead and Co., 432 Park Avenue South, New York 10016. In all other countries from your usual bookseller.

IVORIES FROM NIMRUD 1:2 Equestrian Bridle-Harness

THE BANK OF IRAN AND THE MIDDLE EAST

Ornaments
by

J. J. Orchard
A catalogue of one class of ornamental ivories found at Nimrud during the excavations there by the British School of Archaeology in Iraq in 1949-1963. This is the first volume in a comprehensive series of publications which is intended eventually to cover the whole of the great collection of Nimrud ivories. Published by The British School of Archaeology in Iraq, 31-34 Gordon Square, London, W.C.i. Price: 4 guineas

Head Office: KUCHEH BERLIN, TEHRAN, IRAN

The Bank is associatedwith THE BRITISHBANK OF THE MIDDLE EAST which was itself founded in 1889 as: The Imperial Bank of Persiaand which provided extensive bankingservices in Tehran and the provinces until
1952.

Thus we can draw on an experience of business in Irangoing back for more than 80 years
Also associated with THE HONGKONG AND SHANGHAI BANKING CORPORATION and MERCANTILE BANK LIMITED

The Persian Land Reform

1962-1966
Ann K.

S.

Lambton

The primary aims of the Persian land reform were to break the power of the landlords and to give the peasants responsibility for the running of their own affairs. Professor Lambton's extensive travels in Persia both before and after the land reform began enabled her to assess both the difficulties and the achievements. She examines the law and its implementation and reception in the various provinces against the background of the prevailing climatic, social, and tenurial conditions, and the part played by the rural co-operative societies in the achievement of the aims of the reform. 1 map. 84s.

OxfordUniversityPress

(J. S. D. THORNTON)

J. THORNTON & SON

Established 1835

Heffers
two new catalogues free on requestare:

UNIVERSITY BOOKSELLERS
11 BROADSTREET,OXFORD
Telephone Oxford 42939

ORIENTAL AND AFRICAN STUDIES LINGUISTICS

Specialists in Orientaland Archaeological & Books & Periodicals-Importers Exporters

catalogues issued

High prices paidfor largeor small collections of books on

Writeto: W. Heffer &rSons Ltd

Petty Cury, Cambridge, England

Oriental Subjects

IRAN
Journal
of

the

British

Institute

of

Persian

Studies

VOLUME VIII

1970

CONTENTS . . . . . . . . . Governing Council . . . Obituaries, ProfessorA. J. Arberry; Dr. S. M. Stern . Director's Report . . . . . . . . . Excavations at Sirif: Third Interim Report, by David Whitehouse . . Pyramidal Stamp Seals in the Persian Empire, by John Boardman Persian Lacquer in the Bern Historical Museum, by B. W. Robinson . Charles I and the Antiquities of Persia: The Mission of Nicholas Wilford, . . . . . . . by R. W. Ferrier . . Iletmish or Iltutmish? A Reconsideration of the Name of the Dehli . . . . . . . Sultan, by Simon Digby . Saka Studies: The Ancient Kingdom of Khotan, by H. W. Bailey . Dailamis in Central Iran: The Kdkfiyids of Jibdl and Yazd, by C. E. Bosworth . . . . . . . . . The Archers of the Middle East: The Turco-Iranian Background, by . . . . J. D. Latham Some Kurdish Proverbs, by D. N. MacKenzie . . . . . The Arabic Element in Persian Grammar: A Preliminary Report, by Mohammad Ali Jazayery . . . . . . . The Genesis of Safawid Religious Painting, by J. M. Rogers . . Excavations at BdbdJan, 1968: Third PreliminaryReport, by Clare Goff Excavations at Haftavin Tepe, 1968: First Preliminary Report, by . . . . . . . . . Charles Burney
The Khurtb Pick-axe-Corrigenda, by H. Lechtman of Excavations in Iran, 1968-69 Survey .. .
. .

Page vi vii xi I 19 47 51 57 65 73 97 105 I 15

125

14I
157
173 175

.
.

.
.

Published annuallyby

THE

BRITISH

INSTITUTE

OF PERSIAN

STUDIES

c/o The British Academy, Burlington House, Piccadilly, London, WIV ONS

BRITISH INSTITUTE OF PERSIAN STUDIES GOVERNING COUNCIL President *Sir MAX MALLOWAN, C.B.E., M.A., D.Lit., F.B.A., F.S.A. VicePresident BASIL GRAY, Esq., C.B., C.B.E., F.B.A. Members DEREK ALLEN, Esq., C.B., F.B.A. ProfessorSir HAROLD BAILEY, M.A., D.Phil., F.B.A. R. D. BARNETT, Esq., D.Lit., F.B.A., F.S.A. *Sir MAURICE BOWRA, M.A., D.Litt., Litt.D., LL.D., F.B.A. ProfessorJ. 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., F.B.A. Professor A. K. S. LAMBTON, O.B.E., D.Lit., Ph.D., F.B.A. 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.

Acting Assistant Director

ANTONY HUTT, Esq., M.A. c/o The British Academy, Burlington House, Piccadilly, LONDON, WIV ONS
*Denotes Founder Member

P.O. Box 2617,

Tehran, IRAN

OBITUARIES

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

It is with deep regret that we have to record the death of Professor Arberry on October 2nd 1969, at the age of 64. He was born at Portsmouth on May 12th 1905. Both his parents were keen students of literature, and it is not therefore in the least surprising that he should have inherited this interest in full measure. He was educated at Portsmouth Grammar School, where he showed great promise, winning a number of scholarships. In the autumn of 1924 Arberry went up to Pembroke College, Cambridge, as the major scholar of his year. His academic career was brilliant in the extreme. He obtained first class honours in both parts of the Classical Tripos, with special distinction in literature. Acting on the advice of Dr. (later Professor Sir Ellis) Minns, of Pembroke College, Arberry, after taking his degree, stayed on for another two years at Cambridge, taking first class honours in both parts of the Oriental Languages Tripos. He was awarded the William Browne medal, the E. G. Browne scholarship and the Wright Studentship. In consequence of these outstanding successes, he was made a Research Fellow of Pembroke College. Arberry decided to go to Egypt to pursue his studies there. It so happened that the headship of the Department of Classics at Cairo University was vacant at the time. Arberry applied for, and was given, this post. Beside his teaching obligations Arberry had ample time to pursue his work on Arabic literature. It was while he was in Cairo that he met and married a Roumanian lady, Miss Sarina Simons. vii

Viii

JOURNAL

OF PERSIAN

STUDIES

In 1934 Arberry returned to England, where he became Assistant Librarian in the India Office Library, an occupation which enabled him to pursue his oriental studies to advantage. When the 1939-45 war broke out Arberry was posted to the Postal Censorship, but was later transferredto the Ministry of Information, where he worked under Brendan Bracken. In the Ministry part of his work was to demonstratethe interest that had been taken by past Britishscholarsin the notable achievements of the civilizations of Asia. In 1944 Arberry ended his ten years as a civil servant when he was appointed Professorof Persian in the University of London in succession to that eminent orientalist ProfessorMinorsky. Three years of later, when the Sir Thomas Adams Professorship Arabic at Cambridge became vacant, Arberry was to it and was thus able to return to his " spiritual home ". Furthermore, his Pembroke appointed Fellowship was renewed. Arberry found time from his academic duties to write an impressive number of books on Arabic a and Persiansubjects. In 1958 appeared his Classical Persian Literature, most valuable work. A reviewer said of this book: " ProfessorArberry nobly continues his indefatigable labours to introduce Middle Eastern literature to a wider English public which, in general, refuses to recognize any Persian poet before or since 'Omar Khayyam." Arberry nevertheless devoted two books to 'Omar Khayyam and his truly remarkable English translator Edward FitzGerald. " the mystical movement within Islam " as he himself Arberry was much interested in an Account theMysticsof Islam. describedit. He wrote an excellent book on the subject entitled S.ifism, of to He had previously published another book on the Stifis entitled Introduction theHistoryof S.ifism: S.fitsm. Arberry by no means confined his attention to Persian subjects. He produced a number of works on Arabic literature, perhaps the most notable of which is his masterly two-volume translation of the Koran, entitled The Koran Interpreted. of Arberry also achieved note as an editor, being responsiblein that capacity for TheLegacy Persia. He was chairman of the Editorial Board of the Cambridge History Iran (a large part of which is still in of course of preparation). He was much interested in the British Institute of Persian Studies, of which he had been VicePresident since its inception in 1961; in the early stages of the existence of the Institute his great experience of administration and his wide knowledge of authorities concerned with Iran provided the Institute with invaluable guidance. In the years before his health deteriorated, he attended Council meetings regularly and contributed to this journal. He also formed a plan of editing, translating and has publishing a series of Persian texts, the first of which, entitled the Humdy-Ndma, already appeared. It is much to be hoped that others may follow in his footsteps in this respect. It is regrettedthat bad health prevented Arberryfrom visiting Iran, where his admirable scholarship was widely and highly esteemed. We are deeply grateful for all that he has done on our behalf, and we believe that it caused him every satisfactionthat our Institute has followed on the sound lines that were laid down at its inception. L.L.

OBITUARY

ix

Dr. S. M. Stern, M.A., D.Phil.

The death occurred on October 29th 1969 of Dr. Samuel Miklos Stern, a contributor in the past to this journal and a scholar whose great breadth of interest extended from Muslim Spain and the extreme west of the Islamic world to Iran and the eastern Islamic regions; his contributions to Jewish studies and to questions of Islamic philosophy and theology were also of prime importance. Born on November 22nd 1920 in Tab, Hungary, Stern studied oriental studies and Romance philology at the Hebrew University in Jerusalem, under such masters as Baneth and Goitein. After spending part of the war years in the British censorshipservice in the Middle East, he came to Oxford and read for a doctorate under ProfessorH. A. R. Gibb on the subject of Hispano-Arabic poetry. He of subsequently fulfilled such duties as Secretary to the new edition of the Encyclopaedia Islam,Assistant of Oriental Coins in the Ashmolean Museum, and was at the time of his death a Senior Research Keeper Fellow of All Souls and University Lecturer in the History of Islamic Civilization. One of Stern's especial interests was that of the extremist Shi'ite sect of the Ism1'ilis, both in its manifestationas the Fatimid Caliphs of Egypt and in its manifestationof the Isma'ili sectaries and the and and Culture Bulletinof theSchool Oriental Assassinsin Iran. Two substantial articles of his in Islamic of the activities of Ismi'ili missionariesand propagandistsin Sind and in NorthStudies examined African West Persia, Khurasan and Transoxania respectively. His considerableerudition as a numismatistwas Chronicle the coinage of Oman under the Biyid Abai Kdlij*r on seen in his articles in the Numismatic of Amul. He had assembled much material on the coinage of (with A. D. H. Bivar) and on the coins the Ghirid Sultans of Afghanistan, although it is not yet clear whether this is in publishable form. In Iran IV (1966) he described the inscriptions of the Kharraqan mausoleums, and an article of his, written in conjunction with Miss Elizabeth Beazley, on the castle of Khdnlanjdnnear Isfahan, had been promised for Iran IX (197') and will, it is hoped, eventually appear in this journal. A detailed bibliographyof Stern's many works is now being prepared by Dr. J. D. Latham and Miss Studies. Helen Mitchell, and will be published in the near future in the Journalof Semitic C. E. BOSWORTH

DIRECTOR'S REPORT
NovemberIst 1968 to October 31st 1969 The past year has seen the establishment in Tehran of the French Institute and the American Institute of Iranian Studies, both of which will be concerned, as we ourselves are, with the whole range of Persian studies. Thus Tehran continues to grow in importance as a centre of field research with, above all, ever wider library facilities, that should be of inestimable value to scholars of all nationalities. Visit of ProfessorJ. A. Boyle During the second week in April Professor Boyle spent several days at the Institute while on his way eastwards to Pakistan and India. His visit closed with a most successful open lecture when he read a paper on " Rashid al-Din, the First World Historian ". OtherLecturesat the Institute " The Medes before Cyrus: Excavations at Tepe Nfish-i Jdin ", by Mr. Stronach on January I8th. " The poetry of Omar Khayyam-and some of its translators ", by Major John Bowen on June 12th. InstituteLecturesin London At the Institute's sixth Annual General Meeting, held in the rooms of the British Academy last November, Dr. May H. Beattie delivered an illustrated lecture on " Some Antique Persian Carpets ", while, in a second lecture held in June, Dr. David Whitehouse gave a detailed account of the results of the first three seasons' work at Sirdf. Director At Professor Negahban's invitation, Mr. Stronach contributed to the lecture programme at the Faculty of Archaeology, Tehran University, during the spring term. In April he completed a tour of Sfisa, Kharg, Masjid-i Sulaiman, and other sites in Khizistan with either Achaemenian or Parthian remains. Additional field trips were made to Fdrs and Kurdistan. In outside lectures, Mr. Stronach spoke on the excavations at Tepe Nflsh-i Jdn and Shahr-i Quimis at the Asia Institute's first annual Archaeological Colloquium, held at Shiraz from September 13th to 15th. Assistant Director Following Mr. David Blow's resignation in February, Mr. Antony Hutt took over his new duties as Acting Assistant Director on March Ist. As part of his current research connected with the evolution of the minaret, Mr. Hutt has travelled widely in the regions of Gurgan, Khurdsdn and Kirman. Several unpublished or otherwise little-studied monuments were planned and photographed in each region with special reference to the Seljuq period. Fellows While still a Fellow of the Institute, Mr. Hutt visited a number of early Islamic sites in Afghanistan and was also able to spend a month in India and Pakistan visiting museums and monuments. In India he represented the Institute at the UNESCO Congress on Central Asia held in New Delhi. As a re-elected Fellow, Mr. Andrew Williamson has been engaged in an extensive archaeological survey in Fdrs, Sistdn and Kirmdn. Five sites with pre-Sasanian painted pottery were found near the coast of the Persian Gulf and, with reference to later settlement, particular interest attaches to the discovery of the extensive ruins of early Islamic Sirj n. Each of the great Islamic cities of old Hormuz, New Hormuz and Kish yielded important new material, with the town on the island of Kish proving much larger, and in part much earlier, than previous reports had indicated. Of the two new Fellows elected for I969-70, Mr. Robert Hillenbrand has already resumed his earlier work on Iranian tomb towers of the early Islamic period. Mr. P. A. Andrews, whose fieldwork is only due to begin in the spring of 197o, will be making a survey of nomad tents in Iran-a survey designed to complement similar studies already completed in both Turkey and North Africa. xi

Library Thanks to the ever-increasingsize of the library, which now houses more than 4000 volumes, most of the collection has been re-catalogued. Steps have also been taken to improve the system of borrowing books. Registered readers have been issued with library cards and those books that may be borrowed are currently being provided with book cards. Guests Those staying at the Institute during the past year have included some twenty students engaged in long-term studies in Iran and more than 200 other visitors. Among others who stayed as guests were: Mr. Edward Archer (University of Liverpool); Mr. D. Barrett (Department of Oriental Books, Bodleian Library, Oxford); Dr. Peter Beaumont (Department of Geography, University of Durham, geomorphological and hydrological research within the region of the south Elburz); Dr. and Mrs. A. D. H. Bivar (School of Oriental and African Studies, University of London, on study leave); Mr. Jacob Black (University of London, social anthropological research in Luristan); Mr. John R. Bockstoceand Mr. Vincent C. Piggot Jr. (University of Pennsylvania, archaeological survey in eastern Iran); ProfessorJ. A. Boyle (Department of Persian Studies, University of Manchester); Professor J. A. Brinkman (Oriental Institute, University of Chicago); Miss Mary E. Burkett (Director, Abbot Hall Art Gallery, Kendal, visiting Islamic monuments in north-east Iran); Mr. and Mrs. Charles Burney (University of Manchester, Haftavdn Expedition); Dr. Schuyler van R. Camman (University of Pennsylvania); Miss Elizabeth Carter (Oriental Institute, University of Chicago, archaeological research in Khizistdn); Mr. Vincent Costello (Department of Geography, University of Durham, field research in Kdshin); Mr. Hubert Darke (Faculty of Oriental Studies, University of Cambridge, literary research); Professor Myles Dillon (Dublin Institute for Advanced Studies); Mr. Michael Fischer (Department of Anthropology, University of Chicago, research on Zoroastriancommunities); Dr. Clare Goff (Kingston-on-Hull College of Education, Bdbd J5n Expedition); Professorand Mrs. G. M. A. Hanfmann (Harvard University, archaeological study); Miss Ann Heckle (draughtsman at the Institute and at Tepe Yalhy) ; Professorand Mrs. Frank Hole (Rice University, Expedition to Deh Liirin); Dr. and Mrs. M. Kirkby (Department of Geography, Bristol University, geomorphological research in connection with the excavations at Deh Liran); Mr. Audran Labrousse (Deligation ArchdologiqueFrangaiseen Iran, architecturalstudies); Professorand Mrs. C. C. Lamberg-Karlovsky (Harvard University, Expedition to Tepe Yahyd); Mr. and Mrs. David Marsden (Department of Anthropology, University of Durham, studying social interrelationsbetween village and city in Fars); Dr. Charles McBurney (University of Cambridge, Palaeolithic Reconnaissance in western and northeastern Iran); Dr. Roger Moorey (Ashmolean Museum, Oxford, visit); Miss Margaret Munn-Rankin (Faculty of Oriental Studies, University of Cambridge, archaeological study); Dr. James Neely (Department of Anthropology, University of Texas, survey and excavation in Deh Liiran); Professor Murray Nicol (Harvard University, Expedition to Darvazeh Tepe); Mr. Ralph Pinder Wilson (Department of Oriental Antiquities, British Museum); Mr. Michael Power (Department of Geography, University of Durham, field researchin Isfahin); Dr. and Mrs. Colin Renfrew (Department of Ancient History, University of Sheffield, Rice University Expedition to Deh Liiran); Mr. Richard Salzer (Department of Anthropology, University of California, social anthropological field research in Fars); Miss Jennifer Scarce (Royal Scottish Museum, studying Islamic monuments in Kirman); Professor Philip Smith (Department of Anthropology, University of Montreal, excavations at Ganj
Dareh Tepe); Mr. and Mrs. John D. Speth (University of Michigan Museum of Anthropology, Ann Arbor, middle Palaeolithic cave excavations at Khurramabad); Professor Brian Spooner (Department of Anthropology, University of Pennsylvania, village study in south-west Khurasan); Mr. and Mrs. William Sumner (University of Pennsylvania and American Institute of Iranian Studies, Tehran); Dr. and Mrs. David Whitehouse (University of Oxford, Expedition to Siraf); Dr. G. L. Windfuhr (Department of Near Eastern Language and Literature, University of Michigan, engaged in linguistic studies); Professor Henry T. Wright (University of Michigan Museum of Anthropology, Ann Arbor, archaeological survey in Khiizistan); Professor T. Cuyler Young Jr. (Royal Ontario Museum, Expedition to Godin Tepe). xii

EXCAVATIONS AT SIRAF Third Interim Report By David Whitehouse

The third season of excavations at Siraf took place between October and February 1968-69.1 During the season we continued work at the Great Mosque and in the residential quarter discoveredin 1967. We examined also a second, smaller mosque and completed a detailed plan of Siraf at a scale of 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 permissionto excavate at Sirif. The Director of the ArchaeologicalMuseum gave us much valuable advice. Mr. Taghi Rahbar once more accompanied us to Siraf as the Representativeof the Archaeological Service and we thank him for his unfailing help. The excavations at Siraf are sponsoredby the BritishInstitute of Persian Studies. During the season we received most generoussupportfrom the Calouste Gulbenkian Foundation, the British Museum, the British Academy, the Royal Ontario Museum, the Monro Trust, the Pilkington Glass Museum, the Russell Trust and the Corning Museum of Glass. We received additional support in the form of a munificent anonymous donation, most of which will be used during the fourth campaign, planned to take place between October and February 1969-70. Finally, the British Museum and the Royal Scottish Museum each kindly released a member of staff to join the expedition. Without this wide and varied support, work on the scale achieved would have been impossible.3 During the season we built a permanent excavation house and museum at Siraf. Part of the cost of this was met by a grant from the Archaeological Service of Iran, which we acknowledge with gratitude. The expedition staff was as follows: David Whitehouse (director), Ruth Whitehouse, Gerald Dalby, Peter Donaldson, Clifford Long and Jan Roberts (site supervisors),Jennifer Aldsworth and Ronald Shoesmith (site assistants), Frederick Aldsworth (surveyor), Lars Hesselgren (assistant surveyor), Dianne Clegg, Enid Parson, Jennifer Scarce, Nicholas Lowick and Andrew Williamson (finds assistants), Rosemonde Nairac (conservator), Sonya Stangroom (draftsman), Giles Sholl (photographer) and Edward Harris (quartermaster). Finally, I am personally grateful to the following for advice on the excavation and the material mentioned in this report: Miss Margaret Medley, Miss Mary Tregear, Mr. John Ayers, Mr. Basil Gray, Mr. Ralph Pinder Wilson, Dr. S. M. Stern and Mr. David Stronach. THE EXCAVATION The excavation lasted fourteen weeks, during which we employed up to 200 workmen. Throughout the season work continued at the Great Mosque, providing new informationon the history of the main enclosure and the extension. Outside the main enclosure we uncovered an ablution area and part of a bazaar. Concealed beneath the Great Mosque were the remains of at least one large building which
1

I : 500.2

For preliminary reports, see Iran VII (1969), p. 182 and AntiquityXLIII, no. 170 (1969), pp. 107-8. SThe new plan covers most of the area mapped by Sir Aurel Reconnaissances North in Stein and published in Archaeological India and SouthEasternIran (London 1937), pp. 202-12, Western particularly plan 17. The surveyors began the task of contouring the plan by establishing a series of bench marks. We hope to complete the levelling next season. 3 We thank also the Department of Civil Engineering, Imperial College of Science and the Winchester Research Unit for lend-

ing equipment. As before, the staff of Decca Services, Bushire, gave us both hospitality and assistance. In addition to direct financial support, our sponsors again offered bursaries for competition among members of staff. Bursaries were awarded to Frederick Aldsworth, Gerald Dalby, Clifford Long, Jan Roberts and Giles Sholl. Our sponsors also lent us a long wheelbase Land-Rover. We are indebted to curator and staff of the Pitt Rivers Museum, Oxford, for providing us with storage space for the finds.

JOURNAL

OF

PERSIAN

STUDIES

was already in ruins when the main enclosure was begun c. 800 A.D. In the residential quarter we excavated four more houses, all of which had been abandoned by c. I150. At the smaller mosque (Site G) we discovered a shrine of the late thirteenth or fourteenth century. The three sites are described in the following order: I. Site B. The Great Mosque. Site F. The Residential Quarter. 2. 3. Site G. The Shrine. Site B. The Great Mosque By the end of the season we had uncovered the whole of the Great Mosque, revealing a complex of buildings 55 m. square, comprising the main enclosure, extension and ablution area (Fig. I). Excavation below the floor of the main enclosure yielded new evidence for the form of the earliest mosque, while a detailed examination of the pier bases increased our knowledge of the latest restoration. On the basis of this information, we suggest that the history of the Great Mosque should be revised as follows: I. The original mosque was a rectangular building fourteen bays deep and thirteen bays wide. The courtyard was surrounded on three sides by a single arcade, with a triple arcade in front of the qibla wall. The mosque was enlarged by the addition of a fifteenth bay at the qibla end. Possibly at the 2. same time, an additional bay was inserted on all four sides of the courtyard. Thus, at the end of Period 2, the mosque had a courtyard surrounded on three sides by a double arcade, with a sanctuary five bays deep. 3. The extension was added. The ablution area was built at the same time, or slightly later. 4. Part of the extension was rebuilt. 5. The main enclosure was restored after a partial collapse. The Main Enclosure. Excavation beneath the floor of the main enclosure told us much about the earliest mosque, revealing its plan, the form of the mihrdband the appearance of the arcades. The original plan emerged from an examination of the foundations which support the arcades. We observed at an early stage that the arcades conceal two types of foundation: dry stone and mortared walls. Furthermore, while foundations thought to be original (e.g. arcade 3) are dry stone, those considered secondary (e.g. arcade 6) are mortared (P1. Ic). By the end of the season we had recovered a consistent pattern: foundations 3, 5 and B are dry stone; 15 and M were originally dry stone but have been partly rebuilt using mortar, and 6, 14, C and L are mortared throughout. Clearly, the dry stone foundations are original and the earliest mosque had a single arcade on three sides of the courtyard (arcades 15, A and M) and a sanctuary three bays deep (arcades 3, 4 and 5). During the first season we concluded that bay I is a secondary feature and that the original qibla wall occupied the position of arcade 2, a deduction based on the discovery of a circular buttress at the southeast end of the arcade (P1. IIa).4 Last year we confirmed this by excavating the remains of a mihrdb projecting from foundation 2. The mihrdbhas particular interest because it contains two periods of construction, each with a different form. As it stands, the structure resembles the mihrdbof Period 2 (P1. Ia and b and Fig. 2) and consists of a rectangular projection from the qibla wall. The projection is roughly built and, unlike the qibla wall, contains little or no mortar. Moreover, there is no bonding between the lower part of the mihrdband the qibla wall and it is clear that the former was inserted into the wall, removing all trace of the original structure. However, the absence of a scar on the qibla wall implies that the first mihrdbwas contained in the thickness of the wall and that this original " flat " mihrdbwas replaced by a projecting structure before the mosque was enlarged in Period 2. In my second report I suggested that the earliest piers in the Great Mosque possibly rested directly on the foundations, without the square bases used in Period 2.5 We know now that the conjecture was mistaken. During the excavation of the earliest mihrdbwe exposed the foundation of arcade 3 in bays F and G. Underlying bases 3F, 3G and 3H, and concealed beneath the sanctuary floor, we found the
*SirTf I, p. 9. 6 Siraf II, p. 43.

I.

EXCAVATIONS 1
2 3 4 5 6 7 8

AT
9

SiRAF
10 11 12 13 14 15

A4

STREET

SLOT B

IDfEl
P2
El- -

i DR
I N

-17

G
MBELL H
/

MR

238

I
i

i
/D
71DRR
JARCISTERN -

STEPS

K-----

--------

----

"SLOT
I
El

EXTENI
L-, L ....

STR E

C IS TE S

L__.J

L__ P O

OVEN
4-

FOOTPATH

.
D,-i-

STREET
VAULTED CELLAR 0 15

METRES

HA OD PD JR & DW

PERIODS 1 and 2

plan. Fig. z. The GreatMosque. Simplified

JOURNAL

OF PERSIAN

STUDIES

2G

~foundation

of mihrab
DRY STONEMIHRAB inserted in 1 QIBLAWALL

\ ,

o
MIHRAB ofiche / 2 PERIOD /PERIOD footing
0

hole

slot hole mortared foundation

4 metres
Fig. 2. The GreatMosque. The mihrabs of Periodsi and 2.

PDm DWd

square bases of the piers of Period I. If the arcade is typical of the earliest mosque, the piers had square bases in both Periods I and 2. It follows also that the mosque was not simply enlarged in Period 2; it was partly rebuilt. The original qiblawall and arcade 3 were demolished, while in arcades 15 and M not only were the piers replaced but the foundations were thoroughly restored. The Extension. During the season we continued work in the extension, uncovering the original floorsin all surviving bays and removing the floors in bays 3 and 8-11I. This revealed the original form of the extension and provided evidence for its date (see below, p. 6). We knew already that the extension is a rectangular structure divided into eleven parallel bays. The floors are almost 2 m. below the floor of the main enclosure and the walls dividing the bays are remarkablythick. The jambs of the entrances into the extension are of a different construction from the walls. In an attempt to explain these features, we speculated that the extension may have been designed with floors at the same level as the sanctuary, but that plans were changed and low-level floors and openings inserted. Such a view reconciled the massive partitions, which are comparable with the mortared foundations in the main enclosure, with the low-level floors." Last year, however, the hypothesis was destroyed when we investigated a blocked opening in wall 12 and discovered a second opening in wall I I.1 If the conjecture were correct, the opening in wall 12 would prove to be a secondary feature, cut through a solid foundation. This is not the case; the opening is original. Furthermore, we found also an original opening in wall I I, connecting bays Io and I I (P1.IId). Thus, it appears now that the extension was
6
7

SirdJfII, pp. 44-5. In 1967-68 we discovered two openings in wall i2: the opening mentioned above and an opening between bay ii and the

ablution area. The second opening was rather weathered and we were uncertain whether it was original or inserted. Lack of time prevented us from investigating the first opening.

EXCAVATIONS

AT

SIRXF

built as intended, with rooms at a low level entered from the street. It is possible that a wooden roof existed at the level of the main enclosure and that the partitions supported the arcades of an upper storey, but no trace survives of a superstructureor roof. After an interval, during which the bays were used for storage and ablution, part of the extension was backfilled and floors were constructed, or reconstructed, at a higher level. In bays to and I I we discovered pier bases and plaster floors at the same level as the main enclosure, built after the bays had been partly backfilled with earth and rubble. The project involved replacing several piers in arcades M and N and it was probably at this date that the side of the main enclosure was rebuilt between bays 12 and 15. The Ablution Area. The ablution area fills the angle between the main enclosure and the extension and completes the complex of the Great Mosque. It is a rectangular building, 14 m. long and I I m. across (P1.IIIa). At the centre is a rectangular cistern surroundedby a paved precinct, which contains a latrine. The cistern floor is only m. above the level of present day high tides and the latrine pit is two permanently waterlogged. On theo.3 outer sides of the precinct are narrow ranges of rooms, probably intended for ablution. The precinct has an entrance from the street in the east angle and a second opening leads into the extension. Between the precinct and the main enclosure is a passage leading to the opening into bay I I of the extension, described above. The ablution area was frequently repaired. The ranges of rooms were remodelled and when part of the extension was backfilled the passage was converted into cubicles. The dump of pottery found last year outside the extension had been thrown into the passage after the opening had been blocked, but before the cubicles were begun.8 Clearly, the ablution area was already in use when the extension was rebuilt and, although final proof is lacking, the two structuresmay well be contemporary. TheLatestRestoration. The Great Mosque was extensively restored, probably in the twelfth century. the main enclosurethe latest repairswere carried out with a distinctive type of pink plaster. Throughout If we are correct in assuming that all late repairs with pink plaster are contemporary, the restoration entailed: (i) replasteringthe arcades; (2) building or at least repairing the semicircularbuttressesin the sanctuary; (3) building the rectangular buttresses; (4) repairing large areas of the sanctuaryfloor; and (5) constructing the small rooms in bay I5.9 Restoration was apparently long overdue, for a rectangular buttress to pier CIo incorporates a fragmentary pier identical in size and construction to the piers of Period 2. Evidently, part of the mosque had collapsed. TheFinds. The Great Mosque yielded an important collection of finds, particularlyfrom the makefor Period I of the main enclosure. Among the material from Period i are Islamic lead coins and a up vast amount of pottery. The coins are poorly preserved and so far we have failed to identify either the mints or the rulers concerned.1' However, the script and style of the coins point to a date in the eighth century and their similarity suggests that they were issued by a single mint, possibly at Sirdf itself. Among the pottery is a fragmentary stonewarejar bearing two Arabic names, Yiisuf and Manstr or Maymin (P1. XIIc and d). The names have particular interest because they were incised before the vessel was glazed. The jar is undoubtedly Chinese and we conclude that it was made to order for a Muslim exporter, trading from a port in south China. Another Chinese ware present in Period I consists of bowls with a cream slip and decoration in green and brown." This " painted stoneware " occurs also at Sites A and C, but is rare at Site F, where it had gone out of use before the alleys were filled with rubbish (see below, p. 15). The only abundant type of glazed pottery in Period I was Sasanian-Islamic ware, which has a cream fabric and a vivid green glaze. The best known form is ajar frequently decorated with applied, stamped and incised motifs.'2 Scholars have proposed a wide range of dates for jars of this type and it is important to note that numerous sherds occur in Period I. Other
s Siraf II, p. 46.
10 Lead coins occur also in the make-up for Period 2 and in the 9 To this earliest levels of the shops. It should be noted that the corroded period belong also the impression of the mimbar, recorded in Sfirafll, fig. 2, and the slot which runs continuously lead coins from bay IL, provisionally identified as Indofrom IE to 3E and 3N. We presume that the slot marks the Portuguese in Sirdf II, p. 48, are in fact early Islamic. position of a wooden rail, enclosing the most important part of 1 SirdfI, pp. 17-18 and pl. VIb. the sanctuary. 1i Sirdf I, pl. VIc. 2A

JOURNAL

OF PERSIAN

STUDIES

Sasanian-Islamic pottery from Period I includes a fragmentarybarrel-shaped with incised ornament jar and a Kafic inscriptioncontaining the maker'sname, IHltim. Among the unglazed pottery are sherdsof unusually fine cream ware bearing Kfific inscriptions,vine scrolls and other vegetable motifs moulded in low relief. The sherds are comparable with the finest moulded pottery found at Sfisa.13 The extension, too, yielded a wide variety of finds, including a delicate glass vessel, cm. across, 5"4 decorated with moulded ribs (Pl. XIIa). We also found, in a disturbedlevel, a plaster holder for a lamp cm. high, with a circular depression or incense burner (Pl. XId). The holder is a solid plaster cube, in the top. The sides bear moulded ornament which includes 8.8 a star-shaped motif and a palmette. A fragment of a second plaster holder was found in house N at Site F, in a layer deposited before the introduction of later sgraffiatoware in the eleventh century.14 In 1967-68 we discovered two fragmentarycrenellations among debris from the facade of the main enclosure.15Last year we were fortunateto find a third crenellation, broken but complete, buried under the floor of bays MI 1-12 (Pl. XIb). The crenellation, which is 42 cm. high, is made of plaster and bears vegetable motifs moulded in low relief. Part of the floor of bay M was replaced when the south-eastwall of the mosque was restored (see above, p. 5) and we assume that the crenellation was broken and discarded during demolition prior to rebuilding. It would follow from this that not only the faqade but also the side of the main enclosure were embellished with stucco crenellations. Finally, a shallow pit in the street to the south-east of the ablution area yielded a third fragment of the dedicatory inscription found in 1967. The new fragment provides the opening words of lines 1-3, but does not contain a date. The assembled fragments (P1. XIc) may be translated as follows: Shihdb ad-Dawla 'Izz al-Milla Nasir al-Umma Jamdl. ...
Kd(mri). ... (i) In the name of God the Clement, the Merciful. This is what was ordered by the Amir. ...

(3) al-Ma'MliKdmrfi b. Hazdrasb b.

(2)

Although the make-up for Period I yielded a large collection of coins, most were lead pieces which defy identification. Nevertheless, we also found several bronze and silver coins, the latest of which was minted c. 780. Since the lead coins are attributed on the basis of style to the eighth century, we suggest that the earliest mosque was built soon after c. 780. Last year I reported that the pottery associated with Period 2 in the main enclosure contained none of the tin-glazed pottery found in Period 2 of the sounding. Last season we recovered several thousand additional sherdsfrom the make-up for Period 2 and these, too, did not include a single tin-glazed fragment. Indeed, the only glazed wares commonly found in Period 2 were Dusun stoneware and Sasanian-Islamic pottery and the collection bears a marked resemblanceto the pottery of Period I. If we are correctin believing that tin-glazed wares were developed in the mid-ninth century, Period 2 can hardly be later than c. 850o.1 A cache of seven coins, concealed before the extension was built, provides a lower limit for the date of Period 3. The latest coin was minted in, or after, 1024. Shortly after the extension was completed, later sgraffiatoware came into use.18 Observation here and elsewhere at Sirdf suggests that the ware was introduced c. I05o and for this reason we conclude that the extension was built some time between
1024 and c. 1050.

suggested that Periods I and 2 belonged to the ninth century, while Period 3 was later than c. Iooo.16

TheChronology theGreat Mosque. The abundant coins and pottery found in 1968-69 allow us to be of more precise about the chronology of the Great Mosque than was possible the previous year. I then

The dates of Periods 4 and 5 depend on the chronology of Seljuq pottery with a fritty paste. The alterations to the extension, which comprise Period 4, were carried out before, and the restoration of Period 5 after the introduction of fritty wares at Siraf. The earliest wares with a fritty paste are usually
13

de musulmanes Suseau Musle du Raymond Koechlin, Lesciramiques Louvre(Paris 1928), p. 28, cat. nos. 29 and 30o; cp. pl. IV. 14 For an exceptionally fine plaster holder, see George T. Scanlon, " Fustat Expedition: Preliminary Report 1965. Part II ", Journal of the AmericanResearchCenterin Egypt VI (1967), pp. 65-86, particularly p. 74, fig- 5a. 16StrdfII, p. 44; p. 57, fig. 8 and pl. VIIb. 16 Straf II, p. 46.

The date of the earliest tin-glazed pottery is at present in dispute, although the occurrence at Qairawin of tin-glazed tiles with lustre decoration, almost certainly exported from Baghdad in 862, strongly suggests that the technique of making tin-glazed pottery was established in Iraq by the middle of the ninth century. 18 For a description of " later sgraffiato ware ", see Sirdf p. 15, I, and for an illustrated example, see Sirdf II, pl. VIe.
'1

Pl. Ib. Site B. The mainenclosure

Pl. Ia. Site B. The mihr7dbs periodsI and 2. of

Pl. Ic. Site B. The mainenclosure.Foundations o at thejunctionof arcades 5

at Pl. IIa. Site B. The mainenclosure. Foundations thesoutheastend at of bay I. The qibla wall of period I, reinforced the angle with a rounded is buttress, on the left. Pl. HIb. Site B. The mainenclosure.Part of thepre-mosque between foundationsof bay I. the structure,

Pl. IIc. Site B.

The bazaar. An earthenvare bread oven.

between Pl. lId. Site B. Theextension. The blocked opening bays io and Ii with (left) one of the crosswalls in bay I I.

P1. IIla. Site B. The ablutionarea. Generalview, lookingtowardsthe extension.

Pl. IIIb. Site B. The bazaar. A row of shopsabuttingontothefagade of the main enclosure, from the northwest.

Pl. IVa. Site F. Generalview, lookingsouth.

Pl. IVb. Site F. HouseR, lookingeast.

Pl. Va. Site F. HouseN, lookingsouth.

Pl. Vb. Site F. The mainstreet,lookingeast, with house W and theprivatemosqueon the right.

P1. Vc. The street on the south side of house E, looking towardsthe to entrance houseS.

of P1. VIa. Site F. The courtyard houseN, lookingnorth.

wall in thealleyon the left. to E. P1. VI b. SiteF. The mainentrance house Note thelate blocking

Pl. VIIa. Site F. The alley separatinghousesR and Y and N, S and E, lookingsouth.

P1. VIIb. Site F. Drain in the alley to the east of houseE, as found.

PI. VIIc. Site F. Drain in thealley to theeast of houseE, with thecover stoneremoved.

Pl. VIIIa. Site G. The main entrance.

Pl. VIIIb. Site G. The mihrdb,seenfro entran

Pl. VIIIc. Site G. The mihrdb.

Pl. IXa. Site G. Openingwith an ogival archin the undercroft.

Pl. IXc. Site G. Stepsdownto the u

Pl. IXb. Site G. The undercroft the east side of the mosque,exposed on floor rooms. by the collapseoj theground

outsidethe mosque,seen from above. P1. Xa. Site G. The ruinedtombchamber

Pl. Xb. Site G. An openingin the und in si

Pl. Xc. Site G. The largercistern.

panelfrom houseN. Length92-5 cm. P1. XIa. Site F. Stucco

Pl. XIb. Site B. Stuccocrenellati enclosure.

Pl. XIc. Site B. The dedicator,inscription.Length68 cm. P1. XId. Site B. Plaste Height

Pl. XIIa.

Site B. Glass vesselfrom the extension. Diameter

5"4

cm.

P1. XIIb. Site G. Glasslamp. Height9-6 cm.

Pl. XIIc. Site B. Chinese stoneware with two Islamicinscriptions. jar Diameter42 cm.

Pl. XIId.

Site B. Chinese storeware jar. Detail of inscription.

Pl. XIIe. Site F. Sasanian-Islamic amphora from houseN. Height 60 cm.

Pl. XIIf.

Stone shank from a grapnel anchor. Length 77 cm.

EXCAVATIONS

AT

SIRAF

10

11

12

13

14

15

A B

' C L-

F
G r

238

LL

47

metre

L~j (ERYC13 yA(LATE); PRE-OSQU

REMSU

0
metres

15

or beneath GreatMosque. the Fig. 3. Site B. The eighthcentury earlierstructures

JOURNAL

OF PERSIAN

STUDIES

attributed to the late twelfth century and it is unlikely that they were manufactured before c. I 150o. It is highly improbable that the impoverished community described by Yaqiit in 1218 had sufficient resources to restore the Great Mosque and we assume that Period 5 was completed some time before that date. If pottery with a fritty paste was perfected in the third quarter of the twelfth century, Period
5 belongs to the period c. I1150-75 to 1218.

On the basis of this information, the revised chronology is as follows:

Period I. Shortly after c. 780; perhaps c. 8oo.

Period 2. c. 85o, or possibly earlier.

Period 3. After 1024, but probably before c. 1050. Period 4. Before c. I 150-75. Period 5. After c. 150-75, but before 1218; perhaps c. II75-1200.

The Shops. Outside the Great Mosque is a group of small structures and a narrow street. The structuresabut on to the fagade of the main enclosure and consist of two rows of premises,one on each side of the steps leading into the mosque (P1. IIIb). The structures are flimsily built, with plaster partitions and paved or plaster floors. Few, if any, would have supported an upper storey. The partitions are without openings and each room is entered from the street. All the rooms are small, the largest measuring barely 3x 2 m. internally. Several rooms contain ovens (P1.IIc) and it is clear that the structuresare the small lock-up shops and workshopsof a bazaar. The shops extend along the northeast side of the ablution area and it is tempting to conclude that the Great Mosque was surroundedon the landward side by a bazaar. The excavated shops abut on to the Great Mosque and it is probable that the bazaar developed after the constructionof the earliest mosque. Unfortunately, the evidence is insufficientfor us to determine whether it was built as a piece of town planning or grew up haphazardly during the ninth century. The earliest shops yielded pottery and lead coins comparable with material from Periods I and 2 in the main enclosure, while the latest structures were associated with sgraffiato ware of the eleventh and twelfth centuries. By the time the Seljuq wares reached Sirif, the shops were deserted. TheEarliestOccupation. floors and footings concealed beneath the mosque belong to some of the The earliest structuresfound at Sirif and comprise one of the most important discoveries of 1968-69 (P1. IIb and Fig. 3). The remains consist of one or more large buildings occupying an area of at least 70 x 55 m. The buildings were already in ruins when the firstmosque was built and at one point a small structure was erected after the end of the early occupation but before the construction of the mosque. We found the buildings firstbeneath bays 9-11 of the extension, after removing the floorsin a search for evidence to date Period 3. Subsequently we uncovered footings below the main enclosure, the ablution area and the bazaar. It is clear that the remains of buildings exist beneath a large part of Site B, although erosion has removed all trace of structuressouth of a line between bay IL and extension bay 8. The buildings appear to be roughly contemporary with the earliest deposits at Sites A and
C.19

The buildings have walls of mortared stone which rest on stone foundations set in mud. They have plaster floors. With the exception of the small structure mentioned above, we have no complete plan
and the form and function of the buildings are unknown. Unlike the houses at Site F, which have narrow ranges of rooms surrounding a central yard, the buildings consist of broad agglomerations of rooms linked by connecting doors (e.g. under bay i) or narrow passages (e.g. under bays 8L and 9M). At present we have little evidence for the date of the earliest occupation. Material from below the level of the floors contains little glazed pottery, although scraps of Dusun stoneware show that goods were already imported from China. Among the finds are two sherds with short Pahlavi graffiti. The erosion and the small structure suggest that an appreciable period elapsed between the abandonment of the buildings and the construction of the mosque. If the mosque was begun c. 8oo, the early occupation ended in the eighth century and its origins may well be earlier.
19

Siraf I, p. 2x.

EXCAVATIONS

AT

SIRAF

SiteF. The Residential Quarter In 1967 we excavated an impressive house near the centre of Siraf, some 400 m. west of the Great Mosque.20 Last season we investigated a large area to the east and north-east of the house, revealing four additional houses and parts of five more (P1.IVa and Fig. 4). The area emerged as a prosperous residential quarter bisected by a grid of streets and narrow alleys. HouseE. House E is a rectangular building, 27 m. long and 12 m. wide. We recovered evidence for three periods of construction on the site: (I) a building which had been completely demolished and survives only as a footing on the south side; (2) a rectangular courtyard house; and (3) a phase of repair and re-occupation in the ruins of Period 2. As it stands, house E belongs almost entirely to Period 2 (P1. VIb and Fig. 6). The house has an asymmetrical plan, consisting of a courtyard with rooms on three sides only. Like the building excavated in 1967 (house W), it has both front and rear entrances and, like house N (see below), it has a faqade turned through a few degrees to conform to the line of the street. As in house W, the main entrance is marked by pairs of double pilasters, the bases of which survive at the inner end of the passage. Excluding the entrances, the building has eight ground floor rooms, including a well house near the centre of the west side. A shallow vault supports the floor of the room in the north-east angle, but we have no reason to suppose that the cavity was intended as a cellar. On the east side the courtyard wall projects almost I m. beyond the rooms, thus preservingthe building line of Period I. House Houses N and S are relatively small, each occupying only half the area of houses E and W N. (Fig. 5). House N, which measures approximately 16 x 17 m., is almost square, although the fagade has been turned through 20 to conform with the street (P1. Va and VIa). The plan is roughly symmetrical and comprises nine rooms opening into a central yard. On each side of the entrance is a rectangular room measuring 2 75 x 5'5 m. internally. Flanking the courtyard are two pairs of rectangular rooms, one of which is divided by a narrow partition. Opposite the entrance is another m. wide and 6-2 m. long. The room is remarkablefor its vaulted ceiling, the rectangular room, survives on the south wall, I 85 m. above the floor. Traces of a second vaulted springing of which2"3 ceiling survive in the room which occupies the south-west angle of the house. House N possessedmore than one storey. Lying in the room with a partition were fragments of a second narrow wall. Unlike the partition, the fallen wall contained a doorway and cannot, therefore, be part of the ground floor. It follows that at least part of the house was two or more storeyshigh. The house had been repaired,for the doorway into the room to the east of the entrance had been blocked and the walls on the south side of the yard had been refaced with reddish plaster. HouseS. House S is even smaller than N. It is I4-5 m. square and has a complex history with four major periods of building. In each period the house was remodelled and in constructing Period 4 the builders removed much of the preceding masonry. During the season we uncovered the whole of Period 4, but investigated only parts of Periods 1-3, revealing the sequence but not the complete plan of the individual phases. We know little of Period I. In Periods 2 and 3 the site was occupied by a courtyardhouse with semicircularpilasterson the outer walls and a single entrance near the centre of the east side. In Period 4 the house was almost totally rebuilt. On the north and east sides the walls were reinforcedwith rectangularpilasters. The position of the entrance was changed to provide direct access
2. from the broad street on the south side of house E. After successive rebuilding the floors of Period 4 were more than I m. above the surrounding streets and the house was entered by a flight of steps (P1. Vc). The steps lead into a rectangular passage, through which one entered the central yard. The yard is surrounded by seven rooms, one of which has a narrow partition. In the north-west corner of the yard is a well, which belongs either to Period 4 or, more probably, to the post-medieval occupation which overlies most of the excavated area (see below, p. 15). House R. House R is the most unusual and complicated structure found at Site F. It is a large building with an irregular plan comprising two phases of construction and at least one period of repair (P1. IVb and Fig. 7). In its original form, the complex was probably a single house, much of which
,o Siraf II, pp. 48-53.

10

JOURNAL

OF PERSIAN

STUDIES

ii
House_ E

E1

WD

Hos

r...;

House N

ii

O
(est

oi

HoseY Main,

street

,,

t~(

east )
House

House

(west) _ .::

(
metres

.... G,-m.

'

, :--:--::-:-_-_"--GDm DWd"-

Fig. 4. Site F. General plan.

EXCAVATIONS

AT

SIRAF

11

15

metres GD&CLm DWd

Fig. 5. Site F. HousesN and S.

12

JOURNAL

OF PERSIAN

STUDIES

Pit
0
Dry stone wall

15 metres

GDm DWd
Fig. 6. Site F. HouseE.

survived unchanged when it was rebuilt in Period 2. It possessedtwo entrances: a main entrance from the street to the south and a side entrance from the alley to the west. It appears that neither of the existing south doors is original and in Period I the main entrance was probably between the two, opposite the entrance to house N. The house of Period I had an L- or [-shaped yard and its irregular plan may reflect the form of earlier buildings of which nothing survivesabove ground level. In Period 2 the building was divided into two semi-detached houses, each with an entrance from the street to the is south. The new easthouse a compact structure measuring only 15x 8 5 m., with four rooms on the ground floor. The west houseis altogether larger, with eight or more ground floor rooms and two entrances: a new south door and the old side entrance of Period i. The west house underwent repairs and minor alterations, during which the side entrance was blocked, converting the passageinto an extra ground floor room. Like houses N and W, the building had more than one storey (see below). HouseY. Although we excavated only a small part of house Y, we uncovered enough to suggest that it resembled house W, with a fagade comprising two rooms on either side of an imposing entrance. However, unlike house W, it was carelesslybuilt; the pilastersare placed at irregularintervals along the faCadeand the room in the south-west angle has an irregularplan. This room is unusual in having an entrance from the street, but no apparent opening into the courtyard. House Y was repaired, possibly more than once, and a rectangular pilaster was added to the south-west angle.

EXCAVATIONS

AT

SiRAF

13

?wall

drain

1 Period

Period2

SII
0

Uncertain
5

Late walls
15 metres

Fig. 7. Site F. HouseR.

semicircularpilastersand (2) using rectangular pilasters. Wherever informationon their relative dates exists, the semicircular pilasters are earlier. Thus, house W was built with semicircularpilasters, but repaired with rectangular pilasters;21 houses E and S both were built with semicircular pilasters,
21 Sirdf II, p. 51.

at DomesticArchitecture Sirdf. The houses at Site F display two distinct architectural styles: (I) using

14

JOURNAL

OF PERSIAN

STUDIES

demolished and rebuilt using rectangular pilasters; finally, house Y and the unexcavated building south of house S were built with semicircularpilastersbut, like house W, were repairedwith rectangular pilasters. Furthermore,in the two houses with semicircularpilasterswhere evidence survives (i.e. house W and the second and third phases of S) only the courtyards and entrance passages are paved; the living rooms have plaster floors. In contrast, in three of the four houses with rectangular pilasters (houses E, N and R) courtyards,entrances andliving rooms are paved. A similar change occurs in the buildings at Sites A and C where, although we lack information on the pilasters, we know that all the early floors are made of plaster and all the paved floors are late. It is possible, therefore,that a change from semicircular to rectangular pilasters was accompanied by the adoption of paved floors for living rooms as well as courtyardsand entrances. In short, a new style of building was introduced to Siraf and at Site F its arrival coincided with a spate of rebuilding. It is possible, therefore,that the new style was introduced in a period of reconstruction, perhaps after a disaster like the earthquake of 977At least two of the houses excavated last year were more than one storey high. In house R we excavated a fallen wall nearly 6m. high, with a ceiling scar 5 m. above floor level; while in house N 3" we found a tumbled wall which could not be accommodated in the ground floor plan. These, like house W, were evidently two of the multi-storeyhouses mentioned by the tenth-centurywriter Istakhri. Architectural Decoration. Every house yielded pieces of decorated stucco from the debris which filled the rooms. Although none survived in situ, it was clear that most, if not all the fragments belonged to the rooms in which they were found, and it follows that every house had stucco decoration. Much of the plaster is comparable with the stucco found in house W,22 but the most remarkable pieces consist of triangular panels decorated with stars and rosettes contained in a network of stars and polygons (P1. XIc). Several panels retain a band of ornament on the longest edge, showing that they were set above openings. Few of the rooms at Site F possess niches or windows and we assume that the panels were placed above the doors. When Site F was abandoned, several houses were decorated with this type of stucco. We maintain (see below, p. 15) that houses N, R and W were not only empty, but had already collapsed and become filled with rubble before c. o50o. If this is correct, the triangular panels can hardly be later than the early eleventh century, thus forming the earliest dated examples of star and polygon ornament. TheStreets.The excavation of Site F revealed in detail a street plan similar to the pattern of streets found during the survey of surfaceremains at Sirdf. In both cases, we recovered regular grids of streets and while there is no suggestionthat Sirdfwas laid out with a master-plan,it is clear that whole quarters were deliberately planned, presumablyunder official control. The pattern found at Site F consistsof a main street, or streets, 4 m. wide (P1.Vb), intersected by narrow alleys up to 2 m. across. Neither the streetsnor the alleys were paved and the surfaceconsistssimply of earth. Roofs were drained by columns of earthenwarepipes which carried storm water into stone-lined pits (P1.VIIa). Wherever possiblethe pits were dug in the alleys and only one building, house Y, has a drainage pit in the main street. Although most of the pits received columns of pipes, several were used for the disposalof domestic waste, possiblyfrom latrines, and a drain near the south-east angle of house E was found with its stone cover in position (P1.VIIb and c). As the prosperityof Sir~f declined and the houses fell into decay (see below, p. 15), it became increasingly common for the remaining occupants to tip refuse in the alleys. The
alleys to the west of house E and to the south of house S, for example, became so choked with rubbish that the survivors felt compelled to close them with dry stone walls (P1. VIb). The Finds. The houses and alleys yielded a vast number of finds. Among the pottery, for example, is a three-handled amphora with a dull green glaze, deposited on a floor in house N before the building collapsed (P1. XIIe). Perhaps the most interesting finds, however, are two fragmentary stone shanks from grapnel anchors. They were identified after comparison with a complete example found elsewhere at Sirif by villagers and recovered by the expedition from a modern house between Sites B and C. The complete object (P1. XIIf) is a sandstone block 77 cm. long, tapering towards the top. The lower end is
22 Sirf

IIl, p. 52; p. 58, fig. 9 and pl. VIIc and d.

EXCAVATIONS

AT

SIRAF

15

pierced by two rectangularholes, one above the other, cut at right angles. At the upper end is a circular hole, above which are shallow grooves. The rectangular holes originally contained the wooden flukes, while the cable was attached by the circular hole at the top. The grooves prevented the cable from chafing. Mr. Neville Chittick informs me that similar shanks occur at Mogadishu and Kilwa in East Africa. The only datable examples known to me are the two fragmentsfrom Site F; the earlier piece was found in house R, in a layer preceding the introduction of later sgraffiatoware, and the later piece in house E, associated with later sgraffiatobowls. Thus, grapnel anchors with heavy sandstone shanks were already in use at Sirif by the mid-eleventh century. Site ThePost-Medieval Occupation. F was re-occupied in the fourteenth and fifteenth centuries. The occupation consisted of a scatter of small dry stone buildings which yielded pottery and other finds comparable with the material from Site E.23 In 1967 we discovered traces of buildings overlying house W; last season we found buildings above houses E, S and R, and the alley separating houses E and N. The structuresare difficult to date, largely because they lack the Chinese porcelain found in the more opulent building at Site E. However, unglazed painted wares are common and a layer of agricultural soil beneath a building above house S contained three Islamic coins: two minted by AbUi Sa'id( 1316-35) and a third struckin the early fourteenth century. Inserted into the layer was a hoard of sixty Chinese copper coins, lying in a manner which suggested that they had been strung together on a cord. The coins span a period of more than 6oo years. The earliest pieces were minted by Kao Tsu (618-26), T'ai Tsung using the title T'ai Ping (976-84) and Chih Tao (995-8). The latest coins were minted by Li Tsung using the title Chia Hsi (1237-40), Shun Yu (1241-52) and Tu Tsung (I265-75).24 TheChronology SiteF. The occupation of Site F falls into two distinct periods: (I) the medieval of occupation, which ended with the decline of Siraf, and (2) the re-occupation in the fourteenth century. Although detailed information on the earliest medieval buildings must await excavation below the floorsof the houseswith semicircularpilasters,we know much about the period of decline. The stratified earth and rubble which filled the ruins of houses N, R and W yielded tens of thousandsof potsherds,but not a single piece of later sgraffiatoware. The debris filling houses E and S contained sgraffiatoware, but no Seljuq pottery with a fritty paste. It is unlikely that the houses were deliberately filled with rubble and there is good reason to suppose that the debris accumulated gradually. If, as I suggest (see above, p. 6), later sgraffiato ware came into use c. 1o05 and wares with a fritty paste c. I150-75, houses N, R and W had been abandoned some time beforec. 105o and houses E and S beforec. I 150-75. When Ydqiit described Sirif, Site F was choked with rubble. We found nothing to suggest that the area was re-occupied before the fourteenthcentury. However, a century later a village occupied the whole of a triangle enclosed by the beach, the modern road from the Great Mosque to Site F and the watercoursefrom Shilau valley. We do not know yet how long the area was occupied or what prompted the inhabitants to move along the bay to the modern village. 3. Site G. The Shrine The ruins of Site G occupy a commanding position on the edge of a promontory separating the valleys of Kunirak and Shilau, near the north-west corner of Sir~f.25 The promontory ends in a cliff and the ruins actually project beyond the edge, supported by a massive revetment. The surrounding area contains numerous rock-cut graves and is littered with broken grave covers. Before excavations began, it was clear that the ruins included a mosque and investigation revealed that this was associated with ablution facilities and a walled cemetery. Coins and pottery from the site suggest that the mosque was built between the thirteenth and fifteenth centuries and the typology of the grave covers indicates that the area was already a cemetery in the twelfth century. The mosque, therefore, was built in an awkward position, demanding extensive revetments, in the middle of an old cemetery. It is difficult to escape the conclusion, which accordswith local tradition, that the building is in fact a shrine.
23 Sirdf II, pp. 56-8. 24 Mr. Lowick points out that the composition of the hoard is strikingly similar to that of a larger hoard of Chinese coins from Kajengwa in East Africa; see G. S. P. Freeman Grenville,

21

" Coinage in East Africa before Portuguese Times ", Numis6th maticChronicle, series, XVII (0957), PP. 151-79, particularly p. 164. See Stein's plan, reproduced in Sirdfl, fig. 2, and SirdflII, fig. I.

16

JOURNAL

OF PERSIAN

STUDIES

The mosque itself measuresapproximately 15 x 16 m. and is roughly square, although the qiblaside, which projects beyond the edge of the cliff, is considerably bent (Fig. 8). The focal point is an elegant courtyard three bays wide and two bays deep, measuring 8-5 x 7 m. The centre bay on the qiblaside contains a low mihrdb decorated with cast plaster panels (P1. VIIIb and c). The courtyard is entered a porch, which again bears plaster panels (P1.VIIIa), and beyond this is a passage2 - 2 m. long. through On each side of the porch is a squarish room. The north room contains a flight of steps leading to the minaret. The absence of a heavy footing suggeststhat this took the form of a cage-like structureon the roof, rather than a tower."2 To the south of the courtyard are the remains of a double arcade, running the full depth of the mosque. The building is badly ruined on the south-eastside and nothing survivesof the outer wall above floor level. However, the form of the piers on the south side of the inner arcade confirms beyond doubt that a second arcade existed. Finally, to the south-west of the courtyard is a rectangular room 8 -5 m. long and nearly 3 m. wide. Outside the south-west room, on a shelf of rock half way down the cliff, is a small tomb chamber containing a single grave with a stone cover. The cover is uninscribed, but clearly the grave contains an important person. Unfortunately, we do not know whether the actual grave is older than the mosque. In front of the mosque is a walled cemetery containing more than twenty graves, a crypt and a small mausoleum. The graves are aligned at approximately90goto the direction of Mecca, like the burials of Period 2d at Site A.27 Each grave is marked by a vertical slab at the head and foot and some are outlined with stones. On the south side of the cemetery is a crypt containing two or more stone and plaster graves. The crypt measures 4 x 3 -8 m. internally and was at least I .3 m. deep. It was built 2" afterthe mosquefell into decay and was backfilledbefore the cemetery went out of use. On the opposite side of the cemetery are the remains of a mausoleum, 4- I m. square. The mausoleum is built of mortared stone and is plastered on the inside. It has a symmetrical plan with an opening I .4 m. wide in every wall. At an unknown date the north opening was blocked, creating an internal niche. In one corner, I 12 m. above the floor, is a fragmentary squinch, which establishes that the structure had a dome. The chamber itself contains two stone and plaster graves.28 At the side of the mosque is a triangular yard containing ablution facilities and a range of rooms. The buildings are of two main periods: Period I, in which all the components already existed, and Period 2, in which the rooms were completely rebuilt. The ablution facilitiesoccupy the west side of the yard. They comprise two covered cisterns and a sunken area used for washing. The cisterns are of the standard Sirdfi type, being large rectangulartanks, with rounded ends and a vaulted roof.29They were fed with storm water and at Site G we found a seriesof conduits intended to collect water from the roofs and to carry away any excess. The roof of the smaller cistern forms part of the courtyard floor and the top of the larger tank is barely 0* 5 m. above floor level. Between the tanks is a small washing place with a floor I -2 m. below the level of the yard. Access was provided by a flight of steps and the only facilities were a bench and openings in the cistern walls for drawing water. Outside the yard are at least three rectangular platforms associated with conduits bringing water from the larger cistern. The area may have served for washing corpses before burial in the walled cemetery. On the north side of the yard is a range of narrow rooms. Five rooms existed in Period I, occupying an area up to I I m. long and 5 m. wide. In Period 2 the rooms were demolished and rebuilt with a different plan. The new building comprisestwo rooms separated by an opening in the north wall of the
yard. The north room measures I
26 27

m. and the east room, which has a rhomboidal plan, is x "9 6.6

As in modern mosques in the region; see Sfraf II, p. 48. cularly, p. 297. We cannot at present identify any of the graves at Sirif as necessarily Jewish. SfrdfI, p. 6, fig. 3. Today the Sunnis of TTheri bury their dead with the axis of the body pointing towards Mecca, while the 28 The mausoleum is comparable with a larger building, which still retains part of the dome, on the ridge above Site B; see Stein, Shi'is are buried at 900 to this alignment. The two practices Reconnaissances, 207. For similar, but more elabop. explain the different alignments of graves in the major Archaeological cemeteries at Sirdf, which led Stein, ArchaeologicalReconrate domed mausolea, see William Murray Clevenger, " Some Minor Monuments of ", Iran VI (1968), pp. 57-64. naissances, pp. 2o8-og9, to suggest that some of the graves might in the courtyard of the Great Mosque. be Jewish, a conjecture repeated by Jean Aubin, " La ruine de 29 Cp. the ruined cisternKhurdsmn The type is not, of course, confined to SirIf and is widely used Sirf et les routes du Golfe Persique aux XIe et XIIe si~cles ", Cahiersde CivilisationMddidvales I (1959), pp. 295-3oI, partitoday.

EXCAVATIONS

AT

SIRXF

17

Crypt

Light ~ well ..-Steps

r

downto Undercroft

"c
Mausoleum

Cemetery

Stepsup Minaret

-0

Cistern 1 c

/I,

Courtyard

Area

! %

Period1, possibly rebuilt etc Period1, foundations

Platform

Period 2
S Uncertain

10
metres

Well

plan. Fig. 8. Site G. The later mosque. Simplified

18

JOURNAL

OF PERSIAN

STUDIES

2 2 m. wide and up to 6 m. long. During Period 2 a wall was built across the yard, making the east room inaccessible from the mosque. Among the finds associated with Period 2 is a glass lamp of a type commonly found at Siraf. The lamp (P1.XIIb), which has a shallow body and a tapering base, was intended for suspensionin a metal and at Siraf the earliest fragments ring. The type has a long history. Examples occur at Sdmarra,30 come from the make-up for Period 2 at the Great Mosque. Elsewhereat Site B fragmentsare associated with later sgraffiatoware and the vessel from Site G cannot be earlier than 1264 (see below). The Chronology Site G. We cannot date the structureswith precision. The stucco panels in the of and main entrance clearly belong to a late stage in the history of the site and the form of the mihrdb mosque is difficult to parallel among dated buildings elsewhere. At present our evidence for the date of the mosque consists of two coins and some 500 potsherds associated with the early rooms in the yard. Both coins were minted for Abish bint Sa'd with Abagha as overlord (1264-82). If we exclude two residual fragments, one tin-glazed and the other later sgraffiatoware, the only glazed pottery from the rooms has a thin apple green finish. At the Great Mosque identical wares came into use during the currencyof later sgraffiatoand Seljuq wareswith a fritty paste. Among the unglazed sherds,the painted wares characteristic of fifteenth-centurylevels at Site E are completely lacking. The evidence of the pottery combines with that of the coins to suggest that the rooms were in use before the fifteenthcentury, a view strengthened by the discovery on the site of an unstratified tile of the thirteenth or fourteenth century. Although we know little of the later history of the site, we may summarizethe evidence for the whole occupation as follows: I. A cemetery already existed in the twelfth century. 2. The mosque and courtyard were built in the late thirteenth or fourteenth century. A cemetery developed in front of the complex. 3. The courtyard and the cemetery wall were rebuilt, possibly on the same occasion. 4. The mosque itself was redecorated, perhaps in the eighteenth century.31 Note: A preliminary account of the fourth season of excavations appears on p. 187.

30 CarlJohann Lamm, Das Glas vonSamarra (Berlin 1928), cat. no. 145, pl. IV. 31 The style of ornament on the panels recalls tiles of the Zand dynasty at Shirdz. When Kempthorne visited Sirif in 1835, the

mosque was in ruins; G. B. Kempthorne, "A Narrative of a Visit to the Ruins of Taheri . . .", Trans. BombayGeographical SocietyXIII (1856-57), pp. 125-40.

PYRAMIDAL STAMP SEALS IN THE PERSIAN EMPIRE By John Boardman

The intention of this article is to assemble and discuss a number of seals of the Achaemenid period which, for their shape and style, and sometimes their find places and inscriptions,can be attributed to the western satrapies of the Persian empire. The purpose has not been to compile a corpus of these seals, although I hope that the full range of devices upon them will be represented. Special attention is paid to those with Lydian inscriptions, with certain characteristic linear devices in the field, or with purely Greek devices. This is not, then, simply an enquiry into one aspect of Achaemenid glyptic, which at any rate cannot be properly studied piecemeal, in this way, but there are other important issues relating to Lydian words and names, and to Achaemenid iconography, while some contribution may also be made to the much-discussedproblems of Greco-Persiangems by considering their antecedents, which are generally not taken into account at all. The relevant seals known to me are listed at the end of this paper and referred to by their list number only, italicized. Where the number is asteriskedthis means that the piece is illustrated in the plates. The photographs, most of them taken by Robert L. Wilkins, are of plaster impressionsand are shown at twice or (at the top of plates) four times life size. It has proved necessaryalso to use or copy line drawings of some seals. The seals are pyramidal stamps, roughly rectangular in section, with bevelled corners giving an octagonal face which is lightly convex. On a few the edges are rounded to give an oval face, but these are readily distinguishable from the bigger conoids (see below). The sides rise to a rounded back, pierced near the top through the longer axis, and sometimes grooved for the attachment of a metal, stirrup-likefitting with suspensionloop.' The arms of the fitting are often in the form of ducks' heads, as on 137* and the weight stamp, 188*, which is related to our series. Other parts of the seal may on occasion be sheathed in gold or silver. The shape is one adopted in the Persian empire from the regular Babylonian type, which is often slimmer in proportions. Among other closely related shapes which will be mentioned are the more massive conoid stamps, of similar general proportionsto the pyramidal, but round or oval in section. These too derive from Babylon but their distribution is different from that of the pyramidal and their style always closer to that of most Achaemenid cylinders. Some scaraboidswill be found relevant, a shape virtually unknown to Mesopotamia and Persia hitherto. The home of the
The following abbreviations appear here in addition to those in current use in this journal: Delaporte, Bibl. Nat. L. Delaporte, Catalogue des cylindres Nationale(I 9Io) orientaux, Bibliothique Delaporte, Louvre L. Delaporte, Catalogue des cylindres, cachets et pierres gravees de style orientale,Musie de Louvre(1920-23) sur Lajard F. Lajard, Recherches le Culte publicet les mystiresde Mithra en Orient et en Occident(1847) London WA London, British Museum, Department of Western Asiatic Antiquities London H. B. Walters, Catalogue the Engraved of Gems in the Greekand Roman Department,British Museum (1926) Munich E. Brandt, AntikenGemmen deutschen in Sammlungen I, Staatliche Muinzsammlung Miinchen I (1968) I am greatly indebted to the following scholars and institutions for information, casts, permission to study and publish: M. Karamessini-Oikonomides (Athens); D. K. Hill (Baltimore); C. C. Vermeule, E. B. L. Terrace (Boston); K. Herbert, R. V. West (Bowdoin College); R. V. Nicholls, G. Pollard (Cambridge); M. L. Buhl-Riis (Copenhagen); M. L. Vollenweider (Geneva); M. Schliuter (Hanover); N. Firatli (Istanbul); W. C. Braat (Leiden); K. S. Gorbunova (Leningrad); E. Tankard, D. Slow (Liverpool); R. D. Barnett, D. E. L. Haynes, R. A. Higgins (London, British Museum); Mrs. D. Russell (London); H. Kuthmann, E. Brandt (Munich); A. Gallina (Naples); D. von Bothmer, B. F. Cook (New York); R. W. Hamilton, P. R. Moorey (Oxford); G. le Rider, F. Rosswag (Paris, Bibl. Nat.); P. Amiet (Paris, Louvre); J. Baradat (P6ronne); E. Kohler (Philadelphia); R. Noll (Vienna). And to the following for information or discussion of various points: B. Buchanan, H. A. Cahn, J. Friedrich, O. Masson, P. R. Moorey, E. Porada. The drawings of impressions which are not copied either from Lajard or from Kadmos IV (for 7) were executed by Marion Cox. 1 Good examples are shown in C. D. Curtis, SardisXIII.I, pl. Io, which does little justice to the quality of the metalwork.

19

20

JOURNAL

OF PERSIAN

STUDIES

scaraboid lay in Phoenicia but it is its Greek form, as developed by Greek studios in the early fifth century, which is of importance here. Most of these scaraboidsand a few other shapes carry the so-called Greco-Persianmotifs, which are not our concern. Cylinders,the easternseal type par excellence, will be mentioned only where their motifs or style are closely related to those of the pyramidal stamps. The only other shape to be considered is the " weight stamp ", with high sides and flat round face. The devices on them generally have some Phoenician traits. The shape is a common one for eastern stone weights, here turned to a sphragisticpurpose.2 The principal stamp shapes are shown here in Fig. I.

PYRAMI DAL STAMP

CONOID

STAMPS

WEIGHT STAMP

seal Fig. I. Stamp shapes.

The material of the pyramidal stamps is normally blue chalcedony, which had been favoured for the shape by the Babylonians, but white and cloudy chalcedony are also used, and other materials, as cornelian, agate, rock crystal, haematite, lodestone. SEALS WITH LYDIAN INSCRIPTIONS The first group to be considered is distinguished not by any unity of style, but by the Lydian inscriptionswhich the seals carry. There are ten in all known to me. Gusmani considered four in his Wirterbuch Lydisches fir (1964) and Barnett, having discussed two of these in Jahrbuch kleinasiatische
ForschungI (1950-51),

study and use this important piece. The examples I have added have been overlooked by students of both Achaemenid seals and Lydian epigraphy largely because this class of seals has never been isolated and because their inscriptionshave been variously published as Greek, Aramaic, Phrygian or Lycian. One scaraboid (5*) has to be included here with the pyramidal stamps for its style and inscription,and I am not sure of the shape of lo, which may be another scaraboid. The inscription on i* is crucial to our understanding of this class. It reads mitratalis sadmJl. el Mitratalijis a genitive of the good Persianname Mitratag,which is found at Sardis where it seems to be the name of a priest.3 ef is taken to mean " this " or " this is ". sadmis' proved a problem; " work " has
or " possession " has been suggested and Gusmani declares for " inscription ". Reading on the seal

is publishing another (our i*).

I am much indebted to him for permission to

" this is the sadmel Mitratag" we are bound to take it to refer either to the seal itself or to the linear of device around which the inscription is set. In a way it can be seen to apply to both, and the best translationfor it is the Greek word sima, a sign, mark or token. In Greek it is used of a token of identity or blazon on a shield, a seal or a coin. On a sixth-centuryGreek scarab we find exactly the usage in the inscription " I am the sima of Thersis, do not open me ".4 Even more telling is its other use in Lydian, s'is since the words edsadmilappear at the start of a Lydian epitaph6and at Sardissadme twice associated with the word mru-which is taken to mean stili, the object rather than its function or pattern. Here the
Journal XVI (1966), SOn seals and weights see Israel Exploration pp. 187 if. ' See L. Zgusta on " Iranian Names in Lydian Inscriptions " in Charisteria Orientalia, Rypka, 397 if.; E. Benveniste, Titres J. pp. et nomspropresen Iranienancien(1966), pp. o103f. Greece 961), p. I13, no. 7; ( SL. H. Jeffery, LocalScriptsof Archaic Boardman, ArchaicGreekGems (1968), p. 73, fig. 2. SGusmani, pp. 261 f., no. 26.

PYRAMIDAL

STAMP

SEALS

IN

THE

PERSIAN

EMPIRE

21

or exact parallel is the common formula on Archaic Greek gravestones,sima tod'esti, a similar formula the word."Moreover sima and stili are distinguished. employing es sadm-iis omitted on the other seals where we have simply the possessiveof names. On 2* and 3* the father's name is given as well: " of Bakivasson of Sams " and " of Sivam' son of Ate' ". The last two are particularly common Lydian names and Bakivas is shown at Sardis to be the version of the Greek name Dionysikles.7 On 4*, 5* and 6* there is the same name, Manes.8 On 7 (Fig. 2: firstread as Phrygian) the mane suggeststhat this name is again involved but the sense of the other letters is obscure. 8* (first read as Aramaic over a Pehlvi letter) gives tafulim,an otherwise unknown name. 9 (Fig. 2) is known to me only in Lajard's drawing, and since he read it as Greek (" le nom de Bealpolis, 6crit en caracteres grecs d'une 6poque ancienne ") the transliterationcannot be assured, but it seems that we have another possessiveform of a name, ending -olim.9 On io the name might have been Tete.xo The occurrence of the same name, Manes, on three or four seals, might suggest that this is not a personal name but a title. However, it is exceptionally common in Anatolia. There was a Lydian king Manes, a god Men, and in Athens Lydian slaves were called Manes (Strabo 304) as naturally as we might vii
call an Irishman " Paddy ".

Fig. 2. Impressionsof stamp seals.

On i* and 2* the names encircle a linear device, while on the other seals comparable devices appear in the field, except on 4* where there is an elegant flower. These then are personal seals, and the linear devices which form the centrepiece on I* and 2*, and which are fitted in beside the conventional figure devices on the others, are personal blazons or identifying marks of the men named. In a seal usage such as we shall see this to be, with much reduplication of virtually identical figure motifs, these alone could not serve for immediate identification and the linear devices were required. Where the owner could not bespeak a seal with his name and device, he added his device to a ready-made seal. From our examination of the Achaemenid seals inscribed in Lydian it is now possible to understand the significance of the many linear devices which are seen on other seals. These are assembled later in this article. SEALS WITH OTHER INSCRIPTIONS There are no examples with Greek inscriptionsalthough we shall see some with Greek motifs or in a Greek style. Two with Greek Style devices carry Cypriot inscriptions(13*, i5-Fig. 6) naming owners. There are, however, a number with Aramaic inscriptions. This was the official script of the empire and in the Phrygian satrapy capital, Daskylion, Aramaic inscriptions are seen on the Achaemenid sealings,xxwhile at Sardis there is the famous Lydian-Aramaic bilingual. We have then to reckon with the possibilitythat some seals so inscribed may belong to the Achaemenid period. On the pyramidal seals the inscriptions themselves and the style of the figure devices suggest that most belong where the script was originally at home, in Phoenicia and Syria. The two obvious exceptions, with
L. H. Jeffery, BSA LVII (1962), pp.
I 15ff.

for Archaic

epitaphs. 7 Personennamen Zgusta, Kleinasiatische (1964) for these names. s Ibid., pp. 287 ff.

problems about the various possessive endings for these -lid, -lirn,-li, which I am not competent to discuss. names, -liW, 10Ibid., p. 510o. 11AnatoliaIV (1959), p. 128.

9 There are

22

JOURNAL

OF PERSIAN

STUDIES

Achaemenid devices, are 98* and 158 (Fig. 12), which give owners' names, and we shall see what seem to be single Aramaic letters on other pyramidal stamps. These too lend support to the idea that other Aramaic-inscribedstamps of this shape might be Achaemenid. For these, where we lack the evidence of purely Persianscenes or names, the problem is whether they might better be referredto the period of Babylonian domination. Criteria of material and shape (the Babylonian stamps are generally taller and slimmer than the Achaemenid) need not apply for such " provincial " seals, especially when they are from an area with such a strong tradition in stamp seal usage. Some with inscriptionsalone, set in two lines, with or without dividers in the manner of many earlier Phoenician scarabs,fall into this class
(179-182).

two drop-shaped and one circular motif. These have been described as eyes, but are more probably suggested by the pattern of the winged sun disc, highly stylized.12 Of those with figure devices two in the field which are seen on Phoenician and Babylonian stamps. For these too the doubt must linger whether they might belong to the period of Babylonian domination. Complicating factors here are, on the one hand, the evidence for the use of Babylonian pyramidal stamps in Syria and Phoenicia (and there is one with a Cypriot inscriptionadded),13and the probability that stamps with purely Babylonian devices were still being cut in the Persian period. THE LINEAR DEVICES The Lydian inscription on I* indicated that the linear device on that seal was the personal blazon of Mitratas. On 2* the only motif apart from the inscription was a similar linear device, differently composed. Other linear devices appeared on each of the Lydian-inscribedseals except for 4*. A great many other pyramidal stamps which are not inscribed also carry these devices and in all more than sixty are known to me, and their devices tabulated in Fig. 3. They are referred to here and in the
lists as DI, D2, etc. On I* and 2*, as we have seen, the inscriptions and the devices composed the show bulls (172*, 173) in exactly the style of the earlier inscribed scarabs, and with the star and crescent

181 has a Babylonian device on its side. On others (183, 184) the divider is a panel with

entire decoration. Of the others only 3* seems to have been designed to include two devices, but on all the rest they may well have been added after the original design, and to the order of their new owners. The number, variety and significanceof these devices seem hitherto to have escaped scholarlyattention. Nearly all appear on the pyramidal stamps but I have included in Fig. 3 and in this discussionothers of the same type which appear on other shapes. These seals are listed as I85*"-98*. They include three large conoids decorated in a style exactly like that of the pyramidal stamps; three weight stamps; and four scaraboids, a scarab and a gold ring, of Greek shape and with either wholly Greek or strongly hellenized decoration, such as was current in Lydia, where (at Sardis) three were found. Finally, there are devices on an Achaemenid gold ring from the Oxus Treasure (together with an Aramaic inscription) and two on a cylinder. There is, of course, a great variety of devices to be found in the field on eastern seals, and before any special importance can be attached to those listed here we should discount others, which may on occasion be seen on the Achaemenid stamps, but whose occurrence and significance may not be limited to this class, or for which different explanations may be sought. Obvious examples are the star and crescent or the diamond-shaped device seen on many Assyrian and later seals, but the setting or form of some of our crescent-shaped devices (DIo-DI 3) do seem to earn them consideration here.
The ankh or crux ansata is another of these common eastern symbols closely resembling, although not exactly matching, the linear devices. On the pyramidal seals rampant animals or monsters are often set at either side of a tree. At its simplest this takes the form of a spiky branch (as on 34-Fig. 6, to38*) and is merely an abbreviated version of the usual " tree of life ". We may easily distinguish from these the tree-like D42, which is set before a single creature. However, some of the trees set as central features are stylized in a highly unusual manner. One of them, on 68", is combined with one of the usual linear devices (D37), so we should probably not regard these odder arboreal fancies as anything other than artists' stylizations.
12 Compare the scaraboids, London WA LXIV (1941), pl. 7.70. deutschen Palaistina-Vereins 48503 and A.A.S.O.R. Chypriotes (1961), pl. 60.353Jerusalem II-III (1921-22), p. o05, no. 2; Zeitschrift des 1i 0. Masson, Inscriptions

PYRAMIDAL

STAMP

SEALS

IN

THE

PERSIAN

EMPIRE

23

On the only Lydian-inscribedseal without a linear device, 4*, there is a small lotus and palmette
floral springing in front of the monster. A floral very like this appears also on 13o* and 133*. It is not a

normal eastern bloom but the Greek variety in which a palmette is made to grow within a lotus. In almost exactly this form it appears on silver coins issued by native Macedonian princes at the turn of the sixth century.'4 These were apparently designed for them by Greeks, perhaps of Abdera, where coins of the same size (octodrachms) and in a similar style (compare the bulls) were struck. The coins with a floral device should probably be referredto a period of active Persianinterest in north Greece,'5 but this does not mean that the florals were personal devices. We may now turn to the devices tabulated in Fig. 3. D I-D9 are based on circles with simple additions of straight lines or arcs. The forms invented for several of these devices may well have been suggested by other patterns. Thus, D2-D4 recall the winged sun disc with stylized tail, legs or wings. D5 and D6 look like simple bucrania or the so-called " taurine " device.16 D8 and D9 suggest a floral.

10

II

12

13

14

IS

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

Fig. 3. Table of Linear Devices.

The arc devices, D Io-DI13, are sufficiently unlike the usual crescentson seals, in their placing or form (incurved or lengthened arms), to be included here. DI5-D30 offer a variety of devices based on arcs and straight lines, and in the are based on a motif like an omega.17 The loose, D3I-D35 eachcompositions but are found twice, curving patterns D36 and D37 probably to be explained in the same terms as the others. D38-D42 are rectilinear, the last like a small tree. D43-D47 are composed of the same elements as the others. They are not, however, strictly symmetrical but look like contractions and conflationsof the other devices (compare too D28-D30). Of these D43 and D44 were the centre-pieces to the Lydian inscriptionson 1* and 2*. D5o and D51 are monograms,although not certainly of Greek letters, and D5 at least may be a very late addition to the seal (a conoid). D49 resembles a character in south Arabian script. D54-D64 are single letters, with D59 and D62 possibly monograms. Of these D58, D6I and D63 might be Aramaic since they resemble the script of an Aramaic inscription at
x1

Coins(1966), fig. 388, and E.g. C. M. Kraay-M. Hirmer, Greek pp. 328, 333. Compare also the thoroughly Achaemenid stylization of goats on coins of Aigai in Macedonia in this period, ibid., figs. 380, 381. 1 With Megabazus in 512, Mardonius in 492-1 or before Xerxes' invasion. 16 This is seen on a number of late Achaemenid stone seals which were current throughout the empire, especially in the east; see D. Bivar, Journal of the NumismaticSocietyof India XXIII

17

(1961), p. 314. Since our D6 is on a gold ring from the Oxus Treasure it might be interpreted in the same way, and removed from our table, but there are other west Anatolian and even Greek seals to be found in the east, which will be discussed by the writer in GreekGemsand FingerRings (1970). See I. Fuhr, Ein altorientalisches Symbol(1967). The device on the Middle Assyrian cylinder, H. Frankfort, Cylinder Seals Coll. II, pl. 8.Qc.35), was probably (1939), pl. 35c (Southesk added to it at a later date.

24

JOURNAL

OF PERSIAN

STUDIES

Anatolian and Sardis.18 couldbe a Lydianletteror the Lyciantriskeles, the restresemble Aramaic, D64 or Greek(D54-D56) characters cannotcertainlybe placed. Finally,D65 is a pseudo-cartouche, but in possiblyintroduced this positionin imitationof the commonEgyptianand Phoenician composition of a cartouche device. supported monsters.In which case it need not be a personal by So much for the formof the devices. Some generalconclusions alreadybe drawnabout the can seals which bear them. Their shapesand motifsare all Achaemenid, with one or two Greek. The are with them are Lydian, and such of the linear devicesas resemblescript characters inscriptions as Aramaic,the officialscriptof the empire,or in termsof the local scriptsin the western explained of of satrapies Asia Minor,especiallyLydia. The recordedproveniences the sealswith lineardevices tell the samestory: elevenfromSardis,capitalof the Persiansatrapy,othersfromGreece(4), South Russia (3), " Asia Minor" (i) and Persepolis(I). The pyramidalstampsmay not be exclusively on as so. Anatolian-Persian, we shallsee, but they are dominantly The Lydianinscription i* tellsthat the lineardevicesare personal blazons. We have now to askwhatotherevidencethereis for thisusage in the Persian of whosemaincentres production empire. Thisis easilyfoundon another groupof objects were in the westernsatrapies-Achaemenid coins. The mintingof coins, for Lydiansand probablyby Greeks,began in Lydia in the later seventh had taken Sardisin 546 B.c. centuryand soon spreadthroughthe Greekworld. Once the Persians the the mintsforsilverand gold coinsof the olderLydiantype, but soonintroduced silver they keptup " and gold " Darics" showinga Persiankneelingwith bow and spear. They had movedinto " sigloi an areawhichhadbecomeaccustomed coinage,andtheyfollowed fashion.Whilethe chronology the to of Achaemenid and coinsis stillsomething a problem,19 to lackof datedcontexts of an intelligible of due in types,it seemsprobablethat they werefor the mostpartmintedin the west, certainly development at Sardisand perhapselsewhere.Many of thesecoinsare countermarked smalllineardevicesof with exactlythe type we have seen on the seals; some, indeed,are preciselythe same. There are various of tablesof thesecountermarks"0 whichI haveassembled Fig. 4 a selection thosemost in from published

Fig. 4. Select Countermarks from AchaemenidCoins. For sourcessee note 20.

closely resemblingthe devices on the seals, and composed in exactly the same manner. These, with close variants, account for about two-thirds of the known countermarks. It must be borne in mind that we cannot be sure about the correct orientation of a linear countermarkon a coin, since they are struck without regard to the device. Comparativelyfew resembleAramaic, Lydian or other script characters, but more of the later ones are monograms (compare our D50o,D51), found on the early Hellenistic double Darics.21 These countermarksare generally explained as bankers' or changers' devices. From what we have learnt from the seals we may reasonably suppose that they are personal emblems of some sort, rather than emblems of an office or locality. Another source for linear devices of this type is the silver coinage of the west and south Anatolian states within the Persian empire. On many of their coins we find devices either in the field or over animal bodies, ranging from the obvious Lycian triskelesto more elaborate compositions like those on the seals. These are not countermarks,but were cut in the coin die, and may have local rather than
1sJHS XXXVII (1917), p. 78, fig.; Sardis VI.2, pl. I.I. References in Kraay-Hirmer, Greek Coins, p. 358. 1i NumismaticNotes and MonographsCXXXVI, pp. 19 f., pl. 15 o0 (hoards from " Asia Minor " and near Smyrna); Numismatic Chronicle1914, p. 5, fig. I (hoard from Cilicia); E. Babelon, Les Perses Achiminides (1893), p. xi, pl. 39; British Museum Catalogue Coins: Arabia, p. cxxxvii. of Babelon, Les PersesAchiminides, pp. xviii-xx.

'2

PYRAMIDAL

STAMP

SEALS

IN

THE

PERSIAN

EMPIRE

25

personal significance. Some of these coins also carry figure subjects of a purely Achaemenid type. On Carian coins there also appear various composite symbols resembling Carian script, and Carian inscriptions offer ligatures like some of the devices.22 A selection of these Anatolian coin and script devices is shown at the top of Fig. 5. We may also here recall the florals which appeared both on the pyramidal seals (as 4*, 13o*, 133*) and on coins of Macedonia, presumably under Persian domination. This usage of a letter, monogram or device as a personal emblem supplementing the figure device on a seal or coin is an important one, of great relevance to our understanding of the early development of coinage and choice of coin types. This is a subject for another place, but the immediate origins of this usage are worth a moment's consideration. In the east there seems no precedent since most of the symbols on seals seem to be religious rather than personal emblems. In the Greek world, however, the use of a painted or incised mark by merchants or carriers of pottery is well attested from the earlier seventh century on. There are grafiti on seventh-century vases found at Old Smyrna, mainly in the form of monograms.23 The grafiti on Athenian black figure vases of the later sixth century have been studied by Hackl,24 and these are mainly simple monograms or symbols. Specifically East Greek are the red painted inscriptions found mainly on Rhodian and Chian pottery of the first half of the sixth century.25 Of these marks on Greek vases a sample is offered in Fig. 5 below.

LYCIA

CILICIA

PAMPHYLIA ? COINS

CARIA

SMYRNA

GRAFFITI

TOCRA

MARION DIPINTI
Lycia, Caria, Lycaonia, and note 22. For

Fig. 5. Devicesfrom Coins and Vases. For the Coins see British Museum Catalogues: the Vases see notes 23-25.

A similar use of marks of identity may be observed in the brands on the rumps of horses.26 These appear on Athenian vases in the sixth century but they are usually no more than circles or crosses, often stippled. But in Phrygia, within the Persian period, we see horses on painted clay revetments with a simple triskeles on the rump.27 The beasts are ridden in pursuit of griffins. The device (as our D64) is usually associated with Lycia, which was Bellerophon's base of operations, riding Pegasus against the chimaera. The revetments may tell some related story. The use of identifying devices, letters and monograms, appears then to be Greek from the seventh century on. Its later history is mainly that of countermarks on Greek coins but in glyptic we may notice
22

See E. S. G. Robinson in AnatolianStudies(presented to W. H. 25 Boardman and Hayes, TocraI (1966), p. 45, fig. 22, references on p. 46; Histria II (ed. E. Condurachi, 1966), pl. 64; Swedish Buckler), pp. 269-75, especially no. 5; and the ligatures listed in Pauly-Wissowa, Realencyclopddie, II, Expedition pls. 77.1, 140.2 and p. 856, fig. 298.31 for a Cyprus Suppl. VI, p. 144. 23 Jeffery, BSA LIX (1964), P 41, fig. i. And cf. the monogram painted device on a Chian sixth-century amphora from Marion. on an Athenian SOS amphora from Syracuse, BSA L (1955), 26 Some references in Daremberg-Saglio, Dictionnaire s.v. " equus ", p. 8oo. pp. 68 f., fig. 1.6. 24 In " Merkantilische Inschriften ", Miinchener Archdologische 2~ From Duver, A. Akerstr6m, ArchitektonischenTerrakotten Kleinasiens(1967), p. 219, fig. 70.x, and a convenient illustraStudien (1909). A new study is being prepared by A. W. tion in Archaeological ReportsI964-65, p. 66, fig. 5. Johnston.
4

26

JOURNAL

OF PERSIAN

STUDIES

also two Greek gems of the early fifth century bearing monograms: one from a Black Sea site, the other in Istanbul Museum. A comparable monogram appears on a coin of Dardanos in the Troad, of this period. These all seem to be within or on the borders of the Persian empire.28 This is not, however, the end of the matter. The Greek devices are an Aegean phenomenon of the seventh century and later. The Achaemenid pyramidal seals and the coins belong to Anatolia of the later sixth to fourth centuries. For a very similar later use of devices which closely resemblethose on the seals we look to the north and east and a practice which may derive from horse-brandingby nomads. It is well attested in the kingdom of the Sarmatians, successorsto the Scythians, and in the Parthian and Sasanian kingdoms of Persia, which take us a millennium later than the Achaemenid seals. The devices have been studied recently by Janichen,29 who explores their probable personal, royal and religious significance. Many of the Sarmatian and Parthian devices are exactly matched on the Achaemenid seals and it is impossible not to believe that there is some continuity in the practice. But the Achaemenid seals belong to the western empire, they can be related to earlier Greek usage and the devices seem not to be found in Persia in the Achaemenid period, which is where we would expect to find the evidence for continuity. This is not a problem which can properly be further discussed here. DEVICES AND STYLES The figure devices may tell us more about the homes, affiliations and date of the pyramidal seals within the Persian empire. They are here divided into four classes: I. With purely or mainly Greek devices in the " Greek Style ". 2. With " orientalizing " devices either not met at all or generally in another form on other Achaemenid works. 3. With purely Achaemenid devices in the Achaemenid " Court Style ". 4. With non-Greek, non-Anatolian and non-Achaemenid devices or inscriptions. The lists on pp. 39-45 on which this account is based, may not be exhaustive but they include all examples I have noted in publications and collections, and I believe that they give a reliable indication of the range of figure motifs on this shape. I have already given my reasons for the inclusion of the occasional scaraboid, as 5* for its style and inscription, and some oval conoids.
I. GreekStyle (11*-17*)

Seven pyramidal stamps are decorated in a purely Greek Style, closely matched in execution on Late Archaic and Classical Greek scarabs and scaraboids,and with devices which are wholly or largely Greek in inspiration. They must be the work of Greek artists and they serve to remind us that large areas of the East Greek world, both the mainland cities of Asia Minor and some of the offshoreislands, were for long periods of time within the Persian empire, or in some way dependent upon the Persian King's patronage. Moreover, the states of Phrygia, Lydia, Lycia, Caria, Pamphylia and Cilicia were all in varying degrees hellenized in their arts. The Lycian tomb monuments or the Mausoleum at Halicarnassusare enough to remind us of the activity of Greek artistsin Anatolia in the Persianperiod. In glyptic the main result of this interaction is the series of " Greco-Persian" gems, mainly scaraboids, which will be discussedelsewhere. The Greek Style pyramidal stamps give an even more vivid illustration of the marriage of Greek and eastern arts. The Greek qualities of these seals require no discussion here30 and I simply note what eastern features they display, other, of course, than their shapes. The Hermes on Iz* is an East Greek, with long dress and wing-cap. On I2*" Herakles holds his lion like an easterner too, and seems dressed like one. This might be Cypriot work. The animal fights on 13* and 14 (Fig. 6) are purely Greek in style, but the griffin and stag on 13* are accompanied by a Cypriot inscription; the lion's quarry on i4 is a camel; and on 15 (Fig. 6), also with a Cypriot inscription, the device of two birds tearing the carcase of a calf is one we shall see again often on the
28 For these gems and coins see my ArchaicGreekGems,p. 153. 29 Die Bildzeichen der k6niglichenHoheit bei den iranischenV1lkern (1956). The device on a Lycian coin is noted (pl. 30.3) but not
30

those on the seals. They will be treated again, with the Greco-Persian gems, in chapter VI of my GreekGemsandFinger Rings (1970).

PYRAMIDAL

STAMP

SEALS

IN

THE

PERSIAN

EMPIRE

27

pyramidal stamps and have to discuss further below. The child on 16* is matched on coins from the Troad. The Eros on 17* is shown here because the subject was not recognized in its original publication and it is the only specimen in Greek Style (or near it) with provenience-Sardis. The device is a common one on Late Archaic Greek gems,31and a Greek example found at Sardis32shows how the motif could be transmitted.
2. Orientalizing (3*, I8*-75)

There are a number of motifs which do not seem to be wholly typical of Achaemenid art, as we know it from Persia itself. Although they would all be intelligible enough to an easterner, we shall see that some derive rather from Archaic Greek orientalizing motifs and some from the hellenizing styles of native states in Anatolia. The very simple antithetic groups or single animals also stand in marked contrast to the more complicated assemblageson most earlier eastern seals, and recall rather the Greek manner of decorating vases, coins or gems. We may be sure that this treatment and to some degree the choice of motifs is determined by the fact that these seals were cut in the Anatolian satrapies, especially and perhaps exclusively Lydia. The motifs and rather simple style are basically still Archaic, in Greek terms, and it is likely that these all belong to the later sixth and early fifth centuries. But the differences in style between these and the seals discussedin the next section are sometimesslight. They can perhaps best be described as provincial Achaemenid. The source and degree of this " provincial " influence will become clear as we discuss the principal motifs.
Domestic (8"*)

The conversationon 18* is thoroughly Greek in conception, and hardly to be looked for on a purely eastern work, where the inferior status of the woman would be in some way indicated. The man is dressedas a Mede, with knee-length tunic, leggings and a " busby " headdress. She is not dressedin the usual Persian or Greco-Persian convention for the chiton-like female attire, and the only distinctive feature is the stuff hanging from her arm, which is the Greek himation worn like a stole. The gesture, and the way she holds a flower, seem also more Greek than eastern. But as a group this is the forerunner of the many domestic scenes on later Greco-Persiangems, which may be set in boudoir or bedroom.33
Potniai theronand related (Ig-27)

A winged woman holds two lions-or exceptionally lion griffins (21)-by their tails and head-down, except where the beasts are set along the long axis of the stone and are allowed to rear up (24*). Her dress resembles the Greek chiton with kolpos on Ig but it is usually stylized like the dress for Persian men with close set folds running from a centre line or strip on the skirt. On 24* the dress markings are more Greek in appearance, with splaying folds; more Greek too the bent legs and the goddess' head. The top of the head is plain, with no headdress, but it is rather domed on 25* and 26*, where too the dress is shortened and without folds so that the figures look more like winged Medes than women (compare 18*). This is an odd confusion of types. The lions, with their cross-hatched manes and marked haunches (not ribs) are closer to the normal Achaemenid type of the early period.

14

15

22

34

40

Fig. 6. Impressionsof stamp seals.

31 ArchaicGreekGems,

pp. 72, 99, and finger rings, AntikeKunstX (1967), p. 25 f., N3-5. 32 Sardis XIII.i, no. 93, pls. 9, 11; ArchaicGreekGems,no. 171. 33 E.g. London, Walters, no. 436, pl. 7; Baltimore, Walters Art Gallery 42.517; Southesk Coill.I, pl. I6.Io (A. Furtwaingler, Die

antikenGemmen, 12.11); Cambridge; J. D. Beazley, Lewes pl. und Kameen House Gems (1920), no. 63; G. Lippold, Gemmen (1922), pl. 62.8; Furtwangler, pl. 12.15. On a silver box, Dalton, The Oxus Treasure,p. xxxviii, fig. 19.

28

JOURNAL

OF PERSIAN

STUDIES

The goddess with lions is no stranger to eastern art, but this specific type, walking or running and winged, is more typical of Archaic Greek art, and but for the shape of the stone we might take 24* for Greek. But, of course, many easterners, including Persian heroes, hold lions like this. On 27 we see one crowned, but naked and winged, kneeling to grapple with a lion in a pose which owes something to both east and west. The special relevance of this motif to Sardis is worth noting since it was an Artemis of this type who was worshipped there, to be assimilated to the Persian goddess Anahita when her cult was introduced to Sardis by Artaxerxes II at the end of the fifth century. In the form shown on our seals the winged Artemis appears with two lions, in Greek Style, on a clay revetment from Sardis (Fig. 7), of a series including also rearing antithetic horses (compare our 67) and a Theseus killing the Minotaur in the Archaic Greek pose which exactly matches the Persian hero fighting a lion on our seals. These revetments34 are of the later sixth century, within the Persian period at Sardis.

Fig. 7. Clay revetment from Sardis.

Sirens (28-30) The siren is not a normal Achaemenid creature.35 On 28 its behaviour, playing a lyre, is wholly Greek.36 Closely related are two small conoids on which the siren is " royal ", with bearded, crowned head. The first is sniffing at a flower, like a Greek (29*), and the other, with bow and arrow, is behaving more like a Persian (30). A Greco-Persian scaraboid shows one holding flowers in human arms. These all have plump egg-shaped bodies like sirens on Late Archaic Greek gems from East Greek workshops.37 Lions and winged lions (3*, 3V-42) The lions are of the type we have seen with the winged goddess, and generally rather simpler than the usual Achaemenid creatures, some lacking all body detail except for a bristling mane, others with the cross-hatched mane, marked haunches and knobbly paws of Achaemenid-and many Greekbeasts. There are two examples of a lion alone (31, 32*) and on 3* there is a lioness joined by a lion, its head turned and paw raised to its jaws as though it were chewing a bone. This is a Greek pose for the creature.38 There are other odd features on this interesting seal: the Lydian inscription, two linear devices, and the heavy, Phoenician cross-hatched exergue. It may be a little later than others of this group.
34 For these and the potniamotif see Akerstr6m, op. cit., pp. 88-90, pls. 37, 38.1, 40.1; and G. Radet, Cybbde' (1909) passim. On

the Greek stele from Dorylaion the goddess holds one lion and her " polos " is very like the Persian tiara; E. Akurgal, Die KunstAnatoliens(1961), pp. 241 f., figs. 211. 3* Compare the late Archaic marble siren2Io0, from Kyzikos, now in XVII (1967), pl. 6.3. I have Copenhagen, IstanbulMitteilungen

not seen 28 and it may be that it should be considered with the Greek Style seals. 36 Some male Assyrian sirens, E. Porada, Corpusof Near Eastern Seals I, pl. 92.633, 634. As ArchaicGreekGems,pl. 10.141-4337 38 E.g. Furtwaingler, op. cit., pls. 11.37, 13-44, 45-

PYRAMIDAL

STAMP

SEALS

IN

THE

PERSIAN

EMPIRE

29

Other lions are in pairs, rampant, regardant, with or without a tree or plant between them, and not especially eastern. In Greece such groups are common, usually with a centrepiece, but there is an East Greek electrum coin without,39as on 33*. Finally, there are the winged lions, rampant or sejant (41, 42), a breed rare enough in Greek art but lacking the horns and bird hindquarters of the Achaemenid lion griffin.
Lions attackinganimals (43-59)

The usual type (43-54) shows a lion attacking a bull from the front. Although this motif had a long history in the Near East, at the time when the earliest of these seals were being cut it was most popular in Greece. The scheme is the commonest one for the encounter on Greek gems,40 but on our seals it is rendered far more stiffly because the bull keeps its forelegs straight, and only on 43 (Fig. 8) does it sink to the ground, as in the Greek rendering. These are very stilted and dull versions of an animal fight which Achaemenid sculptors were able to render with much more verve, as at Persepolis. The lions are of the same type as those already discussed, with cross-hatched, bristling and often humped manes. The bulls have weak bodies, without marked dewlaps. By Greek standards none of these need be any later than about 500 B.c. On a fine weight stamp from Sardis the more elaborate attack is shown, with the lion crossing behind the bull's body.41 In other attack scenes the quarry is a goat, once in the style of the commoner fights with bulls (55*), and once in a summary cut style (56). There are also three wilder mel1es with lions and animals, not of the main series (57-59), but related to each other in style. Birdstearing carcase a (6o*-65*) The creature, apparently an antelope, lies stretched out on its back, but its head is raised, except in 62*. On 63-65* there is a branch between the birds making a near-heraldiccomposition and 62* has a three-branched tree here. An earlier version of the scene appears on a bronze bowl from Nimrud42 where two vultures tear a hare, and on a Karatepe relief.43 It reappears on a bronze shield from the Idaean Cave in Crete, and there are sundry other examples of one or two birds with their prey, Phoenician or Syrian.44Its immediate antecedents in the east seem, then, not to be Persian. In Greece such an antithetic composition in the sixth century usually involves lions at a bull or other animal, although two eagles appear seizing live hares on a Caeretan vase.45 But the subject, with a dead fawn, had already appeared on an Athenian vase before the end of the seventh century.46 There is still nothing exactly to match the rather special treatment of the motif which seems confined to seals of this class and shape. A similar victim is seen on 16o* and we have a wholly hellenized version of the whole group
on 15 (Fig. 6).

Otherbeasts(66-75) On 66 (Fig. 8) there is a pair of griffins-wholly Greek monsters in this form, lacking only the forehead knob, but with the long ears. They exactly match, even in pose, the griffins painted on clay revetments from Pazarli in Phrygia47(Fig. Io), which are dated in the second half of the sixth century and could be near contemporaries of the seal. On both the revetments and the seal we are dealing with a provincial, Anatolian Greek version of the monster. Winged horsesrear at a tree on 67. These are creaturesbetter known to Greek than eastern artists48 in this period, although an Archaic Greek would give them a hero-master as centrepiece (compare our
On these groups see P. Jacobsthal, Greek Pins (1956), pp. 77 ff.; 456HS LXXV (1955), P- 4, fig. 2, pl. Ib. Compare later the coin, ibid., fig. 305. For birds as familiars to lions, as on 33, the fine eagles and hares on fifth-century coins of Olympia, see ArchaicGreekGems,p. 125. sometimes in pairs, Kraay-Hirmer, Greek Coins, figs. 173, 40 Ibid., pp. I23 ff., Scheme C. 175-81, 18341 Sardis XIII.I, no. 121, pls. Io, II, the face is concave, the 46 J. D. Beazley, Attic Black-FigureVase-Painters (1956), p. 4, top; S. Karouzou, Aggeia tou Anargyrountos exergue unfinished; ArchaicGreekGems,p. 123, Scheme B. (1965), pl. IF and pp. 42 Cf. R. D. Barnett in Eretz-Israel VIII, pl. 5.I (H. Layard, 92 f. Monuments NinevehII, pl. 62). of 47 E. Akurgal, PhrygischeKunst (1955), pl. 56 and pp. 78f.; 43 P. Matthiae, Studi sui rilievi di Karatepe(1953), pl. 18; for Akerstr6m, op. cit., pl. 87-5. 48 For an Achaemenid seal with winged horses see L. Legrain, prostrate animals cf. Tell HalafIII, pl. I22. a Culture the Babylonians 44 As at Nimrud, Barnett, op. cit., pl. 5; Oxford 1920.20, (1925), pl. 38.848. There are more on of Syrian haematite scarab-pendant (D. Harden, The Phoenicians Assyrian seals, as Delaporte, Bibl. Nat., pl. 22.324; Iraq VI (1963), p. io8a); the seal, Lajard, pl. 56.2. (1939), pl. I .92.
39

30

JOURNAL

OF PERSIAN

STUDIES

43

66

69

70

74

Fig. 8. Impressions stampseals. of

Fig. 9.

113*). The oddly stylized tree here has elements which recall some of the linear devices, as does the tree between goats on 68*. It is rather surprising to find cocks recruited for a similar group on 70 (Fig. 9), but they are so used in Greece, and the fowl become popular, in a Greek form, on the later Greco-Persiangems. The birds on 69 (Fig. 8) look like well-fed versions of those already studied in their attacks on a carcase. In this form, though, they can be found in pairs on Achaemenid cylinders.49

Fig. io. Clay revetment from Pazarli, Phrygia.

The grazing goats, 71*, 72, and mare suckling a foal, 73*, resemble Greek devices. So too the pose of the bull on 74 (Fig. 9), but it is winged, like an Achaemenid, and this is not at all a common feature in Greek art.50 The hippocamp, if such it is, on 75, seems also Greek in form. Court 3. Achaemenid Style Here the motifs are all Persian, or had been adopted into the repertory of Persian court art. The subjects are almost all supernaturaland there are many more groups or figures which appear again on cylinders and conoids, which were the more usual homeland seals. On the conoids too the style is often very close to that of these pyramidal stamps, and in any comprehensivestudy of Achaemenid glyptic it would be necessary to consider many of the conoids with some of these pyramidal stamps. I have, indeed, allowed some with oval faces into the lists, and this is the point at which a distinction by shape alone becomes rather vague. The segregation of the pyramidal stamps in this article is justified rather by the content and style of the many other devices which they bear. This Court Style is that common to many Achaemenid arts and on seals it owes most to earlier Assyrian fashions. There is normally no real body modelling and anatomy is reduced to patterns of manes, ribs and haunch muscles on the barrel-likebodies, with full use of the drill for knobbly paws, joints and snarling muzzles. Human features are reduced to plump cheeks and rolled hair with beady
49Porada, Corpusof Near EasternSeals I, pl. 122.818; Delaporte, compare late Archaic Greek and Etruscan finger ring devices, Louvre pl. g9.A 792; Legrain, op. cit., pl. 36.809-I I; Berytus BSR XXXIV (1966), p. 4; AntikeKunst X (1967), p. 8. II, VIII (1944), pl. IId. Cf. the gold plaque, 0. M. Dalton, The 10 See Archaic Greek Gems, p. 146; and the Greco-Persian, OxusTreasure (1926), pl. 13.47. For the birds with a tree between Furtwingler, III, p. 124, fig. 86.

PYRAMIDAL

STAMP

SEALS

IN

THE

PERSIAN

EMPIRE

31

on Achaemenid coins. Most of the pyramidal seals resemble closely the style of the dated sealings earlier than the mid fifth century, but the few with more advanced features may be from Persepolis,51 later. Some problems of identification and dress52 best considered first. The male engaged with lions are or monstersused regularlyto be called the King, but now more cautiously a " Persianhero " or " royal hero ". He may wear a long dress of various forms. The normal one has loose sleeves and the skirt is divided vertically by a bunch of folds from which other folds splay to either side. On the seals the centre strip may be no more than a simple patterned band or a line and the side folds are often straight and sometimes omitted. Fuller representationsin relief sculpture show that the bunch of folds is formed by pulling the stuff up through the belt, thus raising the hem and gathering the folds.53 This is exactly the manner of wearing a Greek chiton, and it has the same effect. The way it was representedin Archaic Greek art was copied in Achaemenid and the Greeks would not have hesitated to identify the dress as a chiton, even if the structure of the sleeves or upper part was different. This does not mean that the Persians adopted the Greek chiton and it is clear that the dress had a long history in the east.54 The Greeks certainly adopted their chiton from an eastern dress,just as they used an eastern word for it,"6 but long before the Achaemenid period. Nor did the Greeks teach the Persianshow to wear the dress, they simply showed them how to represent it effectively. The bunched folds would be at the front of the dress, where they are normally shown in Greek art, but in Achaemenid the dress pattern is often moved round to fit a profile view, although occasionally the centre strip is moved towardsthe front with straight vertical folds before it. It is the simple front-on-sideview that we see on most of the seals. It is in fact so unreal that it is not surprisingthat the intention of this vertical strip is sometimes forgotten and it is represented illogically, or without showing the necessary rise in hemline which is involved. This may have happened more easily in studios far removed from the Persian capitals.56 A good and explicit example of the dress pulled high through the belt and the sleeve folded back for action appears on the Persepolisreliefsshowing a fight with a lion griffin.57In this the forwardleg is bared to above the knee. The same position for the dress is shown on our seals by short vertical folds covering only the forward thigh (as on 84* and I07*). This rendering may reasonably be taken as more advanced than the simple, symmetrical pattern, but it can be easily mistaken, especially on small or summary seals, for quite a different attire which is anything but " advanced " and probably belongs to the very beginning of the Achaemenid series. This is the old Assyrian and Babylonian dress, split to near the waist and patterned with horizontal or oblique bands, with a short undergarment. ProfessorPorada very plausibly coins the term " proto-Achaemenian" for cylinder seals showing figures with this dress.58 There are a few examples of it on our pyramidal stamps (on 98* and 113*) and all these variants or odd treatments of Persian dress may prove to have some bearing on our understanding of the stones, their date and origin. The hair of the figures with this costume is usually rolled at the neck and may be dressed with only a slim fillet in the Assyrian manner, also seen often in Achaemenid art. Sometimes the fillet is a little broader. The usual headdress is tall and vertically fluted-the " tiara ". This is worn too by the " royal sphinxes " and sometimes identified as the kidaris. A crenellated crown is worn by Darius on
52 Useful remarks on dress by A. S. F. Gow in JHS XLVIII (1928), pp. 142 ff.; B. Goldman in IranicaAntiquaIV (1964),

later seals are the heads and drapery treated more realistically, and the same progression can be observed

eyes and button noses; the dress to a linear pattern of folds. Only on some of what are probably the

or the extreme version on a Persian-Phoenician scaraboid, not without Greek traits, where the whole pattern is shown on one side of a dress split to the waist, E. Babelon, Coll. Pauvertde la Chapelle(1898), pl. 4.36. pp. I33 ff.; and G. Thompson in Iran III (1965), pp. I2I ff. 53E.g. the Persepolis reliefs, E. Porada, The Art of AncientIran 5 Porada, op. cit., pl. 44 left; and cf. Darius at Behistun for the skirt, H. Frankfort, Art and Architecture the Ancient Orient of (1965), pl. 44, left, and the gold plaque, Dalton, op. cit., (1954), pl. I78a. For the " overfall ", Dalton, op. cit., pl. pl. 15.92. 15.92. 54 For this see especially Goldman, loc. cit. and TheirEasternNeighbours(1957), "8Porada, Corpusof Near Eastern Seals I, p. io2; but no. 812, 55T. J. Dunbabin, The Greeks sur emprunts sdmitiques pl. 122, is a rather different case, with London WA 89oo9 p. 58; E. Masson, Recherches lesplus anciens Seals of Western en grec (1967), pp. 27 if. Asia, no. 107); and cf. (D. J. Wiseman, Cylinder London WA 10o3o013 (ibid., no. 109) and L. Speleers, Catalogue 56E.g. on the cylinder, SoutheskColl. II, pl. 8.Qc.34 (London des Intailles, BruxellesII, no. 1474WA 129571), where the vertical is shown as two separate hems; II, 51As E. Schmidt, Persepolis pls. 3, 4.

32

JOURNAL

OF PERSIAN

STUDIES

the Behistun relief and seen occasionally elsewhere.59 The zigzag crown worn by the heroes and sphinxes on some of our seals might be intended for this, but it is a feature we have to return to briefly in a moment. Occasionally, but more often on our stamps than elsewhere in Achaemenid art, the figure with lions or monsters wears Median dress with the " busby ",60 short tunic and leggings.61 The Achaemenid king is never shown thus, so we should not identify the figure on the seals as the King. Nor need the zigzag approximation to the crenellated crown worn with Persian dress indicate the King, although it is also worn by the " Darius " on his famous lion-hunt cylinder in London.62 It may also be a headdress for Ahuramazda, or sphinxes (our 5*), but it seems very doubtful whether the Persian King was ever regarded in life as a divinity.63 " Royal hero " may be a fair description, but what of the Mede? I call the figure a Persian hero or a Median hero, according to dress, including as Persian those with the " Assyrian " dress. On our stamp seals it could be that the artists had no very clear idea about whom they were representing, a confusion evident in their treatment of other motifs. So we are perhaps here absolved from attempting to identify properly the " hero's " counterpart in the Persian homeland. The other creature requiring discussion is the lion griffin.64 This has a respectable Near Eastern ancestry. The type is well established in Middle Assyrian art-a lion's head, generally with long ears and occasionally with bulls' horns, a lion's body and forelegs, wings, hind legs as a bird's and a short feathered tail like an ostrich's. There is virtually no change into neo-Assyrian and neo-Babylonian, and although versions of the creature are seen in Persia, as on the Ltiristan bronzes, it seems to be the Mesopotamian monster which has most to do with the renderings in Achaemenid art. The body remains leonine and this sometimes induces the replacement of the bird's tail and legs by a lion's tail and legs. Oddest of all is the family group on the Boston seal, 115*. Here a lioness griffin is invented by providing a line of dugs along the belly. This is, of course, quite the wrong place for them on a real lioness, and we would not expect a Persian or Mesopotamian artist to make such a mistake. It is seen again, for a lioness, on our 3*, and on an Achaemenid cylinder whose origin in the western empire is attested by its style, its misunderstanding of dress, and the inclusion of a Bes shouldering a horned lion.65 The same mistake with the dugs was regularly made by Greeks,66who had never seen a lioness, in the Bronze Age and in the sixth century, when they tended to make it behave and look like a dog. There may have been lions in Lydia, where the head was used as a royal blazon on the earliest coins. In some instances, especially where tiny horns are worn, it seems possible to identify the monster in Assyria as a Pazuzu or Lamashtu,67 but the identification need not be general, and there is no reason to think that in Achaemenid art the monster had any such specific titles. Indeed it seems probable that it was copied and adopted as a formal, perhaps meaningless monster, suitable for the attentions of a Persian or Median hero. One indication of this is the variety of mutations on the basic monster which are admitted in Achaemenid art, especially on our seals which were produced at some distance from the homeland of both the Persians and the lion griffin. Thus, we may find the bulls' horns replaced by short twisted goats' horns, or by long ibex horns, and the ears are not always emphasized. Occasionally even the eagle head of the true griffin type is admitted, and this would be more familiar to Greeks. The Persian bestiary requires more careful study. A herofights a lion (9-Fig. 2, 76-91) The usual scheme shows the creature rearing onto its hind legs while the hero grasps its neck, or rarely its forelock, and threatens or stabs it with a short sword. The hero wears Median dress nearly as often as he does Persian on our seals. The lion normally claws the hero's thigh with a hind leg, but on
59Cf. Porada, The Art of AncientIran, p. 159, fig. 85 and pl. 45. 60 The soft headdress wrapped round the chin which appears regularly on the later Greco-Persian gems is not seen on the stamps. " Persian " head61 The dress may even be combined with the as on Iro*. dress, 62 London WA 89132; Porada, op. cit. p. 176, fig. 89; Wiseman, op. cit., no. ioo. And Boston 21.1193, with the hunt of a lion griffin, Beazley, Lewes House Gems,pl. 1.8. 63 Cf. Porada, op. cit., p. 158 f. On the Greek evidence see W. Tarn in JHS XLVIII (1928), pp. 20o6ff. answering Taylor; and R. N. Frye, The Heritageof Persia (1963), PP. 95-9764

Its development is discussed by B. Goldman in AJA LXIV (1960), pp. 319 ff., but not all varieties are noted and it is common to find in other works descriptions which do not reveal close observation of details of legs, head and tail. Cf. also H. Kantor in JNES XVI (1957), pp. 11-13, pl. io; T. Mahdloum in SumerXX (1964), pp. 57 ff. 65 London WA 129571; Southesk Coll. II, pl. 8.Qc.34; Arch. Anz. (1928), p. 661, fig. 17a. 66 Cf. W. L. Brown, The EtruscanLion (1960), pp. I66 f. 67 For these see now P. R. S. Moorey in Iraq XXVII (1965), PP. 33 ff.

PYRAMIDAL

STAMP

SEALS

IN

THE

PERSIAN

EMPIRE

33

82* and 83 (Fig. I I) it simply stands and throws back one paw in the expansively belligerent attitude familiar from Assyrian works, but rather less common in Achaemenid.68 83 is a very dry, Court Style piece, and it may be significant that it is the only one with a Persian provenience (" from Persepolis "). The figures of a few of the lions and some of the Median heroes are hardly more detailed than those which were discussed in the last section, but most of the lions are heavily formalized in the Achaemenid Court Style of the Archaic period, with bristling cross-hatched manes and heavily marked haunches. On go a goat is introduced below the pair, probably on the analogy of the animals often shown " supporting " the hero with two beasts on cylinders. And on gI the hero kneels with a bow drawn against a gesticulating lion, but this might be taken as a mortal lion-hunt scene, in which the bow is commonly used. On cylinder seals the hero with a single beast or monster is far less common than the broader antithetic group involving two, often with a pair of creatures beneath.

77

83

85 Fig. Iz. Impressionsof stamp seals.

101

103

A herofights othersingle creatures(92-105) The usual quarry is a lion griffin, treated just like the lions. The hero is always in Persian dress, but on 97* he lacks a " tiara ", and on the fine example, 98*, he wears the older Assyrian type of costume which is most carefully detailed. On 99* there is a supporting lion underneath, and above, the winged sun disc, rarely admitted on these seals. The fight is that rendered on the magnificent Persepolis reliefs,69 where the hero more reasonably grasps the monster by its horns. The only other monsters grappled singly are goat-sphinxes (see below), once with a Mede, once with a Persian, and one winged I I, to lo5) there is a horse, a goat and a bull. These are goat.70 And for ordinary creatures (to3-Fig. uncommon encounters for the hero.

A herowith two creatures (7,

The Persian hero is shown on some of our seals holding two lions upside down by their tails. On Ilo* they are replaced by winged goats and on 7 (Fig. 2) one lion remains upright. The scheme is an extremely common eastern one and often seen on Achaemenid cylinders. It had been used for the winged potnia theronin the Greek Style (see above), and IIi, which shows the Persian hero with wings, may be influenced by these devices.7 It is a similar scheme which is suggested by the fragmentary impression from Egypt (112), where the rearing animals held by their tails appear to be sphinxes. Where the creatures rearing on either side are winged horses held by their harness (113*) the hero is behaving like a Greek Herakles with the horses of Diomedes.72 The Persian winged hero with two lion griffins is also rendered in Classical Greek sculpture of the fourth century.73 The fight with a single animal is recalled on some seals where a second animal is admitted. Thus, there are two upright lions held by their throats on I14 (Fig. 12), where there is the winged sun disc
68

io6--I5*)

On the realism of these man-lion duels see Frankfort, op. cit., pp. 99 f. The hero dispenses with the cloth wrapped around his left arm, which is carried by Persians attacking boars on many Achaemenid and Greco-Persian gems. 69 Porada, op. cit., pl. 44 left, but with a scorpion tail. 70 Also fought on Delaporte, Louvre II, pl. 9o.A 782; and a pair with the hero on Legrain, op. cit., pl. 33.662; J. Menant, Cat.

des cylindres la orientaux, Haye (1878), pl. 7-33; Speleers, op. cit., p. 219, no. 555. 71 The Persian hero with wings fights a lion on our 27, and a crude version of one with four wings and two lions is seen on an agate scarab in Paris, Lajard, pl. 68.21. GreekGems,p. 72 Cf. Archaic 49. Ldwen(I944), PP. 31 f. 73 Cf. C. Simonett, Die gefliigelte

34

JOURNAL

OF PERSIAN

STUDIES

and supporting bull (?) like the lion on 99*. The finest, which also has the sun disc, is 115*. The hero fights a lion griffinin the usual manner but behind him is the monster'smate, providing a family group which, for monsters, is perhaps more Greek than eastern. In Greece even griffins have a private life and may suckle their young.74 But the style of our seal is pure Achaemenid Court, with finely patterned dress, wings and limbs.
Royal sphinxes (5*, zz6*-128)

These fine creaturessit and salute each other on several seals, sometimes over a plant and beneath the winged sun disc (5*, II6*-121). A closely similar composition is seen on gold plaques from Sardis.7s They may also be paired with a lion griffinor winged lion, chest to chest or back to back, facing or with royal sphinx is behaving like the King in the chase, and accordingly has human arms and is constructed with a human torso like a lion-centaur.76 He is drawing his bow at a lion, over the body of a stricken antelope. 127 (Fig. 12) shows a bicorporate royal sphinx, a type of monster more common in Greek and than eastern work of this time.77 It reappearsin Greek form on a later Greco-Persianscaraboid,78 in this class there are other studies of the royal sphinx.79 On 128 the single sphinx lacks a crown and has his wing and body rendered with unusually sensitive modelling, but overhead the Achaemenid winged disc hovers. It is interesting to note that royal winged bulls, which are seen on scaraboidsand cylinders, are not found on the pyramidal stamps, although Persian winged bulls, with or without human features, seem to have made their mark on East Greek glyptic in the Archaic period.80
and others (I29"*-139*, 198*) Goat-sphinxes heads turned away (I22*-124). Once one walks towards an ordinary sphinx (125*). On 126* the

mane. They are normally shown walking along (I29*-132*), once seated (133*), twice seated in pairs and once with their foreparts joined back to back like an Achaemenid figure capital (136*). (134, 135*) One communes with a winged lion (138*). The monster is comparatively rare on other Achaemenid works8xalthough goat-demons have a long history in Persia.82 Finally, there is one walking sphinx with an " ordinary " human head but a bird's tail (139*). Lions and lion griffins (4*, 8*, I4o*-163)

These are like the royal sphinxes except that they have no " tiara ", but a goat's horn, ear and short

Representations of the lion alone with the characteristic Achaemenid Court treatment of shoulder and ribs are remarkablyrare on the seals. A few of the lions fighting heroes are detailed in this manner, and there is one individual study (z4o*) showing also the feather-like marking behind the shoulder which appears on felines in various forms (as on the goat-sphinx on 133*). It seems to represent no single anatomical feature and may even be shown like a small wing.83 Sometimes it is the tuft of hair behind the leg, or a continuation of the chest hair, but elsewhere it seems to suggest belly hair or even the whole bare rib panel. Here too we see that distinctive treatment of the rear haunches with three arcs ending in blobs, which in a less rigidly stylized form was to be seen on Assyrianseals, was remarked
'' E. Kunze, Meisterwerkeder Kunst, Olympia (1948), fig. 34;

Boardman, GreekArt (1964), p. 61, fig. 50. And at the end of the Bronze Age they had been shown with a nest: M. R. at Popham and L. H. Sackett, Excavations Lefkandi (1964-66), cover and p. 18, fig. 35. '" SardisXIII. , pl. t.1. 76 Cf. the cylinder London WA 129565, SoutheskColl. II, pl. 8.Qc.28, where the body is a bull's, the wing has a bird's head at the tip, and the quarry is a goat; the cylinder, W. Nagel, Altorientalische Kunsthandwerk (1963), pl. 68.40, with a spear against a lion griffin; the Babylonian cylinder, Porada, Corpusof Near Eastern Seals I, pl. 113.749, where the body is a horse's and the quarry is a winged lion; the Assyrian cylinders, Frankfort, CylinderSeals, pl. 36a, wingless with a sword against a lion, and pl. 34d, against a bull. Cf. in Greek, ArchaicGreekGems,pl. 2.35. Gems,p. 68 with n. 6. 17 Cf. ArchaicGreek Drouot 78 Pierresgraveesantiques: coll. d'un archiologue-explorateur, (May 8th 1905), pl. 2.21.

78Once Arndt A 1411; Furtwangler, pl. 11.20 (Berlin F.187); cimmerien S. Reinach, Antiquitis du Bosphore (I892), pl. I6.I4; Lajard, pl. 46.1380 ArchaicGreekGems,p. 146; and on a Greco-Persian scaraboid in the Court Style, Oxford x965-362, Ashmolean Museum Exhibition of Antiquities: Spencer-Churchill (x965), pl. 17.35; Pope, Survey,pl. I24G. 81 Cylinders: Legrain, op. cit., pl. 33.661; Delaporte, Bibl. Nat., pl. 33-502; Drouot (April 20othI964), pl. 5.8I (Th6ry Coll.); (May 2oth 1959), pl. 1o.298; MiinzenundMedaillen,Sonderliste K, no. 83; and cf. B. Buchanan, Catalogueof Ancient Near Eastern Seals: Ashmolean Museum I, pl. 44.674. Conoid: Leningrad, Recueil III, pl. 3.12. Greco-Persian scaraboids: Furtw~ngler, III, p. 124, fig. 87 and pl. Ii.i8. 82 See Porada in Dark Ages and Nomads (ed. M. Mellink, 1964); The Art of AncientIran, p. 66, fig. 39 and pl. 15; or the Lfiristan situlae, Berliner Jb. V (1965), p. 33, H I, 2. 83 E.g. on the gold rhyton, Porada, op. cit., pl. 47 left; and cf. E. Akurgal in AntikeKunstX (1967), p. 33 and pl. 8.9.

PYRAMIDAL

STAMP

SEALS

IN

THE

PERSIAN

EMPIRE

35

by Greek artists in the Persian period, and is a hallmark of the Court Style creatures, where it is not confined to felines. Lion griffins are far more common, and sometimes equally elaborately detailed in the Court Style. An excellent example is the Lydian-inscribed 4*, where the creature is standing or walking. It appears like this on several other seals (I42-151*) although only twice again with one paw raised. 151* is notable for the odd linear stylization of the haunches, the tail which is a snake and recalls the Greek chimaera,84 and the crescent, star and triangle tree (or thymiaterion ?) which may together indicate a Syrian or Phoenician provenience. The raised paw appears again for seated lion griffins (I52*"-54*) and there is a single study (155) of the beast, with wings displayed, couchant. In pairs they may be seated, facing or back to back (156, 157). Where they rear up at each other heraldically there is a treeplant between them on 158 (Fig. 12), and on 16o* they stand over the prostrate body of a deer, as though it was a victim.85 This hint at their hunting behaviour seems to be further suggested by the lion griffin holding the forepart of a boar in its forelegs (161, Fig. 12). A composition like this may have

114

127 Fig.
12.

158 Impressionsof stamp seals.

161

inspired the mixed monster on 8* where to the lion griffin's chest are added the heads and necks of a boar and of a goat-sphinx. On 162* it is a winged lion and a real griffin of Greek type, with bird head and long ears, which rear over a crouching calf.8" Both creatures, however, have borrowed the bird tail from the lion griffin, and there is only one study of the normal griffin, with lion body (163). Canonical lion griffins are also seen on several Greco-Persian scaraboids in a comparable style, but better modelled and with variants in construction which make many of them more like ordinary winged lions or griffins. Bes-sphinxes(I64*"-68) This is a beast known in Egyptian art, and borrowed in Phoenicia, whence it may have had some effect on Greek glyptic in the sixth century.87 A simple, rather rough version of the monster is seen on 164*. On the more meticulous 165* a lion head, seen from above, is attached to its chest. It is likely that this form was suggested by the sphinxes with lion heads at their chests, which seem at home in North Syria88 rather earlier. On 166* the Bes head is highly stylized but still unmistakable and the creature is paired with a lion griffin. An even more extreme stylization appears for a Bes head and body (?) which springs from the back of a sphinx with a goat's head on other seals (167*, 168), including one from Cyprus. The Bes element has been taken for a representation of the Ephesian Artemis! These belong here for their style although there is some uncertainty about their shape. Although Bes had been most at home in the Near East in the Syro-Phoenician area this does not mean that these seals were made there rather than in Anatolia. Bes and Bes-sphinxes are accepted on a
84

Easterners normally stylize tail-tips as birds' heads. The snake tail at this date indicates Greek influence.
88

85 So too the griffins on the Pazarli revetments, our Fig. Io. 86 Cf. the monster on the cylinder London WA 89788 (Wiseman, op. cit., no. 113) where the forehead knob is also shown. 87 Cf. ArchaicGreekGems, p. 29. The standing Bes-sphinx with

of dugs on a " limestone cone " shown in Proceedings the Society XXIV (1902), p. 25, fig. 8, is perhaps to of Biblical Archaeology be associated with our stamp seals. As the Carchemish relief, Frankfort, Art and Architecture the of AncientOrient,pl. 161; Tell Halaf III, pl. 88; Gordion ivory, AJA LXVI (1962), pl. 47; North Syrian seals, Iraq VI (1939), pl. 7.56; and Arch. Anz. 1928, pp. 522 f., figs. I, 2.

36

JOURNAL

OF PERSIAN

STUDIES

number of Achaemenid cylinders and conoids,89 so their appearance in any part of the Persian empire need not occasion surprise, and the way the Bes-sphinxes are treated on our seals does seem highly individual. Othersubjects(6*, i69*-I7I) The horse on 169* is purely Achaemenid, exactly matched for build and detail of its trappings on the Persepolis reliefs.90 It is perhaps odd to find it without groom or rider, and the accompanying dog recalls common Greek groups. The hatched border too is regular on Greek Archaic scarabs but met only here on a pyramidal stamp. It seems that Greek practice determined some aspects of this device. The bull on 6* is a magnificent study and again purely Achaemenid, although Barnett rightly draws attention to possibly Greek details of modelling, especially the treatment of the beast's flanks. A comparable study on this scale is the winged goat on a cylinder in London,91 but while this carries all the usual Achaemenid stylizations for parts of an animal body, it lacks any subtlety of relief modelling. The dismounted hunter spearing a boar, his horse waiting behind him (17o, Fig. 13), is an unusually crowded, frieze-like composition for a field of this shape and size. It is more naturally suited to a cylinder, where this whole scene appears,92 although the horse is generally omitted, as it is always on the Greco-Persian scaraboid versions. This is, then, a transference of a whole cylinder composition onto a stamp, while elsewhere we see only extracts or abbreviations. Finally, the royal guard on 171 (Fig. 13) closely resembles monumental studies of these figures, as in glazed brick at Susa.93

170

171 Fig. 13. Impressionsof stamp seals.

186

details Other
The subsidiary ornament on the Orientalizing and Court Style stamps deserves some attention. Only on 169* do we find the hatched border to the intaglio which is usual on contemporary Phoenician and Archaic Greek scarabs. Where ground lines are provided for the figures they are normally a single or double line. The usual elaboration to the ground line is to make it a strip of hatching with equally spaced vertical strokes, but on 122* and 170 (Fig. 13) the strokes are grouped in twos and threes. On 74 (Fig. 9) the ground line is a row of dotted squares, which looks rather a Greek pattern but is seen on one Achaemenid cylinder.94 3* has a full cross-hatched exergue of Phoenician type. The strangest pattern is the alternate hatching in the exergue of 6*. On engraved gems this is best known on Etruscan scarabs of the fifth century but its use can be traced in the Greek world too,95 and it is probably simply a variant on the cross-hatched or zigzag exergues. Here, however, it may be borrowed from Greek practice, and the fine bull above it owes something to Greek treatment of animal
*9Achaemenid cylinders: Legrain, op. cit., pl. 43.925 (Bes holds two Bes-sphinxes); London WA 89352, Wiseman, no. o6 (Bes holds two goats); London WA 89133, Wiseman, no. 103 (Bes beneath sun disc); A. Moortgat, Vorderasiatische Rollsiegel, pl. 89-758 (Persian hero holds two Bes-sphinxes), 764 (Besholds two lion griffins); Delaporte, Bibl. Nat., pl. 33.502 (Besholds two goatsphinxes); Buchanan, op. cit., pl. 44.675 (Bes frontal); London WA 12957I, Southesk Coll.II, pl.8.Qc.34 (shoulders horned lion). Achaemenid conoids, holding two lions: London WA 91895; Bibl. Nat. Io6o, Lajard, pl. 34.14; Geneva 20563. With dog and cock on the impression, Persepolis pl. 13, seal 64. Cf. the gold II, pendant, Coche de la Fert6, Les bijouxantiques(1956), pl. 5-4. Bes MuseumBulletin,forthcoming. on Phoenician seals, see Budapest der 90And cf. the cylinder, Furtwangler, Beschreibung geschnittenen Steine, Berlin, pl. 4.I8o and its Greek Style counterpart, E. and Greeks(1913), p. I93, fig. 85Minns, Scythians 91 London WA 128865, Wiseman, no. i I6. and the Greco92 As London WA 89144, Wiseman, no. iixi, Persian cylinder, Boston o3.10o 1. 93 Porada, op. cit., pl. 42; and cf. the cylinder, Moortgat, op. cit., pl. 90.768. 94 London WA, M. Menant, La glyptique orientaleII (1886), pl. 9-395 ArchaicGreekGems,p. 16.

PYRAMIDAL

STAMP

SEALS

IN

THE

PERSIAN

EMPIRE

37

motifs. On some Court Style stamps an animal, usually a lion, occupies an exergual position (go, 99*, I 4-Fig. 12, 166*); we may except those which represent an animal brought down by others, or those which suggest such a scheme (as 4o-Fig. 6, i6o*, 162*). There are occasional winged sun-discs of the usual Achaemenid type (99*, II4-Fig. 12, 116*, 117*, 128, 157). The florals are usually simple spiky trees (34-Fig. 6, to 38*, 41, 62*-65*, 7o-Fig. 9, 145) but we find stranger stylizations (67, 68*, 116*, 117*) or berry-like blossoms (39*, I58-Fig. 12) recalling the Assyrian interlaces with cones. The lotus patterns (4*, 130*, 133*), similar to subsidiary devices on Macedonian coins, have already been noticed.

Other styles
A number of the seals listed here bear Aramaic inscriptions and we have observed already the possibility that some of these could be earlier than the Persian period. One indeed has a Babylonian device on its side (181). The scheme, subsidiary decoration and style of the bulls on 172* and 173 are very close to those of earlier Aramaic-inscribed seals from Syria and Palestine, but it should be noted that the bull is of the Assyrian or Persian type rather than the Phoenician-Egyptian. The sphinx and griffin on 174 and 175* can probably be placed in Syria; the latter especially for its straight wing, but it seems to have some traits of the lion griffin with pointed ears and bird tail. The grazing goat on 176* is simpler and more like Archaic Greek, while the lion on 177* is executed as on earlier Syrian work, with stick legs. The zigzag exergue beneath it is a Phoenician feature, and especially Cypro-Phoenician.96 Finally, the chariot on 178 is in a wild provincial style. All these devices have in the field Syrian or Mesopotamian symbols like stars, crescents, diamonds. The seals with bulls were cut in cornelian and agate; all the rest in a greyish limestone.97 Chalcedony was preferred for the pyramidal stamps with only Aramaic inscriptions (i79-i84), and this may be a further suggestion of their earlier date. CONCLUSION It remains to summarize what can be deduced about the pyramidal seals from this study of their decoration and devices, and to relate them to other styles and shapes of seals in the Persian empire. The Lydian character of the inscribed seals was obvious from both their find places and their inscriptions. There are ten other examples from Sardis without linear devices, as well as two stones and a ring of Greek Style from Sardis with the linear devices. The Lydian capital has yielded in all twenty-one stones relevant to our enquiry, to which we can reasonably add six others with Lydian inscriptions but no recorded provenience. This accounts for more than half the examples for which a find place is known or alleged. Their further distribution is intelligible in terms of their probable Lydian origin. To the west-the East Greek sites, Ephesus (one), Smyrna (one); the Greek IslandsTenos (one), Lesbos (one); in Greece itself-Pharsalos (one), a Greek stone with linear device from Thessaly and perhaps Athens (one). To the north five reached the Black Sea cities. There are three and an impression from Egypt. Nearer their home in Anatolia there is one each from Kara Hisar, Adana, Tarsus and Cyprus, and we may recall the Cypriot inscriptions on two in Greek Style. Two are " from the coast of Syria " and only one is reported from Persia. There is no significant difference in the distribution of the Orientalizing and the Court Style seals, but the few with Syrian or southerly features must clearly be regarded as a local extension of this mainly Lydian phenomenon, fostered perhaps by earlier or continued usage of this Babylonian shape in the area. No excavated contexts offer very valuable or close dating. 5* was in a grave at Kerch with an Achaemenid cylinder of comparably advanced Court Style and a Greek gold ring of the mid-fifth century, but other jewellery in the grave was apparently no earlier than the end of the century.98 The Bliznitza grave with 86* belongs to a complex of burials of the second half of the fourth century, but it is clear that in the South Russian graves the seals and finger rings were often deposited long after
96Ibid., p. 30 and n. 22.
98

97 Possibly to be associated is a haematite conoid, London

WA 120233, with a rather roughly cut zebu, crescent and star.

Co-finds in Reinach, op. cit., pl. I6.5, 6 (cylinder); I8.9 (gold pl. ring; also Furtwlingler, Die antikenGemmen, Io.2o); text to pl. 8.3 on the jewellery, which matches that from Solocha (B. Goldschmiedekunst (1967), pl. 47c). Segall, Zur griechische

38

JOURNAL

OF PERSIAN

STUDIES

their date of manufacture. The seal from Lesbos, 134, however, was found in a painted clay sarcophagus whose decoration is hardly later than the mid-fifth century.99 The Sardis graves might prove more informative but the accompanying pottery has never been published, and the jewellery which has been published does not allow of much precision in dating. We are forced to rely rather on stylistic considerations. For the Greek Style seals this is comparatively easy, and the examples seem to span the fifth century, the earliest being clearly of about 500 B.c. Such Greek elements as there are in the Orientalizing seals are Archaic of the later sixth and early fifth centuries. The majority of the other seals are in what we may judge to be an unemphatic Archaic Court Style matched by Persepolis sealings of the first half of the fifth century, but there are some with more developed Court Style forms which are very probably later. On the available evidence it could be argued that none were cut after the fifth century and it is very likely that the earliest could belong to the early years of Persian administration in Lydia. Only one probable sealing from a pyramidal stamp is known, from Memphis in Egypt. The only other indications of their use are provided by the inscriptions and the linear devices which are found on about one-third of them. These demonstrate personal ownership and use, identified by the inscription or device. The linear devices are seen to have affinity both with earlier Greek commercial practice and with a Persian usage not well attested for this period in the Persian homeland. The similarity to countermarks on Achaemenid coins suggests official use, but all the names in the Lydian inscriptions are Anatolian, with the exception of the Persian name Mitrata', who need not have been Persianborn.100 However, this may mean simply that Persian owners were satisfied with the linear device alone, or that the official seals of higher Persian officials in Anatolia were cylinders which, if inscribed at all, would carry their names in Aramaic.o01 The non-Anatolian, Persian content of the seals is at any rate clearly demonstrated both by their shape, which derives from earlier Assyrian and Babylonian stamps not current in Anatolia, and by their engraved subjects, nearly all of which can be matched on Achaemenid works. The subjects betray their western origin occasionally in the Orientalizing seals and in rare details of the Greek Style seals. These foreign elements can be explained partly in terms of hellenized Anatolian art, but mainly in terms of the art of the East Greek states on the borders of and sometimes within the Persian empire, and the few pyramidal stamps with wholly Greek subjects in Greek Style demonstrate the link quite clearly. For comparisons with other Achaemenid seals only the Court Style subjects on the pyramidal stamps are relevant. A closely similar style is seen on larger conoids and cylinders, especially the former, some of which we have seen to carry the linear devices while only one cylinder was given them. It is likely that the conoids were also current in the western empire but most are cut in the more advanced Court Style and may be later than the main series of the pyramidal stamps. A comprehensive study of Achaemenid cylinders would demonstrate clearly the relationship of groups of them to the various groups of stamp seals, but this has yet to be undertaken-or at least published. One other shape deserves attention since its record seems closely matched by the pyramidal seals. It is the weight stamp (see above, p. 20).102 Its currency in Anatolia is shown by four examples from Sardis and three which bear the linear devices. The figure subjects resemble those on the Archaic Orientalizing pyramidal stamps, while 189 looks mainly Greek, but its hatched border and the crosshatched exergues look rather to Phoenicia and Syria, where the shape must also have been current and where it had probably been first applied to seal use.
99Praktika tes archaiologikes Etaireias (1926), pp. 151 f.; R. M. British MuseumVIII, p. 45 Cook, CorpusVasorum Antiquorum, with n. 5; AnatoliaX (1966), p. 187, no. 16. 100Thus Zgusta in Charisteria Orientalia,J. Rypka, pp. 397 ff. A Mitratai is named as a priest (?) in a Lydian inscription. 101As the cylinder of the Lydian official Artimas published by D. Bivar in NumismaticChronicle 1961, pp. 19 ff. 102 For examples of the shape other than our i88*-9io* see: SardisXIII. x, no. I21, pls. Io, 11 (agate; lion attacks bull, sun, crescent and vase in field; see above, note 41), and no. 122 chalcedony; suckling goat and bird; summary); Oxford 1892.1408, from Sardis (mottled red and white marble; chariot); Jerusalem, from Sardis, The Bible in Archaeology (Israel Museum I965), pl. 3k, no. 113 (chalcedony; Greek griffin); Munich A 1393, Munich I, pl. 26.239 (chalcedony; griffin on a hatched ground line); Athens, Karapanos Coll., Journ. Int. d'arch. num. XV (1913), pl. 6.460 (chalcedony; Syrian sphinx); Coll. de ClercqII, pl. 3-55, from Amrit (agate; hero fights lion); Leningrad 403 (chalcedony; two lions); Delaporte, Louvre II, pl. 92.A733 (cornelian; Babylonian device and Aramaic inscription).

PYRAMIDAL

STAMP

SEALS

IN

THE

PERSIAN

EMPIRE

39

The relationship with the so-called Greco-Persian seals is more difficult to assess, but possibly more rewarding. These stones, which will be considered in greater detail elsewhere, are mainly scaraboids of the larger variety developed in East Greek workshops in the first half of the fifth century. They seem to have been produced in the southern states of Anatolia, not Lydia. A few scaraboids carry Archaic Court Style subjects like those on our seals, which might be taken for another indication of date since the sequence of scaraboids is easier to establish. More scaraboids have the advanced Court Style subjects (one with a linear device is our 192) and the overwhelming majority are in the mixed GrecoPersian style which is not represented on the pyramidal stamps at all. It may be that the main period of the Greco-Persian scaraboids is at the end of and after our main series. This rather tangential relationship may be further demonstrated by our 5*, which is a scaraboid with an Achaemenid subject in the advanced Court Style, bearing a Lydian inscription and one of the linear devices. zig*, which also has a linear device, is a scaraboid cut in a broadly Orientalizing style but with the eastern crossed animal motif. The linear devices are exceedingly rare on the scaraboids where their place may be seen to be taken by isolated Greek letters. It might be possible in time to demonstrate more clearly in terms of time and place the relationships between these two series of mainly Anatolian Achaemenid seals. At present it seems that seal engraving in the more remote satrapy with its Lydian capital was more thoroughly Persian in outlook than elsewhere in the western empire. And we may observe that the style of our pyramidal seals has much in common with Achaemenid coins, while the Greco-Persian has much in common with the more hellenizing coins of the south Anatolian states. But with an average survival rate for pyramidal stamps of far less than two per annum for their probable period of production it would be wiser not to draw any drastic conclusions about government or personnel from the seal evidence alone.

LIST OF SEALS All are pyramidal stamps of chalcedony unless otherwise stated. For the linear devices in the field, DI, D2, etc., see Fig. 3. The maximum dimension of the engraved face is stated for all illustrated examples, in millimetres. WithLydianInscriptions (i-ro) I.
2.

3. 4. 5.

6. 7. 8.

e' P1. i. London WA 115591, from Sardis. L. 17. Inscribed: mitratalis' sadmei. At the centre D43. Dr. R. D. Barnett is publishing this piece. samlid. At the centre D44. Delaporte, P1. I. Louvre A 1226, from Sardis. L. 12. Inscribed: bakivalid Louvre pl. Io7; W. Buckler, SardisVI.2, p. 72, no. 51, pl. I8 (B. Haussollier); Gusmani, Lydisches II, Wiirterbuch, 75, 190, no. 51. pp. P1. I. Naples 1475. L. 16. A lioness and a lion, its head turned frontal and a paw raised to its muzzle. Above the animals Dio and D2. Inscribed sivdmlim atelis. Cross-hatched exergue. Pl. Bibl. Nat. io86. L. 20. Standing lion griffin with one paw raised. Before it a flower. Inscribed: I. I maneli. Perrot-Chipiez, Histoirede l'Art II, fig. 347; Barnett, Jb. fiir kleinasiatische Forschung (1950-51), pl. 16.2; Gusmani, no. 55. P1. I. Leningrad, from Kerch. Cornelian scaraboid. L. 19. Two royal sphinxes seated, with one paw raised. Between them D58. Inscribed: manelim.S. Reinach, Antiquitisdu Bosphore Cimmerien, 16.Io; pl. und A. Furtwaingler,Die antiken III Gemmen (I900), p. I17; G. Lippold, Gemmen Kameen (1922), pl. 79.2. For the context see above, p. 37. P1. I. Geneva 20564 (once Kenna), acquired at Latakia. L. I8. A bull. Before it D60. Inscribed: manelim.Alternate hatched exergue. Barnett (see 4), pl. I5.5a-c, 16.1; Gusmani, no. 56. Fig. 2. Once Herzfeld. " Agate ". A Persian holds two lions, one inverted. At one side D24. Inscribed: mane.omen IV (?). J. Friedrich, Kadmos (1965), pp. 154-56, fig. I (the inscription as Phrygian, but in a letter Professor Friedrich agrees to my suggestion that it is Lydian). P1. I. Philadelphia CBS.5117, acquired in Bagdad. " Limonite or flint." L. 17 5. A crouching lion griffin with head turned back. From its chest spring the foreparts of a goat-sphinx and a boar. Above is D45. Inscribed: tafulim. Legrain, Culture the Babylonians, 48, 63, no. 1031, the inscription desof pls. cribed as Aramaic.

40 9.

JOURNAL

OF

PERSIAN

STUDIES

Fig. 2. Once Lord Strangford. " Blackish agate." L. 20. A Mede fights a lion. Between them D36. Inscribed: bsadvolim.Hatched ground line. Lajard, pl. 47.3. I . Once Clark, from Sardis. Shape and material not known (scaraboid ?). A man before a bull. Above is D I. Inscribed in the exergue: teteleit. C. Torrey, Annualof the American Schoolof OrientalResearch,
Jerusalem II-III
(1921-22),

p. o107; Gusmani, p. 270, B. GreekStyle (11-i7)

P1. 2. New York 86.1.3. L. 17. Hermes, wearing winged hat, chiton, himation and winged shoes, and holding his caduceus and a flower. Before him a bird. Furtwangler, p. 6.49; G. M. A. Richter, Catalogue of Engraved Gems,New York(1956), pl. 6.3312. P1. 2. Boston 95.80. L. 19. Herakles, wearing a short chiton and holding his club and a small lion by its forelock, looks towards a Gorgon holding two lions by their tails. She is winged, wears a chiton, and has snaky hair. Furtwangler, pl. 6.48; Lippold, pl. 37.1. Musee Danicourt. L. I9. A griffin attacks a stag. Cypriot inscription: a-ke-se-to-ta-mo. 13. P1. 2. PWronne, In the field a frontal face, added later. E. Babelon, La gravure pierresfines en (1894), p. 78, fig. 49; PerrotChipiez, III, p. 652, fig. 462; O. Masson, Inscriptions chypriotes (1961), p. 349, no. 363.
II. pl. 15. Fig. 6. Once de Bellesme. Haematite conoid (?). Two birds at the carcase of a calf. Cypriot inscrip16. tion: a-ru?-la-ti. Syria IV (1923), pl. 29.19; O. Masson, Syria XLIV (1967), pp. 372 f., pl. 20.4. P1. 2. Londonno. 6o9, pl. 10. Cornelian. L. 14. A baby crawls towards an ivy (?) branch and bunch of grapes. Furtwangler, pl. 10.30. 14. Fig. 6. Once Rome market. L. 14. A lion attacks a camel. Lajard,
43.17.

17. P1. 2. Istanbul, from Sardis. L. 8 - 5. Eros flying holding a branch and flower. SardisXIII.I, no. Io8,
pls. 10, I I, the subject not recognized.

Orientalizing (18-75)

18. P1. 2. Boston 01.7602. L. 17. A Mede talks to a woman holding a branch and a fillet. Between them D33. Hatched ground line.
19.
20. 21.
22.

23. 24. 25. 26. 27.

28. 29.

Bibl. Nat. o1032. As the last. Delaporte, Bibl. Nat., pl. 38.638. Baldwin Brett Coll. As the last, but winged lions. von der Osten, BaldwinBrettColl., pl. I2.137. Fig. 6. Bibl. Nat. L. 20o. As 20. Lajard, pl. 49.6; Delaporte, Bibl. Nat., pl. 38.637. P1. 2. Munich A 1396. L. 17. As the last, crested cap and possibly a beard. Hatched ground line. MunichI, pl. 26.237. P1. 2. Bowdoin College 484, from the Black Sea area (once Ouseley). L. 18. A winged goddess holds two rearing lions by their tails. Radet, Cybib', 21, fig. 28; Archaeologische Zeitung (1854), pl. 63-2; K. Herbert, Ancient in BowdoinCollege(1964), no. 484Art P1. 2. Oxford 1889.382, from Syria. L. I8. A winged figure in knee-length dress holds two lions upside down. P1. 2. Cambridge E.I3-.963 (once St. J. Hart). L. 12. A winged Mede holds two lions upside down. Baldwin Brett Coll. Conoid. A kneeling winged man with an angular headdress fights a lion. von der Osten, BaldwinBrettColl., pl. 12.I65. from
Sardis. A siren plays a lyre. Sardis XIII.I, no. 107, pls. Io, I I. Istanbul, New York, N. East Dept. 41.160.268. Agate conoid. L. 13. A royal siren sniffs at a flower. P1. 3.

Istanbul, from Sardis. A winged goddess holds two lions upside down. Sardis XIII. I, no. I I I, pls. 0o, I I.

30. Once Steuart. Agate conoid (?). A royal siren draws a bow. Lajard, pl. 46.15. 3'. Bibl. Nat. 1035. A standing lion with one paw raised. Delaporte, Bibl. Nat., pl. 38.622. 32. P1. 3. Istanbul, from Sardis. L. 8. Cornelian. As the last. SardisXIII.I, no. Io6, pls. 10, I I. 33. P1. 3. Louvre A I230. L. 17. Two rampant lions, heads turned back, birds above them. Between them
D2I.

Fig. 6. Hanover, Kestner Museum 1965.8. Conoid. L. 17. Two rampant lions, heads turned back, a tree between them. Above is D4o. Miinzen und Medaillen: Auktion,Basel XXVIII, no. 630; Kestner Jahresbericht (1964-65), no. 14. 35. Bibl. Nat. Io4o. Agate. As the last. Delaporte, Bibl. Nat., pl. 38.629. 36. P1. 3. Louvre K 506. Whitejasper. L. I6. As the last. Lajard, pl. 26.6; O. Richter, Kypros, 32.I I, pls. 78.5; Lippold, pl. 84.4. 34.

Hatched ground line. Delaporte, Louvre II, pl. 107.

PYRAMIDAL

STAMP

SEALS

IN

THE

PERSIAN

EMPIRE

41

37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49.
50.

51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 6I. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71.
72.

73. 74. 75.

Lyons, Musie des Beaux-arts. As the last. Above is D32. Delaporte in MilangesDussaudII, pls. 2, 5, no. 43. P1. 3. Once Arndt Coll. A 1394. Conoid. As the last. Above are D13 and D28. P1. 3. London WA 89871. L. 18. Two rampant lions, heads turned back, over a flower. Fig. 6. Geneva 19207. L. 14. Two lions rampant, heads turned back, over a hare. M. L. Vollenweider, raisonnidessceaux,Geneve (1967), pl. 41.91. I Catalogue Copenhagen, National Museum 9469. Two winged lions rampant, heads turned back, over a tree. Istanbul, from Sardis. Two seated winged lions, heads turned back. SardisXIII.I, no. I 12, pls. 10, I I. Fig. 8. Once Parot. L. I6. A lion attacks a bull. A bird over the bull. At one side D54. Lajard, pl. 43.I8. P1. 3. Munich 21780. Cornelian. L. 17. As the last, with the bird. MunichI, pl. 27.245 (the lion is not horned). P1. 3. Bibl. Nat., Pauvert de la Chapelle 22. L. 13. A lion attacks a bull. Above is D48. E. Babelon, Coll. Pauvertde la Chapelle (1899), pl. 4.22. New York, Morgan Library. As the last. E. Porada, Corpus Ancient Near EasternSeals I, pl. 126.842. of P1. 3. Bibl. Nat. 1042. Cut down (?). L. 20o. As the last. Imhoof-Blumer and O. Keller, Tier- und Pflanzenbilder MiinzenundGemmen auf (1889), pl. 19-33. of Copenhagen, Thorwaldsen Museum. Cornelian " scarab ". As the last. P. Fossing, Catalogue the Gems(1929), pl. 1.3AntiqueEngraved P1. 3. New York 81.6.I. Cut down. L. 17. As the last. Richter, pl. 4.20. P1. 3. Bibl. Nat. Io41. L. I6. As the last. Lajard, pls. 12.12, 45.I6. Athens, Numismatic Museum, Karapanos Coll. As the last. Journ.Int. d'arch.num.XV (1913), pl. 7-523II, P1. 3. Louvre A 1233. Rock crystal. L. 14. As the last. Delaporte, Louvre pl. 107. Bern, Merz Coll. Cut down. As the last. Louvre A 1231, from Sardis. As the last. Delaporte, Louvre pl. Io7. II, P1. 3. Unknown. Impression in Oxford. L. 14. A lion attacks a goat. New York 26.31.379. A lion attacks a goat. Richter, pl. 9.56. der Berlin F 97, from Athens. A lion leaps over a quadruped, a bird above. Furtwangler, Beschreibung Steine(1896), pl. 3.97geschnittenen Munich A 1391, from Tarsus. As the last and in the same style. MunichI, pl. 27.248. Indiana University, Burton Y. Berry Coll. A lion attacks a goat, with other quadrupeds in the field. IndianaUniversity Museum Art Publication (1965), pp. 12 f., no. Io. V P1. 3. London WA 132357, from Asia Minor. L. 13. Two birds at the carcase of an antelope. Above is D22. New York 4I.I60.444. As the last. Richter, pl. 10.63. P1. 4. Boston 95.83. L. 17. As the last, a triple branch between the birds. London WA I115625. As the last, a branch between the birds. P1. 4. Hanover, Kestner Museum 1928.234. L. 20o. As the last. Above is D9. no. P1. 4. London 236, pl. 5, from Kara Hisar. L. 14. As the last. 8. Bibl. Nat. 1039. L. 17. Two griffins rampant. Lajard, pl. 57.9; Delaporte, Bibl. Nat., pl. Fig. 38.627; 0. Richter, Kypros, pls. 32.12, 78.9. Once Southesk. Two winged horses rampant at a tree. Hatched ground line. Southesk Coll. I, pl. 16.0.7. P1. 4. Cambridge E.15/509.1954. Lodestone. L. 19. Two goats rampant at a tree. Above is D37. Iraq XXI (1959), pl. 11.64Fig. 8. Bibl. Nat. 1048. L. 18. Two birds. Lajard, pl. I.I; Delaporte, Bibl. Nat., pl. 38.630. Fig. 9. New York 23.I6O.28, bought in Athens. L. 22. Two cocks at a tree. Richter, pl. 10.64. P1. 4. London, Mrs. Russell Coll. L. 15. A goat grazing. V Bibl. Nat., de Luynes 328. A goat and branch (? linear device). Arethusa (1928), pl. 7.15. V P1. 4. Bibl. Nat., de Luynes 325. L. I5. A mare suckling a foal. Arethusa (1928), pl. 7.14. Fig. 9. Once Buckler. L. 18. A winged bull. Ground line of dotted squares. Impression in Oxford. Once von Aulock. A hippocamp (?) and fish. von der Osten, vonAulockColl., no. 203. Court Style (76-171)

76. 77.
5A

Istanbul, from Sardis. A Mede fights a lion. Below is D6I. SardisXIII.I, no. 116, pls. Io, I I. Fig. II. London WA 89881. L. 16. As the last. DII between them. Hatched ground line. Lajard, pl. 19.2.

42

JOURNAL

OF PERSIAN

STUDIES

78. Munich. As the last. MunichI, pl. 26.235. 79. Munich A 1400. As the last. MunichI, pl. 26.233. 8o. Once von Aulock. As the last. von der Osten, vonAulockColl., no. 153. 8 . Switzerland, Market. As the last. MiinzenundMedaillen,Sonderliste K, no. Io05. 82. P1. 4. Impression in Oxford. Conoid (?). L. 14. As the last. 83. Fig. I 1. Bibl. Nat., from Persepolis. L. 19. A Persian fights a lion. Between them D4. Lajard, pl. 47.2; M. Menant, Glyptique orientale p. 165, fig. 143; Delaporte, Bibl. Nat., pl. 38.642. II, 84. P1. 4. Leiden, Rijksmuseum A. 1951/4-4. White jasper. L. 20o. As the last. Above is D59. 85. Fig. I I. Once de Montlezun. L. 16. As the last. Between is D50. Lajard, pl. 46.7. 86. P1. 4. Leningrad, from the Great Bliznitza tumulus. L. i6. As the last. D37 above. Compte Rendu,St. and Petersburg 1869, pl. 1.18; E. Minns, Scythians Greeks, 427, fig. 318. p. 87. P1. 4. Oxford 1890.1 I8, from Smyrna. L. 13. White jasper. As the last. 88. Bibl. Nat. As the last. Delaporte, Bibl. Nat., pl. 38.640. 89. Switzerland, Private. Cut down. As the last. V 90. Bibl. Nat., de Luynes 321. As the last, over a kneeling goat. Delaporte, Bibl. Nat., pl. 38.641; Arethusa (1928), pl. 6.26. 91. Istanbul, from Sardis. Cornelian. A kneeling bowman and a lion. Between them D52. SardisXIII.x, no. I18, pls. 10, I I. Once private, Volo, from Pharsalos. A Persian fights a lion griffin. Perrot-Chipiez, Histoirede l'Art V, 92. p. 852, fig. 500. 93. P1. 4. Oxford I890.124, from the coast of Syria. L. I2-5. As the last. Bare legs. 94. Istanbul, from Sardis. As the last. SardisXIII.I, no. I i9, pls. Io, I I. 95. Baldwin Brett Coll. As the last. von der Osten, BaldwinBrettColl., pl. 12.136. 96. Baltimore, Walters Art Gallery 42.154. As the last. Iraq VI (I939), pl. 15.122. 97. P1. 4. London WA I156o6. L. 19. As the last but no headdress. DI8 between. 98. Pl. 4. London WA 132356, from Asia Minor. L. 21. A Persian fights a lion griffin. Aramaic inscription at one side. Hatched ground line. 99. P1. 5. Bibl. Nat. M 7539. Conoid. L. 20o. As the last, a winged sun disc overhead and a lion below. To one side D63. Soo. Baldwin Brett Coll. Rock crystal. A Mede fights a goat-sphinx, von der Osten, pl. 12.138. IoI. Fig. I I. Leningrad. L. 17. A Persian fights a goat-sphinx. Lajard, pl. 51.3. 102. Bern, Merz Coll. A Persian fights a winged goat. 103. Fig. II. Bibl. Nat. L. 16. A Persian fights a horse. Lajard, pl. 29.8; Delaporte, Bibl. Nat., pl. 38.639. 104. Oxford I889.985, from Lower Egypt. Conoid. A Persian fights a goat. 105. Unknown. A Persian fights a bull. Lajard, pl. 15.2. So6. Bibl. Nat. 1030. A Persian holds two lions inverted. Delaporte, Bibl. Nat., pl. 38.636. 107. P1. 5. Munich A 1408. L. 24. As the last. MunichI, pl. 26.236. io8. P1.5. Oxford 1891.324, from Gebelain, Upper Egypt. Cornelian conoid. L. is I6. As the last. Above D55London WA I I5551. As the last. Summary. og9. I o. P1. 5. Oxford 1892.1406, from Ephesus. Grey serpentine. L. 15. A Persian holds two winged goats inverted. Hatched ground line. III. New York L.46.25.8. A winged Persian holds two lions inverted. Hesperia, Suppl. VIII, pl. 31.1. I12. Clay impression from Memphis. The preserved part shows a rearing sphinx held by the tail by a central figure. F. Petrie, Meidumand Memphis(III), pls. 35, 36, no. 25. 113. P1. 5. Bibl. Nat. L. 18. A Persian holds two rearing winged horses by their harness. Between them D5. i 14. Fig. I2. London. L. 22. A figure in short dress and with a low headdress holds two lions upright. Winged sun disc above, and a bull (?) below. Lajard, pl. 44.10. S15. P1.5. Boston 98.70o7. L. 19. A Persian fights a lion griffin. A female lion griffin behind him and a winged sun disc above. I116. P1. 5. Bibl. Nat., de Luynes 198. Cut down. L. 19. Two royal sphinxes seated, a flower between them, V winged sun disc overhead. Delaporte, Bibl. Nat., pl. 38.633; Arethusa (1928), pl. 6.19. I 17. P1. 5. Geneva 65/20272. L. 14. As the last. Beside the flower DI5. Vollenweider, op. cit., pl. 4?I.93. I I8. P1. 5. London WA 115534. L. I8. Two royal sphinxes seated. Above them D64. Lajard, pl. 49.'I. I19. Boston 2 7.653. Conoid. As the last, paws not raised. Between them D8. 120o. P1. 5. Cambridge E.2/5o9.I954. Agate conoid. L. 22. As the last, paws raised. Between them D65. Iraq XXI (1959), pl. I1.63.

PYRAMIDAL

STAMP

SEALS

IN

THE

PERSIAN

EMPIRE

43

121.
122.

123. 124. 125. 126. 127. 128.

129.

130.
131.

132. 133. 134. I35. 136. 137. I38. I39.

140.

141.
142.

I43. I44.
145.

146. 147. 148. I49.

I50o.

15'.
I52.

153. I54. I55.

I56.

Oxford 1933-417, from Sardis. As the last. Pl. 5. Munich A 1398, from Egypt. L. 19. A lion griffin (leonine rear legs) and a royal sphinx, seated back to back, heads turned back. Before the sphinx DI4. Winged sun disc overhead, ground line with spaced double bars. MunichI, pl. 27.243. P1. 5. Istanbul, from Sardis. Rock crystal. L. 15. A royal sphinx and a winged lion with horns sit facing each other. SardisXIII.i, no. 114, pls. 10, I I. Copenhagen, National Museum o1025. Agate. A winged lion and a royal sphinx seated, heads turned back. Between them DI9. P1. 5. Bibl. Nat. M 6560. The back of the stamp sheathed in gold. L. 15. A royal sphinx faces a sphinx, both standing. Between them DI 7. P1. 5. Boston 03.Io12. L. 17. A royal sphinx with human torso draws a bow at a lion. A dead antelope below. Before the lion D62. Fig. 12. Bibl. Nat. L. 15. A bicorporate facing royal sphinx. Lajard, pl. 49.3; Delaporte, Bibl. Nat., LIX (1935), pl. 38.631; Bulletinde Correspondance Hellenique p. 320, fig. 7. Munich A 1397, from Adana. A bearded sphinx, with paw raised, seated beneath a winged sun disc. MunichI, pl. 27.242. P1.6. Liverpool City Museum M 8691. L. 13. A goat-sphinx walking. Before it D35. Lajard, pl. 46.14P1. 6. Bibl. Nat. 5069. L. 15. As the last. A floral before it. Revuenumismatique 1905, P. 290, no. 9 bis and fig. Oxford 1889-377, from the coast of Syria. As the last. P1. 6. Bibl. Nat. M 6598. L. 15. As the last. Before it D25. Pl. 6. Bibl. Nat. M 5990. L. I3. A seated goat-sphinx, a floral before it and D47 behind. RevuenumismatiqueI905, pl. 8.9. Mitylene (?), from Lesbos, found in a clay sarcophagus. Two seated goat-sphinxes. Between them DI6. Praktikates archaiologikes Etaireias1926, p. 153, with fig. 6. Boston 13.231. L. 19. Two seated goat-sphinxes. P1. II, P1. 6. Louvre A 1238. L. I5. Joined foreparts of goat-sphinxes. Delaporte, Louvre pl. 1o7. P1. 6. Istanbul, from Sardis. L. 16. A seated goat-sphinx with paw raised. Behind it D31. Sardis XIII.I, no. Io9, pls. IO, I I. P1. 6. Boston o3.Ioo3. L. 14. A winged lion and a goat-sphinx seated. Between them D38. II, P1. 6. Louvre A I237. L. 14. A sphinx with bird's tail. Before it D39. Delaporte, Louvre pl. Io7. P1. 6. Cambridge E.I 13.1904. Rock crystal. L. 15. A lion standing. Iraq XXI (1959), pl. I1.67. Once Arndt Coll. A 1392. As the last. Istanbul, from Sardis. A lion griffin, standing. In front is D53. SardisXIII.I, no. I 13, pls. o10,II. Istanbul, from Sardis. As the last. In front is DI3. Ibid., no. I Io. Gemmen, P1. 6. Boston 95.84, from Tenos. L. 13. As the last. In front is D42. Furtwangler, Die antiken pl. 12.I. Once Southesk and Robinson. As the last, a tree in front and D36 behind. Southesk Coll. I, pl. 17.0.33; AJA LVII (1953), pl. 25-53. P1. 6. Bibl. Nat., Pauvert de la Chapelle 21. L. 13. As the last. In front is DI2. Babelon, pl. 4.21; Delaporte, Bibl. Nat., pl. 38.625. Istanbul, from Sardis. As the last. SardisXIII.i, no. 115, pls. 10, 11i. Munich 94903. As the last (a replica ?). MunichI, pl. 27.241. P1. 6. London WA II5565 L. I3. As the last, one paw raised. Bibl. Nat. Io37. As the last. Lajard, pl. 26.4; Delaporte, Bibl. Nat., pl. 38.624. P1. 6. Munich A 1395. L. 9. As the last, walking, with a serpent tail. A triangle tree in front and a star and crescent in the field. Hatched ground line. MunichI, pl. 26.240. PI. 6. Bibl. Nat. Cornelian. Cut down to a scarab. L. 13. A seated lion griffin with paw raised. In front DI. Lajard, pl. 68.4. P1. 6. Bowdoin College 492. L. I I. A seated lion griffin with paw raised. Before it D29. Herbert, Art Ancient in BowdoinCollege,no. 492. P1. 6. Bibl. Nat. Io38. L. As the last. Delaporte, Bibl. Nat., pl. 38.623. Io.5. Rock crystal. A crouching lion griffin with head turned back. Iraq XXI (1959), Cambridge E.I 5.1904. pl. Ii.68. Bibl. Nat. Two seated lion griffins with heads turned back. Delaporte, Bibl. Nat., pl. 38.626.

44

JOURNAL

OF PERSIAN

STUDIES

157. Once Millingen. A seated lion griffin with paw raised and a seated winged lion, with both paws raised, back to back. The winged sun disc overhead. Lajard, pl. 43.23158. Fig. 12. Unknown. L. 16. Two lion griffins rampant over a tree plant. Aramaic inscription above. 159. Bibl. Nat., de Luynes 323. Conoid. As the last, without the tree. Delaporte, Bibl. Nat., pl. 38.628; Arethusa (1928), pl. 6.27. V
i6o. P1. 6. Oxford 1892.1407. L. 17. Two rampant lion griffins over the carcase of a calf (?). Lajard, pl. 43.26; O. Richter, Kypros, pls. 32.7, 77.13,
142.4.

161. Fig. 12. Leningrad. Conoid. L. 15. A crouching lion griffin holding the forepart of a boar. In front is D41. Furtwangler, III, p. 124; Lajard, pl. 62.8. 162. P1. 7. Cambridge E.I2.I963, from Asia Minor. L. 22. A winged lion and a griffin rampant over the carcase of a calf. Coll. I, pl. 2.B.7. 163. Once Southesk. Cut down, perhaps a conoid. A crouching griffin. Southesk 164. Pl. 7. Bibl. Nat. Io85a. L. I6. A crouching Bes sphinx, head frontal. I65. P1. 7. Boston 27.665. L. I5. A crouching Bes sphinx with head frontal and the head and neck of a lion, seen from above, issuing from its chest. In front is D56. 166. P1. 7. London WA 11x5596. Rock crystal. L. I5. A crouching Bes sphinx with head frontal, and beyond it the front of a crouching lion griffin. A lion in the exergue. I67. P1. 7. Unknown. Conoid (?). L. 17. A crouching sphinx with a goat's head. From its back rise what seem to be the torso and head of a Bes figure. I68. Once Southesk, from Cyprus. " Green agate-jasper ", perhaps a cut scaraboid. As the last. Southesk Coll. I, pl. 16.0.19, and p. 206, where similar or identical gems are mentioned. I69. P1. 7. London. Crystal. A horse and dog. Above is D26. Hatched border. O. Keller, Die antike TierweltI (1909), pl. 3.I; Imhoof-Blumer and Keller, Tier- undPflanzenbilder Miinzen und Gemmen auf Behind him a horse, over a 170o. Fig. 13. Paris, de Clercq 76. Agate. L. 20. A Mede on foot spears boar. which is a linear device, not clear in the publication, but resembling D58. Ground line with spaced triple bars. L. de Clercq, Coll. de ClercqII, pl. 3.76. 7'1. Fig. 13. Yale. Felspar. L. 18. A Persian guard with bowcase and spear. Newell Coll., pl. 31.462.
Other Styles or Inscriptions (172-184) (1889), pl. 16.52.

P1. 7. Bibl. Nat. M 618I. Cornelian. L. 15. A bull on a block ground line. Star, crescent and diamond in the field. Aramaic inscription below. I73. Once Steuart. Agate. A bull, as the last, with a crescent and plant. Aramaic inscriptions above and LXIV Vereins below. Lajard, pl. 43.1; Zeitschrift deutschen des Paldistina(194I), pl. 6.45de Clercq 54, from Amrit. Grey limestone. A sphinx with a plant and triangle. Coll. deClercq II, 174. Paris, pl. 3-54172.

175.

P1. 7. London WA 115621. Grey limestone.

L. 14. A lion griffin with crescent and tree.

I76. P1. 7. London, University College, from Egypt (?). Grey limestone. L. 15. A grazing goat, with Use, pl. 14.223. of angular crescent or " ka " arms and star. F. Petrie, Objects Everyday 177. P1. 7. London WA 132894. Grey limestone. L. 17. A lion with a tree, crescent above and a quadruped below. Zigzag in the exergue. 178. Paris, de Clercq 76 quater. Grey limestone. A chariot, with star, crescent and disc in the field. Coll. de ClercqII, pl. 7.76 quater.
S79. Bibl. Nat. M 578I. Aramaic inscription in two lines. I80. London WA 102972. Aramaic inscription in two lines. I81. London WA 113200. Aramaic inscription in two lines. On the side of the stone a Babylonian horned

deity.
182.

Jerusalem.

Aramaic inscription in two lines. Israel Exploration Journal XV (1965), PP. 230-2,

pl. 40F.

I83.
184.

Yale. Aramaic inscription in two lines, divided by panels with two drop-shaped and one circular motifs. A.A.S.O.R. Jerusalem II-III (1921-22), pp. IO5f., no. 3; von der Osten, Newell Coll., pl. 33.539.
Paris, Louvre A I 139. As the last. Delaporte, Louvre II, pl. Io4.

OtherShapeswith LinearDevices(185-198) . conoids Large I85. P1. 7. Unknown. L. 17. Two rampant goats, back to back. Flowers in the field. Between them D5I.

2 3

6
Plate r.

11

13

16

12

17

18

23

24 Plate2.

25

26

45

60

29

32

33

36

39

38

44

47

49

50

52

55

Plate 3.

82

84

86

62

64

65

68

71

73

87

93

97

98

Plate 4.

115

126

99

107

108

110

113

116

117

118

120

122
Plate 5.

123

125

133

137

137

129

130

132

135

136

138

139

140

144 152

146

149

151

153
Plate 6.

154

160

167

169

162

164

165

166

172

175

176

177

185

190
Plate 7.

191

188

194

195

194

196

198
Plate 8.

PYRAMIDAL

STAMP

SEALS

IN

THE

PERSIAN

EMPIRE

45

186. Fig. 13. Munich A 1407, from Kotiaion in Phrygia. L. 23. A Persian fights a lion. Crescent above; 187. Oxford 1911.132, bought in Aleppo. Rock crystal. A Persian holds two lions inverted. In the field D55Weightstamps 188. P1. 8. Boston 66.1077. Two rampant lions. At one side D2o. BostonMuseumrear (1966), p. 43 below. 189. Munich A 1283. Agate. A bull; above it D23. Cross-hatched exergue and hatched border. MunichI,
pl. 35-307. between them D49. Munich I, pl. 25.226.

Pl. 7. Vienna IXB.123. Cut down. L. 12. A lion attacks a bull. Above is DI8. Cross-hatchedexergue. sur Lajard, Recherches le Cultede Venus, 17.2. pl. Scaraboids I90.
191.
192.

Pl. 7. Boston 01.7578.

L. 16. Two leaping stags, crossed.

Between them D34.

Furtwangler,

III,

p. io6, fig. 73. Boston 23-584, from Greece. Agate. A zebu bull. Before it DI I. Furtwangler, pl. 12.6;
Lewes House Gems (1920), pl. 4.64New York 42.11.24, from Thessaly. Cornelian. A grazing horse. Below is D27.

J. D. Beazley,

193.

Hatched border.

Richter, pl. 18.Io6; Boardman, ArchaicGreek Gems,no. 564. 194. P1. 8. Istanbul, from Sardis. Haematite. L. 17. A lioness. Above is D46. SardisXIII.I, no. 99, pls.
9, 1I ; Archaic Greek Gems, no. 431.-

Scarab 195. Pl. 8. Istanbul, from Sardis. Cornelian. L. 13. A boar. Above it is D30o. SardisXIII.I, no. 98, pls.
9, I 1; Archaic GreekGems, no. 497.

Goldrings 196. Pl. 8. Istanbul, from Sardis. L. 19. A lion and a bird. Before the lion is D3. SardisXIII.I, no. 90, pls. 9, II; Boardman, AntikeKunstX (1967), p. 26, N17. 197. London WA 124oo6, from the Oxus Treasure. A royal winged bull. Aramaic inscription above, and D6 in front. 0. M. Dalton, The Oxus Treasure, I05, fig. 54, pl. i6. no. Cylinder
198, P1. 8. Boston o0.7609. H. 19. Goat-sphinxes, with bulls' bodies and forked tails, at a palm tree. A

winged sun disc. By the animals D7 and D57.

PERSIAN LACQUER IN THE BERN HISTORICAL MUSEUM By B. W. Robinson

The oriental collections of the Bern Historical Museum deserve to be more widely known. On the Islamic side they comprise mainly the enormously rich collections of arms and armour from Persia, a Turkey, India and Central Asia formed from about 1868 onwards by Henri Moser-Charlottenfels,1 of metalwork mostly from the same source and covering the same areas, and the representativerange painted lacquer which will be described in the present article. For the benefit of orientalists with a rather wider horizon it is worth mentioning that the Museum also has in store a very considerable Japanese collection including swords2and sword-furniturein abundance, armour, lacquer, porcelain, prints and illustrated books. The following figures will give some idea of the extent of the Persian lacquer collection, most of which is in store: forty-three qalamddns (pen-boxes) including one fine example of carved Abddeh work, one of steel with gold overlay, and one Indian specimen; eighteen mirror-cases,including one example of painted enamel on gold; five boxes and caskets; four spectacles-cases; and two panels. Among these are sixteen signed pieces representingthe work of thirteen different artists,dating from the mid-eighteenth to the early twentieth century. It is, of course, neither possible nor desirableto give here a detailed description of each piece; rather more than half may be classified as of good or very good quality, and we shall confine our attention to these. A number of the other pieces, it must be admitted, are poor or coarse work of the kind that has given such a bad name to the art of the Qatjarperiod. The four earliest signed examples bear the kind of punning signature, usually in the form of an invocation, that was common among Persian painters, especially in lacquer, throughout the eighteenth and well into the nineteenth century. Thus the painter Muhammad Zamdn would sign Td SZdhib al-Zamdn," O Lord of Time ", a stock invocation to 'Ali, and the work of Sidiq usually bears the invocation Yd al-Wa'da, " O he who speaks truly in his promise ", often addressed to Ja'far the sixth Shi'ite Imtm. S.ddiq Two al-S.diq, pieces in the Bern Museum bear the above-quoted signature of Muhammad Zamdn. No. 652 is dated I1161/1748, and this is disconcerting, because it falls just mid-way between the two hitherto recorded Muhammad Zamins. The first and better known artist of this name was a court painter sent to study in Italy under the second 'Abbas. To him is largely due the European veneer that overspreads the face of Persian painting from the end of the Safawid period. The second is known as a talented miniaturist and lacquer painter, who also worked in oils, at the beginning of Fath 'Ali Shdh's of reign. Unless there is an errorin the date, this qalamddn I i6 I /1748 must be the work of a third artist of the same name; it is of very good quality, decorated with hunting scenes. The other qalamddn bearing the same invocation, No. MK 1072, is also of good quality and probably the work of the later Muhammad Zamin under Fath 'Ali Shdh; under a rather dark varnish it shows a standing lady in a landscape, with pastoral scenes round the sides. The formula az-ba'dMuhammad ashrafast, " after Muhammad 'All is the noblest ", seems to 'Ali have been used as a punning signature by the painter Ashraf,and it occurs on three very similar mirrorcases with bird and flower designs of exquisite quality.3 The Bern qalamddn, 21-1912, on which the No.
1

The oriental arms and armour collections of Henri MoserCharlottenfels have been described and catalogued in a long series of articles beginning in 1928 and continued over twenty years, by R. Zeller and E. Rohrer, in the Jahrbuchof the Bern Historical Museum. They were also the subject of a luxurious folio album produced by Moser himself, Collection Henri MoserCharlottenfels:OrientalArmsand Armour(Leipzig 1912). 2 B. W. Robinson, " Japanese Blades in the Historical Museum,

Bern, Switzerland " in the Journal of the T6-ken Society of Great Britain, no. 3 (January 1967), gives a summary list. 3 (a) Dated II53/1740; Edinburgh, Royal Scottish Museum, no. 1921-1433. (b) Dated i i6o/1746; London, Victoria and Albert Museum, no. 758-1876. (c) Dated 1169/1756; London, collection of John PopeHennessy Esq.

47

48

JOURNAL

OF PERSIAN

STUDIES

same formula is found, is similarly painted with birds and flowers, and may be dated to the middle years of the eighteenth century. The most distinguished and influential painter of the eighteenth century seems to have been Persia sometimes called Muhammad Sadiq or Mulla Benjamin, who was American Consul in S.diq, in the I880s and made extensive enquiries intoS.diq. the history and traditions of Persian painting, formed the impressionthat he was " one of the most noted painters of modern Persia ",4 and it may have been this reputation that prompted Texier's informants a generation earlier to attribute to him the famous fresco in the Chehel Sitin at Isfahan representing Nadir Shah's victory over the hosts of the Grand is, however, found as late as 1797, and Mughal at Karnal in 1738.5 Signed and dated work by is a very long working life, though admittedly not entirely out of the question. We do at S.diq sixty years any rate know that he worked for Karim Khan at Shiraz during the third quarter of the eighteenth century, as oil paintings signed by him with his usual punning invocation are still to be seen in situ in one or two niches in the Pars Museum of Shiraz, which was originally one of the great Vakil's garden pavilions. The invocation YdSddiqal-Wa'daappears on two mirror-casesat Bern. No. 72/I13,painted with Christian subjects-the Annunciation, the Presentation in the Temple and the Adoration of the Magi-is of excellent quality, and is dated 121 [o] /1795. The other, No. 641, is dated to the following year, but the painting, though bold and effective, is not really of the first quality. It may well be the work of a pupil. The last of these punning invocation-signaturesto be found in the Bern collection, Td Shdh-iNajaf, " O King of Najaf", is another ejaculation customarily addressed to 'Ali, Najaf being the town that grew up round his tomb near Kilfa, where he was murdered. It was used as a signature by Najaf, or Najaf 'Ali, perhaps the best known and most prolific of the Persian lacquer painters." His father was the painter Aqa Bdbd of Isfahan, his younger brother was the celebrated Painter Laureate (naqqdsh bdshi)Muhammad Isma'il (of whom more later), and his three sons Kazim, Ja'far and Ahmad were No. all talented artists in lacquer and enamel. The Bern Museum has one qalamddn, MK 987, signed in the usual way; it is finely painted with dervishes and portraits of youths under a rather by Najaf dated I3O8/1891, is signed kamtarin dark varnish. No. MK 984, another finely painted qalamddn ("the humble ") Najaf; this might be a grandson of the famous Najaf. Several unsigned pieces in the very with love-scenes, and, more collection are very close to the latter's style, such as No. 654, a qalamddn No. MK 989, painted with a standing girl and a caged bird. This latter is an almost exact especially, in duplicate of a qalamddn the Victoria and Albert Museum (No. 849-1889) bearing the usual inscription Yd Shdh-iNajaf. A qalamddn signed by Najaf's second son Ja'far (No. 20-1912) and dated 1278/ of fine quality and follows his father's style pretty closely (P1. VIc). The painted enamel 1861, is portrait of Nasr al-Din Shah on the gold mirrorNo. 381, which seems to date from the I85os, may also be by him, though it is signed simply Ja'far; his normal signature was Ja'far ibn Najaf 'Ali. But the two finest painted lacquer pieces in the Bern Historical Museum-perhaps, indeed, the finest outside Persia itself-are both the work of Najaf's younger brother Muhammad Isma'il.7 The
more spectacular of the two, No. 71/13, is a box or casket measuring 32.3 x 26.3 x I I.6 cm., signed and Ismd'ilnaqqdsh-bdshi, dated 1282/1865. The top of the lid (P1.I) gives a panoramic view Muhammad of the Siege of Herit by Muhammad Shah Qajar in 1837-38. It comprises literally hundreds of tiny

figures, yet the composition is ordered and coherent. In the background rise the walls and minarets of the city, and immediately in front of them is part of the Afghan army drawn up in a wedge-shaped

formation. The central space is taken up by a complicated mblhe in which elephants are involved, and on either side and in the foreground are serried ranks of Persian troops and artillery, with some tents on the left. In the immediate foreground, slightly left of centre, the Shah is seen on horseback receiving a tribute of Afghan heads (we are told that he specially ordered the first Afghan prisoner taken to be bayoneted in his presence). Other scenes from the same campaign, of equally meticulous
SS. G. W. Benjamin, Persia and the Persians (London 1887), p. 319. See also B. W. Robinson, Persian Miniature Painting, Victoria and Albert Museum (London 1967), no. 93, p. 77. 6 Charles Texier, Description de l'Armdnie, &c. (Paris 1852), p. 124. " 6 See Amir Mas'fid Sipahram, Aqq Najaf Isfah~ni
7

qalamdmns~z " in Honar va Mardonz,no. 31, p. 65. For a general account of Muhammad Ismd'il and his works, see B. W. Robinson, "A Lacquer Mirror-case of 1854 " in Iran V (1967).

PERSIAN

LACQUER

IN

THE

BERN

HISTORICAL

MUSEUM

49

execution, occupy the front and back surfacesof the box; on the front is a review (P1.IIa), and on the back the Persian bombardment of Ghairidn (P1. IIb). Muhammad Shah's stocky and unattractive figure (what a contrast to the regal dignity of his grandfatherand predecessorFath 'All Shih!) appears in both scenes; at the review his head is shielded by what appearsto be an umbrella from the Burlington Arcade, as he gingerly fingers the muzzle of a cannon, and at the Ghairiin assault he is accompanied
Mirza Aghdsi. The by his vizier, the " ignorant and fanatical ",8 " incapable and crotchety "9 H two sides of the box (P1.IIIa and b), however, are painted with scenes from the campaigns of Mirza .jji

Muhammad Khan Qaj~r against the Turkmans some twenty years later than the Herat affair, and it seems highly probable that this superbcasketwas commissionedby that prince.10The beaten Turkmans are shown fleeing before the invincible Persian phalanx with its blazing artillery, whilst the prince receives, on one side, the humble submissionof their chiefs and, on the other, pinioned prisonersand the inevitable severed heads. On the interior of the lid (P1. IV) is an oval medallion framed in rococo scrolls and surroundedwith flowers, within which Muhammad Shah is representedenthroned with his on court, Mirzd Aghdsi (a faithful likeness"1) his right. The figures here are a trifle larger than those on the exterior surfacesof the box. The interior base is superbly painted with floral designs, and .hIjji even the exterior base is covered with a fine pattern of arabesque scrolls. It is difficult to imagine that a finer or more elaborate example of Persian lacquer painting than this box exists anywhere. Muhammad Ismd'il's other work in the Bern Historical Museum is a splendid mirror-case, No. the box it is of equal or perhaps even finer quality, and the condition is impeccable. It is signed and dated 1288/1871, so must be one of the artist'slast works. The back of the mirror-caseand the exterior surface of its sliding cover are painted most exquisitely with an identical floral design of an oval medallion of interlacing gold scrollworkwithin a floral border. But it is upon the inner surface of the cover that Muhammad Isma'il has exercised his utmost skill (P1. V). In the centre sits Hazrat 'Ali holding his bifurcated sword Zi'l-Fiqdr acrosshis knees, a representationprobably inspired by the fine painting by Abii'l-lHasanGhaffari Sani'al-Mulkin the Museum of Decorative Arts, Tehran.12 Behind his haloed head is a dissolving cloud of angels' faces evidently derived, via Sani' al-Mulk (who studied in Italy between 1846 and 1850) from Raphael's Sistine Madonna; Sani' al-Mulk used this device in at least two portraits of royal children.13 'Ali's two young sons, the Imams IHasanand IHusayn,are representedhaloed and attended by angels in a pair of pendant medallions below him. On either side and a darkof 'All are full-length figuresof an old man with folded arms, possibly his father Abil skinned man holding a battle-axe, probably intended for the Abyssinian Bilal, the first of the Muslims T.lib, appointed to give the call to prayer. Below them again on either side, just above the medallions containing IHasanand lHusayn, are two half-figures, of which the one on the right, represented in armour, is most probably Mukhtar, the avenger of the slaughtered innocents. Above the central figure are three more cartouches framed in gilt rococo scrollwork; in each of the outer ones are two angels, one blowing a trumpet and the other offering a bunch of flowers, whilst in the middle one, at the head of the whole composition, is a miniature representationof the Mi'raj, or Heavenly Ascent, of the Prophet mounted on Buraq, conducted by Gabriel, and attended by other angels. The whole is contained in a border enclosing panels of verses, ten couplets in all. This splendid mirror-caseforms a worthy culmination to the career of Persia's foremost lacquer painter.
Other artists represented in the Bern Historical Museum's collection, in chronological order, are: Muhammad'Ali: a fine qalamddndated I258/I842 decorated with birds and flowers, No. MK 985. No other lacquer painting by him has so far come to light, but he may possibly be identifiable with " the most humble (kamtarin)Muhammad 'Ali " who signed an oil painting of the court of Fath 'Ali Shih, dated 1250/1834, in the Museum of Art of the Peoples of Asia, Moscow.
8 Sir P. M. Sykes, History of Persia, 2nd edn. (London 1921), vol. II, p. 328. 9E. G. Browne, Literary History of Persia (Cambridge 1953), vol. IV, p. 127. 10For this suggestion I am indebted to Mr. Yahya Zoka, formerly Director of the Ethnographical Museum, Tehran, and now Director of the newly-founded Carpet Museum.
6A

73/13, measuring 27 X 17.3 cm. Though the painting is not so intricate and complicated as that on

xxCf. Browne, op. cit., pl. X.

12

Reproduced by Mr. Zoka in the second of his two admirable articles on Sani' al-Mulk in Honar va Mardom, no. i I, P. 23. 13 Ibid., pp. 22, 23.

50

JOURNAL

OF PERSIAN

STUDIES

Mulla Murad: a box dated 1291/1874, No. 633, painted with birds, flowers and the heads of young people in medallions, on a backgroundimitating tortoise-shell,said to have been made for Nasir al-Din Shah. This is the only example of the artist's work that has so far come to notice. 'Abbas Shirazi: a very fine qalamddn dated 1291 /1874, No. MK 973, painted in markedlyEuropean with family and pastoral scenes, deer and two portraits of young girls (P1.VIa). 'Abbas was a style well-known qalamddn-sdz; examples of his work were exhibited at Cairo in 1935,14 and another, two dated 1286/1869, is in a private collection in Tehran. Jawad al-Imami: a qalamddndated [I]3Io/i893, No. 24-1912, finely painted in black and gold on a cream-coloured background with birds and flowers. This artist, who is sometimes distinguished by the appellation 'Abd al-Raji, also decorated with birds and flowers a fine spectacles-case, dated number of talented painters throughout the nineteenth century. Simirumi: a qalamddn dated 1320/1902, No. MK 978, extremely well painted in completely Eurowith a Persian lady in European dress and European landscapes. The artist's name, which pean style derives from Simirum, a village near Isfahan, was borne by three painters at about this time. The one most likely to have painted this qalamddn Muhammad Ibrahim Simirumi, who learned European is methods of painting in Russia, and later workedfor the Zill al-Sultan, eldest son of Nisir al-Din Shah, the vandal governor of Isfahan who cut down the magnificent trees in the Chahar Bagh for firewood and despoiled or destroyeda number of the Safawid palaces with which the city was formerlyadorned.15 has Finally, one or two of the anonymous pieces are worth a mention. No. 23-1912, a qalamddn, a highly original design of girls and cupids half hidden amongst a mass of vegetation; No. 19-1912, a is box, is covered with varieties of marbling, beautifully executed; No. 65o, a qalamddn, a fairly early example of an attempt at archaeological accuracy in the representationof Sasanian kings, which can be seen in the illustration of lithographed editions of the Shdhndma from the mid-nineteenth century onwards; No. 640, a mirror-case,though by no means of the first quality, is neverthelessa rarity in that the inside of the cover is painted with an ingeniously contrived but highly improbable erotic scene; No. MK 982 (P1. VIb), a qalamddn with girls' heads in engaging Victorian poke-bonnets, is similar to some of the work of Muhammad Isma'il and may have been painted by one of his pupils; and No. 631, a casket, the exterior showing the reception of an European ambassadorat the court of Fath 'All Shah, has the interior painted with a prince hunting, who is labelled " Chingiz Shah ", together with an inscription underneath stating that the work was executed in the reign of " Chingiz Shah"
in the year 735/1335. " If you throw a stone ", runs the Persian proverb, " throw far! "
1283/1867,

in the Museum of Decorative Arts, Tehran (No. 7624). The Imami family produced a

14 G. Wiet, Expositiond'Artpersan de (SocietiedesAmisl'Art)janvier-fivrier z935 (with Albumof plates) (Cairo n935),nos. P. 92 and P. 222, two qalamdans dated 1293/1876 and 1289/ 1872 respectively.

15

Once again I must thank Mr. Zoka for this information on the Simirumi painters, kindly communicated in a letter.

P1. I. Lid of box No. 711/3 by Muhammad Ismda'l, 1282/1865.

Siege of Herat.

P1. IIa. Front of box No. 71/13:

Muhammad Shah at a review.

Pl. IIb. Back of box No. 711/3:

Bombardmentof Ghfirian.

P1. IIIa. Right-hand side of box No. 71/13:

Rout of the Turkmans.

side P1. IIIb. Left-hand of box No. 71/13:

Vaevictis.

P1. IV. Interior of lid, box No. 70/13:

Court of Muhammad Shah.

cover Pl. V. Interior mirror-case No. 73/13 by Muhammad Ismd'dl, of 1288/1871. Hazrat 'AlTandothersacred personages.

Pl. VIa. QalamdSn by 'Abbas Shirdzi, 129i/I874,

No. MK 973.

c. P1. VIb. QalamdSn, styleof Muhammad Ismd'Fl, i86o, No. MK 982.

P1. VIc. Qalamdan by Ja'far b. Najaf 'AlT, 1278/1861, No. 20-1912.

CHARLES I AND THE ANTIQUITIES OF PERSIA THE MISSION OF NICHOLAS WILFORD By R. W. Ferrier
On January 2oth 1638, at a consultation held in Bandar 'Abbis " the death of Nicholas Wilford who departed this life the 26 December last past was made known unto the English merchants".1 Decisions were taken about his belongings, for he died intestate and his servantJohn Amberseene was not thought to be trusted,2and about carrying out the responsibilitieswhich had been entrusted to him by his master, Charles I, by whom he had been entertained as a groom of his chamber on March 27th
1637.3 Earlier, on March 23rd 1637, the Committees of the East India Company at a meeting of their Court, considered the request to transport " Nicholas Wilford, skillful in the art of painting-recommended from the Earle Marshall ... on the Jonas for Persia being sent out by his Majesty upon some special business to the King of Persia ".4 The Jonas arrived at Swalley, the port for Surat, on October I8th and early in December Wilford, in company with the Persian merchant Allee Ballee who had travelled with him, set sail in the Blessing for Bandar 'Abbas, arriving there on January i8th 1638. The reason for Nicholas Wilford's journey and the services expected of him in it are indicated in the following: " Charles R. Memoriall by way of Instructions with or trusty servant Nicholas Wilford shall principally attend for or service in Persia. Ffirst after he hath delivered or letters into ye Emperor of Persia and ye pictures of or selfes, or Queene and children wich we send unto him, he shall seeke to gett his favor by ye best means he can, as even by employing his art to seve him, as he shall best like an also to apply himselfe to gett ye love and favor of those yt are in authority under him, yt so he may have their helpe in seeing ye Antiquities of yt Country and in procuring such as are fitt for us to be brought away to ye sea side on landThe Antiquities we principally desire are Antique statuaes whose heads or peeces Baise Relioug, subscriptions or such like in Marble or brasse, any great or small Vessels of Aggats, Brasse or Marble any Antique Coins or gravings in any kind of Aggat or any Mettall any Instruments for ... or such like As also any excellent auntient Bookes in ye Persian Arabian or Greeke Language. Secondly he is exactly to desyne all auntient Buildings and Edifices such as are either famous for Antiquity or Architecture and particularly to note their forme and eligancy and if amongst such (or in any other places) he shall find any Statuaes, Inscriptions, Pillars, etc., such as are (by being fixed to some Building) imoveable or otherwise in respect of their greatness then so to designe them as to be able to render a Just accompt both of their Magnitude, position and proportion. Thirdly he is to take notice of their manufactures with the Excellencyes of them as Cloth of Gould Silke with its colours and dies. Ffourthly whatever his dilligence or Industry shall acquire for us in their kinds, especially Statuaes Inscriptions Aggats, etc., he shall with all dilligence case up and so with a particular of them consigne them to ye Agents of or East Indian Companie for ye time being resident in those parts, by them to be safely conveyed unto us. Ffithly he is not to faile by all Conveniencies on land through Turky as also by sea to Communicate with or trusty servant Wm Petty in Italy either att Rome Fflorence or Venice both the worth, forme and Antiquity of all such statuaes etc as being ours on whose judgement and on things of yt nature we much rely.
1
2

(I)ndia (O)ffice Records, January 20oth1637/8. E/3/16/x615, I.O., E/3/I6/I658, January i5th 1638/9. 3 I.O., E/3/i6/I6o9, December 26th 1637.

4 B (Court Minutes of the East India Company)/18, p. 297,

March 23rd 1637-

51

52

JOURNAL

OF PERSIAN

STUDIES

But above all he is not to faile by all ye waies and meanes prescribed to acquaint here from time to time or right trusty and welbeloved Cozen and Counsell Thomas Earle of Arundell and Surrey Earle Marshall of England with whose eldest sonne ye Lord Maltravers with all ye particulars and occurrencies of this Journey and employment yt so from him we may recieve therein full and ample satisfaction. Sixthly and lastly we having taken a Roiall determination for ye planting a Colony on ye Iland of St. Lawrence otherwise Madagascar wich by reason of its propinquity to Persia and those places may much advance or trade in ye Easterne Indies amongst whome we shall employ or trusty servant Captn. Bond with whome or will and pleasure is yet he keepe all fitting correspondence such as may be advantageous to us in ye furthering of or Roiall designes."5 From this it is clear that Wilford's responsibilities were twofold, one artistic, the other commercial. It was not the first time that Charles I had shown particular interest in both aspects regarding Persia. The literature of the time contains many references to Persia, but the interest of the first two Stuarts in Persia had been stimulated by the adventures and activities of the Sherley brothers, Anthony and Robert. In particular, Robert Sherley's two unsuccessful visits to London in 1611-13 and 1624-27 to persuade the City merchants to use his services as an intermediary with the Persian king, Shah 'Abb~s, had excited interest.6 The Levant Company was unenthusiastic about trading with the Grand Signior's traditional enemies, and the East India Company, with their factory already established in Persia since 1617, were satisfied with the services of their own servants. So in spite of royal interest and suggestion of royal participation in the silk trade, Sir Robert Sherley failed to convince the London merchants of either his sincerity or ability. It is interesting to note that both Sir Robert Sherley and his wife, Teresa, were painted in Persian dress by Van Dyck just prior to their English visit when they were in Rome.7 Not only did these admirable portraits come into Charles I's collection and have particular artistic and historical distinction in their own right, but it was through the patronage and interest of Thomas Howard that Van Dyck was introduced into the royal circle. Also, Inigo Jones was in this period including characters dressed in Persian style in his Court masks. The arrival of another emissary of Shdh 'Abbds, Naqd 'Ali Beg, with an Armenian merchant, Khwaja Shdhsuwar in attendance, produced a pantomime-like situation in which rival protestations of masquerading were disputed and Sherley was assaulted in the royal presence by the Persian. Charles I was scarcely persuaded to grant an audience to the Persian, and hardly spoke at it; he was determined to have the insult righted and his propositions about the silk trade presented in person to the Persian king. So he despatched an embassy under Sir Dodmore Cotton.8 Unfortunately marred by the death of the three ambassadors, the mission was nevertheless distinguished by the pen of Thomas Herbert, the secretary.

His first edition of A Description thePersianMonarchy beinge Orientall the now Iles partsof Indyes, andother of
the GreaterAsia and Africk was published in 1634 and contained a description of Persepolis on pages 56 to 6o. He was a very close friend of Charles I and present at his " tragicall beheading ". Thomas Herbert was enthusiastic about Persepolis, " where nature and art seem to conspire towards the creating amazement and pleasure both in sense and intellect ". He felt that it was a " great pity that some illustrious prince or noble person valueing rarities has not ere this sent some painter or other like artist to take a full and proper draught of this so ancient an monument ". There is little reason to doubt his statement that he himself made such a proposal to " that great maecenas of antiquity, the late noble Lord Thomas, Earl of Arundel ". It was a Mr. Northgate who recommended Wilford, who " could both design and copy well ", to the Earl.sa Not long after the unfortunate embassy of Sir Dodmore Cotton, Charles I was writing in I630 to Shah Safi commending " our Cosin, Subiect and Servant William, Earle of Denbigh " who wished to
6
6

8 Sir Thomas Herbert, Travelsin Persia 1627-29, ed. Sir William I.O., G/29/x. Sir Anthony Foster (London 1928); The Journal of RobertStodart, ed. Sir ed. Sherleyand his Persian Adventure, Sir E. Denison Ross (London 1933); Boies Penrose, The Sherleian Odyssey E. Denison Ross (London 1931). Cotton's instructions are in the Coddrington Library, All Souls College, Oxford, Owen (London 1938). Artists in 17th CenturyRome, Wildenstein Gallery (1955), PP. Wynne MSS., vol. XII, ff. 8aHerbert, op.cit., pp. o9f. 144-5. Van 45-8; Sir Anthony Dyck,Thos. Agnew and Sons Ltd. (1968), pp. 21-3. Both paintings are at Petworth House, Sussex.

CHARLES I AND ANTIQUITIES:

MISSION OF NICHOLAS WILFORD

53

" sattisffy his heart with the fulness of your Princely presence, and the glory of your royall Court and greatness ".9 The Earl of Denbigh, however, never reached the Persian Court. Early evidence of Charles I's interest in the arts and knowledge of Persia is indicated by his request to the East India Company on February 2 Ist 163410 asking them to obtain Arabic and Persian manuscripts, a request which the Persian factors acknowledged on November 27th 1634: " Our Soveraignes requiries of your worships to furnish him with some varyties of Persian and Arabian manuscripts we shall have regard to ".11 President Methwold from Goa, whither he had gone to negotiate a truce with the Portuguese in the East, informed the Company of his limited success in satisfying the wishes of the King and drew their attentions to the problem of reading the script: " Wee are exceedingly grieved that we cannot in all points accomplish his Majestes Royal pleasure. There is no want of Persian bookes of all sorts, most men of quality in this citty and kingdome being either Persians borne, discended from them, or educated in the knowledge of that language; so that Persian bookes are plentifully to be had and we have sent Io such, of severall subjects, although we do believe that there are few in England that will understand them; for however the character resembles the Arabique (every letter carryeing the same denomination and pronunciation) yet for want of those pricks, both above and below, which point out the vowells, and are all waies used in the Arabian character, the Persian is very difficulty read and understood but by them which are conversant therein. But we will hope that some industrious young man will make use of the opportunity he may injoy, and attayne to so much perfection as to give some light at least to direct more able linguists."12 Later he apologized for not being able to obtain enough " Arabian bookes " but advised that through the help of friends he had sent into the Red Sea for more. " In the meantime we believe these will find worke enough before they will be well understood."13 Benefiting from the interest of Archbishop Laud and his protdge, Edmund Pococke, the first Professor of Arabic at Oxford, and that of others, the Bodleian Library was beginning at this time its collection of Oriental manuscripts. Coinciding with this request were the Persian king's demands for colours for painting. These were forwarded by President Methwold to the Company on April 28th 1636,14 and were probably sent with the Swan to Persia from Surat, November 24th 1637.15 Artists, too, were seeking work at the Persian Court, such as " Clement Evans a working goldsmith expert in all the work incident to that trade ",16 " and Benjamin Webb ... a very sufficient workman in the art of lymming ".17 These two men never reached Persia, but a clockmaker named Festy was credited with the clock in the Bazaar in Isfahan."s There was therefore precedent enough for Charles I to send Nicholas Wilford to Persia for " seeing ye Antiquities of yt Country and in procuring such as are fitt for us to be brought away ". However, that was but the one aspect of his journey, the other was advancing " or trade in ye Easterne Indies ". Charles I at this time was engaged indirectly or actively in a number of enterprises which both infringed the privileges of the East India Company and endangered its existence in the East. The Peace concluded with the Portuguese had opened up new possibilities of trade. As President Methwold wrote, " never since it was a trade were theare so many conveniencies conducing thereunto as at present ", but he was amazed at the action of Charles I, " letting Cobbe loose unto the Redde Sea by commission under his privy signett and countenancing Captain Weddall by his more publique letters patent ".19 The project for Madagascar was longer maturing and finally proved a failure, but in the interim it caused the East India Company some anxiety. One settlement on Mauritius by the Earl of Southampton had been unsuccessful, " that other intended by the Earle of Arundell on St. Laurence will, wee believe, have like issue, since that of the two is the most unlikely to produce ought worthy the paines and cost.
9 (P)ublic (R)ecord (O)ffice S.P. 102/40, pt. I.
February I634.
16 B/I8,

10 B/2I

17 B/I8,

11I.O., E/3/15/I538, November 27th 1634. 12 I.O., E/3/15/1543A, December 29th i634. 13 I.O., E/3/15/1543B, January i9th 1634/5. 14 I.O., E/3/I5/I547, April 28th 1636. January 13th 1637/8. 15 I.O., E/3/16/1614,
7

is John Ogilby, Asia, The First Part beingan Accurate of Description Persia and the several thereof... (London 1673), p. 15provinces 19 P.R.O. Co. 77, vol. VI, no. 36. This refers to the Courteen Association, a group of courtiers, merchants ana disgruntled former servants of the East India Company, with whom the King was connected.

p. 14I, March 2nd 1635/6. p. 143, March 4th I635/6.

54

JOURNAL

OF PERSIAN

STUDIES

Somewhat else is doubtless intended besides plantations. God grant they prejudice not your affaires, and then it shall not much trouble us what silence they meet with on these rare undertakings. "20 Such was the interest of Charles I in Eastern trade generally. Of Persian trade in particular he had knowledge, because he wrote in 1631 and 1635 on behalf of the East India Company to Shah Safi remonstrating with him because he had broken his contract with the Company.21 Shah Safi replied politely that " since there hath passed their letters of love and goodwill betwixt us, lett me intreat you that itt be not hereafter omitted ",22 but the Company failed to obtain compliance with the stipulated contract. It seems that the actual impetus for his intended trade to Persia was provided by the Persian merchant Allee Ballee,23 a merchant from the Persian king who had been in Venice before arriving in London in July 1635. He expected the East India Company to visit him, " hee being a person of quality and hath brought over a faire estate with him ",24 and took " exception that he hath not his respect by the Company to invite him and his followers one day to dinner ".25 Prior to Allee Ballee's sailing he had been negotiating with the Company about " the transportation of 20 tunns of goods for the accompt of his Master, the King of Persia, for seven persons ".26 The Company, which believed that his request " cannot bee but very chargeable and discommodious to them ", were demanding ?Iooo.27 The Persian ambassador, as he styled himself, who was as concerned about his reputation as his pocket, suggested " that he might go from hence with honour and reputation and that the world might not take notice of any freight he paid the Company ... which he conceived in som disrespect unto him being an ambassador, he wold therefore be content in a secret and private way to present the Company with a present of 500 rials ".28 The result of these protracted negotiations was an advance by Allee Ballee of 1500 rials and a bill for 500 more " and hopes they will expect no greater summe from him ",29 which they did but failed to get. " His impudent bearing ... his insolences " needed at Surat " more then ordinary man-like patience "; the Persian factors considered him, " a man of little esteeme scarce known here amongst the King's servants of better sort ".30 He did not long survive his travelling companion, Wilford; he died, probably at Ldr in June 1638, having disposed of most of his goods in Surat and Bandar 'Abbas. His treatment by the East India Company in London rankled with him and he felt deeply slighted. He claimed that it was Shah Safi's intention, as he wrote to Charles I, that " the whole trade of his Contry for Cristendom should pass through the hands of your M:ties Subjects ". " But I perceive," he continued, " though it may bee your M:ties pleasure yett, there is a neglect in some of your Subjects."31 There was no doubt in the minds of the factors that Allee Ballee was responsible for the proposed royal trade, though they doubted his authority for it, sc. " his Maties letters to Shaw Suffee proposing certaine conditions for bartering Cloath, Tynn and Spices for silke. I have in their severall transcripts sent you whereby your Worships may perceive his Maties willingness to enter upon this trade whereof you are justly weary: but this proposition preceding in an apprehension from Allee Ballee his insincerities and indeavouring by large promises for reducing the whole trade of silk into England to ingratiate himself with his Matie (which being nothing inabled thereto) I believe hee little intends will as suddenly vanish as contrived."32 However, as he was carrying pictures similar to those which were to be presented to the Persian king on behalf of Charles I by Wilford, the newly-appointed Agent, Thomas Merry, was instructed to proceed to the Court at Isfahin, for " which therefore is first shonne will take off the curiosity and rarity of the other ".33 This Merry did at his audience on April I6th 1638, when he introduced himself and proffered the presents of the Company and the letters and presents of Charles I:
I.O., E/3/17/1725, December 9th I63921 P.R.O. SP. 102/40, pt. I.
22 23

20

28

Ibid.

24

I.e. 'Ali Bali? B/I8, p. 193, July 27th 163525 B/x8, p. 263, February i5th 1636/7. 26 B/i8, p. 268, February 22nd I636/7.

B/i8, p. 263, February I5th 1636/7. p. 288, March I5th 1636/7. 29 B/i8, p. 300, March 29th 1637. 30 I.O., E/3/I6/I646, August 3oth 1638. 31 P.R.O. SP. 102/40, pt. I. 32 I.O., E/3/I6/I622, February 3rd 1637/8. 33 I.O., E/3/16/162I, February I3th I637/8.
B/I8,

27

CHARLES I AND ANTIQUITIES:

MISSION OF NICHOLAS WILFORD

55

the Kings Matie letters and pictures together with the pishcash of the Company amt to 112 tomands thereabouts ... not till then the King viewed the pictures with a serious eye as if he were much taken with them and after dinner we were told hee was viewing of them in a private room (whether for the time they were sent) a long
time together and the next day the Ettamon Dowlett34 sent to know the names of the Queene and the children as also to have the letters translated out of English into Persian."35

" On Sunday last at ye Kings feast of sacrifice ... the Agent according to Custome was invited and presented

Finally, the factors who had not " bene by or Honorable Emploiers or otherwise at all or perfectly acquainted with the scope and intent of his designment ", notified the Earl of Arundell of what had been effected on behalf of his Majesty by the presentation of letters and pictures. " And after which time unto the date of my said letter [31 July to E.I.Coy.], I solicited him for answer thereun to but was delayed from time to time according as I advised your lordship. And the like soliciation I have contrieved since uppon all fit occasions, not only to the time of the Emperor's departure from hence towards Bagdat (which was the 8th September) but having occasion to send some of our people to follow the standinge all which I was from tyme to tyme put of with delatory promises that the Kinge would not be mindful to answer them. So or people receved advice this answeare from Ettamon Dowlett at the King's Camp that the Emperor had writt to his Matie by his servant Musta bege (of whose despatch towards Poland with the Polonian Ambassador in July last I made some mention to your lordship in my former) and I am informed that there was another messenger latelie despatch after hin with some new instructions that it may be letters to his Majestie are sent by him: however my dilligence (though weake and unworthy) indeavours weare employed to the utmost in this solicitation."36 Amongst the possessions of Nicholas Wilford which were detailed, is that of his chest: " A box of writings being the Kings Maties Memoriall also 3 quire paper. Secretary Windebanks warrant for post horses. 2 Lrs of his Mtie of England to ye King of Persia. Herberts booke-A Booke of flowers stampd. Dr. Halls works-A Booke of Common Praier. Quintus Curtius, A table booke. Ovids Metamorphosis in Ffrench. A Cover of a booke with draughts. A booke calld Rex Redux. A booke of perspectives. A booke of Artichecture. A Red Cloth Cloke. A black Cloth Suite. A short sleeve Jerkin. A branch taffety shute. A hammer and pinsers. A Cotte with lining. 3 prs shortes. 2 whole shirts 2 halfe shirts. 2 plaine handkerchiefs. 3 pr ould cuffs I pr ould russett boots 2 prs spurs 2 looking glasses. I paper of red ? chossee. 2 Pallates for paint. I pr red silk garters.
12

King's camp I enrolled them to be very much myndfull and dilligent in continuing the same allegation notwith-

sibilles in Black and Wte. Io printed stamps.

36

34I.e. I'tim/id al-Daula. 5 I.O., E/3/I6/I63I, April 2ist I638.

I.O., E/3/16/1649, September I2th 1638.

56

JOURNAL

OF PERSIAN

STUDIES

The 4 seasons of the Yeare. A picture of a Gentlewoman. A picture of a Spaniard. An imperfect picture."37 This is interesting for it must be the first reference to a copy of Thomas Herbert's book in Persia and secondly the mentioning of " A booke of flowers stampd ", and " Cover of a booke with draughts " and " A booke of perspectives " would indicate evidence for the existence of a possible further source of European patterns in Persia which could have been utilized by Persian artists at this period.38 Though his mission was terminated prematurely by his death, and though from papers mentioned in his inventory Wilford was knowledgeable about Persia (having even corresponded with an English Merchant there, Thomas Adler, on January 22nd I634), it is most unlikely that his commercial overtures would have been successful. It is however, a tantalizing prospect and matter of regret that he was never able " to render a just accompt both of their magnitude, position and proportions " of the Antiquities of Persia.

37
38

I.O., E/3/I6/I6o9, December 29th 1637. "

See Basil Gray,

An Album of Designs for Persian Textiles "

in Aus der Welt der Islamischen Kunst,Festchrift Ernst Kiihnel, fiir pp. 219-24, for other evidence.

ILETMISH OR ILTUTMISH? A RECONSIDERATION OF THE NAME OF THE DEHLI SULTAN By Simon Digby
In the history of Muslim India, the knowledge of any dialect of Turkish has been confined to small groups and has not greatly flourishedfor any length of time.' The names of the slave Sultans of Dehli of the early thirteenth century necessarilyappear exotic. Whilst those of other Sultans, of Aybak and Balban, are sufficiently simple to have been preserved correctly, the cumbrous Arabic spelling of the
personal name of the second independent Sultan, Shams al-Din Iltutmish or Iletmish (607-33/12I 1-36)

has given rise to doubt as to its correctform and meaning. In this name the medial dotted letters have been peculiarly liable to scribal corruption or to compressionthrough the exigencies of monumental or numismatic layout. The form of the name common in the historiographical tradition of the seventeenth century and later was tlI. We know from eighteenth- and nineteenth-century English transliterations that this was pronounced then-as it is today upon the lips of old-fashionedIndian Muslims who refer with a tendency to swallow the or to the monuments attributed to this Sultin-as Altamish Altamash,
1 Evidence of Turkish as a spoken language in the Dehli Sultanate is scanty, in spite of the pride of descent which distinguished the great Turki slaves of the early thirteenth century, among them Sultan Shams al-Din Iltutmish himself (see Jfizjdni, Tabaqdt-ied. NJVsiri, W. Nassau Lees [Calcutta 1864], p. 165). Probably the only literary remains in Turkish from the early Dehli Sultanate are the macaronic baitsin Turkish and Persian in the first part of the genealogical work of Fakhr-i-Mudabbir (see Ta'rikh-i-Fakhru'ddin ed. Sir E. Denison Ross Mub&rakshdh, [London 19271, p. 46). The poet Amir Khusrau, writing in the late thirteenth and early fourteenth centuries, refers to the speaking of Turkish by immigrants to the Sultanate, by this time mostly free Turks passing out from territories controlled by the Mongol confederacy; he also alludes to the many Tdziks " who have learnt Turkish with industry and erudition in India " (Dibdcha-i-Ghurrat al-Kamdl, British Museum Add. 21I,04, folio 155, tr. M. Borah in BSOS VII [19331, P- 326). Elsewhere Khusrau has some interesting remarks on the development and use of Turkish (Nuh Sipihr, ed. M. Wahid Mirza [Calcutta 1948], pp. 176, 178, 179): (p. 176)
-,%I0

'

A JT

5;

(p. 178)

TY. 3.I .(p.179)

l,/~4j~!?Zj~ ?J~ ~-t.T" ?s-~ The considerable Turkish lexical element in the Persian of the Dehli Sultanate still awaits study. Mention should be made of the dictionary of Shaykh Mulhammad b. Ldd Dihlavi, Mu'ayyid al-Futald (lithographed, 2 vols. [Kanpfir, Naval Kishor 1883,

reprinted 1899]; see also Blochmann, "Contributions to Persian Lexicography ", JASB XXXVII [1868], pp. 1-72). This dictionary, composed in 925/1519 before the Mughal invasion brought fresh groups of Turkish-speaking immigrants into Northern India, has Turkish as well as Arabic and Persian sections noticed separately under each letter of the alphabet. This would seem to indicate that a knowledge of Turkish still survived at Dehli long after the destruction of the unified Dehli Sultanate by Amir Taimfir in 1398 A.D. Barani, writing in Dehli in the middle of the fourteenth century, laments the disappearance from the market of great Turkish royal slaves ed. (Ta'rikh-i-Fir5zshdhi, S. A. Khan [Calcutta 1862], p. 314); while the term umard'-iatrdk,referring to the royal slave household, was evidently applied in the late fourteenth century to the slaves mainly of Eastern Indian provenance, commonly known in urban Dehli as " the Hind6stdnis " (cf. Muhammed British Museum Or. Bihamad-Khani, Ta'rikh-i-Muhammadf, 137, folio 425b and the account of the same incident of the massacre of the royal slaves in Sirhindi, Ta'rikh-i-Mubdrakshdhi, ed. Hidayat Hosain [Calcutta 1931], p. 150). In the Mughal period, Eastern Turki remained a domestic language spoken in the royal family for a considerable time. In Jahangir's reign (1605-27) the favour and esteem enjoyed by Captain Hawkins and Sir Thomas Roe were largely the result of their previous knowledge of Ottoman Turkish and their ability to converse in it. In the Mughal royal family a knowledge of ChaghatH'i Turki remained even in the late ed. Chandraeighteenth century (see Azfari, Urdu introduction, pp. i-ii; sekharan et al. [Madras 1957], Wdqi'dt-i-Azfarf, Survey,p. 643). Storey, Persian Literature,A Bio-Bibliographical Apart from the diplomatic contacts between the Turkish and Mughal states, merchants of Ottoman Turkish descent had long-standing connections with the ports of India, notably the great Chelebi family of seventeenth-century Sfirat. The writer hopes to publish shortly an article identifying a type of furniture made in Western India during the sixteenth and seventeenth centuries for the Ottoman Turkish market.

57

58

JOURNAL

OF PERSIAN

STUDIES

last indecisive vowel.2 The Turkish term in seventeenth- and eighteenth-century Indian meant the advance guard of the centre ,;.Jl the order of battle; this was pronounced altamish.3 in usage It may be supposed that the pronunciation of the name of the Sultan had by this time been influenced by a false identification with this familiar word. The memory of a different pronunciation of the Sultan's name lingered uncertainly in late sixteenthcentury India. The historian 'Abd al-Qadir Badaytini, who came from a Hind6stani environment preserving pre-Mughal traditions, gave a folk-etymology of the name as meaning " born on the night of the eclipse of the moon ".4 According to the reconstruction proposed by Sir John Redhouse when consulted by Edward Thomas, this should yield in Turkish Ay-tutulmasior Ay-tutulmish,5 forms in themselves unacceptable for the name of the Sultan, but strongly suggesting that the form which the etymology sought to explain was Iltutmishrather than Iletmishor Altamish. Il-tutmish, literally " grasper " or " holder " of the " folk " or " realm ", a name nearly corresponding in sense with the Persian Jahdngir, provided a name which accorded well with the epigraphic and numismatic evidence which had already been marshalled by Edward Thomas in 1870. Three centuries before the Dehli Sultan, this name was apparently borne by a Turkish governor of Ray.6 It was proposed as the reading of the name of the Dehli Sultan also by S. Lane Poole in 1884.7 It gained increasing acceptance among writers on Indo-Muslim history, achieved due consideration in the Encyclopaedia Islam' and passed thence to the Turkish Islam Ansiklopedisiin an article written by Fuat of K6prilui. In I95o Hikmet Bayur, author of a History of India in Turkish, in which he had used the form Iletmish, was provoked by a review of his work by Fuat K6pruilti to justify this usage. He published a long article citing evidence, whose value is examined in this present article, to prove that the correct form of the name was the one which he had earlier used without explanation." A thick expanse of evidence, if uncontested, tends to gain acceptance regardless of its intrinsic credibility. Accordingly we find Bayur's conclusions communicated by Professor Bernard Lewis to Dr. Peter Hardy, who wrote in the following decade upon the historians of the Dehli Sultanate;9 whence, by a recognizable process of
2 Lt.-Col. Alexander Dow, The History of Hindostan, 3rd edn. " Altumish ". Sir H. (London 1792), vol. I, pp. 185-92: Elliott and C. Dowson, The Historyof India as Told by Its Own Historians, 8 vols. (London 1865-78), see index to vol. VIII: " Shamsu-d din Altamsh "; it is difficult to believe that Sir Henry Elliott did not take the opinion of his Munshis and such learned Muslim acquaintances as NavvAb Ziyd' al-Din of Lohdrii. M. Elphinstone, Historyof India, Ist edn. (1938), 5th edn., pp. 371-5, has " Altamish ". Edward Thomas uses " Altamish " in his heading, see below. 3 W. Irvine, The Army of the Indian Mughals (London 1903; photo-reprint Dehli 1964), pp. 224, 226. Irvine spells Iltmish, turc-orientale citing Pavet de Courteille's Dictionnaire (Paris 1871), p. 31; the latter, however, only gives the unvocalized ,..I, " soixante ". Irvine also notes the form A in the eighteenth-century historian Khdfi Khdn (Muntakhab al-Lubdb, ed. K. Ahmad [Calcutta 1860-74], vol. II, p. 876) which, however, is likely to indicate only a common vagary of Dakhni pronunciation. For the customary Indian pronunciation, see Muhammad Pddishdh, Farhang-i-AnandRaj, repr. (Teheran 1335 Shamsi), vol. I, p. 401: bi'l-fath wa td'ifawqdni wa kasri-mimwa sukan-i-shin-i-mu'jama. ed. S'Abd al-Q8dir Badayfini, Muntakhab al-tavdrikh, A. 'Ali et al. (1864-69), vol. I, p. 62; English tr. by G. Ranking (1898), vol. I, pp. 88-9. of 6 E. Thomas, Chronicles the Pathan Kings of Dehli, revised edn. (London 1870), p. 44, n. I. 6 E. de Zambaur, Manuel de gindalogie et de chronologie pour l'histoirede l'Islam (Hanover 1927), p. 44. A portion of Bayur's argument is to the effect that Iltutmish must be a royal title, not a personal name, while the name borne by the Sultan is clearly a personal one, not a title; and that Iletmish, in the sense of " kidnapped ", is more suitable for a Turkish slave. However, names with royal or conquering associations used as personal names are not unknown in the Islamic world, both in the past and today. In view of the story given by Jaizjiniof the childhood of the Sultdn, which is that " like Joseph " (Tasufvdr) he was sold into captivity by jealous brethren, the name Iletmish does not appear especially appropriate. If Iltutmish was not in fact a personal name bestowed, without particular significance being attached to its royal connotations, in childhood, it may well have been considered suitable for Shams al-Din in his very promising youth, when, according to Jfizjdni, a slave merchant asked such a price for him in Ghaznin that Mu'izz al-Din Muhammad b. Sdm did not purchase him and forbade any other buyer to do so; after which, his general in Hind6stan, Qutb al-Din Aybak, had to seek a dispensation permitting him to purchase him at Dehli (JfizjSni, op. cit., pp. 165, 167-8). 7 S. Lane Poole, The Coins of the Sultans of Dehli in the British Museum(London I884), p. xxix. Lane Poole also mentions the conjecture of Ahmad Wafiq Pasha, communicated to him by Sir John Redhouse, that the name might be read .L;-.?I (sic), meaning "the kidnapped ", or "the slave who was carried off". This appears to be the historical origin of Hikmet Bayur's reading. 8 Hikmet Bayur, " Sultan Iletmi?'in adi hakkinda ", Tiirk Tarih Kurumu BelletenXIV (I950), pp. 567-88. 9 P. Hardy, Historiansof MedievalIndia, Ist impression (London 196o), preface, p. v: " Failure to examine this article in time has made inevitable the adoption in this monograph of the conventional but incorrect form." In the 2nd impression (1967) the whole " Postscriptum " has been omitted.

DEHLI SULTAN:

RECONSIDERATION

OF THE NAME-ILETMISH

OR ILTUTMISH ?

59

cultural diffusion, the form has entered into subsequent works on Indo-Muslim history'0 and has attained consecration in the Encyclopaedia Islam2 (s.v. Dihli). The lack of accessibility of articles in of modern post-Ottoman Turkish to historians of Muslim India appears to have prevented any detailed consideration of Bayur's arguments of the type here attempted.11 We may first state some considerations of spelling and palaeography which bear upon the argument. I and ~ I, or ~ I1 and AJ-II, The first vowel of the name is of indeterminate quantity. The forms are interchangeable, the close unrounded Turkish i being written either way, and considered either a long or a short vowel in Persian according to prosodic expediency. Beyond this permissible variation, we should also note the general tendency, both in transcription and in monumental and numismatic epigraphy, towards the omission of letters and the shortening of words rather than towards the introduction of superfluous letters. The abrasion of ~MLI and to ,l is not surprising; but the o1 of IZ4I, .I both in epigraphs and manuscripts (the latter often copied some sporadic appearance centuries later), supposedly corrupted from a correct form .J1--together with the fact that the intrusive letter is invariably a reduplication of the td-is surely an improbable process. Bayur cites three types of evidence for his contention that the correct name is Iletmish: the readings of the name found in manuscripts or editions of historical texts; monumental epigraphs; and coinlegends. Much of his argument is repetitive and cannot be answered in such prolix detail here. I. ReadingsCitedfrom Manuscriptsand Printed Texts As may be expected from our sketch of the historiography of the name, the form tl is common in subsequent manuscripts of late sixteenth- and seventeenth-century general extremely histories of India, notably those of Badayfini, Nizdm al-Din Ahmad and Farishta. From the etymology of the name cited above it would seem that at least one of these late historians, Badayfini, did not intend to spell the name in this manner, though it is so spelt in surviving manuscripts of his work. Bayur's evidence of near contemporary sources is of greater interest. (I) The Tdj al-ma'dthir of IHasan Niz~mi. For this Bayur relies on the abstract in English in Elliott and Dowson, which several times has Altamsh.12 This must, he argues, represent A..-Jin the manuscript used by Sir Henry Elliott, not now accessible. However, in available manuscripts of this prolix and ornate work, which is as yet unedited, the personal name of the Sultan only occurs once, in the notice of his succession. In the British Museum manuscript (Add. 7623), transcribed in 71I A.H. less than a century after the death of the Sultfn, the name appears (on folio 88a) in the form There are two " teeth " between ldm and mim: the first bears two dots, placed one above the ,tpL. while the second is undotted, as is also the yd following the alif. According to Barthold, the other, St. Petersburg manuscript of the same work, transcribed in 829 A.H., reads 5.1.13

(2) The Tabaqdt-i-Ndsiri of Jfizjdni. Besides the nineteenth-century printed edition from Calcutta,14 Bayur has consulted a manuscript of unspecified age in a library in Istanbul. Like the in the places where the name is mentioned; but it also bears careful marginal printed edition it reads ,l which he rejects. The oldest manuscript accessible to myself, British Museum Add. corrections, .I, 26,189, assigned by Rieu on satisfactory palaeographical grounds to the fourteenth century,15 reads on folio Ia JI (with two dots over the second " tooth "), on folio I76a again t}.JI, and on folio I79b JI (unpointed); on folio 175b where it appears in the illuminated heading announcing the beginning , of the reign, it is clearly written ,,I. The gilding of this heading is certainly contemporary, and the of the occasion is an adequate reason for properly pointing the name. formality
10

Aziz Ahmad, Studies in Islamic Culturein the Indian Environment Chughtai, Painting During the Sultanate Period (Lahore (Oxford 1964), p. 6 [and now in the same author's An Intellectual 1963). History of Islam in India (Edinburgh 1969), passim-Editor]. 1x The writer wishes to express his thanks to Dr. Turhan The suggestion does not seem to have attracted the attention of Gandjei, without whose unstinted help in translating and disother writers on the Dehli Sultanate, who continue to write cussing Hikmet Bayur's article the present paper could never have been written. Iltutmish. See K. A. Nizami, SomeAspectsof Politics and Religion in India During the Thirteenth Century(Asia Publishing House 12 Elliott and Dowson, op. cit., vol. II, pp. 237 f.1961), and Studies in Medieval Indian History and Culture 13 W. Barthold, " Iltutmyi ", ZDMG LXI (1907), pp. 192-3(Allahabad 1966); Asit K. Sen, People and Politics in Early 14 Jfizjani,op. cit. Medieval India (12o6-1598) vol. I, pp. 71-2, (Calcutta 1963); M. A. 15 Rieu, Catalogue the Persian Manuscripts, of

60

JOURNAL

OF PERSIAN

STUDIES

The readings on folios I76a and I79b are those in the texts of two qasidas in praise of the Sultan reproduced by Jiizjini. Sir E. Denison Ross drew attention to these verses, pointing out that the scansion required Iltutmish (- - u) rather than Iletmish (u - u).16 Bayur rightly observes that Iletmish (- - u) will serve as well: but this would surely require a spelling within the qasidas of phi, and he has failed to provide variant manuscripts with this reading. (3) The Javdmi' al-hikdydtof 'Aufi. Bayur cites the evidence of a manuscript in Istanbul, again of M. NizSmu'd-din, who had examined a great number of the existing unspecified age, which reads ~t. of the work, refers consistently to Iltutmish,"7 but unfortunately no text edited by him has manuscripts yet appeared. M. Mu'in's edition of the opening anecdotes, with the preamble and dedication to this Sultan, is based upon three manuscripts of the thirteenth or fourteenth centuries and one of the sixteenth. The name is there printed a of two manuscripts, one that of 717 A.H.; ~ , reading evidently without vocalization is noted as the reading of the two other manuscripts consulted.18 (4) The Addb al-harb wa' 1-shujd'aor Adab al-mulak of Fakhr-i-Mudabbir. Bayur attaches considerable importance to the reading in the India Office manuscript of this work, even though palaeographically it should be assigned to the sixteenth or seventeenth century.19 The dedication to the Sultan, which Bayur reproduces on fig. 2 of his article, reads ,t L. However, in the British Museum manuscript of this work (Add. 16,853), which is possibly sixteenth century,20 on folio 9b the name appears as ,-LI (yd undotted). Finally we may examine other near-contemporary works in the historical tradition of the Dehli Sultanate. Ibn Battfita's " Sultan Lalmish " is of no help to us, except as an indication of the difficulty of recalling a strange name after a lapse of years.21 Ziya' al-Din Barani, whose Ta'rikh-i-Fjirjshdhi often recalls the practices of the reign of Sultan Shams al-Din-no less than thirty-seven times in his account of the first two reigns of his chronicle-omits the personal name except possibly on one occasion. There it is given in some manuscripts as pJi; but its omission here also in other manuscripts makes it probable that this is a copyist's interpolation.22 The most important addition to the corpus of historians of the Dehli Sultanate since the nineteenth century labours of Elliott and Dowson is the Futizhal-saldtlinof 'Isami. This is a geste of the Muslim of rulers of India in mutaqdrib metre modelled upon the Shdhndma Firdausi. It was composed in 750 and 751 A.H. (1349-50 A.D.) at the court of the seceding Sultan of the Deccan, 'AlI' al-Din Hasan Bahmanshah. As is abundantly clear from his narrative, 'Isami was a descendant of the Dehli nobility of the previous century. His grandfather was a sipah-sdldr of Sultan Balban (1265-85 A.D.). His poem
16 Sir E. Denison Ross in BSOS VII (1932), p. I 10 . Since writing the body of this paper the republished edition of Jfizjdni's Tabaqdt-i-Ndsiri(Kabul 1342-43 Shamsi/1963-64, 2 vols.) has become accessible to the writer. Habibi in his ta'liqdt has two long notes (nos. 49 and 79, vol. II, pp. 376-8, 417-8) on the form of the name concluding that " in the period men of learning used to read (mikhvdnda and) the name with two tds, but other spellings of it were also common ". Most of Habibi's note no. 49 is taken up with the same metrical argument as that advanced by Sir E. Denison Ross. Iabibi quotes the same qasidasfrom Jfizjdni's text, and also applies the argument to a baytwith the name in a qasidaofTfaj al-Din " Sangriza " and to al-saldtin (for which see occurrences of the name in the this below). Bayur's objections to Futi.hmetrical argument are partially valid. The objection to Bayur's own arguments -I is that the form o5,tLI which would have to against it replace 0, is not found in manuscript readings at these places. Only after this is validity restored to the arguments of Ross and Habibi regarding the prosodic necessity of .l. Habibi also cites information or opinions generally in favour of Iltutmish from Lane Poole, Zambaur, Hodivala and others. Bayur's article does not appear to have reached him; this could be expected from the account in his introduction of the difficulties he experienced in obtaining less recondite works. In ;Iabibi's text the preferred reading is But the I. text was edited from a single manuscript of unspecified age available in Qandahar, collated with the Calcutta edition and the variants noted by Raverty in his translation (see IHabibi'sintroduction, op. cit., vol. I, pp. 2-3). In the portions of Juzjini published for the first time in the original by rIabibi there are references to a contemporary namesake of Sultan Shams al-Din Iltutmish, a prince and military adventurer of Sistdn who actually came to Hind6stdn, the Malik Farbzor Malik-i Nimraz Iltutmish (see op. cit., vol. I, pp. 283, 299). 17M. Niz~mu'd-din, Introduction theJawdmi'u'l-IHikdydt to (London 1929), passim. 18 'Aufi, Javdmi' al-hikdydt pt. i, ed. M. Mu'in, Intishardt-iDdnishgdh (Tehran 1335 Shamsi), p. 5. For details of manuscripts used, see introduction, pp. 50-1, 62-8. 19India Office Manuscript 647, Eth6, p. 1493, no. 1767. Eth6 gives no opinion as to the date of this manuscript. 20 Rieu, Catalogue,vol. II, pp. 487-8, who also assigns it to the sixteenth century. 21 Ibn Battuta, Voyages,eds. C. Defr6mery and B. Sanguinetti (Paris x855), vol. III, pp. I54, 164. 22 vol. I, ed. Shaikh 'Abd al-Rashid Barani, Ta'rtkh-i-FIrjzshdhi, ('Aligarh 1957), references as in indices s.v. (Sultin) Shams al-Din. The solitary and suspect reference is on p. 25. The editor does not indicate which manuscripts he is using.

Pl. Ia. Akrh'i din kd jhompra, Ajmer. The northernminaretfrom the courtyardof the mosque.

P1. Ib. The name of the Sultdn, Iltutmish, in the inscription on the northernminaret, north side.

Pl. Ha. Silver tankd with Kzlfic script Arabic legend. Al-sult.fn/al-mu'azzam shams/al-dunya wa'l-din/[abu'l-]muz.affar Reverse. (After H. N. Wright.) Iltutmish/as-sultnim yamin khalifat/allth nasir [amir]/[al-mu'minin].

P1. IIb. Billon coin with Nigart script legend dated V.S. 1283 and 624 A.H. [Sr]i sulti[n]/Lititimi/-si (After H. N. Wright.)

sam(vata)

1283. Reverse.

DEHLI SULTAN:

RECONSIDERATION

OF THE NAME-ILETMISH

OR ILTUTMISH?

61

conserves a valuable mass of oral tradition supplementing the surviving prose chronicles of the Dehli Sultanate. The Futi7hal-saldatin has been twice published, once in a lithograph based upon the India

Madras I948), who also made use of a second known manuscript in a private collection in Haidardbdd, Deccan. It may be observed that the same metrical considerations hold good in the case of this poem as in the two qasidasreproduced in the Tabaqdt-i-Ndsiri, which have been discussed above. Either ,?I or ~t lacks a necessary heavy first syllable. is required: ,. Usha throughout prints the name as .-'-. It occurs in the poem no less than sixteen times. The India Office manuscript (Ethd no. 895, India Office 3089) is available for collation; it is written in a fine and fairly careful small nasta'liq. It may date from the fifteenth century and is not likely to be later than the sixteenth.23 In one instance here the personal name is omitted in a shorter version of the rubric than that given by Usha, evidently from the HaidarabSd manuscript, although he does not note the variant (India Office manuscript, folio 82b; corresponding to Usha, p. 130). In two instances the name is spelt in rubrics t.-Ji (folios 8ob and 89b; corresponding to Usha, pp. 126 and 143). In the remaining thirteen instances, including every occasion where the name occurs in the verses, it is clearly written and dotted in the manuscript as tall (folios 53a, 59b, 64a (bis), 64b, 7ob, 7Ia, 74b (bis), 82a, 82b, 84a, 86a and 9ob; corresponding to Usha, pp. 74, 78, 94, 95 (bis), Io7 (bis), I14 (bis), 128, nowhere occurs. The evidence therefore is very strong that in 132, 136 and 145). The reading the fourteenth century, in circles descended from the nobility of the earlier Dehli Sultanate, the name t,"-I was remembered as Iltutmish.

Office manuscript alone by A. Mahdi Hussain (Agra 1938) and once by A. S. Usha (University of

II.

MonumentalEpigraphs Bayur surveys the inscriptions of the reign of Iltutmish as published by Horovitz,24 from whose article several have been reproduced in the RipertoireChronologique d'EpigraphieArabe.25 Bayur reproduces a , page from the latter publication, showing that Horovitz has actually read in one place %.41,26 and also many of the plates of Horovitz's first article. With regard to the page reproduced by Bayur from the Ripertoire it may be noted that the " (sic) " which follow 0"..I and Z0JlJIare those of the editor of the publication, not of Horovitz. The plate published by the latter shows two " teeth " between the ldm and the mim and his translation reads Iltutmish. The omission of the second td in the text of the inscription as printed in Epigraphia Indo-Moslemicais quite possibly a misprint. Elsewhere, the readings Iltutmish, Iltutmish, are constantly adopted by Horovitz. Bayur on the one hand complains that the dots of two tds are not visible on the inscriptions, and on the other hand explains away the fact that there are two, often sharp " teeth " visible in the inscriptions between the ldm and the mim, maintaining that the second of these is the beginning of the letter mim. (There is no evidence elsewhere in the inscriptions of the Dehli Sultanate of this epigraphical peculiarity of the letter mim.) " Bayur, however, maintains firstly that there are two teeth " in place of the one visible between the mim and rd of amir, and between the two ni7nsof mu'mininin Horovitz's pl. XVIII, his own fig. 13; and secondly that additional " teeth " are also visible in al-musliminin Horovitz's pl. XXIX, his own fig. 12, and in iftikhdr, the word preceding this on the same inscription (though he erroneously refers here to Horovitz, pl. XVIII, his own fig. i3). Of these examples, the case of iftikhdr is doubtful, but from an examination of the original plate what appears to be an additional " tooth " is most probably the reflection in the inked estampage of some damage to the surface at the corner of the inscription. In other cases cited by Bayur, no additional " teeth " are visible to the present writer.
23

24

Eth6, p. 559, no. 895, assigns it to " the ioth century of the Hijrah ". The small, vigorous and accurate nasta'liq hand makes me incline towards a slightly earlier date. The main body of the inscriptions of the reign are published by J. Horovitz, " Inscriptions of Muhammad Ibn Sam, Qutbuddin Aibeg and Iltutmish " in EpigraphiaIndo-Moslemica(Calcutta of which 1911-I2), pp. I2-54 and pls. I-XX, XXIII-XXX,

25

concern pp. 21-34 and pls. XIII, XV-XX, XXV-XXX inscriptions of the reign of Iltutmish. d'EpigraphieArabe, ed. E. Combe et al., RipertoireChronologique

26

J. Horovitz, " Inscriptions of the Turk Sultans of Dehli " in

at on VIII: VI. Inscription the backwall of a mosque Gangarampur, Malda (the inscription is of the reign of Ndsir al-Din Ma.hmfid).
Epigraphia Indo-Moslemica (Calcutta
1913-14),

vol. XI, p. 2.

p. 21 and pl.

62

JOURNAL

OF PERSIAN

STUDIES

The most remarkable of Bayur's explanations occurs with regard to the monumental epigraph which conflicts most strongly with his thesis (Horovitz, pl. XXVI; Bayur, fig. I6-a blurred reproduction of an initially unsatisfactory inked estampage). In this inscription the name of the Sultan "-I-the two tds having visible dots and theyd lacking them (see my P1. Ib). appears unmistakably as A out that the inscription is " upside down " and suggests that this is evidence of interference Bayur points and of restoration during the period of British rule in India. According to Bayur, Horovitz rightly ., from the position in which it appears in the illustration, restores the phrase jt11j ,LJ1 that is, after the name of the Sultan, and places it before the name of the Sultan, the position in which this title is found in other inscriptions. The inscription is unworthy of credence because it has suffered interference. He does not state plainly that the second td has been added by the conservation department of the Archaeological Survey of India, but if his argument is pursued, it is impossible to resist this inference. However, Bayur has not realized that Horovitz's pl. XXVI, which is the photograph of cut strips of the inked paper estampage, is not an accurate representation of the physical appearance of this inscription. The inscription runs in a thin band around the whole circumference of the lower of two stringcourses upon the northern minaret of the Arhdi din kdjhomprd at Ajmer (P1. Ia).27 Even in Horovitz's plate the estampage is of poor quality; this is probably the result of the difficulties of making it by leaning out perilously over the top of this minaret.28 But as it runs around the entire circumference of the minaret, there is no question of the inscription being restored " upside-down " during the conservabefore ala .1 without tion of the Archaeological Survey of India, and one may read i, WU8L ".&jI emendation. The Archaeological Survey has consolidated the upper platform of the minaret at the height to which it survived by the beginning of this century; and, as may be seen to the right of P1. Ia, a single carved and inscribed block of this string course which had fallen away has been replaced with an uncarved stone block. There is no evidence whatsoever that the Archaeological Survey recarved, or indeed had the ability to recarve, the name in the inscription.29 LegendsuponCoinage Bayur's arguments upon coin-legends follow the same lines as his arguments upon the stone epigraphs. He admits that two " teeth " are often visible between the ldm and the mim, and then dismisses this as without significance, because nowhere on the coins are two pairs of dots over the two " teeth " visible. This absence of dots upon coins will appear unconvincing evidence to anyone conversant with Muslim numismatics. Muslim coinages derive from the undotted or scantily dotted issues of the Umayyads and 'Abbasids, and it is difficult to find a completely and correctly dotted Muslim coin of a date before the sixteenth century. Bayur puts much emphasis on a small copper coin reproduced by him
27 Arkai is written incorrectly for arhd'f in Bayur. The name

means in Hind6stani " the hut of two and a half days ", a popular name bestowed afterwards on this great congregational mosque of the first Dehli Sultans. The name has been thought to refer to the occupation of the deserted mosque for brief periods (perhaps during the 'Urs of one Panjaba Shah) by wandering faqirs; while according to another account, the building was erected by spirits in this time. See H. B. Sarda, Ajmer (Ajmer 1941), p. 69; and Percy Brown, Indian Architecture: IslamicPeriod,2nd edn. (Bombay n.d.), pp. 12-13. Yet another popular explanation is that, the building being originally a Hindu sacred institution, the conversion into a mosque by Shihdb al-Din Gh5ri (Mu'izz al-Din Muhammad b. Sam) was accomplished in this time. See Munshi Muhammad Akbar Jahdn, Ahsan al-siyar, Mufid-i-'Amm Press (Agra p. 88. 1320/1902-03), 28 In 1963 the writer tried, without success, to photograph the personal name of the Sultan by leaning over the top of the minaret. The photograph here published of the name (P1. Ib) was taken with a telephoto lens from the roof above the northern side arches of the mosque. It was against the light, the sun being directly behind the minaret; in spite of this, the two tds

remain sufficiently clear. Like the rest of the lettering they are cut in very deep relief. 29 The value as evidence of Bayur's fig3. 8 and 9 also calls for comment. These are reproductions of the lithographed eyecopies made by (Sir) Syed Ahmad Khan and published in the latter's celebrated tadhkira of the buildings of Dehli, Athdr al-sanddid. They appear to have been reproduced by Bayur (figs. 8 and 9) from the lithographed plates of the 4th edn. (ed. Muhammad Rahmatu'llah " Ra'd ", Nami Press (Kanpfir 1904), concluding portion, pp. 38, 40). One may incidentally compare the eye-copies on pp. 3 i, 34 of the same edition'where two " teeth " are visible. The Athdral-sanddidand its illustrations were originally published in Dehli in 1847, and Sir Syed was engaged in the preparation of the work in the years immediately preceding this, many years before the conjectures noted earlier in this article had been put forward regarding the correct form of the name. The inscriptions are on the third and fourth storeys of this very high minaret, and Sir Syed's friend and biographer, Mawlana Altaf Husayn " Hali " has vividly described the difficulties which Sir Syed, not the most skilful of draughtsmen, experienced in making these copies seated in a hanging basket edn., Mufid-i-'Amm Press [Agra], p. 45). (HIli, Haydt-i-javd,a2nd

DEHLI

SULTAN:

RECONSIDERATION

OF THE

NAME-ILETMISH

OR ILTUTMISH?

63

on a plate taken from an early article of Edward Thomas (Bayur's fig. 5, coin no. 23), on which two large circular dots appear, one above each " tooth " between the ldm and the mim. However, on this coin-face the name is written in such a way as to omit (upon the die) the tail of the shin; the legend is almost as stylized as that upon some other small copper 'adlis reproduced in the same plate, where it is possibly deliberately intended to be readable as either Iltutmishor Shams (cf. nos. 21 and 22 on Bayur's plate with nos. 14 and I5). Besides this early article of Edward Thomas, Bayur cites the British Museum and Bodleian coin of catalogues,30 but makes no reference either to Edward Thomas' Chronicles the Pathan Kings of Dehli31 or to H. Nelson Wright's definitive work on the coinage of the Dehli Sultans,32 which, although published fourteen years before Bayur's article, may not have been available in Turkish libraries. As we have seen, Thomas was the first to discuss the problem of the name of the Sultan. Besides quoting the folk-etymology given by Badayini, he drew attention to the Nagari coin-legend whose significance we shall examine below.33 Lane Poole-to whose work Bayur had access-also drew attention to it.34 H. N. Wright, whose work Bayur did not consult, discussed in detail the variations of the name on the coinage, quoting most earlier contributions to the subject.35 Wright concludes: " Though no less than four forms of the name appear on the coins, viz. J1 and MLLI,'-;4I the four is by far the commoner." the last of In the coin trays of the British Museum the present writer has examined the silver tankasbearing the Sultan's name (including a posthumous coin issued in the reign of his daughter Sultan Ra2iyya). The examination yielded the following results: Nos. 153, 154, 159, I6o, 161, 163 and I64 read oI. No. 156 reads JI. Nos. I55, 157, 158 and 162 read J.l36 Attention should perhaps be drawn to the unique silver tanka published by Edward Thomas after his main work and also reproduced by Wright.37 The legends are in square kific script; this feature sets it apart from any other issue of the reign and indeed of the Sultanate, and demonstrates with peculiar clarity (if such a demonstration be needed) that the second " tooth " is not a part of the mim (P1. IIa). There remains the Nagari coin-legend (P1. IIb), discussed by Thomas, Lane-Poole and Wright, but ignored by Bayur (pl. 38).38 As is quite clear from the reproductions, as well as upon the two specimens in the British Museum collection examined by myself,39there are unquestionably two ts in the middle of the name: this reads f"f"id" Lititimisi, and represents a crude North Indian attempt to record the alien sounds. The present writer doubts Wright's opinion that the full legend should read Ilititimisi: the initial vowel and consonant have probably suffered metathesis in the vernacular pronunciation.40 But
30 S. Lane Poole, op. cit., and the same, Catalogueof the Moham-

in medanCoins Preserved the Bodleian Libraryof Oxford (Oxford 1888). 31 E. Thomas, op. cit. 32 H. Nelson Wright, The Sultans of Dehli: Their Coinage and Metrology(Oxford 1936). 33Thomas, op. cit., p. 44. 34 Lane Poole, Sultansof Dehli, loc. cit. 35 Wright, op. cit., p. 7o. 36 The numbers given are those now to be found on the paper discs upon the trays, corresponding to those of the forthcoming augmented British Museum catalogue. " The Initial Coinage of Bengal, Pt. II: Embracing 37 E. Thomas, the Preliminary Period Between A.H. 614-34 (A.D. 1217" 1236-7) in JRAS, New Series (1873), vol. VI, p. 350, no. 8. Wright, op. cit., p. 17, no. 49K and pl. XXII, then in Berlin; P1. IIa of this article. The author is most grateful to Miss Janice Cornwell of the photographic staff of the School of Oriental Studies, for enlarging and preparing Pls. IIa and b at very short notice. 38 Bayur, fig. 5, no. 14, reverse = Thomas, Chronicles,p. 74, no. 44; Wright, op. cit., p. 3o, nos. 121, 122; cf. also 122A.

39 Lane Poole, op. cit., p. 15, nos. 46, 47, pl. II, which figures these actual specimens. 40 Wright, op. cit., p. 3o, reading the reverse of nos. 121, 122: cf. the actual legend as visible on my own Pl. IIb, Wright's pl. II, also pl. XX. Cf. also such Nagari renderings as amiralimaumnindhi for amiru'l-mu'minin upon other coins of the reign (Wright, p. 26, nos. 68-72). On Pl. IIb the beginning of the name is not clearly visible. In medieval Indian minting the bolt of the coin was almost invariably smaller than the surface of the die; there is no doubt of the syllable Li from other specimens, although these still leave it uncertain whether there was an initial vowel. The Nagari inscription of this issue of Iltutmish is recalled in the Dravya pariksha, a remarkable treatise in Prakrit verse by Thakkura Pherfi, mint-master of 'Ald' al-Din Khilji (625-7 15/1296-13 16) at Dehli, writing in the reign of his son Qutb al-Din Mubarak (716-20/1316-20): " The coins called Titimisi,Kuwwakhdni, Khalifati,Adhachandd and Sikandari,which are all Samsi coins, contain 3 tolds of silver (per Ioo specimens); their rate is 34 per Tanka." See V. S. Agarwala (sic), "A Unique Treatise on Medieval Indian Coins " in H. K. Shirwani (sic), ed., Dr. Ghulamrazdani Volume(Hyderabad 1966), p. 97. Commemoration

64

JOURNAL

OF PERSIAN

STUDIES

it remains highly improbable that the ti syllable should have been mysteriously reduplicated, if the original name was Iletmish.41 Thus we find that in spite of the length at which Bayur has presented his arguments for a reading which he was possibly originally driven by criticism to defend, a large preponderanceof all the types of evidence which he surveys,viz. manuscriptreadings, monumental inscriptionsand coin-legends, in fact support the reading which he was attempting to refute. POSTSCRIPT While this article was in the press Mr. John Burton-Page called the writer's attention to a curious qasidain the metre ramal-imuthamman-imahdhaf, evidently addressed to Iltutmish by one in 626 A.H. al-din " Bazzaz " Dehlavi on the occasion of the arrival of the Caliphal manshfir Mawlanm Burhin "I No other form except the highly improbable (1229 A.D.). In this the name is given as
will suit the metre, which requires u-. .l.

Bu'l-muzaffar iltutmish k' d bi-hukmaz tir tir tir Bugdhardnaddarfavdlash (sic) me-biydyad tir.

The scansion of this couplet is more obvious than the sense.42

41

A. B. M. Habibullah in The Foundationof Muslim Rule, 2nd revised edn. (Allahabad I96I), pp. ioI and IIO, note 77, refers to the Manglina Sanskrit inscription mentioning Lititimishi (sic) of Joginipara (= Dehli); the correct Suratrdna reference for the place of publication of this inscription is The Indian Antiquary,vol. XLI (not XVI as in Habibullah) (1912), p. 87 (article by Pandit Rama Karna, " Manglana Stone Inscription of Jayatrasimha "). The name regrettably appears there as svaratdnasri (sama)saddna(rather corruptly written and pedantically restored by the editor as suratrdna

sri shamsuddin). Probably Professor Habibullah was led astray by his memory of the coin-legend. The present writer has failed as yet to find any alternative published monumental epigraph with a rendering of the name N.igari Iltutmish. 42 Nazir Ahmad, " Some little known Indo-Persian poets of the thirteenth century " in H. K. Shirwani (sic) ed., Dr. Ghulam Volume (Hyderabad 1966), p. 163, razdani Commemoration quoting from [an unspecified manuscript or printed edition of] " Mu'nis al-ahrdr,p. I08o ".

SAKA STUDIES: THE ANCIENT KINGDOM OF KHOTAN By H. W. Bailey

To understand and to interpret according to the meaning of the composers, hence in historical wise, the many ancient documents in languages which we conventionally call Iranian (or Eranian or Iranic, from the name of Eranshahr or Iran, one of the ancient centres of the region where these languages have been recorded for over 2500 years), intense research has been necessary. So much has been lost, not only from the ancient vocabulary of the texts themselves, but also from the vocabularies of modern dialects, such as those archaic ones in Afghanistan, particularly in the Pamir mountains, and in Ossetic of the Caucasus, for here vestiges of the old language survive. But larger help can be found also in the words borrowed or quoted by peoples adjacent to Iran as foreign words (often proper names): in the Akkadian of ancient later Babylon; in the Elamite of Persepolis; in ancient Egyptian; in Greek and Latin abundantly; in Armenian from Arsacid times; in the Georgian, Abkhazian and Daghestanian languages of the Caucasus; in the Aramaic and Syriac of Syria; in early Arab poems and later Muslim books; and further east, in the languages of Kuci (Kucha) and Agni (Qara-shahr) (languages frequently called Tokharian, from a widely extended and imprecisely geographical use of the name), and in Chinese and Tibetan. The reliability in turn of these various sources of information depends upon the scientific level of their publication, and here much remains to be done. The oldest consecutive Iranian texts for which dates are certain are in the Hakhamanishiya or Achaemenid inscriptions. To the texts in Old Persian in C. Bartholomae's dictionary of 1904, the AltiranischesWi'rterbuch its supplement, and in R. G. Kent's Old Persian,some new recently-discovered and are now to be added, including one of Xerxes still only incompletely published by B. inscriptions Qarib, a text similar to that of Dareios at Naqsh-i Rustam (DNb). Many names evidently of Persian origin have been found in Elamite tablets from Persepolis, for a definitive edition of which we still wait upon R. T. Hallock. It is fascinating to have Yamakhshaita (written yamakshedda),the later Jamshid, from one of the Elamite tablets, with Yama in a spelling older than the Yima of the Avesta. A new study of all the Old Persian texts is expected shortly from the pen of E. Benveniste. Next in time, though we have no external dates, the earliest part of the sacred books of the Zardushti of Iran and the Parsis of India, the Gathas, has the marks of great age, and is placed (although the evidence is not decisive) at various dates between times from Iooo to 550 B.C. The Avesta itself is a small part of a large collection of texts which were extant till the end of the Sdsdnid period. The name Avesta, in older spelling 'pst'k *apastdkor *apistdk,has been traced to *upastdvaka-" praise " or apastdyaka " commandment ", without final decision. The Avestan texts have excited great interest in Europe for two hundred years. Intense application has been devoted to their study and in general lines the understanding of these Avestan texts is assured. But in many details further insights are necessary. Apart from the religious aspect, some of the ancient poetry of the Avesta is admirable and there are in it many ingredients of the later epic. Another less ancient study which made some progress last century is that of the Zoroastrian or Zardushti Pahlavi texts of the Sasanid period, which were later transmitted in Iran and among the Parsis who emigrated to India. Here work had advanced slowly, the tradition was insecure, and the highly-developed script, which kept only fourteen signs distinct out of an original twenty-two, seems to have been satisfactory only while the language was current in Sasanid times. We should, however, note that in Arabic script krd suffices for five different spellings in Sasanid words, where krt, kwrt, grt, gwrt represent in Firsi kard, -gird (as in Yazdgird), kurd,gard, gird and gurd. This state of near-defeat was suddenly transformed by the unexpected discovery of many manuscripts written in clear scripts in Central Asia, that is, in Chinese Turkestan, from Kashghar in the west, eastwards to Shachou, and from Kucha in the north, south to Khotan, at the end of last century and 65

66

JOURNAL

OF

PERSIAN

STUDIES

the beginning of this century. The story of these investigations and discoveries is told in J. A. Dabbs, History of the Discoveryand Explorationof ChineseTurkestan(1963). The famous names here were M. Aurel Stein, A. Griinwedel, A. von le Coq, Sven Hedin, P. K. Kozlov and P. Pelliot. Sven Hedin, in a preface to a book of 1931, reported the discovery of heaps of manuscripts in six languages, one probably unknown, from Qara-Khoto; these are still stored away and likely to delight a later generation of scholars. But much was also brought to the museums of Pekin, Tokyo, Kyoto, Delhi, Stockholm, St. Petersburg (Leningrad), Paris, Berlin, London and the universities of Harvard and Yale in the United States. A great part of these texts, an astonishing number in relation to the experts in the field, has now been brought out in numerous publications. From western Iran, the religious books of the Manicheans in Persian and Pahlavanig (so the eastern language is named in the texts themselves, for which we may say Pahlavi or " later Parthian "), were found, though often in disappointingly broken pieces. The script was a clear form of the west Asian Syrian alphabet where each letter was distinct. It had, however, no consistent means to indicate all vowels, so that a fully vocalized text is partly conjectural. Some of these texts were written also in Sogdian script and other words and titles have been found written in the Chinese syllabary. Mani himself was said to have written only one book in Iranian, the Shahparakan Shdhpuhr I (who died in for 273), and the rest of his scriptures in eastern Aramaic. But he encouraged translation, and among other languages, a great part of his writings has been found in the Coptic language of Egypt. The two dialects of Sisanid Persian and Parthian are easily distinguished by phonetic differences, though the vocabularies are basically the same. The Central Asian texts proved invaluable in the interpretation of Zoroastrian Pahlavi texts, with their ambiguous written forms. A part of the Hebrew Psalter was discovered in a Christian Persian translation from Syriac, in a more archaic Syrian script and full of heterograms of Aramaic words as in Pahlavi texts. With excellent facsimiles, it was interpreted by Kaj Barr. The same scholar also edited a fragment of a glossary of Aramaic verbs explained in Sasdnid Iranian, of the type of the Zardushtifrahang or glossarial text, the Frahang i pahlavik. An unknown quantity of these Central Asian western Iranian texts remains unpublished more than half a century after their discovery. Of great importance for western Iranian are all the Sdsdnid inscriptions, for which the Corpus iranicarum will eventually be a complete repository. But progress there has been slow since inscriptionum the Council was inaugurated in 1954From Central Asia, and more recently also from a castle, Mug in the Zarafshan valley near Samarkand, have come many manuscripts in the Sogdian language, which has ceased to be spoken except in the one valley of YaghnTb in ancient Sogdiana (Sugda, Sughd). Ancient letters, not yet satisfactorily interpreted in full, were discovered by Aurel Stein and published by H. Reichelt. They date from around 312-313 A.D. Manuscripts from the scriptoria of the three communities of Buddhists (the most archaic dialect of the three), Manichaeans and Christians, are fairly abundant. They have been partly published, many with facsimile plates. A Sogdian inscription on a pts'k or monument of the Uigur Turkish emperor Ai tingridi qut bulmi' alpu bilgai payi Uigur qayan (ruled 808-82I A.D.) was found with Chinese and Turkish parallel texts in Qara-Balgasun. The Buddhist Sogdian is particularly valuable for its early spelling, where the words have largely forms equivalent to the Achaemenid period in west Iran. A smaller amount of Iranian dialectal material belongs to the Ku6an period of Afghanistan from about 100 B.C. to about 400 A.D. Here there are coins with Greek writing, and there is the recent discovery of inscriptions in Greek script which have perhaps too confidently been called Bactrian from the ancient name of the region. The Tokhari people had come to the west by ioo B.c., and the ancient Bactria eventually received the new name of Tokhiristan. On Hiuan Tsang's visit in 641 A.D. he spoke of Tokhiristin and the local language written in Greek script. Inscriptions found in ancient Bactria could easily be in Tokhitri. A new source of eleventh-century Iranian was first made known in the twenties of this century from the ancient Chorasmia, later Khvirazm and modern Khiva, in the language called in Arabic khuwdrizmi. This new material is in manuscripts in Arabic script and is of Muslim origin. The famous lexical work of Zamakhshari, the Muqaddimat al-adab, was glossed in Chorasmian. A vocabulary of Chorasmian

SAKA

STUDIES:

THE

ANCIENT

KINGDOM

OF

KHOTAN

67

technical terms with Persian and Arabic equivalents has also been found. Al-Bairfini was from Khvarazm and his own name is explained by al-Sam'dni as the Persian btiran Chorasmian anb'~iak for " outside (the city) ", a connexion of Sogdian bik " outside ". Much remains to be published before this Chorasmian dialect will be fully understood. A recent work of J. Benzing (1968), offering the transcription of the Chorasmian glosses and a translation of the Arabic text, will greatly assist towards its interpretation. The frequent absence of diacritic dots at times makes the Chorasmian uncertain. Earlier (1951) a facsimile volume of the Muqaddima was printed in Istanbul. A book by A. Freiman called Khorezmijskijjazyk " Chorasmian language ", also published in 1951, discussed many problems. Otherwise we depend on a number of articles on special points in periodicals. Less copious materials are also preserved as loan-words in Central Asian languages. In the kharostri or kharosthi script (a name in which I have been inclined to see khshathra- (xsalra) " empire " and pishtra- (piftra-) " writing ") the Prakrit and Buddhist Sanskrit manuscripts of Kroraina (modern Lou-lan) which can be dated about 300 A.D., contain many Iranian names, as Tiravhara, like Tiravharna in a Ktbul inscription, and Livirazma, Lyiparasma, but also such common words as divira " writer, secretary ", the dpywr of Sasanid inscriptions; aipista " lucerne ", Akkadian aspastu, Farsi aspast; nacira = ic) " hunt ", Sasanid inscription nklyrpty" hunt-master ", Sanskrit nascirapati-; avimdhama (" penalty ", Sogdian 3nd'm*vinddm; nok-sari " new year ", Armenian loan-word nava-sard,where nok " new " is like Pahlavi nwk, Ossetic ndudg, nog; cozbo " an official title ", Tumshuq Saka cazbd-; gamni, gamhavara" treasure, treasury ", Old Persian ganja-. Among the Yueh-chih (Ue-tsi) in 90 A.D. we hear in a Chinese book of a title za or .a for their subordinate ruler, the same word with the same meaning as the Khotan Saka sau. This people was identified by Kumarajiva of Kuci, who lived 344-413 A.D., with the Tokhdra of Buddhist Sanskrit texts, and, since there is no evidence of this Ia being a loan-word here, the Tokhdra of Buddhist texts were likely to be Iranian-speaking, which is supported by the presence of the sound fricative kh (as in Arabic khd), attested by the various spellings of the name, a sound absent from the languages of Kuci and Agni. Tokhiristan replacing the ancient name of Bactria, had in al-Bairini's time a language cited as tukhdri,but the few words quoted are plant names and do not suffice to assert that this was an Iranian dialect. The fragments of Buddhist literature found in Central Asia in Greek script of about Hiian Tsang's time would seem to belong to Tokhdristdn. In the languages of ancient Kuci (Kucha) and Agni (Qara-shahr), some Iranian words have been noted. The word amok " art " is not found in the Saka of Khotan (it is in Pahlavi dmik " learning "), but other words indicate a Saka dialect. There occur Kuci asdm (pronounced ashdn), Agni dsdm " worthy ", the Khotan Saka dsana- and Kuci asanike, Agni dsdnik " the arhant, the worthy monk " is from Khotan with the suffix -ik. Kuci perne,Agni parn- " fortune ", translating Sanskrit laksmi, is nearer to Sogdianfarn- than to Khotan phdrra-. Kuci ahkwas is the Khotan amgusda-" assa fetida ". Central Asian Turkish texts also show many Iranian words. Here we have animan " meeting ", Pahlavi hanjaman; anwant " cause ", Sogdian nflnt; mobag " teacher ", from d-mal- "to teach "; ton " garment ", Khotan thauna-, Ossetic tund; vaksik " spirit ", Pahlavi vaxv; and others. In the book of Mahmiid al-Kdshghari, the Diwdn lughdt al-turk, sixteen words are quoted from a dialect spoken in villages around Kashghar which he called kaniaki. This name is in Tibetan Ga-hjag (pronounced Ganjag, from Kantak), a name for Kashghar. An obscure context in a fragment from which seems likely to be Konow 8, Murtuq, near Qara-Khoja, has the word (edited Sten k"dIcake, 3) this same name. From Ga-hjag came a princess to marry the king Vijaya Sirnha of Khotan, as we read in the Annals of Li (Khotan). Among these sixteen words of Kaniaki only one, the word kiinbd means " flax " Sname of a plant ", can be traced. In Sogdian kenba, and in Khotan, kuzmbad Saka has been mentioned several times above. It now demands attention. It is a field in which alone I can speak with confidence since it has from 1934 kept me busy for thirty-six years. Here too the documents belong to Central Asia, from two regions, Khotan (though many of the manuscripts were found in Tun-huang), and Tumshuq, a ruined site near Maralbashi, formerly Bariuq. In the many texts the language is called hvatanau,hvatiinau,hvatanoin the older language of Khotan, and later " the language of Khotan ", with the suffix -dva-, and in the phrase hvanau, in the " in the hvam.no language " with suffix -aka-. The name of the kingdom is Khotanahvam.nye Khotanese pharijsahvam.znau,

68

JOURNAL

OF PERSIAN

STUDIES

kharostri document (No. 661) dated in the reign (ksuna) of the khotanamaharayarayatirayahinaza dheva " vij'ida sirzmhasya the great king of Khotan, the king above kings, the army commander, the god, Simha ". In the local texts, the older ones of about 500 A.D. or earlier have the spelling Hvatana, Vijita which later is replaced by HIvannaand Hvam. This later form is known from the Huanna of Hiian Tsang on his visit in 641 A.D., which he reported as the indigenous pronunciation of his time. In Chinese we find Yii-t'ien (whence the Khotan writers also took Trttind) and in Tibetan rvu-then and Hu-then. Here the older -t- has been kept as in Fdrsi Khotan. In Buddhist Sanskrit texts this Hvatana was given an archaic Indianized form Gostana. This was then etymologized as a Sanskrit word with go- " earth " and stana- " breast ". Then this gostana in the Tibetan Annals of Li (Khotan) was rendered by Sa-nu as the name of a boy destined to be a king of Khotan. A tale was then constructed how the boy was suckled by a breast arisen out of the earth. In the Buddhist Sanskrit of Kuci the name was written Korttana. In western Iran the name Khotan is known to all readers of Firdasi's Shdhndmathrough the legends of Siyavash, though these stories are not known to the documents from Khotan. The Persian Geography of the tenth century, the Hudad al-'Alam, knew the region of Khotan and the regions beyond with accuracy. A tale, however, of Rustam has been found in a Sogdian manuscript. In the writing of the documents of Khotan and Tumshuq the script employed was the Indian variety of north-west India, a form of syllabary which we call the Brihmi (a name taken from an ancient Indian text). It expresses all the vowels, and these texts are the earliest Iranian texts fully vocalized. The Brahmi syllabary, however, was inadequate in consonant symbols. Eventually, twelve new signs were invented. But in Khotanese a system of conjunct signs was used, so that, for example, they wroteys for z, and js for dz, making a writing of great complexity. The two dialects of Khotan and Tumshuq are clearly of one language but have striking differences. Such a word as fra- " forward " is replaced in Khotan by ha-, but in Tumshuq by ra-; hence the difference of Khotan hada- " given " and Tumshuq rorda-, from an older fra- with the base bar- " to bear ". Other differences are evident in Khotan diadasu" eleven " and Tumshuq sowarsanaand Khotan pamtsdsd " fifty " and Tumshuq patsasu. The nearest living dialects to these Saka languages are those still spoken in the Pamir mountains, particularly the Wakhi of Wakhan. The Greek writer Ktasias in the time of Artaxerxes II (404-358 B.C.) is quoted as stating that R5xanakE was the name of a polis " town " (as distinct from an unfortified k6mi " village ") where the Saka had a royal seat, a basileion. This name is still known in the Pamir in the region R6shndn, R6shin, where the language is one of the large Shughni group in the Vaksu (Oxus) region. Though direct contemporary statement is not found, it is likely that the two languages of Khotan and Tumshuq were spoken by two tribes of the Saka who about 200 B.C. or earlier settled with a monarchical or oecarchical system in this region of the Taklamakan. From the second century B.c. there is Chinese information on Khotan: no major invasion is recorded. The Saka are named in the Achaemenid inscriptions. Dareios (Persepolis h 5-6) stated hacd sakaibil tayaiypara Sugdam" from the Saka who are beyond Sugda ". He listed also (Pers. 18) Ganddra Sakd Maka; similarly Xerxes (Pers. h 26) cited Dahi Saka, where the Dahd belong to the east of the Caspian Sea, in a district later called Dihistan. They are an ancient people already named in the Zardushti text, the Avesta. Herodotos (vii 64) wrote oi yfip lIpeat advrav tobg ~K65 i KaXouom " LaaKy for the Persians call all the Skuthai Sakai ". From Chinese Annals we hear of the Sak (later pronunciation S6 and Sai) who had a kingdom in the second century B.c. north-west of Kashghar. The Saka people penetrated into ancient Drangiana (Zranka) and their name has remained attached to the land ever since. It is known to Isidoros Kharakenos (in the time of Augustus) as XaKatcavfi, and later in Shahpuhr I's inscription at Naqsh-i Rustam (Parth. 2) skstn, Greek Xr~yracvn; of the people and of their king, which gave in Latin Agathias (third century A.D.) used E'saravov Segestani. But on the Roman map the Tabula Peutingeriana, Saga was written. In the kharostri inscription on the Mathura Lion Capital, sakastana (k = gh) has s-, but Sanskrit writers replaced this by i- in Sakasthana and Saka and knew also Sakdnilipi " writing of the Saka ". In Prakrit we find Saka and Saga. In Iran itself, Zardushti Pahlavi had (Greater Bundahishn 78, 15) sgst'n, or one may read the later

SAKA

STUDIES:

THE

ANCIENT

KINGDOM

OF KHOTAN

69

syst'n; the Sdsdnid inscriptions have sk'n MLKA sakdn-ldh,sk'n MLKTA sakdn bdnbisn,Persian of Turfan sg'n b'nbyin, Greek leyav naa;Sogdian Christian sgst'n, Syriac sgstn, with adjective sgzyq-, The local texts of Khotan and Tumshuq do not use the name Saka of themselves. But in Central Asian Buddhist Sanskrit a Saka-raja " a king of Saka " is named which was perhaps the king of Yarkand, of which place the ancient name was something like Saka, preservedonly in Chinese syllables and a corrupted Greek text. In Khotan texts an adjective sakdna-occurs in personal names as in Sakdfia Sanirii, that is, " Sanira of Sakana " and Sakdm hivi " belonging to Sakana ". The Brahmanas of India wrote of a northern dvipa-" continent, region " which they called Sakadvipa, from which they reported two immigrationsinto north-westIndia, some 2000 years ago, of sunworshippers. The former group, worshippersof Mihira (a later Iranian form of Mithra), had a famous temple in Mfilasthana (Multan) with a golden statue described by Hiian Tsang at his visit in 641 A.D. The maga,their priests, named from the ancient title magu-,were admitted among the Brahmanas as Maga-brahmanas, the highest intellectual class. The second group, who revered a teacher Jara'astra, " a Sanskrit form of Zardusht, had priests called bhojaka ministrants of the god ", and they, being unable to accept the Vedas, were not admitted among the Brahmanas. Sakastdn, Seistan, became famous in the Iranian epic as the home of the hero Rustam (older form R6tastahm, meaning probably " strong in body "). He is called sagzi " of the Saka " and in Armenian Rostom sagvik. He belongs to Zdbul, which is named in SanskritJdguda. There seems little doubt that a wide-ranging tribe of the Saka came to Central Asia, and that the language of Khotan is a Saka dialect. I have adopted that view here. This wide-ranging of the speakersof Iranian languages is vividly shown in the period of European migrations in the fourth and fifth centuries A.D. Two tribes known to the Greeks and Romans, the Sarmatae and the Alani, penetrated to France and Portugal. The Sarmatae occur first in the name Sairima in the Avesta, whence it came into Pahlavi as sim, that is, Sarm or Salm, and into the Arabic books and the Firsi epic in the person of Fret6n's firstson. These Sarmatae from the region to the north of the Black Sea drove their cousins the Scythians westwardsand southwards. Then in the year 173 A.D., victor, of Sarmaticus. Of these Sarmatae, 8ooo were taken into the Roman army and from them 5500 cavalry were sent to the Roman province of Britannia as veterani.An inscription put up by them in northern Britain can now be seen in St. John's College, Cambridge. In France too the road through " the city of Rheims was once called the via Sarmatarum the road of the Sarmatae ". Yet another and still more important tribe, called Alani by the Greeks and Romans, maintained a great state north of the Black Sea and the Caucasus till the eleventh century. As Christiansthey were welcomed in Byzantion (Istanbul); an archbishop is mentioned, and the emperor Konstantinos " Porphyrogennetoslaid down a protocol for the reception of the exousiastis ruler " of Alania at his Court. The Alani who went west in the fourth century were allotted Lusitania (modern Portugal) by
the Roman emperor of the West, Honorius, in 409 A.D. the Roman emperor Marcus Aurelius (161-180 A.D.) defeated the Sarmatae and took the epithet, as Arab sajastdn, sijistdn, sagdn Mdh,Farsi seistdn, with adjective sagzi. Armenian has kept the older Sakastan, Sakstan and Sagastan, with adjective sag'ik and sadik.

From the north of the Caucasus, the Alani intermarriedwith the rulers of Sarir, and in the twelfth century one of them, David Soslan, became the consort of the Georgian Queen Tamar during the
Georgians' golden age. The name Soslan is later found as the name of a famous hero in the Ossetic Nart epic. The Mongol invasion destroyed the Alan state. Some of the Alan tribe called As were taken by the Mongols in the reign of Qubilai Khan to the imperial capital city Khan-baliq, modern Pekin, to serve as a palace guard. They are called Asutai in the Mongol Chronicle. The Alan were for centuries familiar to the Greeks and Romans and also to the Muslim historians. The name was given to the Caspian Sea, the bahral-LUn. Princes of the Arani, another form of the name, are mentioned in a letter of as late as 1459 A.D. In the Caucasus the name is not yet quite forgotten. The Megrel of Mingrelia in modern western Georgia call a brave youth an alani k'ol'i " an Alan man " A tournament is called alanuroba" a festival of the Alans " (-oba is a suffix to name festivals). They use also the name Alan of the Qaradai Turks, their neighbours. As a family name, Alan has survived in Abkhazian.

70

JOURNAL

OF PERSIAN

STUDIES

The Chinese also have preserved the name A-lan. It seems too that the Alan name was taken far to the east. According to a Chinese document dated 886 A.D., a god A-lan was honoured in the nonChinese city of Qomul (Qamul, Khamil) beyond Turfan. There also it was reported that amongst the inhabitants was a temple of the Mazdeans, the worshippers of Hormizd. It seems desirable to adduce here a feature of the phonetic system of Saka dialects. In Qomul the final syllable -ul recalls also the tribe in the Kanchou region named Cimuda, Cumuda, according to Khotanese Saka documents, who were named also by Al-Kdshghari umul; he stated that in the eleventh century they spoke Turkish but had also a language of their own. Now in the west, also, where we know Saka people lived, we find names ending in -ul in Zdbul, Sanskrit Jdguda, and Kabul, Greek Ka3oupa. To interpret an ending -ul we have to note also the Hfina royal name Mihirakula, with Mihira as an older Mithra; and this -kula- can be compared with the kert of the west Iranian name Yazdkert, in Armenian Yazkert, Turfan Persian yzdygyrd, Farsi Yazdgird. Further, a common noun must be cited. This is a word for " basket ", showing two forms *kamrta-and *camrta-,in Khotan Saka Pamir Yidgha kamio, Shughni cemud, Orosori camag (here c = ts) and Sangli iumol. The various spellings here with -1, -d and -1 show an unusual replacement of older -rta-. I am inclined khama.da-, to connect all these names and words as Saka dialect forms containing -rta-, although as always, where names are concerned certainty cannot be reached. Note that in Khotan Saka -d- and -1- alternate in the foreign word halalka-" drum " from Sanskrit hudukka-,and a local word lade " goodness " is once written salai. For older *krta- (as in Old Persian; Avestan karata-) Khotan Saka has kida-, hi.daka-, guqda-, yida-, yuda-. The language of Khotan is attested over a period of some centuries and great changes can be observed. The older type from 300 A.D. or earlier is particularly employed in religious translations. It has kept syllables which later are absent. Thus the Greeks cited Median turapog,which is older Khotan ttatara-, but later ttara- " partridge ", Farsi tadharv; similarly the older satd " hundred " occurs as se, sse, sa, ssa, but the numerals tcahora" four ", pamjsa " five " and hasta " eight " are not changed. At the time we first meet the Saka of Khotan and Tumshuq, it is still a highly-inflected language. The noun has seven cases. The differences of the two dialects can at once be seen: Khotan has nom. sing. and gen. sing. -a and -i, gen. plur. -dnu, but Tumshuq has nom. sing. -i, gen. sing. -d, gen. plur. -and, -enu. A trace of the old dual survives in duve sate " two hundred ". The verb has an inherited present indicative, conjunctive and optative; there are a few survivals of a preterite; but the past is expressed by the use of the participle -ta- with the verb " to be ", as in dtd md " I came " from older *dgatah ahmi. The transitive active has a participle form -tdnt- as inyuddndi sta " you made ". The full details of these inflections can now be read in R. E. Emmerick, Saka Grammatical Studies (1968). The Saka languages retained some of the archaic Iranian religious vocabulary even after the Saka had become Buddhists. It is used with new meanings. The old religious word *Iuanta- gave ysamas'andai " world" from zam- " earth "; without zam- the word &sandd was used for " earth ", and ssandrdmatarenders the Buddhist ri " goddess of fortune ". In Avestan the word was spanta drmaitil " the beneficent earth ". Tumshuq bdrza-, Khotan balysa- served to translate tathdgata, and buddha-. The ttaira hariiysa " the peak of Hardbrz " corresponds to Buddhist Sumeru the world mountain, in Avestan hard baraz, Farsi alburz. The Khotan urmaysddnmeans " sun ", but is connected with Ahuramazddh. So too gyasta-, Tumshuq jezda- is the old yazata- " a being to receive worship ", and gyaysnais the act of worship. Khotan phdrra-, Sogdian farn, correspond to Avestan xvaranah- and mean " fortune ". The epithet of kings and learned men kavi- occurs as Khotan kai, plural kd, used of the drya- " monks " and the rrispara " princes ". The full phrase zdy spanddrmatfor " earth " occurs in Sogdian, where Khotan has only the double phrases ysama-4sandaiand ilandrdmata. The contents of the texts from Khotan are of many kinds. The vihdras (Sogdian 3rx'r, Firsi buxdr) or monasteries as centres of literary activity produced the translations of the Buddhist texts from Prakrit or Buddhist Sanskrit. But it is clear that the royal court of Khotan encouraged the development of literature. The poem called Jdtakastava" praise of the Buddha's births ", originally in Buddhist Sanskrit, was rendered into Khotan Saka verse at the behest of the great king Vida fira; he and his family are praised in the introduction. The prince composed a desandor profession text in verse. Tci.m-ttehi

SAKA

STUDIES:

THE

ANCIENT

KINGDOM

OF KHOTAN

71

The Annals of Li (Khotan), preserved in a Tibetan translation, record that there were sixteen vihdras the Mahdsdnghika Buddhist school in Khotan. It gives also a statistical statement of the of to sixty-eight larger vihdras, ninety-five of middle size and 148 smaller vihdras be found in Khotan itself; there were others in the adjacent towns and districts. The literature of Buddhism translated in Khotan covered most of the genres familiar in Buddhist use elsewhere. There is mention of the various ritual and philosophic treatises. In translation we find sometimes whole texts and sometimes fragmentary remains of large books. They were fortunate to have behind them the long tradition of Indian book-making. Each folio was numbered and in the religious poems the verses are also numbered throughout. Large books existed. A solitary folio numbered 611 has survived. Other folios have numbers 255, 134-148, 152 and one text, the Book of Zambasta, has about 4000 lines of verse on 440 folios. Religious texts attested in Khotan are the well-known Buddhist Sanskrit Vajracchedikd-si7tra, Aparimitdyuh-satra, Saddharmapundarika-satra, Bhaisajyaguru-vaidigryaprabha-rdja-satra, Bhadracaryd-desand, two texts, an Amitdyuh poem, a kind of Sumukha-sitra, Suvarnabhdsa-sfitra, Samghdta-satra, Avalokitefvara and Kaniska, the two imperial text Karma-vibhanga on the penalties for actions. Tales exist of Adoka patrons of Buddhism, giving in brief form the tale of Adokaand Ya'as and of Adokaand Kunala. Of and stapa,and the tale of Kaniska and Advaghosa. Jdtakatales Kaniska, we have the tale of his vihdra relates fifty stories of the were as popular here as elsewhere in the Buddhist world. The Jdtakastava Buddha's powers of endurance. The tale of Sudhana and Manohard is told in three manuscripts in verse. Rima and Sita are transferredto the Buddhist world by the claim that Rama was Buddha in a previousjdtaka. The Book of Zambasta is a miscellany of philosophy and tales of the Bodhisattva. The ascent to idealist philosophy of the make a colourful tale. The citta-mdtra Tusita and the descent at SrmkdAya navada is boldly thrust forward in strong assertions. No reference to Vasubandhu's AbhidharVijfi has makos'a been noticed, in contrast to Kuci, where portions in translation have been found. The " Dharmapada of subhdsita moral saying " is representedby only two versesin Khotanese, which are type and also cited in avaddna tales. But in Kuci both the Uddnavarga a commentary are known. The was found in Khotan. The Sanskrit text has now been edited from parts of two kharostriDharmapada hundred different manuscripts. A text named after Manju'ri is peculiar in that it contains later versions of some part of the Book of Zambasta. The text has not been found in Sanskrit. But a taste for literature existed outside the vihdras, though we are indebted to them for the written traces which have reached us. In several manuscripts there are lyrical verses. The monkish scribe " tended to end the lyrical passages with a denunciation of kamaguna amorousness". These verses, not yet fully interpreted, show a delightful pleasure in gardens with flowering trees and crystal though fountains. The lovers drink the durausa drink, in which we see the old name of the ritual drink of the In the du~raoSa. the Sta6l-Holstein miscellany also we have a love poem copied into the manuAvesta, script at the end of an official report. A traveller'sdiary of a visit from Vi (probably Khotan) to the Adhisthana capital city of Kashmir, the modern Srinagar on the Vitasti river, shows that this type of record was known in Khotanese. To have a report dated to the reign of the Kashmir king Abhimanyugupta, who ruled 958-972 A.D., is a striking historical trouvaille.
On the technical side we find medical texts. At first, Buddhists were adverse to medical studies, as P. Demidville has pointed out in the Buddhist Encyclopedia Hab6girin (under byJ " illness "), but eventually they ascribed medical teaching to the Buddha. In one text, a bilingual Buddhist Sanskrit text with Khotan translation, which, to provide a title, I called Jivaka-pustaka " the Book of Jivaka ", represents the Buddha teaching Jivaka, the personal physician of the Buddha in the tales. A further medical text translated into Khotanese with an indigenous introduction is the Siddhasdraof Ravigupta, of which the original Sanskrit and a Tibetan version are also known, and some part also in Turkish. Other anonymous medical texts have been printed. Most interesting historically are the many reports called hasda aurdsa " reports of information " sent by hada- " messengers, envoys " to the Court of Khotan. They report the activities of potentially

72

JOURNAL

OF PERSIAN

STUDIES

dangerous enemies, the Turks and their allies, between the cities of Chinese Turkestan Si-chou, Shachou and Kan-chou. One text (P 5538a) is a message from the court of Khotan from King Vis Sira to his mother's brother, the Chinese ruler at Sha-chou, the Ta-uang " great king ", Sha-chou being at that time in the tenth century an independent Chinese state. In these documents, jade has an honoured place as a gift for kings. It was the supreme export of Khotan. Many documents dated in the regnal years of Khotan kings have given us the names of kings of Khotan otherwise known only in the Tibetan Annals of Li or hidden under Chinese names in the Chinese dynastic records. Here we see the kings of the Visa (Vijita, Vijaya) family called Simha, Sambhava, Vdham, Dharma, Kirti, Sara, Samgrama, Vikram. The Annals of Li mention other had married a names. In these same Annals we even learn that Hphrom Gesar's daughter the Kaisar of Rome .Huronga King Vijaya Samgrama of Khotan. This Fr6m Kasar has his name from (Byzantion), a name which later plays a great part in the Gesariad of Tibet and Mongolia. The Sogdians and the Parthian used From beside the Persian Hrim to refer to Rome in Byzantion. A document with signatures called a khala-vi " inventory ", probably a Tibetan compound skal-bye, contains a list of gifts made to Arya-Ratanavaraksa and Praketu and to be taken to Sha-chou. Here our interpretation is at a low ebb since many of the things presented are given unknown names. Private letters form another section. The letter of an anonymous person to his old teachers resident in a monastery in Phema (in Marco Polo's description called Pem) is an eloquent expression of gratitude. A fine panegyric to one of the kings called Visa Samgrama is the most eloquent of Khotanese texts. The writer heaps upon him all the epithets of majesty, secular and religious, and calls him as a climax a fifth world-regent. " In one Buddhist Sanskrit text, the Sitdtapatrd-dhdrani the formula of the lady with the white umbrella " the scribe has according to custom inserted his patron's name, and one is suddenly confronted with raksa raksa mama radjid that is, " protection, protection be mine, the king vijitti sambhavasya, Sambhava ". It is a usual formula in these prayers. Here Vijita Sambhava has, as it were, Vijita come to life before our eyes-a king of the tenth century, whose very name was unknown to historians before this present century. Later when all the texts have been printed and translated, we may hope that a fully-qualified historian will relate the history of the Sakas of Chinese Turkestan from their own documents, their struggles with many enemies, with Chinese, Tibetans, and last and most deadly, the Turks of Kashghar, who by 100ooo A.D. had put an end to the Buddhist kingdom of Khotan after it had had a duration of
some 1200 years.
BIBLIOGRAPHY

Recent books where further details are to be found:

J. A. Dabbs, History of the Discoveryand Explorationof Chinese Turkestan(1963). Texts (1961), vol. IV (an introduction on H. W. Bailey, Khotanese the Kingdom of the Sakas in Khotan); and " Languages of the Saka " in Spuler's Handbuch Orientalistik(1958). der R. E. Emmerick, The Bookof Zambasta,English translation of the longest poetical text (1968). M. J. Dresden, The Jdtakastava,English translation
(i955).

" The panegyric of Visa Samgrima ", translated in Asia Major, new series, vol. XI, pp. Io1-I9. " The title Ia of the Yueh-chih " in Sylvain LUvi,Journalasiatique
(1913), part 2, pp. 330 ff. Sogdian ancient letters in H. Reichelt, SoghdischeTexte; and in W. B. Henning, BSOAS XII (1948), pp. 6oI-I I. einerPehleviThe Pahlavi Psalter, edited by Kaj Barr, Bruchstiicke der Ubersetzung Psalmen (1933). Al-dthdr al-bdqiya 'an al-qurfinal-khdliya of al-Bairint, edited by C. Ed. Sachau (1878, 1923).

Bhadracaryddedand, English translaJ. P. Asmussen, The Khotanese

tion (1961). "Annals of Li " in R. E. Emmerick, Tibetan Texts Concerning Khotan(1967).

DAILAMIS IN CENTRAL IRAN: THE KAKUYIDS OF JIBAL AND YAZD By C. E. Bosworth

I

The period of the disintegrationof the Caliphate in the Iranian lands, roughly spanning the third/ ninth to the fifth/eleventh centuries,provides as subject-matterfor the researchesof the Islamic historian a fascinating mosaic of successorstates. An especially interesting feature of the period in Persia is that the crumbling of 'Abbasid authority gave an opportunity for the resurgenceof Iranian elements which had been hitherto submergedunder the facade of unity within Islamic faith and culture. Amongst these elements, Dailami, Kurdish and Lur ones from the northernand westernparts of Iran, and Kufichi and Baltich ones from the south-easternregion, are especially notable. The political domination achieved in the fourth/tenth century by the mountain people of Dailam, the region situated at the south-western corner of the Caspian Sea, led the late ProfessorV. Minorskyto characterizethe period as " the Dailami interlude " of Iranian history, one lying between the Arab conquest of Iran and the introduction of the Islamic faith on the one hand, and the subsequent hegemony of Turkish dynasties. Since the Dailamis, as they first appear in Islamic history, were either imperfectly Islamized or even downright hostile to orthodox Islam, the first decades of Dailami rule (i.e. those of the first half of the fourth/tenth century) show a certain faltering in the processof Islamic acculturationin Iran; but this soon passes, and we find the second generation of Biyid Amirs or the later members of the Ziy~rid family, for instance, enthusiastic promoters of both the traditional Arab-Islamic culture and the nascent New Persian one. One of the minor Dailami dynasties which arose in the early fifth/eleventh century under the wing of the Btiyids of western Iran, was that of the KkTkiyids, whose centre of power was firstly in Isfahan, Hamadhin and other towns ofJibdl, and then latterly in Yazd in northernFars. Their period of virtual independence was not long, being cut short by the Seljuq leader Toghril Beg's capture of their capital Isfahan in 443/1051. Yet because the Seljuqs did not adopt as their model the monolithic " power state " which the Ghaznavids had favoured and did not deliberately curb all vestiges of provincial autonomy, the Ktkfiyids could survive and find a small niche within the Great Seljuq empire as local feudatories or muqta's Yazd. Within Yazd, the power of the Kakiiyid family endured for almost as of long as that of the Great Seljuq Sultans themselves and then it merged, in a smooth transition, into the rule of a line of Atabegs of Yazd, one of the several prominent dynasties of Turkish soldierswhich arose out of the decline of the Seljuq empire. II The founder of the Kdktiyid dynasty in Isfahan, and the most outstanding figure among its rulers, 'Ala' ad-Daula Abti Ja'far Muhammad b. Dushmanziydr,was the son of a Dailami soldier in the service of the Bfiyids of Ray and Jibil, AbiI 1-'AbbasRustam Dushmanziydr b. Marzubdn.1 We hear of this last being granted the district of Shahriyvr (apparently the Shahriyar-Kfihor Jabal Qarin lying in the Elburz Mountains near Firrim and to the north of Ray), with the title of Ispahbadh, and of his efforts on behalf of Majd ad-Daula to protect the in the Caspian lands against the local Bfiyid position Bawandid and Ziyarid princes.2
1

The kunya of Abfi 1-'Abbas is given only in the anonymous Mujmal at-tawdrikhwa I-qisas, ed. Malik ash-Shu'ari' Bahir (Tehran 1318/1939), p. 402. Regarding the name Dushmanziyar, this is the form usually, though not invariably, found in the literary sources; coins always have Dushmanzdr.

Ibn Isfandiyar, Ta'rdkh-i Tabaristdn, abridged tr. by E. G. Browne, Gibb Memorial Series, vol. II (Leiden-London 1905), pp. 228, 230-I, 238; Zahir ad-Din Mar'ashi, Ta'rikh-i Tabaristdnu Riiydnu Mdzandardn,ed. B. Dorn, in Muhammeddes der zur Kiistenldnder Kaspischen anischeQuellen Geschichte siidlichen Meers, vol. I (St. Petersburg 1850), pp. 195, 209.

73

74

JOURNAL

OF PERSIAN

STUDIES

Rustam Dushmanziyar's son 'Ald' ad-Daula Muhammad is frequently called Ibn Kikfiya or Pisar-i Kakfi in the historical and literary texts. Several of these sources explain kdkayaas a hypocoristic from the dialect word kdki7 " maternal uncle ", Rustam Dushmanziyar being the uncle of the famous Sayyida, mother of Majd ad-Daula Rustam b. Fakhr ad-Daula and effective ruler of the northern Btiyid amirate during the greater part of the nominal reign of her son.3 Sayyida was thus 'Ala' ad-Daula Muhammad's first cousin, and not nephew, as has sometimes been said by Western writers. It is uncertain when Rustam Dushmanziyar died, but at some time before 398/1007-08 his son Muhammad was governing IsfahSn on behalf of his kinswoman Sayyida, and in this way begins the Kkfikyid hold on that city which was to last substantially down to Toghril Beg's conquest over forty years later.4 The local historian of Isfahan, Mufaddal b. Sa'd al-Mafarrukhi, states that 'Ala' ad-Daula Muhammad reigned for " forty-odd years ", and since he died in 433/1041-42, this dating would push back his control of Isfahdn to at least 393/Ioo3.5 The internal weakness of the Bfiyid amirate of Ray and Jibil, with the ineffective Majd ad-Daula dominated by his mother Sayyida, enabled 'Ala' ad-Daula Muhammad to extend his power westwards and northwards from Isfah~n into areas which had never been subdued by the Bfiyids and which were in practice independent under local Kurdish chiefs, those of the Shadhanjan 'Annazids being the most prominent. He also acted as the military defender of the Bfiyids against ambitious Dailami adventurers from the Elburz region, who often obtained help from local dynasties of Tabaristan like the Bawandids and ZiySrids. These events can be followed in considerable detail from Ibn al-Athir's chronicle. In we find 'Ala' ad-Daula Muhammad intervening at Hamadhan, where Majd ad-Daula's 411/1020-2I1 brother and subordinate Shams ad-Daula was unable to prevent tension arising between his Kurdish and his Turkish troops.6 This tension culminated in a rebellion of the officers (al-quwwdd al-qiLhiyya) Turks; 'Ala' ad-Daula Muhammad was called in, and managed successfully to quell it at a place called Burjin or Burjain.7 Three years later, he was provided with a pretext to intervene, with permanent advantage for himself, in Hamadhan and the Kurdish regions of Jibal. After the death of Shams ad-Daula in 412/I021-22, Hamadhan passed to his son Sama' ad-Daula, but much of the real seems to have been in the hands of the Kurdish Vizier Taj al-Mulk Abfi Nasr b. Bahram. When power the Dailami feudatory of the nearby town of Burtijird, Farhadh b. Mardawij, fled to 'Ala' ad-Daula Muhammad in Isfahan for refuge, the latter marched towards Hamadhan with an army and eventually toppled Sama' ad-Daula's rule there, annexing Hamadhin, Dinawar and Shabfir-Khwast8 to his own dominions; meanwhile, 'Ala' ad-Daula Muhammad's nominal suzerain, Majd ad-Daula, was impotent to intervene and save his brother.9
a In Kurdish we have kdk(a), and in Luri and New Persian kdkd, meaning " brother " or " uncle " in a friendly, jocular sense, something like " old fellow "; the word is certainly a babytalk one like " dada " and " mama ", and is used only as a vocative. The corresponding Tabari word is gaga, which is not an exact semantic correspondence with kdki "maternal uncle ", and kdka with this meaning is not attested in Zaza, the only true Dailami dialect which we have; however, the presence of gaga in Tabari is probably near enough for our purposes (I am grateful to Dr. D. N. MacKenzie for this etymological note). It should perhaps be further noted that H. C. Rabino di Borgomale propounded an alternative etymology for kdkaya. He suggested that the dynasty came originally from Tunakdbun on the borders of Gilan and and Gibb Mem. Ser., Tabaristan (see his Mdzandardn Astardbdd, N.S., vol. VII [London 1928], pp. 21-4), in the local dialect of which kdkii means " clan " or " office ". He also noted a place or mountain south of Ldhijdn called Kikfi, and surmised that the family name might have arisen from a place name, as seems to have been the case with the Bfiyids (" Les dynasties locales du Gilin et du Daylam ", JA CCXXXVII [1949], pp. 313-14). But all in all, the explanation that kakiiyais a hypocoristic from kdkiiseems the most probable. 4 Ibn al-Athir, al-Kdmilft t-ta'rikh, ed. C. J. Tornberg (Leiden 1851-76), vol. IX, p. 146.
SKitdb MahdsinIsfahdn,Arabic original, ed. Jalal ad-Din Tehran

(Tehran 1312/1933), p. -oo; later Persian version of Husain b. Muhammad Awi, Tarjama-yi Isfahdn, ed. 'Abbas Iqbdl (Tehran 1328/1949), P- 95- mah.dsin 6 Almost every Bufyid ruler, from the time of the original three sons of Buiyawho established the amirates in the first half of the fourth/tenth century onwards, had to cope with dissensions within their forces arising between the Iranian, principally Dailami, elements, and the Turkish ones; cf. Bosworth, " Military Organisation under the Bfsyids of Persia and Iraq", OriensXVIII-XIX (1967), pp. 153 ff7 Ibn al-Athir, vol. IX, p. 226. Is Burjin perhaps the well-known town of Burj, one of the igharain, the two towns originally granted to the 'Abbasid general Abfi Dulaf al-'Ijli? Cf. G. Le Strange, The Lands of the EasternCaliphate(Cambridge 1905), pp. 197-8. 8 Shabir-Khwdst should, in the opinion of Minorsky, IHudad al-'dlam, Gibb Mem. Ser., N.S., vol. XI (London 1937), p. 283, followed by G. C. Miles, " The Coinage of the Kdkwayhid Dynasty ", Iraq V (1938), p. 103, be identified with Khurramdbad in Luristin. 9 Ibn al-Athir, vol. IX, pp. 232-3. Burfjird, apparently restored to its former governor Farhadh b. Mardawij, also acknowledged Kakuyid overlordship now. Apart of course from the Isfahdn coins (whose series begins in 407/1ox6-i7), Burijird is one of

DAILAMIS

IN

CENTRAL

IRAN:

THE

KXKOYIDS

OF JIBAL

AND

YAZD

75

Over the next few years, 'Ala' ad-Daula Muhammad was much occupied with maintaining his new conquests in the west and with holding the Kurds in check. Elated with the Hamadhdn conquest of he had at the same time gone on to attack the 'Annazid chief Abfi sh-Shauk Faris b. 414/O023, Muhammad, and had only desisted at the request of the Bfiyid Amir of Iraq, Khizistdn and Fars, Musharrif ad-Daula.10 In 417/1026 there was sharp fighting between the Kurdish tribal group of the Jiiraqan, modern Gfiran,11 and the Kdkfiyid governor, 'Ala' ad-Daula Muhammad s cousin Abi and the Kurds in that vicinity.12 In the following Ja'far, who had been appointed over Shibfir-Khwast who had a Dailami leader called 'Ali b. 'Imrdn,13 year, 418/I027, supported the Jiiraqan Kurds against 'Ala' ad-Daula Muhammad, succeeded in provoking a conflagration which affected the whole of northernIran. He called in the help of the Bawandid Ispahbadh of Tabaristan, the Ziydrid Mantichihr b. Qabiis and another Dailami leader, Walkin b. Wandrin, and aroused the Ispahbadh's hopes of overrunning the whole of Jibal. Hamadhin was temporarily lost, but 'Ala' ad-Daula Muhammad eventually won a great victory at Nihawand and captured his enemy the Ispahbadh, holding him prisoner till he died in Rajab 419/August 1028. The scene of warfare now shifted to Ray, which Man-ichihr besieged without success, and in the end a general peace was made.14 In the course of all these campaignings, the inability of Majd ad-Daula to move outside his capital of Ray amply demonstrated that 'Ala' ad-Daula Muhammad and not the Bilyid was the most forceful figure in Jibal at this time. Nevertheless, as often happened in the Islamic world, the habit of deference to a nominal suzerain-even after the substance of power had passed elsewhere-persisted here in the case of the Kakuyid. We possess coins of 'Ala' ad-Daula Muhammad minted at Isfahan, Burijird, Asadabad, Hamadhin, Jurbidhq~n, Qirmisin or Kirmanshah, Shabiir-Khwast (probably Khurramabtd, see above), al-Karaj (sc. Karaj of Rfidhrawar), al-Muhammadiyya or Ray, Mah al-Ktifa or Dinawar, al-Qasr (possibly Qasr al-Lusiis at or near Kinkawar, see above) and Yazd.15 Down to the
Ibn al-Athir, vol. IX, p. 233. For the 'Annizids, see the excellent article by Minorsky in E12 s.v. 11 On the phonological process connecting the names Jfiraqdn (originally *Givbdrakin?) and Gfiran, see Minorsky, "The Gfiran ", BSOS XI (1943-46), pp. 77-8. In this same article, pp. 81-3, Minorsky surveyed the r61leof the Jfiraqan in the warfare and politics of western Iran at this time. The mentions in Ibn al-Athir of the relations of the Jfiraqdn (a name which the chronicler writes incorrectly as Jfizaqdn) with 'Ala' ad-Daula Muhammad indicate that in the first half of the fifth/eleventh century, a part at least of the tribe occupied the northern fringes of Luristdn, in particular, the region around
10

Ibn al-Athir, vol. IX, pp. 247-8. 13 Possibly this man should be identified with the otherwise unknown 'Ali b. 'Umar who in 415/1024-25 minted coins, on which he acknowledged the overlordship of both Majd adDaula and 'Ala' ad-Daula Muhammad. The two coins surviving from this date are especially interesting in that they were minted at a place with the very vague name of al-Qasr. Miles very plausibly identifies this al-Qasr with the well12

Shbfibr-Khwdst.

known Qasr al-Lusfis at or near Kinkawar (see Le Strange, The Lands of the EasternCaliphate,p. 188), a place known from Ibn al-Athir's narrative to have been held by 'Ali b. 'Imrin round about this time. See the detailed discussion by Miles in his " A Hoard of Kdkwayhid Dirhams ", pp. 187-8. 14 Ibn al-Athir, vol. IX, pp. 251-2. 15 The whole corpus of the then extant Kakfiyid coins (amounting in number to seventy-one) was surveyed, with full reference to earlier literature and coin descriptions, by Miles in his article " The Coinage of the Kakwayhid Dynasty ". To the coins described there should be added a further acquisition by the American Numismatic Society, plus a few other coins and lead seals, covered by Miles in his " Notes on Kakwayhid Coins ". The number of extant Kakfiyid coins has now been considerably swelled by the acquisition in 1964 by the ANS of fortyfive coins from a hoard of Kakfyid dirhams, dating from between 414/1023 and 418/1027 and found somewhere near Kirmanshah. A point of particular interest is that several of the coins provide the names of local Kurdish and Dailami chiefs ruling in the western Jibal-Kurdistan region as subordinates of the Kikuyids, including, for instance, the 'Anndzids. there are two further dirhams of his in the hoard acquired by the ANS. Unfortunately, the mints and dates are effaced, but Miles places them as probably not later than 416/1o25 (" A Hoard of Kdkwayhid Dirhams ", pp. 18o-I, 192-3). The obvious probability is that the mint was Burujird. The present author doubts whether this Farhadh had any connection with the Ziyarids. The name of his father alone, Mardawij, seems to have led various western authorities (including Zambaur, et Manuel de gindalogie de chronologie, 21 1) to attach him to the p. rulers of Tabaristan and Gurgan; but on chronological grounds, it is quite impossible that Farhadh, who died in 425/1034, could be the son of Mardawij b. Ziyar, who died in 323/935.

the first places known-at least on presently-available evidence -as a Kakfiyid mint; a dirham of 415/1024 has on it the names of Majd ad-Daula, Samd' ad-Daula, Muhammad b. Dushmanzdr and Farhddh b. Mardawij. The name of Sama' ad-Daula continued to appear on the coins of several mints in western JibSl for some years after his deposition from Hamadhin, even on a coin minted at Hamadhan as late as 421 / 1030, indicating that he was still alive then (the date of his death does not seem to be recorded in the chronicles). See Miles, " The Coinage of the Kdkwayhid Dynasty ", pp. 97-00oo, o102; idem, " Notes on Kdkwayhid Coins" in American NumismaticSocietyMuseumNotes, vol. IX (I96o), p. 235; and idem, " A Hoard of Kakwayhid Dirhams " in ANS Museum Notes, vol. XII (1966), pp. 168,I7o--8, I8I-3. As for Farhddh,

76

JOURNAL

OF PERSIAN

STUDIES

Ghaznavid conquest of Ray in 420/1029, he generally mentions on his coins a Bilyid overlord, Majd ad-Daula, Shams ad-Daula or Sama' ad-Daula, depending on the minting-place. Until 418/1027, he seems generally to have been content with his simple name of Muhammad b. Dushmanziyar, without any honorific titles.16 But long before this, he had acquired an impressive string of alqdb,as splendid as any of the titles borne by his nominal suzerains the Bfyids. His independence of action is shown by the fact that h" had obtained titles in 409/1018-I9 directly from the Caliph al-Qadir, without the intermediacy of the Biyids and without their being able to assert any right of control.17 The Mujmal at-tawdrikh relates in its section on the history of the Biyids that in 409/1OI8-19 Ibn Kakiiya sent a mission to Baghdad under one Abi 1l-Fatdl Nasriya. He returned with an investiture patent for the territorieswhich the had conquered, a crown, a jewelled collar and a standard, together with Kakfiyid the titles of 'Adud ad-Din, 'Ala' ad-Daula (this becoming the laqabby which he is generally known18), Fakhr al-Milla and Taj al-Umma, and the designation of IHusamAmir al-Mu'minin " Sword of the Commander of the Faithful ". The same source also states that 'Ala' ad-Daula Muhammad's second son Abii Kalijar Garshisp became governor of Hamadhan for his father, and had the titles of alMu'ayyad Falak ad-Daula and Ghiyath al-Milla; this information must, of course, relate to the post
414/Io23-24

When Mahmfid of Ghazna marched westwardsinto Jibal in the spring of 420/ 1029, occupied Ray, carried off Majd ad-Daula and his son Abi Dulaf into captivity and initiated a purge of all the heretics and schismatics there, a totally new factor was injected into the politics of northern Iran.20 It was the express intention of the Sultan and of his son Mas'iid, to whom he entrusted these newly-subdued territories, that Ray should be the bridgehead for a great push into the heart of the Near East: the remaining Bfiyid amirateswere to be overthrown, the 'Abbasid Caliph released from his tutelage under the Shi'i Dailamis, and a confrontation to be made with the Ismd'ili Fatimids in Syria.21 Whilst Mahmfid was still at Ray, he sent out troops north-westwardstowards Dailam and Azerbaijan; alone amongst local rulers there, the Musafirid lord of Tarum, Zanjan, Abhar, Sarjahan and Shahrazfir, Ibrahim b. Marzuban, opposed him. The reduction of these outlying regions was undertaken by Mas'iid b. Mahmiid, who after bringing the Musafiridsto obedience,22turned southwardsagainst the Kdkfiyids. The Kakfiyid governor of Hamadhan was driven out, and then in Muharram 42 I /January 1030, 'Ala' ad-Daula Muhammad himself was forced by Mas'tid to evacuate Isfahan and retreat to Khfizistan, where he endeavoured to secure help from the Bilyid Amirs Abi Kalijar 'Imad ad-Din and Jalal ad-Daula.23 The news of Sultan Mahmiid's death in Ghazna, which reached Mas'fid in Jumada I
421/May 1030, was 'Ald' ad-Daula Muhammad's salvation, for Mas'fid was now compelled to leave

state of affairs, when Hamadhan had passed into Kkfikyid hands.19

the west and prepare to contest the successionto the Sultanate with his brother Muhammad, who had the advantage of being already on the spot in Afghanistan. At the instigation of Jalal ad-Daula, the 'Abbasid Caliph in Baghdad had already interceded for 'Ala' ad-Daula Muhammad, and an agreement
16

17

Miles, "A Hoard of KTkwayhid Dirhams ", pp. 171-3, 18 The two honorifics 'Adud ad-Din and 'Ala' ad-Daula are those two which appear most frequently on his coins. 185-6 (issues of al-Karaj and Mah al-Kfifa from 418/10o27 with all four of his alqdb). But in "The Coinage of the Kdkwayhid 19Mujmal at-tawdrikh,pp. 402-03. It should be noted that the unknown author of this history was a native of Asadabad and Dynasty ", p. Ioo, n. 2, he notes that there are not really therefore well acquainted with the history of western Jibdl. enough specimens surviving from the period before 419/1028 ed. to make a categorical assertion about the absence of his alqab 20 Gardizi, Zainal-akhbdr, M. Ndzim (Berlin I928), pp. 9', 97; from his coins. Mujmal at-tawdrfkh,pp. 403-04; Ibn al-Jauzi, al-MuntaZamff The right of direct access to the Caliphs was frequently withta'rikh al-mulfik wa 1-umam (Hyderabad 1357-59/1938-41), held by powerful overlords from their governors and subvol. VIII, pp. 38-40; Ibn al-Athir, vol. IX, pp. 261-2; The Life and Times of Sultdn Mahmid of Ghazna (Camordinates. Thus when in 422/1 03 Sultan Mas'fid of Ghazna " was negotiating with an envoy from Baghdad over recognition bridge N.zim, 1931), pp. 81-3; Bosworth, The Imperial Policy of the Early Ghaznawids ", pp. 69-72. of the new Caliph al-Qa'im, one of his stipulations for this 21 Cf. Sultan Mas'fid's declaration of future policy made to the recognition was that the Caliph should not have direct relations with the Ghaznavids' enemies, the diplomatic Qarakhanid ruler Yfisuf Qadir Khan immediately after his assumption of the throne in Ghazna in 421/1030 (Baihaqi, Qarakhanids of Transoxania, but that the practice of Ta'rtkh-i Mas'izdi, ed. Q. Ghani and 'A. A. Fayy4d [Tehran Mahmfid's reign, that all presents, honorific titles, etc. should be sent to them only through the intermediacy of the Ghazna1324/19451, PP. 79-80, cf. also p. 218, Russian translation by A. K. Arends, Itoria Mas'uda [Tashkent 1962], pp. 98-9, cf. vids, should continue (see further on this, Bosworth, " The p. 214; Bosworth, op. cit., p. 73). Imperial Policy of the Early Ghaznawids ", Islamic Studies, 22 Baihaqi, p. 218, tr. p. 214; Ibn al-Athir,vol. IX, pp. 262-3. Journal of the Central Institute of Islamic Research,Karachi I/3 23 Baihaqi, loc. cit.; Ibn al-Athir, vol. IX, pp. 279, 284. [x962], pp. 64-5).

DAILAMIS

IN

CENTRAL

IRAN:

THE

KAKUYIDS

OF JIBAL

AND

YAZD

77

was now made at Ray that 'Ala' ad-Daula Muhammad should be Mas'fid's representative (khalifat) and tributary in Isfahan, paying as tribute 20,ooo Harawi dinars per annum, plus io,ooo sets of clothing from the workshops of the Isfah~ n district, the usual Nauraz and Mihrgan presents, Arab horses, mules with riding saddles and all sorts of travelling equipment. The covenant was sealed by the Caliph's sending 'Ala' ad-Daula Muhammad an investiture charter for the region of Isfahan, together with a robe of honour.24 On returning thus to his old capital, 'Ala' ad-Daula Muhammad seems to have extended his authority beyond his centres of Isfahan and Hamadhan in Jibal. There is extant a dirham of his struck in Yazd in 421/10o3o0, acknowledging only the Caliph al-Qadir as suzerain; Yazd became in this way

the farthest point in the east reached by the Kktkiyids. G. C. Miles is doubtless correct in suggesting that Yazd was one of " the other towns " mentioned in Ibn al-Athir as seized by 'Ala' ad-Daula Muhammad at this time.25 Since the Ghaznavids are not acknowledged on the coin as overlords, its minting probably fell within a period when 'Ala' ad-Daula Muhammad withdrew his obedience from the Sultan. With Mas'id involved in events on the eastern fringes of the Ghaznavid empire, the until a Kdkfiyid Amir was even able now to capture Ray and the region of Demavand for a while, considerGhaznavid force ejected him. In the courseof the fighting, he was badly wounded, and spent a able time recuperating in one of his castles near Hamadhdn.26 It was during this brief occupation of Ray in 421/I o30 that 'Ala' ad-Daula Muhammad issued a dirham which blazoned forth all four of his honorifics and acknowledged only the Caliph al-Qadir.27 The decade of the lo30 was one of extreme confusion in northern Iran. The central r61eplayed there for almost a century by the Bfiyid amirate of Ray and Jibal, however shaky its authority may latterly have become, was never replaced. The region was too distant from Nishapiir and Ghazna for the Ghaznavids to control it properly. Ghaznavid rule in Ray, at first welcomed by the populace as a relief from the excesses of Majd ad-Daula's uncontrollable Dailami troops, soon became unpopular and oppressive, and it lasted for no more than nine years or so. The prime factor making for great instability and insecuritythroughout the whole countrysidewas, of course, the appearanceof marauding bands of Oghuz Turkmens, who ranged across northern Iran from the desert fringes of Khurasan and the Qara Qum desert to Azerbaijan and the borders of Byzantium. 'Ala' ad-Daula Muhammad was quick to see how these nomads might be used as a kind of third force in his perpetual struggle to preserve Hence Kkfikyid independence against other local rivals and against the overbearing Ghaznavids. and Ibn al-Athir recordfor these years a successionof periodsof warfarebetween 'Ala' ad-Daula Baihaqi Muhammad and the Ghaznavid representativesin Ray, Abti Sahl IIamdaini and Tash-Farrash,each followed by a brief and uneasy peace; in all these phases of warfare, the presence of the Turkmens was an increasingly important factor. At the end of 423-beginning of 424/winter 1032-33, 'Ala' ad-Daula Muhammad had recovered from his wounds, and now attempted to re-establish his position in his old dominions. With the help of his former associate Farhadh b. Mardawij, he rebelled against the Ghaznavids. He was initially driven back to Burijird and Shbfibr-Khwist, where the Jfiraqan Kurds now sheltered him, but he was neverthelessafterwardsable to re-occupy Hamadhan, Burijird, Karaj and Isfahrn.28 The rebellion in northern India of his Turkish general Ahmad Inaltigin compelled Sultan Mas'uidin 424/1033 to make peace with 'Ala' ad-Daula Muhammad and recognize him as his tributary in Isfahdn, on the usual terms of an annual tribute.29But in 425/1034 'Ala' ad-Daula Muhammad renounced his promisesonce
again; unfortunately for his hopes of securing his independence, his ally Farhadh b. Mardawij was killed in battle and he himself had firstly to shut himself up in a castle between Isfahan and Jurbadhqan and then to flee to idhaj in Abi Kalijar 'Imad ad-Din's province of Khizistan, allowing Abai Sahl
24
25

Baihaqi, pp. 14-17, tr. pp. 52-4; Ibn al-Athir, vol. IX, p. 279. Ibn al-Athir, vol. IX, pp. 284-5; Miles, " The Coinage of the Kakwayhid Dynasty ", pp. 93, 104. 26 Ibn al-Athir, loc. cit. 27 Miles, The Numismatic History of Rayy (New York 1938), pp. 187-9; idem, " The Coinage of the Kikwayhid Dynasty ", p. 98. Miles makes the interesting point that 'AlI' ad-Daula

Muhammad was the last person to designate this mint by the Arabic name of al-Muhammadiyya; succeeding rulers in Ray, from the Ghaznavids and Seljuqs onwards, reverted to the old Iranian name of Ray. bis2s Baihaqi, p. 361, tr. p. 326; Ibn al-Athir, vol. IX, pp. 288 290. 29 Ibid., vol. IX, p. 291.

78

JOURNAL

OF PERSIAN

STUDIES

Hamdiini's army to plunder his treasury in Isfahan. Amongst other things, the library of 'Al~' adDaula Muhammad's protg6JIbn Sina (see below, p. 81) was seized and carried off to Ghazna, where it remained until the Ghfirid Sultan 'Ala' ad-Din Muhammad sacked Ghazna in 545/ Jihdn-stiz I 150-51 and destroyed it.30 For some time, 'Ald' ad-Daula Muhammad seems to have laid low after these reverses; the Ghaznavid general Tash-Farrashwas able in Rabi' I 426/January-February 1035 to go to Hamadhan, which was then free of all opposing forces.31Not till the end of 427/autumn 10o36 do we hear of fresh fighting. 'Ala' ad-Daula Muhammad had obviously returned from Khaizistansome time before. Hearing of the appearance of a Ghaznavid foraging party in the vicinity of Isfahan, he sent a force and defeated the Ghaznavid detachment; he then tried to recapture Isfahan itself from Abti Sahl Hamdini, but failed because of the treachery of the Turkish element in his army. He had to flee, firstly to Burfijirdand then northwardsto Tarum in Dailam, whose ruler Ibrahim b. Marzuban was, however, too afraid of the Ghaznavids to admit him.32 'Ala' ad-Daula Muhammad was at this time supplementing his armies, which normally had a Dailami nucleus, with Turkmen auxiliaries recruitedfrom the so-called " Iraqi " Turkmens, enemies of the Seljuq family, who had been driven westwardsfrom the northern fringes of Khurasan in 419/1028 by Sultan Mahmfid.33 Baihaqi reports that in DhI 1-Qa'da 427/September 1036 a force of Turkmens under their leader Mansfir had turned on 'Ala' ad-Daula Muhammad and defeated him (presumably this refers to the treachery of the Turks mentioned in Ibn al-Athir's account), and that a group of Turkmen cavalry, whom the Kaktiyid Amir had invited from Khurasan and enrolled as mercenaries, had returned to Khurasan via the southern edge of the Central Desert and Tabas.34 In any event, the result of all this was that yet another peace was made between 'Ala' ad-Daula Muhammad and the Ghaznavids, who in the face of increasing Turkmen pressureon both Khurasan and the region of Ray, could not afford to leave Jibal in a state of suspended warfare. We possess a coin minted by 'Ala' ad-Daula Muhammad in Isfahan in 427/1036 which once more acknowledges Sultan Mas'fd as Amir had asked suzerain.35 The news came to Mas'fid in Safar 428/December 1036 that the Kakfiyid for peace and a restorationof his old tributarystatus, and that the Caliph al-Qa'im's Vizier Abil Tahir Muhammad b. Ayytb had also been interceding on 'Ala' ad-Daula Muhammad's behalf, doubtless at the instigation of the Bfiyids in Iraq. The Sultan agreed, but swore to extirpate the whole Kakilyid dynasty if there were any further trouble.36 Yet the Ghaznavids were by now in no position to enforce such blood-curdling threats, for Turkmen incursions were making the chief communications artery between northern Khurasan and Ray, the road which runs along the southern edge of the Elburz Mountains and to the north of the Dasht-i Kavir, increasingly difficult to keep open. By Jumada I 420/February-March 1037 it was clear that 'Ala' ad-Daula Muhammad's submission was once more a sham and had merely been a ruse to gain time. Because, says Baihaqi, he had much money and extensive treasuries,he had gathered together a numerous army, which included many of the " Iraqi " Turkmens,followersof the chiefs Qizil and Yaghmur and the group known as the BalkhanKtihiyan, sc. those who had originally fled before Mahmiid of Ghazna and taken refuge in the hilly Balkhan-Kiih area to the east of the Caspian and who had now moved westwards before the Seljuqs. Ibn al-Athir mentions the figure of over 1500 men under Qizil's leadership,whilst the rest of the " Iraqi " Turkmens passed on to Azerbaijan, where they harrassedthe Rawwadids and other local Kurdish and Dailami chiefs.37 It was difficult for 'Ala' ad-Daula Muhammad to keep discipline amongst the
anarchically-inclined Turkmens in his service; they speedily turned to plundering and brigandage and were clearly an unreliable element.38 Nevertheless, he retained some Turkmens in his service, and it was through their support that he was able to step in when the Ghaznavid garrison in Ray was at last
s0 Ibid., vol. IX, pp. 296-7. "1 Baihaqi, p. 444, tr. p. 394. 32 Ibn al-Athir, vol. IX, p. 304. 83 Cf. Cl. Cahen, " Le Malik-nameh et l'histoire des origines seljukides", Oriens II (2949), pp. 55-8; Bosworth, The Ghaznavids,theirEmpirein Afghanistanand EasternIran 994-o1040o (Edinburgh I963), pp. 224-5, 234-53" Baihaqi, p. 501, tr. pp. 443-4. 35 Miles, "The Coinage of the Kikwayhid Dynasty", pp. 97, 1o .
36 Baihaqi, pp. 51o-I I, tr. p. 452.
37

38

Ibid., pp. 521-2, tr. p. 462; Ibn al-Athir, vol. IX, pp. 269-71. Ibid., vol. IX, p. 269. The Turkmens' refractoriness is perhaps the basis of the report by Abfi Sahl Hamdfni to the Sultan in Dhfi 1-Qa'da 428/August 1037 that Ibn Kkfikya's position was weak and his Turkmens unreliable-an over-optimistic appreciation of the real state of affairs, as events speedily showed (Baihaqi, p. 530, tr. p. 470).

DAILAMIS

IN

CENTRAL

IRAN:

THE

KAKOJYIDS

OF JIBAL

AND

YAZD

79

forced to evacuate the city and retire via Gurgdn to Khurdsdn. The great battle which took place outside Ray between, on the one side, Abil Sahl and Tash-Farrash's army of 3000 cavalry and a column of elephants, and on the other side, 5000 Turkmens, is described by Ibn al-Athir and .Hamdini's placed by him in the year 427/1036, which must be a mistake for 429/1037-38. In the course of this fighting, the Kurdish troops of the Ghaznavids withdrew from the fray, Tash-Farrash was killed and the elephants captured. AbUi Sahl Hamdfini had to shut himself up for the winter in Tabarak, the citadel of Ray, and after the Turkmens had won a further victory over a relieving force sent from Gurgan, he finally abandoned the city. This withdrawal took place around Jumada I 429/February-March 1038, as the news reached Mas'Uid in Ghazna during the following month of Jumada II.39 Very soon after Abii Sahl Hamdiini's departure, 'Ala' ad-Daula Muhammad and his Turkmen auxiliaries occupied Ray.40 A dinar minted by him at al-Muhammadiyya (= Ray) in 429/1037-38, and perhaps a dirham also, still acknowledges Mas'Uid as suzerain, although a dinar minted by him at Isfahan in 428/1036-37 was an independent issue, acknowledging no overlord. This Ray issue might perhaps relate to Ibn al-Athir's information that 'Ala' ad-Daula Muhammad wrote from Ray to Abfi Sahl Hamdfini, at that time in Tabaristan en route for Nishaprir, asking for the formal grant to him of Ray, a request to which the Ghaznavid governor agreed.41 There can have been little faith at this point that Ghaznavid rule would ever be re-established in Jibal, but possibly the Kakfiyid wished to retain some connection with the Sultan as a possible bargaining counter with the incoming Turkmens. He retained Ray for some time, until the Turkmens in the area combined with its former Bilyid ruler Majd ad-Daula's son Abfi KIlijar Fand-Khusrau and the Dailami ruler of Sawa, Kamrawa, to eject him. He fled from Ray, leaving the city exposed to a savage plundering by the Turkmens, and he reached the safety of Isfahan and its walls, his Oghuz pursuers now being deflected to plunder Karaj.42 The Turkmen pattern of overrunning northern Iran was not in any way one of systematic conquest, with set battles and sieges of towns and strongholds-an approach for which the Oghuz were technically and psychologically unfitted. Rather, they starved the towns into submission by devastating the surrounding countryside and pasturing their flocks on it, thus cutting off the towns from their natural hinterlands. In the course of the Seljuq conquests, it was rare for a town to be taken by storm. Hence a stout ring of walls and fortifications was still an asset in fending off marauders and preventing direct plundering of a town, and in 429/1037-38 'Ala' ad-Daula Muhammad constructed walls around Isfahan. Their circuit came to more than 15,ooo paces, not including the defences of the suburbs of Kamdydn,43 Barayln,43 Sunbulan, Kharj n, Farsan, the Garden of 'Abd al-'Aziz, Jarwayan,43 Ishkahan and Lunban. The twelve gates in the walls were each provided with iron doors, through which a war elephant, fully-equipped with its howdah and battle pennants, could pass. Mafarrukhi cites some of his own Arabic verses on the wonders of these fortifications: A wall whose battlements reach up to the height of Capella, and whose buttresses reach beyond the celestial girdle of Gemini.
39
40

Ibid., pp. 534-5, tr. p. 474; Ibn al-Athir, vol. IX, p. 268. army of Dailamis and Kurds, and tried to recapture Ray; he Baihaqi, p. 546, tr. p. 484; Ibn al-Athir, vol. IX, p. 269. was, however, defeated by the Ghaznavid commander or 41 Ibn al-Athir, loc. cit.; Miles, The NumismaticHistory of Rayy, shihna left there by Mas'fid, and retired from the scene (Ibn al-Athir, vol. IX, p. 284). A decade or so later, Abfi pp. 192-3; idem, " The Coinage of the Kdkwayhid Dynasty ", Kdlijtr Fand-Khusrau was set up temporarily in Ray as a puppet of the PP. 97, 98, 101o,104. 42 Ibn al-Athir, vol. IX, p. 270. The re-appearance of a scion of Seljuqs. A coin of his-which Miles says provides enough the Bfiyids is interesting. According to Gardizi, pp. 91, 97, cf. material for a monograph in itself-is extant from 432//1040-41, and gives him the titles of ShThanshah, Sharaf al-Mulafk and Nazim, The Life and TimesofSul.tdn Mahmfid Ghazna,p. 83, the of Fakhr Allah (Miles, The NumismaticHistory of Rayy, p. deposed Majd ad-Daula was exiled by Sultan Mahmfid to 194). He had now clearly thrown in his lot with the Seljuqs. We last India, but fetched back by Mas'fid to Ghazna, where he lived a life of comfort and honour. His son Abia Kdlij*r Fandhear of him in 439/1047-48, when he is said to have seized the Khusrau escaped Mahmfid's clutches and took refuge in a town of Amid in Diyvrbakr from Toghril's governor (Ibn fortress called Qasrdn (presumably a different place from the al-Athir, vol. IX, p. 370). district near Ray of this name, and more probably the Qasrdn 43The use of alif-madda,as here, to indicate the spellingyadoccurs of SirajRn in Kirmdn, cf. Le Strange, The Lands qf the Eastern also in the anonymous geographical treatise of the late fourth/ tenth century, the Hudfidal-'alam; see Minorsky, " Addenda Caliphate, pp. 216, 301). When in 421/1030 Mas'fid left western Persia to contest the succession with his brother to the IHudfid al-'Alam ", BSOAS XVII (x955), P- 251 = Muhammad, he came out from his hiding-place, collected an al-'dlam, 2nd edition (London 1970), p. liv. .Hudiid

80

JOURNAL

OF PERSIAN

STUDIES

Looking down from its towers, amongst the constellations of the revolving firmament, you can see the stars when they speed past by night. Even if one were to bring along the giant Gog, these towers would not be conquered by day, and the excellence of the wall's construction would render useless the efforts of those trying to sap and mine it. Moreover, there is a moat round it, whose waters have been stirred up like the swollen waves of the sea, and there is no ocean or Nile river which can compare with it. The cost of these walls weighed heavily on the people of Isfahan, but in those insecure days, they must have been a worthwhile protection; unlike the other main centre of Kakfiyid power at this time, Hamadhan, Isfahan was never sacked by the Turkmens.44 Hamadhan was during these years under severe pressure from the Turkmens. 'Ala' ad-Daula Muhammad's second son Abai Kalijar Garshisp had become governor of Hamadhan and Nihawand at some unspecified date, presumably with interruptions during the intervals of Ghaznavid occupation, such as those of 42I/1030, 423/1032 and 426/1035.45 In 429/1037-38 bands of " Iraqi " Turkmens expelled from Azerbaijan by the Rawwadid Wahstidan b. Mamlan and other local Kurdish leaders, moved southwards, and under their chiefs Mansfir and G6ktash, besieged Hamadhan. After fierce fighting, Abif Kalijar Garshasp was forced to yield and make peace with the Oghuz, and to cement the agreement, he married one of G6ktash's daughters.46 But in the following year, the Turkmens turned from the spoliation of Ray (see above, p. 79) to besiege Hamadhan again. Abui Kalijar Garshasp and the chief merchants and notables withdrew to Kinkawar, and the town was now abandoned to a frightful sacking by the Turkmens of G6ktash, Qizil and Buqa, and by the Dailamis of the Bfiyid pretender Abil Kalijir Fand-Khusrau; the ravages of the Dailamis are described as being worse than those of the Turkmens. The attackers were only ultimately driven off through the efforts of the Kurdish 'Annazid chief Abai 1-Fath b. Abi sh-Shauk of Dinawar and of 'Ala' ad-Daula Muhammad himself, who came out from Isfahan with an army.47 'Ala' ad-Daula Muhammad died in Muharram 433/September 1041, still in the battle-saddle. He had been campaigning in 432/1041 against Abfi sh-Shauk, after an appeal to him by the 'Annazid's brother Muhalhil. Abii sh-Shauk remained intransigent, and threatened to cede his lands and fortresses to the Amir of Iraq, Jalal ad-Daula, if ever Abli sh-Shauk himself fell into 'Ala' ad-Daula Bfiyid Muhammad's hands; eventually, the two rulers made peace. It was whilst returning from the peace negotiations that the KkSkiyid Amir fell sick and died, after a life of almost ceaseless campaignings and activity, punctuated by spells of exile amongst neighbouring rulers, usually his Bfyid kinsman.48 To have guided the destinies of the Kkfikyid principality for over forty years, at a time when these territories were occupying a somewhat precarious position between the greater powers of the Bfiyids, Ghaznavids and Seljuqs, was itself no mean feat, and 'Ala' ad-Daula Muhammad must be regarded as one of the most adroit and capable rulers of his generation. The references to " Pisar-i Kakfi " in the Ghaznavid historian Baihaqi amply corroborate this judgment. Although at various points during the Ghaznavid occupation of Ray, 'Ali' ad-Daula Muhammad made his submission to Sultan Mas'fid, he was always regarded with distrust, as someone ready to renounce allegiance whenever occasion presented itself sc. (nim-dushmani " half an enemy ", in the description of one of the Sultan's advisers)49. Especially revealing is the characterization of the Kakiyid Amir's shrewdness, and the advantages which he derived from his strategic position and financial resources, given by Abfi Sahl IHamdfini, who became civil governor of Ray and 'Iraq 'Ajami for the Ghaznavids in Jumida II 424/May 1033, and who over a period of five years had good opportunity to test 'AlI' ad-Daula Muhammad's good faith and intentions:
"4Mgfarrukhi, Kitab Mahdsin Isfahdn, pp. 81, Ioo-oI, Persian version of Awi, pp. 51, 13, cf. E. G. Browne, " Account of a Rare Manuscript History of Isfahan ", JRAS (1901), pp. 23-4, 51I of offprint. The Bfiyid Abfi Kalijar 'Imdd ad-Din also at this time put walls round his capital Shirdz, their construction to 440/1048-49; see extending over the period 436/1044-45 Ibn al-Balkhi, Fdrs-ndma, ed. G. Le Strange and R. A. Nicholson, Gibb Mem. Ser., N.S., vol. I (London 1921), p. 133; Ydqfit, Mu'jam al-bulddn (Beirut I374-76/1955-57), vol. III, p. 351 s.v. "Shirz "; Ibn al-Athir, vol. IX, P. 359. 45 Baihaqi, p. 444, tr. p. 394; Ibn al-Athir, vol. IX, pp. 279, 288 bis-29o. 46 Ibid., vol. IX, p. 270. 47 Ibid., vol. IX, p. 271. 48 Ibid., vol. IX, p. 338; cf. Minorsky, EPl Art. "'Annazids". 4' Baihaqi, p. 263, tr. p. 253.

DAILAMIS

IN

CENTRAL

IRAN:

THE

KAKIJYIDS

OF JIBAL

AND

YAZD

81

Pisar-i Kakfi is at present [sc. in Jumad II 424] governing Isfahan, Hamadhan and part of Jibal, and must be regarded as a shrewd and slippery opponent. He has not only wealth and soldiers, but also skill, craft and deceitfulness. So long as our teeth are not bared against him, so that he may receive according to his merits and be deprived of his rule and territories [he will remain dangerous]. Or else he must submit and send his son [as a hostage] to the [Sultan's] exalted court, and become an obedient slave and pay regularly each year the substantial sum of tribute laid upon him, and then the other local rulers might look to his example and stay pacific. Unless measures like these are put in hand, Ray and Jibdl will never be peaceful.50 By the time of the Kakfiyids' rise to power, the Dailamis had, as was noted at the beginning of this article, emerged from some of their pristine barbarism and grossness. The vicissitudes of his reign and the constant struggle to maintain his position in Jibal can have given 'Ala' ad-Daula Muhammad little time to assume the paternalistic duties of the typical Muslim prince, or to gather round himself a literary circle such as many of the Btiyids and Ghaznavids collected. His greatest claim to fame in this direction is the fact that he gave refuge to Ibn Sina, after the great philosopher and scientist had been in the official service of the Bilyid Shams ad-Daula of Hamadhan; he wrote his Persian encyclopaedia of the sciences, the Ddnish-ndma-yi'Ald'i, for the Kikfiyid Amir, and he died in 428/1037 whilst accompanying his patron on a journey from Isfahan to Hamadhan.51 In his obituary notice of Ibn Sina, Ibn al-Athir asserts that there was a bond of sympathy between Ibn Sina and 'Ala' ad-Daula Muhammad, because the latter held unorthodox beliefs (kdnafdsid al-i'tiqdd) and therefore allowed and encouraged Ibn Sina to compose books on infidelity.52 Nothing, in fact, seems to be known about the Kikfiyid Amir's religious sympathies, although the connections of his family with the Bfiyids make it probable that, like so many other Dailami rulers, his inclinations were Shi'i rather than Sunni.

III
'Ald' ad-Daula Muhammad's eldest son was Zahir ad-Din Shams al-Mulfik Abi Mansfir Faramurz, who on his father's death secured the allegiance of the Kakfiyid troops in Isfahan and succeeded to his father's position there; his brother Abil Kalijar Garshasp proceeded to Nihawand and established himself in western Jibal as Faramurz's subordinate.53 Abti main task was to assert Manstir Faramurz's his paramountcy, as the eldest son, over other members of the Kakfiyid family. When the castellan of Natanz refused to yield up the treasure which 'Ald' ad-Daula Muhammad had stored there, he was joined by Abti Mansfir Faramurz's discontented younger brother Abil Harb, and the two rebels called in a band of Oghuz from Ray, who seized Qajin (? Qashan) and handed it over to Abi Despite was defeated in battle, and fled to the Bilyid court .Harb. help from disaffected Kurds in the region, Abti of Shiraz, where he dangled before Abil Kalijar 'Imad ad-Din the prospect of acquiring Isfahin. .Harb Desultory fighting took place in the neighbourhood of Isfahdn between the forces of Abi Mansfir Faramurz and the Bniyid Amir, until peace was at last made; Abi IHarb retained control of the fortress of Natanz, but yielded a proportion of its treasure to his eldest brother. Abli Mansfir Faramurz was now free to assert his authority in the western parts of his principality. Occupying HamadhIn and Burfijird, he agreed to leave his other brother Abti Kalijar Garshisp as governor there, with his own name acknowledged in the khutba.54

There remained the larger question of Abti Mansfir Faramurz'srelations with the Seljuqs, who now

controlled Ray. In the year of 'Ali' ad-Daula Muhammad's death, Ibrahim b. Inal seized Ray for the Seljuqs, and in the following year, 434/1042-43, Toghril arrived and took it over from his half-brother; also, Majd ad-Daula's son Abti Kalijar Fana-Khusrau was now removed from the fortress of Tabarak. The devastated parts of the city were restored, and for the next nine years, Ray served as Toghril's
50Ibid., p. 392, tr. p. 352. 61Ibn al-Athir, vol. IX, p. 31o; Nizami 'Ariidi Samarqandi, Chahdr maqdla,ed. M. M. Qazwini, Gibb. Mem. Ser., vol. XI/i (London I9Io), pp. 82-3, revised translation of Browne, Gibb Mem. Ser., vol. XI/2 (London 1921), pp. 92-3, cf. p. 163, note XXX, showing Nizdmi 'Arfdi's errors regarding Ibn Sink's relationship to 'Ala ad-Daula Muhammad; S. M. Afnan, Avicenna,his Life and Works (London 1958), pp. 67 ff. On the Danish-ndma-yi'Ald'f, see J. Rypka et al., History of PersianLiterature(Dordrecht 1968), pp. 150, 480. Ibn al-Athir, loc. cit.

52 63

Ibn al-Athir, vol. IX, p. 339"4Ibid.

82

JOURNAL

OF PERSIAN

STUDIES

capital; he started minting coins there as soon as he had installed himself.55 Abti Mansiir Faramurz had already had some contact with the Seljuqs, although the circumstances surrounding this episode are rather mysterious. He had been present with Toghril and the other Seljuq leaders on the battlefield of Dandanq~n in Rama~dn 431/May 1040, when the Ghaznavids were finally hurled out of Khurasdn, and the victorious Seljuq granted to him Ray and Isfahan because, so he said, " You have suffered many troubles ". This must refer to the fact that Abti Mansfir Faramurz had been previously observed by one of the Ghaznavid side in bonds and set on a camel.56 How did he come to be a captive of the Turkmens out there in the desert between Sarakhs and Merv ? One can only surmise that he had been captured in Jibdl, either by the Oghuz or by the Ghaznavids before they had abandoned the province, and brought to Khurdsdn, where he eventually escaped or passed into Seljuq hands. During the next eighteen months, he must have found his way back to Jibal, for he was, as we have seen, on hand to take over in Isfahdn at his father's death. Only five coins of Abfi Mansir Faramurz are extant, those three with decipherable mints all being from Isfahdn; but the first of these, dating from 434/1042-43, shows that at the outset, he acknowledged Toghril as suzerain, for Toghril soon after his arrival in Ray sent an expedition to Isfahdn to secure Abai Mansfr Faramurz's obedience and payment of tribute.57 Over the next years, however, his allegiance was to oscillate between that to Toghril and that to the Biyid Amir Abi Kdlij*r 'Imdd ad-Din (see below). Hamadhin had already been a target for Ibrdhim Inal's Oghuz once he had secured Ray in 433/ He had proceeded to Burijird and Hamadhan, upon which Abi Kalijar Garshasp had fled 1041-42. to a fortress near Shdbiir-Khwtst. The population of Hamadhdn expressed no enthusiasm for Kkfikyid rule, and had no expectation that Abai Klijar Garshdsp would ever return: The people of the town said to him [sc. to Ibrdhim Inal], " If you have come to seek our submission and that which the ruling power normally collects from its subjects, then we will hand it over to you and will observe this obligation. But first search out this fellow who was formerly over us and is now opposing you (they meant Garshdsp), for we have no confidence in his coming back to us. So if you are able either to capture him or drive him off for good, we will give our obedience to you. They escaped a further sacking by handing over a substantial sum to Ibrahim Inal. The people of the Shdbfir-Khwist district suffered considerably, however, from the depredations of the Oghuz before they returned to Ray.58 Abti Kdlijar Garshdsp came back to Hamadhan as soon as the Oghuz had left it (it is presumably this to which Ibn al-Jauzi refers when he says under the events of 433/1041-42 that " AbUi Kqlijar entered Hamadhin and drove off the Ghuzz from it "59). He was there when Toghril came to Ray, and he communicated with the Seljuq leader to secure legitimation of his rule in Hamadhdn. After Toghril had obtained the obedience of Abii Kdlijar Garshdsp's brother Abti Fardmurz in Isfahdn, he marched on Hamadhdn in person. He demanded that the fortress Mansfir of Kinkawar be handed over to him, but its castellan refused to surrender it, and Toghril therefore dispossessed Abfi Kdlijar Garshisp of Hamadhan and appointed a governor of his own there (434/ Io042-43) .60 It was not until 436/1044-45 that Abii Kilijir Garshdsp left the security of Kinkawar and returned to Hamadhin, expelling the representative of the Seljuqs and now giving his ultimate allegiance to the Biyid Abai Kdlijar 'Imdd ad-Din and not to the Seljuq Amir.61 Toghril was now exasperated with V Abi KVlijar Garshisp, and in 437/1045-46 sent Ibrahim Inal against HamadhVn once more. The threat of a re-appearance of Oghuz bands in western Jibal and Kurdistan brought Abi Kalijar Garshasp and the 'Annazid Abi sh-Shauk together into an alliance. Nevertheless, their Kurds and Dailamis had to fall back before the Oghuz, whose bands penetrated as far west as Kirmanshih, Hulwan and Khaniqin before turning southwards to Luristan and to Sirwan and Saimara on the borders of Khtizistan, where their presence alarmed Abi Kalijar 'Imad ad-Din.62 Hamadhan now
15

Ibn al-Athir, vol. IX, pp. 339, 347; Miles, The Numismatic 58 Ibn al-Athir, vol. IX, p. 347. 59 al-Muntazam, vol. VIII, p. io8. Historyof Rayy, pp. 196-7. 60 Ibn al-Athir, vol. IX, p. 348. 5 Baihaqi, pp. 627-8, tr. pp. 553-4. 61 Ibid., p. 359. 67 Ibn al-Athir, vol. IX, p. 348; Miles, " The Coinage of the 62 Ibid., p. 360. Kikwayhid Dynasty ", pp. 97, 99, 102.

DAILAMIS

IN

CENTRAL

IRAN:

THE

KXKUYIDS

OF JIBAL

AND

YAZD

83

passed firmly under Seljuq control, and two years later, in the autumn of 439/1047, the fortress of Kinkawar, which had long held out under Abti KIlijar Garshasp's lieutenant 'Ukbar b. Faris, surrendered under ameinor safe-conduct to Ibrahim b. Inal.63 Abii Kalijir Garshasp seems to have spent the last few years of his life mainly in exile amongst the Bfiyids. Unlike his brother Abii Mansiir Faramurz, he never became reconciled to the Seljuqs, and when occasion presented itself, he made sorties into the remaining Kaktiyid territories, hoping to lead a revanche against the Seljuqs. Thus he was in Isfahan at some time before 441/1049, and was in communication with the Ghaznavid Sultan Maudfid b. Mas'fid, who was endeavouring to organize a grand anti-Seljuq coalition in Iran and thus re-establish the former Ghaznavid position in Khurasan; it appears that Abi Kalijatr Garshasp actually despatched a body of troops eastwards for this purpose, but that they never got beyond the Central Desert of Iran.64 He died in Khiizistan in 443 / 1051-52, having governed that province for the Bilyid Amir Abti Mansfir Fiiladh-Sutiin b. Abi Kl-ijar.65 Ab-i Mans&irFaramurz was as restive under Seljuq suzerainty as his brother Ab-i Kalijar Garshasp, and like him, endeavoured to adopt a balancing r61e between Toghril on the one hand and the Bilyid Abti Kalijar 'Imad ad-Din on the other. At some time before 435/1043-44 he had come to an understanding with the Bfiyid Amir, on a basis of mutual non-aggression, but in that year he saw a chance of personal aggrandisement and renounced the compact. He seized two Bfiyid fortresses on the frontiers of Kirman, but Abfi K~lijar 'Imad ad-Din came with a superior force, captured Abarqfih from the Kakilyids and defeated AbiMMansfir Faramurz's army, thus regaining the disputed strongpoints.66 Abii Mansfir Faramurz had naturally transferred his allegiance back to the Seljuqs when he had fallen out with the Bilyids. But he was disappointed in his expectations of them, and when Toghril returned to KhurSsan, he deemed it prudent in 437/1045-46 to submit to the nearer potentate, Abi Kalijar 'Imdd ad-Din. In face of the threat from the Oghuz to all of them, a general peace was made between the Biiyids and the local rulers of Jibdl, viz. AbiMKalijar 'Imdd ad-Din, AbiMMansir Faramurz, and the 'Annazid brothers Abii sh-Shauk and Muhalhil.67 Nevertheless, western Jibdl remained very disturbed, with Ibrahim Inal's followers swarming over the countryside; and after AbiMsh-Shauk's death in Ramadin 437/April 10o46, his son Su'da threw in his lot with the Oghuz.68 Toghril appeared before the walls of Isfahan in 438/1046-47, and after a long siege, AbiMMansir Faramurz submitted and promised to pay tribute as before and acknowledge the Seljuqs in the khutbaof Isfahan.69 To fill out Ibn al-Athir's narrative of the events of these years, on which the previous paragraphs are substantially based, we possess a small item of information from the early eighth/fourteenth century author Hindiishah Nakhchavani. In his general history (which is a Persian amplification of Ibn atTiqtaqa's well-known Kitdb al-Fakhri, see Storey, Persian Literature,vol. I, pp. 81, 1233), he speaks of a Persian minister who served Toghril at this time, one Abfi 1-Fath Rdzi, who had been for most of his career in the service of 'Ald' ad-Daula Muhammad in Isfahan and then in that of the Kakfiyid Amir's son AbiMMansfir Faramurz. The latter sent him on a mission to Toghril, who was impressed by his ability and took him into his own service. AbMi Mansfir Faramurz was incensed at this, confiscated AbiM Rdzi's possessions and allowed his house in Isfahan to be plundered. When Toghril came to 1-Fath Isfahin and besieged AbMiMansir Faramurz there, Abfi 1-Fath Razi was eventually left as the Seljuq Amir's agent, responsible for collecting the Ioo,ooo dinars' tribute laid upon the Kaktiyid, a task which he managed very successfully. However, he later left Toghril's service, with the Amir's permission, and went to serve the Biiyid Abi KIlijar 'Imad ad-Din as Vizier, remaining in office till his dismissal in o48.v70 Sha'bin 439/January-February
Ibid., p. 366. Ibid., pp. 381-2. Ibid., p. 398. 66 Ibid., pp. 355, 361. This is the first specific mention which we have of Abarqfih as being in Kdkfiyid hands. 67 Ibid., pp. 361-2. " 68 Ibid., pp. 362-3, cf. Minorsky, E1 Art. 'Annazids ". 69 Ibn al-Athir, vol. IX, p. 365. 70 Hindfishdh b. Sanjar Nakhchavdni, Tajdrib as-salaf, ed. 'Abbas Iqbdl (Tehran 1313/1934), pp. 260-I. The whole episode
64
65 63

raises a number of interesting points, some of which are discussed by Iqbdl in his book Vizdrat dar 'ahd-i salktin-buzurg-i saljiqi (Tehran 1338/1959), PP- 37-9. Briefly, he relates Hindfishdh's mention of Toghril's coming to Isfahan to the year 434/1042-43, when Abfi Mansfir Fardmurz first acknowledged him as suzerain and agreed to pay tribute (see above). Abfi 1-Fath Rdzi is not mentioned in any of the lists of Toghril's Viziers (see H. Bowen, " Notes on Some Early Seljuqid Viziers ", BSOAS XX [1957], pp. Io5-Io, to whose sources on [continued nextpage

84

JOURNAL

OF PERSIAN

STUDIES

So far as we know, Abfi Manstir Fardmurz remained faithful to the Seljuqs for the five years after and the two the last ones of his rule in Isfahan; coins minted by him in 438/1046-47 438/1046-47, as sovereign."7 The end of Kakfiyid rule in Jibal was, however, following years acknowledge Toghril near at hand. The tergiversations and transfers of allegiance of Abfi Mansfir Faramurz must have exasperated Toghril, as had those of his brother Abi Kdlijar Garshdsp in Hamadhan a few years beforehand. Moreover, the semi-independent position of the Kdkfiyid principality blocked the effective extension of Seljuq authority over the Kurdish chiefs of Kurdistan and Luristdn and, above all, over the remaining Bfiyid territories in Iraq, Khfizistdn and Fars. The periodic alliances of the Kdkfiyids with their Bfiyid kinsmen, in face of the common Seljuq threat, meant that Toghril could never rely on an independent, completely friendly Kakilyid power. In Ibn al-Athir's words, Abfi Mansir b. 'Ala' ad-Daula, the ruler of Isbahan, never kept to a single pattern of behaviour with Sultan Toghril, but was constantly changing his attitude. At one time he would show obedience to him and adhere to his cause, but on another occasion he would throw off his allegiance and transfer it to al-Malik within his heart a plan of mischief against him, and when he returned this time from Khurdsan to resume the province of Jibdl from his brother Ibrahim Inal and assert his own rule over it, as we have related, he turned aside to Isbahin with the intention of seizing it from Abfi Mansfir.72 Toghril invaded Abfi Mansfir Fardmurz's dominions, and in Muharram 442/June 1050 occupied the countryside around Isfahan and besieged the town itself. Yet once more there was demonstrated the Turkmens' lack of aptitude for siege warfare. Isfahan held out for almost a year, with the population reduced to such extremities that they had to pull down the timbers of the Friday mosque and use them for winter fuel. In Muharram 443/May-June 1051 Abi! Mans&irFaramurz finally surrendered. The population was favourably treated by Toghril, who now made Isfahan his capital, razing part of its walls,73 and transferring thither his treasury and armoury from Ray. His Turkmen troops were given land grants (iqttd'dt)in the countryside of Jibdl. To compensate for the loss of his ancestral territories, Abfi Mansfir Faramurz was given the two towns of Yazd and AbarqiIh in northern Fars, both of which had at times been controlled by the Kdkfiyids (see above, pp. 77, 83)-4
ar-Rahim [sc. the Bfiyid Amir of Iraq, Khfizistdn and Fars, reigned 440-47/1048-55]. Now Toghril concealed

IV
The second phase of the Kdkiiyid dynasty's history, that of their r61e as local governors in Yazd and Abarqfih, now begins. We can set forth briefly what little is known about the history of Abarqfih at this time. It was noted above, p. 83, that Abti Mansfir Faramurz was in control of the town in 435/1043-44, in which year the Bilyids of Fars captured it from him, this being presumablya temporary conquest only. It is possible that Abarqfih had passed into Kakiyid hands as part of 'Ala' ad-Daula Muhammad's conquestsof 421/ Io30, together with Yazd (see above, p. 77). Apart from Ibn al-Athir's information about Abfi Mansfir Faramurz's rule in Abarqilh, the only evidence which relates to the town at this period lies in its oldest surviving monument, a mausoleum of characteristic northern Persian type, the Gunbadh-i 'Ali. The Ktific Arabic inscription on this building records that it was
1 Miles, " The Coinage of the KakwayhidDynasty", pp. 97,
72 Ibn

99,

102.

'3An interestingindication that the Turkmensstill felt uneasy

al-Athir,vol. IX, p. 384.

14

about being confinedby walls, for the walls of Isfahdnwould

have been a valuable protection for a town which was henceforth to serve as the Seljuq capital. Mujmal at-tawdrikh,p. 407; Ibn al-Athir, vol. IX, pp. 384-5; Ja'far b. Muhammad Ja'fari, Ta'rikh-i Tazd, ed. Iraj Afshtr (Tehran 1338/196o), p. 19. Ja'far, called Ibn Fasdnjus, whose kunya is given by Ibn alAthir, vol. IX, p. 358, as Abi 1-Faraj, which is near enough to Abfi 1-Fath. Ibn Fasdnjus served Abfi Kdlij*r 'Imid ad-Din when the latter entered Baghdad in 436/1045, but was dismissed in 439/1047-48 (Ibn al-Athir, vol. IX, pp. 358-9, 370-I). In any event, it seems that we have here a new name to include among Toghril's Viziers.

continued from previous page] should be added the Nasd'im al-ash(dr Ndsir ad-Din Kirmani, of now edited by Jaldl ad-Din IHusaini Urmawi [Tehran 1338/ 1959]), but the informal nature of the office amongst the Seljuqs at this early period makes an omission not wholly surprising. Iqbdl plausibly identifies him with the man later known to us as the Bfiyid Vizier Dhfi s-Sa'ddst Muhlammad b.

DAILAMIS

IN

CENTRAL

IRAN:

THE

KAKOYIDS

OF JIBAL

AND

YAZD

85

erected in 448/1056-57 by 'Amid ad-Din Shams ad-Daula Abti 'Ali Hazdrasp b. Saif ad-Daula alHIasanb. Nasr b. al-Hasan b. al-Firfizdn, as a burial place for himself and his mother, the Noble Lady ... (this name is, alas, indecipherable). This 'Amid ad-Din Shams ad-Daula is the grandson of the Dailami chief Nasr b. Hasan b. Firiizan, who played an active part in the warfare of the late fourth/tenth century in the Caspian region between the Bfiyids and Ziydrids. The Firizanid family were originally lords of Ishkawar or Shukir in the western part of Tabaristdn, and amongst them was numbered the adventurer Mikdn b. Kdki, rebel against the Sdmdnids,who was killed in 329/940-41' It is possible that when the Ziyarid Qabiis b. Vushmagir consolidated his power in Tabaristdn and Gurgdn, the Firiizinid family abandoned their old home and moved southwards to the territoriesof their kinsmen the Bilyids and b. Firiizdn had married a daughter of the Bfiyid Amir Kikfiyids (Hasan Rukn ad-Daula and Rukn ad-Daula had married a daughter of the Firfizanid, so that they were each father-in-law of the other; and Rukn ad-Daula was also the father-in-law of Sayyida, 'AlI' ad-Daula Muhammad Ibn Kkfikya'scousin). Thus the family may have remained quietly in Abarqfih, acknowledging perhaps the supremacy of the Kdkfiyids,down to 'Amid ad-Din Shams ad-Daula's death some time after 448/1056-57. Unfortunately, we know nothing of the subsequent history of the Firiizdnids or Kdkilyids in Abarqfih; it may be that the town reverted early to the Seljuq Sultans.76 In turning to the Kktkiyidsof Yazd, we do not, it is true, come up against the almost complete lack of information relating to their rule in Abarqfih. Nevertheless, the history of these Yazd Kdkiyids is much more obscure than that of their predecessorsin Isfahdn and Hamadhdn, whose fortunes can, as has been shown above, be traced quite closely from Ibn al-Athir, with contributory material from such sources as the Ghaznavid historian Baihaqi. For this later phase of Kikfiyid history, we have only very scattered mentions in Ibn al-Athir and in such Seljuq sources as Aniishirvin b. Khdlid, supplemented by occasional referencesin local historiesof the adjacent province of Kirmdn, but most important of all, by quite extensive mentions of them in the local historiesof Yazd itself. The first of these latter worksis Ja'far b. Muhammad b. Hasan Ja'fari's Ta'rikh-iYazd (often referredto by the subsequent historians of Yazd as the Ta'rikh-iqadim-iYazd),written in the first half of the ninth/fifteenth century." Not long Ahmad b. Husain b. 'Ali Kitib wrote his Ta'rikh-ijadid-i Yazd,and in the second half of the eleventh/ seventeenth century, Muhammad Mufid Mustaufi Bafqi put together his Jdmi'-i Mufidi, a vast compendium of historical and topographical information on his native town of Yazd; both these works draw considerably on Ja'fari's original work.78 Despite help from these sources, it is not possible to elucidate the exact chronology of the KikSiyid governors in Yazd; even the date of Abi Mansfir Fardmurz's death is not known. Numismatics, often so useful for these questions, provide no help; no coins of the Kakilyids are extant after one of Abfi Mansir Fardmurz minted in Isfahtn in 440/ and it is possible that subsequent Kikfiyids did not hold rights of minting gold and silver 10o48-49,19 coinage. We know little about AbUiMansir Fardmurz'slife in Yazd beyond his building activities there as described in the Ta'rikh-iYazd and the other local histories (see below), supplemented by one or two mentions in the chronicles. He was obviously an honoured and respected figure in Seljuq circles. He appears in these later years with a further laqab, Shams al-Mulfik, which may have been acquired at
75 See Ibn Isfandiyir, tr. pp. 225-6, 228-31; Rabino, Mdzandardn 77Not included in Storey's section on the local histories of Yazd,

after this, in the latter part of the reign of the Qara Qoyunlu Amir Jahan-Shah (841-72/1438-67),

and Astardbdd,p. 14o; and B. Spuler, El1 Art. " Firfiznids ". Spuler's statement here that the last member of the family to be mentioned is Kandr b. Firfizin b. Hasan, in 388/998, therefore requires modification. The Gunbadh-i 'Ali is described and discussed at length by Andr6 Godard in his article "Abarkiah (Province de Yazd) ", Athdr-e'Irdn I (1936), pp. 48-53 (I am grateful to Dr. G. FehtrvWrifor this reference). 76 Ibn al-Balkhi in his Fdrs-ndma (written in the early years of the sixth/twelfth century), p. 124, mentions Abarqfih, but makes no reference to the Knkfiyids as being there; on the other hand, he does not mention them either when he deals with Yazd, ibid., p. 122. 9

78

Persian Literature, vol. I, pp. 1293-4, but mentioned by Rypka in his Historyof IranianLiterature, 447p. See Storey, op. cit., vol. I, pp. 352, 1293-4, and Rypka, loc. cit., All these local histories have either been edited or are in process of being edited by the Persian scholar Iraj Afshir, who is himself a Yazdi and who promises a work on the history and town. for topography of the Afshar I am most grateful to Mr. his kindness in supplying me with copies of these texts at short notice. Miles, " The Coinage of the Kakwayhid Dynasty ", p. 97,
no. 21.

79

86

JOURNAL

OF PERSIAN

STUDIES

this later period.8s In 448/1056 Chaghri Beg Da'tid had married one of his daughters to the 'Abbasid Caliph al-Qa'im, and in 453/106I Chaghri's brother Toghril aspired to a much greater prize, an 'Abbasid bride for himself. That a Caliph's daughter should be given to a rough Turkmen was an unheard-of conception for the time, and al-Qa?'im objected strongly to the proposal. Hence to prepare the way, Toghril sent a delegation to Baghdad which was led by his Vizier, the 'Amid al-Mulk Kunduri, and which comprised some of his leading Turkish commanders and " the outstanding potentates of the Dailamis and Gil, including Malik Shams al-Multik Faramurz b. 'AlI' ad-Daula b. Kakfiya, Jalal ad-Daula Surkhab b. Kam-rawa,81 and Muntakhab al-Din Nasir al-Muluk Mantichihr b. Asfarsitan ".82 The Caliph yielded to political and financial pressure; but not till 455/1063, just before his death, was Toghril able to go to Baghdad and meet his bride. Again on this occasion, Abti Mansiir Faramurz was in the retinue which accompanied the Sultan, together with the Btiyid Abti 'Ali FanaKhusrau (or Kai-Khusrau) b. Abi Kalijar 'Imad ad-Din, lord of Kirmanshah and later of Naubandajan,83 Hazdrasp84 and the 'Annazid Surkhib b. Badr b. Muhalhil.85 Abti Mansfir Faramurz must have died shortly after this. The next member of the family whom we know of is his son Mu'ayyid ad-Daula or 'Ald' ad-Daula 'Adud ad-Din 'AlI, who like his father was high in the favour of the Seljuq Sultans, and in particular, of Malik-Shah. These later Kktkiyids all enjoyed a certain status at the Seljuq court as scions of a former dynasty of reigning princes, and regularly had marriage links with the royal family. Thus in 469/1076-77 'AlI' ad-Daula 'Ali married Arslan Khatun in Antishirvan b. bint Chaghri Beg, whose first husband, the Caliph al-Qa'im, had died in 467/1075; Kh~lid's words, " She exchanged the Qurashi for a Dailami, and the Imam for a barbarian (ummi) ".86 Yet this last sneer was far from being merited. The more leisured existence of the Kakfiyids of Yazd, compared with the hectic, war-filled lives of their forefathers in Jibil, gave them time to cultivate and encourage the arts of peace. Afdal ad-Din Kirmini says in his history of his native province, the 'Iqd al-'uld li-l-mauqif al-a'ld, that " 'Ala' ad-Daula, who was the ruler of Yazd, continually sought to attract the eminent men of both Khurasan and Iraq, encouraged them with all sorts of promises and expressions of favour, and brought them to Yazd ".87 'Ald' ad-Daula 'Ala's patronage of the great Seljuq poet Muhammad b. 'Abd al-Malik Nishaptiri, called Mu'izzi, is the subject of an anecdote in the Chahdr maqdla. According to this, Amir 'Ald' ad-Daula 'Ali, who was himself a great lover of poetry and the confidant of Malik-Shah, introduced Mu'izzi to the Sultan and thus started him on the path to fame (the takhallus or nom-de-plume of Mu'izzi derives from Malik-Shah's laqab Mu'izz ad-Dunyi wa
d-Din).88

The Diwdn of Mu'izzi contains three poems dedicated to the Kakfiyid prince. The first qasida is headed " In praise of Amir 'Ali' ad-Daula 'Ali b. Shams al-Mulfik, son-in-law of Chaghri Beg ", and speaks of him in these terms: The most exalted Amir, splendour of the world, 'All, who is the heart-encouraging one and the monarch who cherishes religion too. My verse has become exalted on account of the two 'Alis, one a leader in war, the other made great by his followers and devotees. 'All b. Abi Talib in heaven, and 'All b. Shams al-Mulfik here below. The first one is the son-in-law of Chaghri Beg, the other was the son-in-law of the Prophet. One is a fountain of noble qualities, because of his exalted birth, the other is custodian of the fountain of Kauthar.
80 wa-nukhbat Bundari, Zubdatan-nusra al-'usra,ed. M. T. Houtsma in Recueilde textesrelatifsa 'histoire Seljoucides, II (Leiden des vol. 1889), p. 133; Nisir ad-Din Kirmfni, Nasd'im al-asZhdr, 22. p. None of Abfi Mansfir Fardmurz's coins bear his honorifics. 81 Perhaps the son of Abfi KdlijAr 'Imdd ad-Din's son Abfi Talib Kdm-rawd, see Bowen, " The Last Buwayhids ", JRAS (1929), pp. 24I-2. 82 Bundari, p. 19; Ndsir ad-Din Kirmdni, p. 22. 83 See on him, and his honoured position amongst the Seljuqs, Bowen, op. cit., pp. 242-5. 84 Presumably the Kurdish chief Tatj al-Mulfik Hazarasp b. Bankir, muqta'of Basra in the time of the Bfiyid Feil1dh-Sutfin b. Abi K~lijir 'Imid ad-Din, cf. Bowen, op. cit., pp. 242-4. 85 Bundari, p. 25. For the background to all these marriage History of Iran, vol. V negotiations, see Bosworth in Cambridge The Saljuq and Mongol Periods (Cambridge 1968), pp. 48-9. On Surkhab b. Badr, see Minorsky, EI2 Art. "'Annazids "; he was an influential figure in the local politics of Kurdistan and Luristdn down to the end of the century. 86 Bundari, p. 52; Ibn al-Athir, vol. X, p. 72. 87 Ed.'Ali Muhammad 'Amiri Ng'ini (Tehran 1311/1932), p. 102. 88 Chahdr maqala, pp. 40-3, tr. pp. 45-8.

DAILAMIS

IN

CENTRAL

IRAN:

THE

KAKfYIDS

OF JIBAL

AND

YAZD

87

One of them went on the Khaibar expedition and broke down the gate of the town, the other has foes as fierce as those at Khaibar. All that fame which an illustrious name brings, comes from the celebrity of the name 'AIA'ad-Daula. The Caliph made his (sc. 'Ald' ad-Daula's) famous name his sword, a sword where auspicious fortune is made his servant.89 The second qa?ida is headed " In praise of Amir 'Adud ad-Din 'All b. Shams ad-Din Faramurz-i Kakiya " and is equally fulsome: You may ask, " What source of pride is there for me, the lover, when you, O Turk, never strike my heart and never concern yourself with such matters? But this is a greater source of pride for me, the fact that in the places of feasting and pleasure, all my eloquent praise is of you, Shams-i Mulfikin, 'Acdudad-Din. The noble and free-born prince, 'Ali b. Faramurz, the military protector and the boon-companion of the Supreme Sultan. His grandfather and father also had the honorifics of'Adud and Shams, and both these titles have acquired a new splendour and adornment because of him.90 The third qasida is " In description of a sword and in praise of Amir 'Ali b. Faramurz ", and eulogizes the Kdkfiyid prince thus: The most exalted Amir, 'Ali b. Faramurz, the emperor-like one, a Rustam in his ways of action, a Ma'n in his praiseworthy qualities, and like Sam in his exalted dignity. An Afrfsiydb in his royal power, the Siydvush of his age, the IsfandiyArof his time, and the Manfichihr of his house. He is the divinely-aided one, the just monarch (or: He is the one whose kingly power is divinely-aided, the just one); because of his illustrious forebearsand his monarchical power, when he wants a thing to be done, it is! My good fortune has become highly-exalted because of the noble virtues of the two 'Alis, and by singing the praises of the two 'Alis my nature has become excited by joy. The one chosen by the Prophet has become joyful-hearted through my eulogies, and the one honoured by Chaghri Beg has become merry because of them. The first one became powerful against the opponent of Islam, and the second has become successful in all the regions of Islam. The first one was skilful in warfare on the Chosen One's behalf, and the second has been a valiant hero in the affairs of the kingdom on behalf of the monarch (sc. the Seljuq Sultan).91 Noteworthy in all three extracts is the great play made with the name and titles of 'Ald' ad-Daula 'Ali,9l and the emphasis on his noble ancestry and descent from a family of kings. One discerns clearly that the social position of the Kdkiyids, as descendants of one of the noble houses of the Dailamis, was high at the Seljuq court, even if their political power was now very circumscribed. Indeed, the general position at this time of the Kdkiyids, of the scions of the Bfiyids and of various other Dailami and Kurdish chiefs, shows that one cannot altogether consider the Seljuq conquest of Iran as marking a sharp break with the past. There is little sign of any vengeful animus on the part of the Seljuqs against the Dailamis; the dispossessed Bfiyids were, on the whole, favourably treated; and Dailami families and chiefs continued down to the second half of the sixth/twelfth century to hold positions of local importance in the fairly decentralized structure of the Seljuq Sultanate.93 It may be that in the light of such facts as these, one ought to reconsider the whole of the generally-accepted view of the history of the fifth/ eleventh century, that the Seljuqs came to deliver the 'Abbasid Caliphs from the Btiyids, that the
89Diwdn, ed. 'Abbds IqbWl (Tehran 1318/1939), p. 121 11. 9' Also, Dailami commandersand soldiers continued to be an 2681-8. importantelement in the military and political structuresof 90 Ibid., p. 510o 11. 11975-8. the Iranianlandsdown to the end of the sixth/twelfthcentury; 91 Ibid., p. 522 11. I2239-45. see on the diaspora of Dailami mercenariesthroughoutthe 92 The fondness of the Kkfikyids for a limited number of personal Islamic lands, Bosworth, " Military Organisationunder the names (e.g. 'Ali, Garshisp, Abaf K~lij*r) and honorifics ('Ala' ad-Daula, 'A~lud ad-Din, Shams al-Mulfik) is also notable. Bfiyidsof Persiaand Iraq ", pp. 158-9.

88

JOURNAL

OF PERSIAN

STUDIES

Caliphs were eager to throw off the harsh yoke of the Dailamis and that the 'Abbasids welcomed Toghril and his Turkmens in Iraq with open arms; the need for such a revision of history has already been stressedon more than one occasion by ProfessorGeorge Makdisi.94 'Ala' ad-Daula 'Ali is one of the few Kakikyidsof Yazd whose date of death is known with any certainty. He espoused the cause of Malik-Shah's brother Tutush when the latter in 448/1095 raised a rebellion against Sultan Berk-yaruqb. Malik-Shah, and was killed in the battle near Ray which ended Tutush's bid for the throne.95 His son 'Ala' ad-Daula 'Adud ad-Din Abfi Kdlijar Garshasp also held Yazd and was high in the favour of Sultan Muhammad b. Malik-Shdh, having married the sister of Muhammad and Sanjar. It is recorded by Ibn al-Athir that in 501/I Io8 Abai Kalijar Garshasp b. 'Ali and also the sons of Borsuq were amongst the military leaders in the army of Sultan Muhammad when the Seljuq marched against the Mazyadid Amir of Hilla, Saif ad-Daula Sadaqa.96 When Muhammad died in 511/111I8, 'Ala' ad-Daula Garshaspwas at first again highly regarded, as one of the group of descendantsof Dailami rulers, at the court of his son and successor,Mahmfid; according to Anfishirvan b. Khalid, the new Sultan treated him like one of his own brothers. This favour was, however, short-lived. 'Ala' ad-Daula Garshaspwas aware of a group of enemies at court, and tended to stay in Yazd. His infrequent attendances at the court in Isfahan aroused Mahmfid's suspicions,and he sent a military force to Yazd in order to arrest him. 'Ala' ad-Daula Garshasp was surprised by this force, seized and jailed at Farrazin, a castle near Karaj in Jibal.97 Yazd was now granted to the Sultan's cupbearer Qaraja, later governor and Atabeg of Fdrs.98 'Ala' ad-Daula Garshasp finally managed to escape from custody and returned to Yazd, where the local populace received him with joy. He sought the protection of Sanjar in Khurasan, and played an important part in urging him to march westwards against his nephew Mahmild, giving Sanjar information about the roads through Iran and about the personal rivalries and dissensionsin Mahmiid's entourage. Hence in the battle which took place at Sawa in Jumada I 513/August II 19 between the two Seljuq monarchs, 'Ala' ad-Daula Garshaspwas one of the five kings said to have fought in Sanjar'sarmy.99 Presumablyhe was restored to Yazd when the victorious Sanjar was able to impose his conditions on Mahmfid. There is mentioned in Muhammad b. Ibrahim's history of the Seljuqs of Kirman a dispute within the Kaikiyid family in Yazd arising during the later years of the Seljuq Amir of Kirman, Arslan-Shah
b. Kirman-Shah (495-537/1101-42), in which Arslan-Shah intervened to restore the dispossessed party:

and Ald' ad-Daula. Amir' Ali b. Faramurz fled for refuge to the court of Kirman, and contracted a marriage alliance. Amir Arslan-Shah handed Yazd over to him and they sent Amir Muhammad b. Kai Arslan as military governor of Yazd.0oo The difficulties raised by this laconic scrap of information were noted by Houtsma.10l The 'Ala' ad-Daula mentioned could well be 'Ala' ad-Daula Garshasp b. 'Ali, but the mention of'All b. Faramurz is .ertainly anachronistic; unfortunately, we have no other source to throw light on it. Concerning the end of 'Ala' ad-Daula Garshasp's reign, we have only the information in the Ta'rikh-i razd of Ja'fari, substantially repeated by later historians of the town, that Amir 'Ali b. Faramurz was killed fighting at Sanjar's side against the Qara Khitai, sc. at the Battle of the Qatwan Steppe in 536/I I41.102 This can
94

In the latter part of his reign [sc. of Arslan-Shah's] Yazd some causes of emnity arosebetweenhim in

at-tawdrikh, a contemporary source, gives an exact date for E.g. in his article " The Topography of Eleventh Century 'Ald' ad-Daula Garshasp's arrest, Tuesday 2ist Rabi' I 513, Baghdad: Materials and Notes (II) ", Arabica VI (1959), with his escape falling in Rajab of that year. If this dating is de pp. 297-8, and in his book Ibn 'Aqil et la risurgence l'Islam traditionaliste XIe sidcle(Ve sidclede l'Hegire) (Damascus 1963), au accepted, he cannot have gone to Sanjar's court and then taken part in the Battle of Sawa, which took place in Jumida I pp. 77 ff. 513; in any case, the 21st Rabi' I 513 was a Wednesday and 9 Mujmal at-tawdrikh,p. 409; Ibn al-Athir, vol. X, p. 312. 96 Ibid., cf. Bosworth in Cambridge not a Tuesday. It would be more in harmony with the other History of Iran, vol. V, p. I 5. 97 Cf. Le Strange, The Lands of the EasternCaliphate, sources to take the year in question as the previous one, 512, p. 198. in which the 2Ist Rabi' I actually fell on a Friday. 9s He was later killed by Sanjar in 526/1132, see Bosworth, op. 100 Ta'rikh-i Saljiqiydn-i Kirmdn, ed. Houtsma in Recueil de textes cit., pp. 124, 243. vol. I (Leiden I886), p. 26. 99Mujmal at-tawdrikh, 414; Bunddri, pp. I33-4; Ibn al-Athir, relatifs a l'histoiredes Seljoucides, p. vol. X, p. 387. The whole episode is cited by Anfishirvqn b. 101" Zur Geschichte der Selguqen von Kerman ", ZDMG XXXIX (1885), pp. 374-5Khilid as yet another illustration of the blunders and misb. Muhammad. The Mujmal 102 Ta'rfkh-i razd, pp. 21-2. management of Ma.hmid

DAILAMIS

IN

CENTRAL

IRAN:

THE

K.AKfYIDS

OF JIBAL

AND

YAZD

89

hardly refer to the real 'Ali b. Faramurz who died, as we saw above, fighting for Tutush b. Alp Arslan fifty years before this. Either 'Ala' ad-Daula Garshdsp must be meant, or else the whole item of information has been confused with the death in battle of the real 'Ali b. Farimurz. None of the other sources dealing with Sanjar's defeat at the hands of the Qara Khitai mention the presence there of the Kakikyid prince.

V
It is time now to look at the local histories of Yazd and see what additional information about the Kdkfiyids can be extracted from them. The amount of this information is quite large, but as we shall see, it is nowhere near contemporary evidence, and the problems of evaluation are considerable. We may conveniently start with the relevant section in the oldest of these texts, Ja'far b. Muhammad Ja'fari's Ta'rikh-i Yazd. The passage in question runs as follows:103
[p. 19] Sectionfour

worksthere and Toghril'sgrantof Tazd to AbuiMan;r 'Ald' ad-DaulaFardmurz, theconstructional downto the timeof theAtabegs of Abi Man;r and his descendants, When Toghril came to Isfahan, Abfi Mansfir sent presents. In turn, Toghril sent [an envoy] to Abfi Mansfir with the following message: "Although you come from a family of rulers, you do not possess much of an army. I, on the other hand, have I8,ooo falconers alone, and I have with me an army too great to be numbered; and Isfahin is now the place where my army is encamped. Hand over the city to me, and I will give you in exchange any place in 'Irdq ['Ajami] which you like". Abli Mansfir replied, "Give me the city of Yazd, and let me transfermy residence thither". Toghril further gave his brother's daughter Arslan Khatun in marriage to Abfi Mansfir, and wrote out an investiture charter which stated, " We have established Yazd as Abfi Mansfir's centre for worship and pious exercises (Ddr al-'Ibddax04), and have despatched him to take up residence there ". Abfi Mansfir came to Yazd with his retinue of Dailamis, and they constructed there a lofty palace. His four chief commanders-Abfi Mas'lid Bihishti, Abfi Ya'qfib, Abfi Ja'far and Abfi Yfisuf-constructed a fortified wall round the town, with four gates, each with iron doors. The first was called the Gate of Kiya, the second the Gate of the Qatariyin,105 the third the Gate of Mihrijird, and the fourth the Gate of the New four commanders was inscribed on the door of the gates, together with the date.106 They also built the Friday mosque in the inner city.x07 Arslan Khatun built a mosque and minaret in Durda. It was the first minaret ever built in Yazd, and it has remained standing till the year 807/[I404-0o5.108 Kiyd Narsfi had a qanat dug, known as the qandt of Narsfi-dbbd, and he founded the village of Narsfi-abid.109 Abfi Ya'qfib constructed a qanit known as the Ya'qfibi qandt, and founded a village, also named after himself. Kiyd Narsfi also laid out a cemetery on the
103
104

Palace. These walls and gates were constructed [p. 20] in the year 432/[I040-41].

The name of each of the

Ibid., pp. I9-23. According to Ah.mad b. IHusain b. 'All Kitib's Ta'rikh-i jadid-i Tazd, ed. Afsh~r (Tehran 1345/1966), p. 59, followed by Muhammad Mufid MustaufiBSfqi in his Jami'-i Mufidi, ed. Afshdr (Tehran 1340-42/I961-63), vol. III, p. 735, the name D~r al-'Ibqda became a synonym for the town of Yazd. 105 Could this name be a reminiscence of the followers of the leader of the Azraqi sub-sect of the KhawArij, the famous warrior Qatari b. al-Fujd'a? During the Caliphate of 'Abd al-Malik, the Azdriqa controlled most of southern Persia, including Fars and Kirmdn, until Qatari was forced by internal division in his community to move northwards, via the fringes of the Great Desert, to TabaristAn and his final death in 78/697-98, cf. G. Levi della Vida in EIl s.v. It would not be at all surprising if Qatari at some time in his career held or passed through Yazd. The name of the Gate of the QatariyAn was certainly long-lasting in Yazd. It is mentioned several times in the Jdmi' al-khairdt, the Waqfnamaof Sayyid Rukn ad-Din Husaini Yazdi (d. 732/1331-32), Vizier to the Il-Khanid Sultan Abf Sa'id (ed. M. T. Danish-

Pazhfih and I. Afshar [Tehran 1341/1962], pp. 66-7, etc.). This work, which enumerates all the charitable works done by the minister in his home town, mentions a large number of the places, gardens, qandts, etc., which appear in this passage by Ja'fari on the Kikfiyids. 106 Amending the text here from dar to bar-a. 107 This information about Abfi Mansfir Fardmurz's arrival in Yazd and the building operations of his four commanders is attributed in the two later histories of Yazd, the Ta'rikh-i jadfd-i razd and the Jami'-i Mufidf, to the Amir Abfi Ja'far 'A1i' ad-Daula b. Majd ad-Daula Abfi Kdlijar (sic); see further, below. 10s However, Muhammad Mufid adds that in 832/1428-29 the foundations and supports of the mosque showed signs of weakness, and the building became ruinous (Jdami'-iMufizdf, vol. III, p. 737). 109 According to Muhammad Mufid, op. cit., vol. I, p. 78, in his time the village of Narsfi-abdd was largely inhabited by IHusaini sayyids.

90

JOURNAL

OF PERSIAN

STUDIES

and built a lofty mausoleum there, [at the place] which is known today as " the edge of the sandy desert,110 highway of Fakhr-i Jallid ". Kiyd Shuja' ad-Din built a madrasa and a mausoleum, also on the edge of the sandy desert, at the side of the place where there is the cool well of the desert fringe, dug out by Amir Ahmad Tabrizi. Lild Raihln built a cemetery and a lofty mausoleum, adjacent to the refuse-dump, and the [other] amirs and commanders of 'Ala' ad-Daula for their part erected many buildings. Khwaja Sawab, who was major-domo of Arslan Khatun's palace, constructed the qanat of Sawab; the whole town of Yazd benefited from that water, and most of the buildings of the town were supplied with its waters. Two slave girls of Arslan Khatun, called Abr (? Abarr) and Mubaraka, constructed two interconnected qanats, named after themselves. When the office of Sultan of the Seljuq dynasty passed to Sultan Malik-Shah b. Alp Arslan, 'Ala' ad-Daula Abfi Mansfir became the Sultan's courtier in Isfahan, whilst his son 'Ali b. Faramurz [p. 21] went to become Sultan Sanjar's companion in Khurasan and became his son-in-law. When 'Ali ad-Daula Abfi Mansfir Faramurz died in Isfahan, the Sultan granted Yazd to the former's son 'AlI' ad-Daula Garshasp. Garshasp brought back his father's body to Yazd,"' built a madrasa with two minarets (the one thus called " the madrasa with two minarets "), a lofty mausoleum with a subterranean vault, and he buried his father there. Now, his still-intact body, exactly as it was, is laid to rest on a platform in that vault. This burial-place was constructed in the year 523/[ I129]. Likewise, Garshisp built a Friday mosque in the quarter of Durda, together with a lofty assembly-hall and a library. He brought water from the stream which flows in the qanat of Zirich to the mosque. One has to descend about seventy steps to this qanat. Malik-Shah died in Baghdad. Warfare broke out amongst his sons Berk-yaruq, Mahmfid and Muhammad, and the fighting between them lasted for a considerable time. Muhammad b. Malik-Shah eventually secured firm control of the kingdom. After him, power passed to his son Abfi 1-Qasim Mahmfid; he became involved in warfare with his paternal uncle Sanjar, and fell into the latter's hands as a captive. Sanjar pardoned him; he gave him the hand of his daughter, together with an elephant with a golden litter; bestowed on him the town of Khuwar of Ray; and then finally sent him back to Isfahan. When Sultan Sanjar fought with the Khan of Khita and the Seljuq army was defeated, about 50,000 of the Sultan's most prominent warriors were killed, and the pagans captured Malik Taj ad-Din, ruler of SistAn, together with the Sultan's harem, and carried them off to Khita; it was a year later before they were sent back. 'All [p. 22] b. Faramurz (read: Garshasp b. 'All) was killed in that battle, so Sanjar handed Yazd over to his daughters. Garshasp had been governor of Yazd for about forty years. The Garshispi Garden is one of his surviving monuments, together with the market running round the two opposite ends of the old Friday mosque. After that, his two daughters ruled in Yazd. Section five Shdh Mentionof theAtabegsof razd, from Ruknad-DfnSam and 'Izz ad-Din Langardownto Trisuf worksin Tazd b. SalghurShdh,and theirconstructional After Garshisp's death, and when the Seljuq Sultanate had passed to Sultan Arslan b. Toghril, Yazd remained in the hands of 'Ald' ad-Daula Garshasp's daughters. The Sultan ordered that one of the great commanders of Garshisp's army should become Atabeg for the two daughters and look after the safety and security of the state. So Rukn ad-Din Sim b. Langar was made Atabeg. He acted as Atabeg for a while, and was responsible for building a madrasa called " the madrasa of the Atabeg ". But he lacked the necessary competence and judgement for running the state, so the prominent commanders resolved to set up his brother 'Izz ad-Din Langar (or read with iddfat, 'Izz ad-Din-i Langar?) as Atabeg. The latter was a brave young warrior, and during the period of Seljuq dominion had exercised many positions of responsibility; he had displayed great valour in the fighting with Mengfi-bars, and had stormed and captured the walls of the city of Shiraz from him.112 He now took charge of the government of Yazd. ...
[p. 23]

110

Sar-rfg. This appears in the Jmi' al-khairft, pp. 34, 37, 175, etc., as both Sar-rig and Ra's ar-Ramal, and is the name given Muf id, quoting to a quarter outside the town. Mul.ammad from the Ta'rikh-ijadid-i razd, says of the Masjid-i Jum'a of Sar-rig (a building dating from the early eighth/fourteenth century) that there was there a caravanserai with the cemetery adjoining it (Jdmi'-i Muffdi, vol. III, p. 650). "x It will be remembered that 'Ali b. Faramurz died in battle fighting for Tutush b. Alp Arslan, see above, p. 88.

112

MengiA-barshad become ruler of Firs after Sanjar had killed Qaracha Saqi in 526/1132, and had secured from Sultan Toghril b. Muhammad the position of Atabeg to his son Alp Akhbdrad-daulaas-saljaqiyya, Arslan ($adr ad-Din ed. M. Iqb1l [Lahoreal-I.usaini, iox). In 532/1138, however, 1933], p. he was killed in the Battle of Panj-angusht, when the Seljuq Dq'fid b. Muhammad and the ex-Caliph ar-Rdshid were defeated by Sultan Mas'fid b. Muhammad, see Bosworth in Historyof lran, vol. V, p. 129. Cambridge

DAILAMIS

IN

CENTRAL

IRAN:

THE

KAKKUYIDS

OF JIBAL

AND

YAZD

91

The information in this passage-and indeed in all the local histories of Yazd-on the chronology of the Kakfiyid family is clearly confused, and one does not feel much confidence in the accuracy of the dates given in it. That of 432/1040-41 for the constructionof the walls of Yazd is apparently too early, did since the Kakikyids not transferto Yazd from Isfahan till after443 / 1051 (unlessthere is a conceivable for this apparent anachronism in the suggestion made below, p. 93). It is otherwise explanation Faramurz suvived long enough to become the boon-companion of Malikunknown that Abfi Mans&ir at least, he is not heard of after 455/1063 (above, p. 86). It is doubtless his son 'Ala' ad-Daula Shah; 'Ali who is meant here. Moving on to the dynasty's last days, if a Kakilyid was present at the Battle of the Qatwan Steppe, it was more likely Garshaspb. 'Ali, and certainly not 'Ali b. Faramurz. With the confusing nomenclature of the Kakikyids,arising from their use of a limited number of names and honorifics,it is not surprisingthat Ja'far b. Muhammad, writing three centuries or so later, should have been led astray. The local historians of Yazd who wrote after him, sc. Ahmad b. Iusain b. 'Ali Katib and Muhammad Mufid Mustaufi Bafqi, drew very largely on the original Ta'rikh-irazd. Naturally, Ja'fari's errorsand incongruitiesare repeated and magnified, understandablyin the case of Muhammad Mufid, who was writing two-and-a-half centuries after Ja'fari and some six centuries after the events described. It seems that Muhammad Mufid drew his material via Ahmad b. Husain rather than directly from Ja'fari. The confusing nomenclature and titulature of the Kakiyids, already mentioned above, led these later historiansinto errorsmore serious than those found in the Ta'rikh-i Tazdof Ja'fari, in so far as we know the correctform of these names and titles from contemporaryand near-contemporarysourceslike Aniishirvan b. Khalid and Ibn al-Athir. The chronology of reigns and the the Mujmalat-tawdrikh, of events become even more aberrant, and it seems that there was a distinct decline in accurate dating historical knowledge about the Kakikyidsin the course of the generation or two separatingJa'fari from Ahmad b. Husain. Thus the first Kakfiyid known to Ahmad b. IHusainis " Abai Ja'far 'Ala' ad-Daula b. Majd adDaula Abai Kalinjar (read Kalijar) "; the Kktkiyids' and Btiyids' common Dailami origin and their kinship ties account perhaps for this hybrid name. In Ahmad b. IIusain's account, it is this Amir who and occupies himself transfersfrom Isfahan to Yazd, where he henceforth establishes his Ddr al-'Ibddd with pious and charitable works; but the Seljuq Sultan responsiblefor this is transformedfrom Toghril -who is never mentioned at all in this connection-into Malik-Shah. It seems to have been MalikShah and Sanjarwho, to later generationsin Iran, stood out above all other Sultans of the Great Seljuq dynasty. Muhammad Mufid follows his own account (derived from Ahmad b. Husain) of the transfer of the Kkafiyids to Yazd, with a lengthy digressionon Malik-Shah'srelationswith his celebrated Vizier Mustaufi's Nizam al-Mulk, drawing on such sources as Zamakhshari's Rabi' al-abrdr, Mirkhwand'sRaudatas-safd'and Khwandamir's Habibas-siyar; this digressionhas no .Iamdallah Ta'rikh-i guzida, connection with the history of Yazd, but the author apparently felt that he could not pass by the Seljuq period without giving this story.113The dating of the transfer to Yazd is placed by both the later historiansof Yazd well beyond Malik-Shah's actual death, to 504/I Iio-i i, by which time this Sultan had been dead for almost two decades.'14 'Ala' ad-Daula's wife Arslan Khatun, constructorof a new mosque and minaret at Durda in Yazd,
and patron and employer of various other persons active in constructional and irrigation works in the town, is identified in the two later histories of Yazd with the daughter of Malik-Shah's paternal uncle Sulaimin-Shah, given in marriage at the time of 'Ala' ad-Daula's transfer to Yazd as a token of the Sultan's thanks at the Kikiyid's complaisance in the matter.'l5 It is conceivable that there was some liaison arranged in Malik-Shah's reign between a Kakiyid and a Seljuq princess, the Sulaiman-Shah named here presumably being Alp Arslan's brother Sulaiman b. Chaghri Beg, a nonentity whom the Vizier Kunduri vainly tried to set up as Sultan on the death of Toghril in 455/10o63.11 But it is more
113Jami'-i
114

Mufidf, vol. I, pp. 45 ff. Ta'rfkh-ijadid-i Yazd, p. 59; Jami'-i Mufidf, vol. I, p. 44, vol. III, pp. 734-5-

1'5

Ta'rikh-i jadid-i Tazd, pp. 60, 62; Jdmi'-i Mufidi, vol. I, pp. 76, 79-80, vol. III, pp. 644, 709, 712, 734-5, 737116 See Bosworth in Cambridge Historyof Iran, vol. V, pp. 54-5.

92

JOURNAL

OF PERSIAN

STUDIES

likely that this information relates to the marriage in 469/1076-77 of 'Ala' ad-Daula 'All b. Faramurz and Chaghri Beg's daughter Arslan Khatun (who would, of course, be Sulaiman's sister and not daughter), authenticated from its being mentioned by Aniishirvan b. Khalid and Ibn al-Athir (see above, p. 86). Arslan Khatun's life as consort in Yazd made a lasting impression on the historical consciousness of the Yazdis, and her public and charitable works are described at length in all the histories of the town, with many details in the Ta'rikh-ijadid-iYazd and the Jdmi'-i Mufidi additional to those of the Ta'rikh-irazd. Thus it is said that she regularly distributed clothing to the indigent, and each day held two communal meals, one for the upper classes and one for the lower, the highways and byways being scoured for those in need. She constructed the mosque and minaret at Durda, as mentioned by Ja'fari (above, p. 89), and the furtherdetail is given that she had a band of tiles inscribed with her titles placed round the minaret just below the muezzins' balcony and chamber (the qafasa-yi
mandr). According to Ahmad b. Husain, the minaret collapsed in 832/1429. By Muhammad Mufid's

time, the mosque itself was partly ruined, and another mosque built by Arslan Khatun in the inner city of Yazd had disappearedwithout trace."' The buildings of both Arslan Khatun and her husband are all placed by the two later historians of Yazd within the first third of the sixth/twelfth century. Ahmad b. IHusainmoves on now to speak about the Amir 'Ali-yi GarshAsp'Ala' ad-Daula, the son (understand " grandson ") of the original 'Ala' ad-Daula b. Majd ad-Daula Abfi Kalijar and Arslan Khatun, and also about the latter couple's daughter Ata Khatun. It is mentioned that this Amir 'Ali was Mu'izzi's mamdah, it seems that the historical 'Ali b. Faramurz, who was certainly the patron and of Mu'izzi (see above, pp. 86-7), is meant here. Ahmad b. Husain goes on to say that Ata Khatun married Malik-Shah's son Mahmad and had a son by him called Ata Khan. Malik-Shah's historical son Mahmiid was, of course, raised briefly to the throne after his father's death in 485/1092 by MalikShah's widow Terken Khatun and her personal intendant or wakil, Taj al-Mulk Abi 1-Ghand'im; but this Mahmfadwas only four years old at that time and died two years later.118It seems obvious that Malik-Shah's grandson, the later Sultan Mahmid b. Muhammad, is really intended. Ahmad b. IHusain's mysteriousAta Khan-whose existence is unattested in the more nearly-contemporarysources -is said to have resided in Yazd and to have been active as a builder there.119Also mentioned at this juncture is a Terken Khatun (not to be identified with the famous wife of Malik-Shah, mentioned above, unless we have another conflation of persons and confusion of names, which is by no means unlikely), described as the daughter of Sultan Mahmid and a slave concubine; Terken Khatun's hand was sought by one of the Seljuqs of Kirman, but besides this marriage, she also built a madrasa, named after her, in Yazd.120 However, mention of Ata Khan and Ata Khatun fades out; the formeris said to have been killed at Isfahan in the civil warfare between Malik-Shah's sons Berk-yaruqand Muhammad.121 We now have fairly extensive mention of Amir 'Ali's son Faramurz, who must be identified with the Garshaspb. 'Ali of the earlier sources; confusion of the two personal names Garshaspand Faramurz seems to have been chronic in the Yazd historians. Ahmad b. Husain and Muhammad Mufid simply repeat here substantially the information of Ja'fari, that Fardmurzwas Sanjar's boon-companion and married one of the Sultan's cousins ('ammzdda);we know from Ibn al-Athir that Garshisp b. 'Ali was actually Sanjar's brother-in-law (see above, p. 88). There is no reference to Garshasp'stribulations at the hands of Sultan Mahmad b. Muhammad, but the information of the Ta'rikh-iYazd that he died fighting the Qara Khitai is further retailed. He had, it seems, been a great builder in Yazd, laying out the Gard-i Faramurzor districtofFaramurz, still thus called and still populous in Muhammad Mufid's own day.122
We are finally left with the story of the transition to Atabeg rule in Yazd, caused by the failure of the male line of the Kakiyids there after Garshasp b. 'Ali's death, this story being given in the same words as those of Ja'fari.x1s
117

Ta'rikh-i jadid-i Yazd, p. 60o; Jdmi'-i Mufidi, vol. I, p. 76, vol. III, pp. 735-6. Historyof Iran, vol. V, pp. 103-5. 1'8 See Bosworth in Cambridge 119 Ta'rfkh-ijadid-iYazd, pp. 62-6; Jmi'-i Mufidf, vol. I, pp. 80-I. 120 Ta'rfkh-ijiadfd-irazd, pp. 63-4; Jdmi'-i Mufidf, vol. I, loc. cit. 121 Ta'rikh-ijadid-i Tazd, p. 65; Jdmi'-i Mufidi, vol. I, p. 81.

a12Ta'rikh-ijadfd-i razd, loc. cit.; Jdmi'-i Mufidt, vol. I, pp. 81-2, vol. III, pp. 714-15. No doubt the Gard-i Fardmurz was, in reality, named after Abfi Mansfir Fardmurz b. 'Ala' ad-Daula Muhammad, the first Kakfiyid to settle at Yazd. 12a Ta'rikh-i jadid-i Tazd, pp. 65 f.f; Jdmi'-i Muftdt, vol. I, pp. 82 ff.

DAILAMIS

IN

CENTRAL

IRAN:

THE

KAKOYIDS

OF JIBXL

AND

YAZD

93

The value of the passage on the later Kikiiyids cited from the Ta'rikh-irazd lies as much in the rich information it gives on the topography of the town and on building activities there, as in its purely historical information; and this topographical information is amplified by that in the subsequent local histories. Before the establishment of the Kakiiyids there, Yazd had been a town of very moderate standing, making little impact on the wider scene of Iranian history. We do, however, learn from Istakhri (wrote c. 340/951) that in the fourth/tenth century it was already a town with some fortifications and a citadel (hisdr),this last having two iron gates, the Bab Izad and the Bab al-Masjid, this second one being adjacent to the Friday mosque. The Kakfiyids were thus not the first to build protective walls round Yazd, although their ones were more elaborate than those previously existing.124 Much use was also already made ofqandts to provide Yazd with a water supply and irrigateits fields and orchards,so that despite its proximity to the Great Desert, it produced fruit of excellent quality, with a surplus for export. In his useful notes to the Ta'rikh-irazd, Iraj Afshar cites an article by Jawad Majd-zada Sahbd, with a postscript by 'Abbas Iqbal, which deals with the question of the gates of Yazd and of the four Amirs of Abt Mansir Faramurz who erected them.125Sahbd describes an inscription from one of the Yazd gates, that of Mihrijird; the gate itself has now been demolished, but parts of an iron plate from it have been preserved by the local authorities.'26 One of these fragments has on it a Kaific Arabic inscription, which Sahba gives and which may be translated into English as follows: The FortunateAmir Abfi n-Najm al-Bihishtiand the VictoriousAmir Abfi Ya'qfib, [that is] Badr and of Ishdq the two sons of Yindl, clientsof the Commander the Faithful(may God prolongtheir existence!),
ordered this to be constructed and erected in the year 432 /[1o40-41].

Abi Ishlq al-Isfahdni, Blacksmith. the

[This is] the work of Muhammad b.

In regard to this, Iqbdl comments that Ja'fari's " Abji Mas'fld Bihishti" must be a mistake, and this seems probable; the present author suggeststhat al-Mas'adand al-Muzafar should be taken as they are translated into English above, sc. as descriptive epithets and not as personal names, echoing the wellknown practice amongst the Sdmanids of referringto the Amirs in this fashion. Iqbal also adduces a relevant piece of evidence from Ibn al-Athir, who under the year 435/1043-44 describes the fighting between Abfi Mansfir Fardmurz and the Bilyid AbQiKalijar 'Imdd ad-Din over the possessionof two fortresseson the borders of Kirman; in this warfare, the Kdkiyid army was defeated and the commander of the forces, the Amir Islhq b. Yindl, captured by Abil Kalijar 'Imid ad-Din.127One would very much like to see the inscriptionitself, since the text as given by Sahbd contains several grammatical and textual incongruities. The date in the inscription accords well-perhaps too well-with that in the literary source of the Ta'rikh-iYazd. We have noted that Abil MansoirFaramurz did not transferhis residence from Isfahin to Yazd till more than a decade after then, but it is by no means impossible that Yazd was in Kiakiyid hands before 443/o05I1, perhaps from the time when 'Ala' ad-Daula Muhammad minted a coin there, sc. in 421/1030 (see above, p. 77). Unfortunately, the literary sources are wholly silent about this. For the present, the question must rest undecided, but the present author feels that more evidence is necessary before we can posit an extensive building programme like that ascribed to the Kakfiyids and their servants in Yazd at a pre-443/105i date.'28 Other sections of the Ta'rikh-iTazd provide a certain amount of additional information on the constructional and irrigational activities of Ab_ Far-murz's amirs and courtiers. Amir i Mans-ir
124

Kitdb Masdlik al-mamalik (Cairo 1381/x961), p. 77; Ibn Hauqal, Kitdb al-ar4, ed. J. H. Kramers (Leiden 193839), vol. II, p. .Sarat cf. Le Strange, The Lands of the Eastern 280; Caliphate,p. 285. 125 The article appeared in Yadgar PP. 72-80. Unfortunately, 1/3, I have not been able to see this article directly. 126 Ahmad b. Ifusain says that " The names of these four commanders were incised in Kfific script on iron plates on the four gates " (Ta'rtkh-ijadfd-i Tazd, p. 61). 127 Ibn al-Athir, vol. IX, p. 355. 128 The most celebrated, and indeed, the only building which survives today in Yazd from this period is the shrine of the

Duvizdah ImAm. Schroeder says that it is "for a historian, one of the most important buildings in Persia "; its importance to the historian of Persian architecture lies in its use at an early date of squinches to carry the dome on the main walls. The Kfific inscription on one of the building's panels shows that its date of construction was 429/1037-38, i.e. shortly before Yazd became the new centre of the Klkfiyids. See A. U. Pope and Phyllis Ackerman, eds., Surveyof Persian Art (Oxford 1938-39), vol. III, pp. IooI-o4 (E. Schroeder), 1279-80 (Pope), with plates in vol. VIII, pls. 273-4; and Pope, Persian Architecture(London 1965), p. Ioo and fig. 351-

94

JOURNAL

OF PERSIAN

STUDIES

Bihishti, whatever the correct form of his name may be, is obviously the constructor of the Bihishti Garden. Ja'fari says that " In the year 747[/1346-47], Muhammad b. Muzaffar (sc. the founder of the Muzaffarid dynasty of southern Persia and Iraq) took possession of various properties outside the town gates, such as the quarter of the chest makers, the [street of the] hyacinths, the Bihishti Garden, the madrasa of the Atabeg Sam, the Khalaf madrasa, the street of [Muhammad] Jalal, the hospital, the [quarter of] Ilchi Khan, the quarter of the Kiya Gate, the shrine of Sharaf ad-Din Khidr [S.hibi] and the garden of the jujube trees ".129 In his section on the qandt of Khwaja Sawab, the author mentions that in his time, sc. the early ninth/fifteenth century, it was dried up, and most of the water which had flowed from the town to the Bihishti garden no longer ran.130 With regard to Amir Abil Ya'qiib Ishdq, Ja'fari mentions a Ya'qfbi cemetery outside the town and devotes a section to describing the Ya'qfibi qandt. This ran outside the town, originating from the village of Mihrijird, and supplied sweet water to the Ya'qfibi garden, orchards and vegetable gardens. These latter tracts were contiguous with other stretches of cultivated land which were used for market gardening, those of Murydbdd and Salghurabad; they were supplied with streams of running water, and had many irrigation contrivances; the vegetable patches grew carrots, turnips, cotton and millet.131 Of the other qandts built by notables of the Kkfikyid ruling class in Yazd, Ja'fari speaks in detail about those of Arslan Khatun's two slave girls, and that of Abti Manstir Faramurz's general Kiya Narsii. The qanits of Abr and Mubaraka were interconnected, and arose from the same source. The qandt of Abr appeared on the surface in the quarter of the Zoroastrians, and this was a drawback, because the Majfis controlled the fountain-head where it rose to the surface. But the water was extremely sweet and light, and the cognoscenti of the waters of Yazd considered it to be the lightest of all. These two qandts were still very much used in the eighth/fourteenth century, as attests the Jdmi' al-khairdt. Its author says that they are so well-known that he does not need to describe their courses in detail; and he also mentions a mill of Mubaraka, near the Garden of Amir Sam.132 The Narsii-dbdd qandt was built by the general Narsfi; its source was in the Mihrijird district, and it ran through the town, watering a large number of vegetable gardens en route, to the quarter of Narsi-abad.133 Lastly amongst this topographical information, one may note that the village which gave its name to the qandt of Zdrich was one of those around which Muhammad b. Muzaffar built fortifications in 740/1339-40. The other villages named are those of Khdtiindbid, Dailamdbdd and Turkdbdd, and the obvious inference is that these settlements date from Kakliyid times, the first being a foundation of Arslan Khatun, whose charitable and constructional activities have been extensively mentioned above, and the other two being settlements of the Kdkiyids' Dailami and Turkish troops.134 VI The opening sentences of Section Five of the Ta'rikh-i Yazd describe for us the final disappearance of the Kikfiyids of Yazd in the third quarter of the sixth/twelfth century, or rather, their metamorphosis into a line of Turkish Atabegs. This last process was a familiar one all over the Middle East at this time; one might, for instance, compare it with the succession of the Ahmadill Atabegs to the heritage of the Rawwddids in Azerbaijan. Indeed, many Muslim historians found it difficult to separate an original family or dynasty from its Atabeg epigoni. Through the marriage of Garshdsp b. 'Ali b. Faramurz's daughter to the first Atabeg Rukn ad-Din Sim, it became true that Kiakiyid blood flowed in the veins of these Atabegs ofYazd; hence IHamdallah Mustaufi, reporting the death in battle against the Mongols at Isfahan, whilst fighting at the side of the Khwarazm-Shth Jalal ad-Din, of 'Ali' ad-Daula Ata Khan Yazdi, is careful to stress his ancestry back to the founder of the line, 'Alt' ad-Daula Muhammad b. Dushmanziyar.135
129Ta'rikh-i Tazd, p. 33, repeated in Jami'-i Muftdf, vol. III, pp. 737-8, with more exact topographical details. 130 Ta'rtkh-i Tazd, p. 154. 131 Ibid., pp. 37, 148, 152. 132 Ibid., pp. 151-2; Jdmi' al-khairdt,pp. 14, 149. 133 Ta'rikh-i Tazd, pp. 153-4.
134

135

Ibid., p. 35. Zarich was apparently a district as well as a village, for in ibid., p. 153, we find mentioned " Firfizibdd of Zarich ". Its qandt was still used in the eighth/fourteenth century, see the Jdmi' al-khairdt,pp. 12, 176. Ta'rikh-i guztda, ed. 'Abd al-Husain Nava'i (Tehran 1339/ 1960), p. 499, abridged translation by Browne and Nicholson, Gibb Mem. Ser., vol. XIV/2 (London 1913), p. 118.

DAILAMiS

IN

CENTRAL

IRAN:

THE

KAXKOYIDS

OF JIBXL

AND

YAZD

95

Genealogical table of the Kakiiyids, showing their connections with the Bilyids, the Seljuqs and the Atabegs of Yazd Marzubin Rustam Dushmanziyar
I. Muhammad, t433/1041-42

Fulin
Abi Ja'far

Fulana
Sayyida, t419/1028 = Fakhr ad-Daula,

Bilyid Amir of Ray and Jibdl

2.

Faramurz

3. Garshdsp, Abia ;Iarb

t443/1051-2

Fuldna =
Mas'iid of Ghazna

Majd ad-Daula, Amir of Ray

4. 'Ali, t488/io95 = Arslan Khdtiin bt. Chaghri Beg

5. Garshasp = Fulana bt. Sultan Malik-Shah

Fani-Khusrau

Abti Dulaf

Fulana

Fulina = Rukn ad-Din Sam, first Atabeg of Yazd 'AIa' ad-Daula Ata-Khan

THE ARCHERS OF THE MIDDLE EAST: THE TURCO IRANIAN BACKGROUND By J. D. Latham
& Antoine Boudot-Lamotte's recently published Contribution l'e'tude l'archerie musulmane' carries de an illustration of a mounted hunter taken from a floor fresco in the Umayyad palace of Qasr al-Hayr al-Gharbi in Syria.2 The same illustration-seemingly reversed in error-appears in F. Gabrieli's Muhammad theConquests Islam.3 For specialistsin the history of Oriental bows the most interesting and of feature of the frescois the evident Sistnid design of the composite bow which the huntsman holds at full draw. To conclude from evidence such as this that the influence of archers and bowyers of the eastern caliphate had taken firm root in Umayyad Syria as early as the first half of the eighth century would be a gratuitous assumption. The simple truth is that in the sphere of archery we are as yet in no position to pronounce on the origins and diffusion of Iranian influence in the Muslim Arab world. Until a combination of linguistic competence, patient scholarship, and technical knowledge can be brought to bear on all materials available for a study of the problem we must, for the greater part, speak of the situation as we know it in the later Middle Ages. Whatever the military skills of the Arabian conquerors who laid the foundations of the Muslim empire, there is nothing-bombast and pious fiction apart-to suggest unrivalled pre-eminence in their use of the bow. On the contrary, there is every reason to believe that both as archers and bowyers the successorsof the Arab warrior class were superior to those whom they displaced. Moving westward from the old Iranian world, successivewaves of newcomers to the Arab heartlands brought with them weapons and techniques which undoubtedly contributed in no small measure to their success in the field. Already in the high 'Abbdsid period the courage and martial prowess of a new military 61ite evoked the admiration of contemporarieswho have left impressivereports of their skill as bowmen and horse-archers. Embellished though some accounts may be, the substance of much that is said is consistent with demonstrably acceptable evidence from technical sources.4 Among writers of this period who touched on such matters is the celebrated Arab belletrist and polymath, al-J~ihiz (776-869). In his observationson the military training of the Khurasanianshe recordsthat boys practisedvaulting on to horses'backs and men played polo, after which came " shooting at the sitting quarry, the birjas,6 and the bird on the wing ".6 Of the Turks he writes that " if a thousand of their horsejoin battle and let off a single shower of arrows, they can mow down a thousand [Arab] horse. No army can withstand this kind of assault. The Khdrijitesand the Beduin have no skill worth mentioning in shooting from horseback, but the Turk can shoot at beasts, birds, hoops, men, sitting quarry, dummies and birds on the wing, and do so at full gallop to fore or to rear, to left or to right, upwards or downwards,7loosing ten arrows before the Kharijite can nock one ".8 In flight, he avers, the Turk spelt certain death since he was as accurate in retreat as attack.9 More succinct, but no less enthusiasticis Ibn al-Faqih's comment that the Turks were sharp-shootersof the first order.10
1 (Institut 2 Pl. XIV.

bow.
4

Frangais de Damas, Damascus i968.) This plate, unfortunately, does not show the whole

3 (London 1968), p. 96.

See J. D. Latham and W. F. Paterson, Saracen Archery (London x970). This is an English version of a fourteenth-century Mamlfik treatise (see below p. 99 and n. 3x). Each chapter is followed by a full explanation of its subject matter (with diagrams and illustrations) and interpretations of technical data. See also W. F. Paterson, " The Archers of Islam " in JESHO IX (1966), pp. 69-87. SThis was a ring or hoop mounted on a lance, pole or the like, by means of a wooden truck. It was used in lance exercises (see

9 Ibid., p. 29. Cf. Latham, op. cit., pp. 260 f., 265 if. 10 Kitab al-bulddn,ed. M. J. de Goeje (Leiden 1885), PP. 329 f.

8 Op. cit., p. 28.

Kitdb al-makhzin jdmi' al-funan, Bibl. Nationale MS. Arabe 2826, fos. 104a-b) and as a target for horse-archers' practice (Latham and Paterson, op. cit., pp. 76, 83 f.). Al-MaqrizI describes a similar target as the qabaq, but the latter, strictly speaking, was a gourd, cage or other object affixed to the top of a mast (see Latham, " Notes on Mamlfik horse-archers " in BSOAS XXXII (1969), p. 259). Mandqib at-Turk in Tria Opuscula,ed. G. van Vloten (Leiden 1903), PP. If. x Ibid., p. 28. See Latham, op. cit., pp. 265 f.

97

98

JOURNAL

OF PERSIAN

STUDIES

Since it is not my principal concern here to assemble and demonstrate the essential veracity of testimony relating to the dexterity of Central Asian archers, but rather to consider the Turco-Iranian impact as reflected in Arabic archery literature," let us now move directly to consideration of the earliest treatise of which we have a published Arabic text.12 This dates from about the last quarter of the twelfth century and forms part of a larger work on arms, armour, and weapons of war, viz. the Tabsiratarbdbal-albdb.13Written for Saladin by a certain Mardi b. 'Ali t-Tarsiisi, the Tabsirais a forerunner of the kind ofjihdd and furzsiyyaliterature which became so popular from the fourteenth century onwards. An examination of the section on archery reveals among acknowledged authorities and exponents of the art a preponderance of persons whose nisbasbespeak their role as repositories, transmitters,and continuators of shooting techniques current in the Iranian cultural orbit. Links with towns in Khurdsdnand Transoxiana, whether direct or indirect, are clearly indicated in the following

list:14 Hayawayh al-Balkhi (" of Balkh "),15 Tahir b. Muhammad (?) al-Balkhi,16 Abfi Hashim alBdwardi (" of Abiward "),17 Ishiq al-Bukhdri (" of Bukhar ",,),18 Abfi 1-Hasan al-Kdghidi 1-Harawi (" of Harit "),19 (Abil Ja'far) Muhammad b. al-Hasan al-Harawi,20 Zarr~d al-Harawi,21 Abi 1l-Fath Sa'id b. Khafif (HIafif?) as-Samarqandi(" of Samarqand "),22 Abti 'Abd Alldh as-Sarakhsi,23 Abfi Mtisa 1-IHarrdnis-Sarakhsi (" of Sarakhs "),24 Muhammad b. Itirek as-Sarakhsi,25 Abil 'Abd Allih at-Tirmidhi (" of Tirmidh ,").26 Some of these names together with one or two others (e.g. Muhammad b. Yaisuf ar-Rdzi (" of Rayy ",))26a also occur in thirteenth- or fourteenth-century manuals, but beyond their

bare names nothing in the way of biographical detail appears to have been recorded about their owners. The dates at which they flourished, then, remain obscure, and the most we can say at present is that, as far as those mentioned by Mardi are concerned, the second half of the twelfth century can be taken as a rough terminus quem.To suppose that they lived no earlier than the middle of the eighth ante century A.D.would be a reasonable conjecture since it was only from early 'Abbisid times that large scale recruitment of troops and officials from Khursdin and Transoxiana began. It is interesting to remark that apart from Tihir al-Balkhi, AbUiHdshim al-Bdwardi, and one whose name is excluded from the foregoing list only because his nisbais unknown, Ishdq ar-Raffa',26b none of the authorities cited by Mardi survived the passage of time. To judge from Islamic archery literature in general and the Arabic literature in particular, what seems to have happened is that theoretical teaching crystallized into three systems of which Tihir, Abfi Hashim, and Ishiq emerged as founders. The name most closely associated with this development is that of a certain at-Tabari, author of a work of fundamental significancein the history of archery literature, viz. the Kitdbal-wddih
xx On this subject see principally J. Hein, " Bogenhandwerk und

Tabsira, p. I17. Bogensport bei den Osmanen " in Der Islam XIV (1925), 19 See Tabsira, index s.v. KAgidi. The nisba " Harawi " does not und die Literatur pp. 298 ff.; H. Ritter, " La ParuredesCavaliers occur in that work, but it can be established from A. L. S. M. uber die ritterlichen Kiinste " in ibid., XVIII (1929), pp. 116Lutful-Huq's edition of an important fourteenth-century 54; and G. T. Scanlon, A Muslim Manual of War (Cairo 1961), manual, viz. Nihdvat as-su'l wa-l-umniyaft ta'lim a'mdl alpp. 1-21. furusiyya (SOAS Ph.D. thesis, 226), p. 86. For permission to 12 So far it is the earliest of all treatises known to me. Cf. also use and quote from the Nihdya (for further details see my Boudot-Lamotte, op. cit., p. 34remarks in JSS X (1965), p. 253) I am deeply grateful to both 13 All but the section on archery is dealt with by C. Cahen in the editor and the University of London. " Un trait6 d'armurerie compos6 pour Saladin " in BEO XII 20 Tabsira,pp. 83, 127. Cf. Nihdya, pp. 61, 64, 75; Hein, op. cit., (1947-48 [1948]), pp. 103-63. Boudot-Lamotte's Archerie P. 313, etc. musulmane (above, p. 97 and n. I) contains the text and transla- 21 Tabsira, p. 133. Cf. ibid., p. 123 and Nihdya, p. 69 where the tion of the remainder (hereafter referred to as the Tabsira). corrupt r.ddddshould be emended. 14 Boudot-Lamotte's index is somewhat awkward to use and, in 22 Tabsira,pp. 83, 87, 89, 93, 15. Cf. Hein, op. cit., p. 313 where addition, contains certain errors and omissions. For this " Khafif"/" is given as " Hanif". reason I cite page references. The reader is referred to the 23 .Hafif" Tabsira, p. I21. Arabic text. 24 Ibid., p. 81. Cf. Nihdya, pp. 68 f., 75; Hein, op. cit., pp. 307, 15 Tabsira, pp. 83, III, 113, 135. 16 313, etc. Ibid., p. 133 (and passim as " Tdhir "). 17 Ibid., p. 79 (and passim as Abfi H~shim). Because Abfi 25 Tabsira, pp. 81, 1 I3, 135. Hdshim's nisba is frequently corrupted in archery literature 26 Ibid., pp. 83, 89, 99, 113, 125, I33. (one finds "BBrfidi ", " Nawardi ", and especially 26a Hein, op. cit., pp. 307, 313. " Mdwardi "). I have discussed it in detail in Latham and 26b Not " ar-Raqqi " as in N. A. Faris and R. P. Elmer, Arab Paterson, op. cit., p. 39. In his introduction to the Tabsira Archery(Princeton 1945), passim. This work is a translation of (p. 27) M. Boudot-Lamotte accepts my reasoning (personally an anonymous Arabic treatise written in Morocco about A.D. communicated). 1500.

18

THE

ARCHERS

OF THE

MIDDLE

EAST:

THE

TURCO-IRANIAN

BACKGROUND

99

ft r-ramy(sometimes referred to as Kitdbal-iddah).27For all the importance of his work we know very

little about the man. What is certain, however, is that he is not, as some later Muslim authors would have us believe,28the same person as the great historian and scholar of the same name. The most plausible suggestion so far put forward is that we have to do with one Ahmad b. 'Abd Allah Muhibb ad-Din at-Tabari who flourished in the thirteenth century and whose date of death is given as 694/ 1294-95.29 From his nisbawe can take it that he originated from or was connected with Tabaristan, and from what he himself says of his training and mentors we can legitimately deduce the regional source of his knowledge. He was trained, he tells us, in the technique of Abti Hashim by Muhammad b. 'Isd s-Samarqandi, in that of Tahir al-Balkhi by 'Abd ar-Rahman al-Farawi (" of Farava ", now Kizil Arvat), and in that of Ishdq ar-Raffa' by Abai Sa'id al-Khwarazmi (" of Khwarazm ").30 Without a systematicexamination of all extant material on the subject the full measureof at-Tabari's achievement cannot be adequately assessed,but there is much to suggest that his position in archery was in some respects analogous to that of ash-Shdfi'iin jurisprudence. Such, at any rate, is the impression left not only by his own work, but also by many of the post-Tabari manuals. Of these, that most widely esteemed in aftertimesand, on that and other accounts, the most important is the Kitdbghunyat at-tulldb written about A.D. 1368 by the TaybughI 1-Ashrafi1-Baklamishi ft ma'rifatraminy an-nushshdb Mamlfik l-Yinanin.31Speaking, in the early part of his treatise, of Tahir, Abi Hashim, and Ishlq in the same breath as Bahrdm Gfir (the Sasanid Emperor Vahram V, A.D.420-38), he observes: " These were the men who won renown for their methods of shooting, and from olden times right up to the present day people have adhered to their teachings. At-Tabari it was who transmitted their techniques, though his own system was eclectic (ikhtiydri).'"32 a Later in the Ghunya chapter is devoted to what are termed the Here Abi Hishim, of or systems (madhdhib) the imdms, Great Masters, of archery (a'immat ar-ramy).33 Tahir, and Ishdq are presented as ideal specimens of somatic types-tall and thin, short and fat, and of medium build, respectively-for each of which there is a perfect style of shooting. The fact that other authors give different and discrepant accounts of these archers' tenets need not be considered here. What is particularly significant is that Taybughd expounds his teaching within the frameworkadopted by at-Tabari. Knowing the Islamic predilection for a personal originator of almost any teaching or doctrine, we must be highly sceptical of the authenticity of ascriptionswhich cannot be proved. The emergence of personal schools of archery, we may suspect, representsthe organization of teaching into a coherent system in which names enjoying the sanction of tradition are used either to legitimize current practice or to promulgate new methods. It seems probable that by at-Tabari's time theoretical teaching had crystallized into three main systems. Coming face to face with the obvious dilemma of the individual who fell neatly into none of the three stereotypedclassesof physique for which the systemscatered, he would have solved his problem in a characteristicallyIslamic fashion. Skilfully avoiding innovation which could well discredithim, he could reconcile the claims of individuality with those of" orthodoxy " and tradition by application of the eclectic principle. In so doing he would satisfy an obvious need and set a useful example, thereby earning himself-albeit unintentionally-recognition as an imdm. However that may be, Taybugha's four systems, or schools, based on allegiance to the teachings of a master are strikingly reminiscent of the four orthodox schools of law. Since the systems he represents as being those of AbaiHashim, Tahir, and Isha~q not those described by at-Tabari, we may well are surmisethat the respectableframeworkof" orthodoxy " underpinned by time-honourednames-three
27

The two MSS. of this work which I have actually seen are (i) British Museum MS. Or. 9454, and (ii) Istanbul MS. Veliytiddin Ef. 3175. 28 See Hein, op. cit., p. 299. Cf. Ritter, op. cit., pp. 136, 141. 29 Ibid., p. 300. Or. 9454, fos. 57b f., Veliyuddin Ef., fos. 43b f. In both 30 MSS. the nisba " Farawi " has been corrupted, as, presumably, Wdd.ih., also in Berlin MS. 5538 for Hein (op. cit., p. 300) calls him " al-FdzAri " (i.e. of the tribe Fazara in North Arabia), a reading encountered elsewhere in Arabic treatises on the subject. There can be little doubt that " Fardwi " is correct, since at-Tabari actually specifies that it derives from the name

of a place on the marches of Khurdsdn in territory " contiguous with the land of the Turks ". The readings " al-Qariwi " and " Qardwah " in the Istanbul MS. are only one diacritical point away from accuracy. 31 British Museum MS. Add. 23489. This is the treatise referred to in n. 4 above. Add. 23489 is only one of seventeen known Arabic MSS. which either purport to be Taybughd's Ghunya, or are, with or without acknowledgement, modified or expanded versions of it. See Latham and Paterson, op. cit., pp. xxxv-ix. 32 Add. 23489, fo. 37b. 33 Ibid., fos. io6a-io8a.

100

JOURNAL

OF PERSIAN

STUDIES

of them with an undeniably Iranian ring-is being used as a cloak for the promulgationof what is either his own teaching or current practice. Throughout Arabic technical literatureon our subject authorsnot uncommonly reveal their Iranian heritage in one of two direct ways. The first and more direct is the ascription of such-and-such a or weapon, method, or technique to archers of specific towns and villages such as Mary or Balkh,34 of such as Khurasdn or Transoxiana.35 By the second, styles of shooting and points of technique regions are attributed to the ancient Persiansand their kings. At times the referenceis simply to " the Persians" or " the at or kings " (al-mulak as-Sdsdniyya) " the Chosroes" (al-Akdsira), others it is to indiviSsminid Ardashir b. Babak (ArdashirI, son of Pipak, A.D. 226-41), Bahram Gfir (above), and duals, notably The following example is typical: " There are those Shiptir Dhi 1l-Aktif (Shdipir II, A.D. 3o9-79).36 who maintain that aim should be taken from inside [the bow],37this being an ancient method used by the Persians (al-Furs)right down to the time of the Sisanid kings. [As for the external method] this was first used by Ardashir b. Bdbak."38Although Taybugha, unlike the earlier writers and their .... imitators, is more than sparing with his pre-Islamic Persian kings, he does not entirely disregardthem. Hence, when stressingthe value of binocular aiming39to a horseman or a warrior in armour, he is not loath to add: " It is the ancient form of aim used by the Chosroes."40In the same context, moreover, a revealing comment leaves little doubt as to the standing of the Iranian criterion in disputed issues: " The internal method is that in which the arrowhead (i.e. as seen between bow and string) is used, and we cannot draw conclusionson the subject [as those who take the contrary view have done] from a picture of Bahram Gir because a picture on a wall or in a book is not large enough to show such
detail.""41

Islam!42

It almost goes without saying that ascriptionsto ancient Persian kings cannot be taken at their face value. In the absence of evidence to the contrary the function of their names is essentially that of an imprimatur. Within a slightly different frameworkthe spirit of the device is perhaps best reflected by a manifestly absurd genealogy: in a Maghribi treatise of the late sixteenth century (?) Tahir al-Balkhiclearly a Muslim-is representedas the grandson of ShaptirII who died two and a half centuriesbefore In view of what has already been said Iranian influences may reasonably be expected to reveal themselvesin the technical vocabulary of Arabic archery literature. The existence of such influences is indeed a fact, but detection is not always a simple matter. The calque, for instance, presents problems which are unlikely to be solved in the present state of our knowledge. To illustrate, a regular Arabic term for the limb of a bow is baytal-qawswhich correspondsexactly to the Persian khdne-ye the kamdn, first element in both cases meaning " house " and the second " bow ". Is the Arabic a straight translation from the Persian? If so, from what period does it date ? The straight answer to both questions is that as yet we cannot say. With the loan-word we are on firmer ground, though here again there are difficulties. All else apart, there is the thorny problem of distortion. In the later Middle Ages the Arabic technical literature confronts us with a number of terms, many of them in common use and of pivotal importance for our understandingof the subject matter, which are patent distortionsof Persian or Turkish words (to discern which is sometimes a problem in itself), but for which one finds no entries in classical lexicons, whether Arabic, Persian, or Turkish. The proportion of such words varies from one treatise to another. In the Ghunya, instance, it is comparatively high, while in the for Tabsirait is relatively low. For this phenomenon there are explanations which need not concern
us here. Suffice it to conclude this brief introduction to an ill-explored field with a select list of

8' See, for example, Tabsira, p. 139; Wddiih (Or. 9454), fo. 34a;

3"See, for example, Tabsira, pp. 131, 139; Nihdya, p. 69.

With the " external " method the line of sight did not run between bow and string, and this meant that the left edge of the bow would be seen in line with the target. Nihdya, pp. 63, 69; Ghunya (Add. 23489), fo. 2ob; Arab Archery, pp. 5I, 62, 84, 112. as Nihdya, pp. 63 f. 36 See, for example, JNihdya,pp. 63 f., 65 f., 74; Arab Archery, "9 I.e. aiming with both eyes open. 40 Add. PP. 47, 11423489, fo. 49b. "7I.e. between the bow and the string. In practice this meant 41 Ibid., fos. 5ob-5ia. that the right edge of the bow would be aligned with the target. 42 Arab Archery, I14. p.

THE

ARCHERS

OF THE

MIDDLE

EAST:

THE

TURCO-IRANIAN

BACKGROUND

101

terms from this Turco-Iranian vocabulary which for one reason or another seem to me to merit
attention:42a

n.43 Amaji,adj.; amafjiyya, amdjfyydt, The adjective is used to qualify arrowsand arrowheadsused in pl. the shooting at butts. In the Ghunya noun denotes an arrow with a blunt head used by horse-archers " Asbdsaldri, adj. as n.45 Name of an unidentified type of arrow-guide.46 [P sipahsdldr commander ".] A device for strengthening a weak point in the limb of a bow so n.47 Baghsaq/bughsuq/barsaq/barshaq, that the bow may be braced. It is defined in one source as a " strap, or cord, about I spans (c. I I in.) long. Once tied at the ends, the bow could be inserted into it so that the weak limb might be held firm ".48 It would seem, then, to have been some kind of loop which could be slipped aux over the siya (on which see s.v. dustdr) and tightened. It is recorded in Dozy's Suppliment dictionnairesarabes(2nd edn.) in the sense of" baldric " (I, p. 86o). [P barshak.] shdh,n.49 Name of an unidentified type of arrow-guide. [P proper name.] Bayram and n.50 Daymak/dimak, Boudot-Lamotteprefersdaymak cites a sourceaccording to which " il designerait la partie renflde de l'enferron de la fleche (unbfibat al-/hadid) ".51 Commenting that Mardi uses the
word in quite a different sense, he writes: ". . . il semblerait qu'il s'agisse pour lui d'une partie for shooting at ground targets,44 presumably butts of earth or the like. [P amdj " butt ".]

m6diane de l'arc ".62 This conjecture is correct. The part intended is in fact the " arrow-pass", i.e. the point at the top of the handle against which the arrow rests when nocked and in place for we shooting. In some recensionsof the Ghunya are told: " The point which the arrow passes when

The vocalization and the etymology are problematic, but the word could conceivably be the diminutive of dim [P] in the sense of " face ". In Ar. the belly, or inner surface of the braced bow, handle included, was termed wajh" face " as well as the more usual batn.54Unlike the back (gahr) which was not visible to the shooter unless he reversedhis bow, the arrow-passon the lateral surface of the handle could at most times be seen almost as easily as the " face ", of which it was in fact a limited extension. We cannot exclude the possibility of a connexion with dim " leather ", since in some extant oriental bows a small piece of shagreen has been affixed for the purpose of protecting the arrow-pass against wear from friction, but I have found no evidence that this was done in medieval times. (See also s.v. tadmik.) distdrbeing rare.) ArabArchery: Dustdr/dastdr/distar, (The vocalization is most commonly dustdr, n.55 " What lies over the ibranjaq, So it with the grip is called by the Persiansdustdr."56 also, connecting de mieux,Boudot-Lamotte.57This meaningless definition along with Faris'sidentification with faute dastdr[P] " turban "58 led Elmer to wonder whether some kind of binding was intended.59 In some recensionsof the Ghunya, however, we read: " The section from the base of the siya to the handle is called bayt (' limb '). The Persians call it dustdr."60 Since the siya is the almost rigid, lever-like of the limb,61the dustdr must be the " workingpart of the limb ", i.e. the section taking the extremity bend when the bow is drawn. Taybught's specificationsfor a composite bow confirm this since his
42a

shot is termed kabid al-qaws (' liver, centre, of the bow '), or daymak/dimak,or niyd.t(' heart ').,"53

Unless there are indications to the contrary the MS. references in the notes which follow (Add. 23489 excepted) are to what are in effect modified or expanded versions of the Ghunya. They do not, of course, always bear that title. 4a Add. 23489, fo. 74a; Oxford Bodleian Library MS. Huntington 208 (n.d.), fo. 49a; Istanbul Topkapi Sarayi Miizesi Kittiiphanesi, III. Ahmet, 2608 (A.D. 1403), fo. 30a; British Museum MS. Or. 1358 (n.d.), fo. 5Ib. 44Add. 23489, fo. 74a, et al.
45 Ibid., fo.
46

Basically, an arrow-guide is a simple piece of equipment designed to enable the archer to draw an arrow inside the bow and therefore shoot arrows or darts that would be otherwise too short for his draw. See Arab Archery, pp. 124 ff., 174 f. 47Bibl. Nationale MS. Arabe 616o (A.D. 1397), fo. 51b; Istanbul Aya Sofya MS. 3800, fo. 93a; Add. 23489, fos. 8oa-b. 48 Istanbul Aya Sofya MS. 4193 (A.D. 1400) fo. 66a. 4, Add. 23489 fo. 10o3b. 10

Io03b.

50Ibid. fo. 88a; III Ahmet 2608 fo. 22a; Or. 1358 fo. i6b; Tabsira, pp. 87, 121, 131, 141, 153; Nihdya, pp. 67, 7551 Op. cit., p. x58. 52 Ibid. p. 15953 Or. 1358 fo. 40b; Bibl. Nationale MS. Arabe 2833 (sixteenth century) fo. I5b; Istanbul Aya Sofya MS. 3314 (fifteenth or sixteenth century) fo. 16b. 54 Add. 23489, fo. 78a; Or. 1358, fo. 40b; Tabsira, p. 71; (Or. 9454), fo. 26b; Nihdya, 15 Add. 23489, fos. 22b, 23b; Wdd.ih p. 6i. p. .5ss Arab Archery, 57 Op. cit., p. 159p. "8Arab Archery, 29. 69 Ibid., p. 16360 III Ahmet 2608, fo. 22a; Or. 1358, fo. i6b; Arabe 2833, fo. i5b; Aya Sofya 3314, fo. I6b. 6x See Paterson (above n. 4), JESHO IX, pp. 72-9, figs. I and 2, pls. II and III.

102

JOURNAL

OF PERSIAN

STUDIES

basic components are five: the grip or handle, two dustdrs, and two siyas.62 Dustdr is sometimes loosely applied to the whole limb.63 The etymology is obscure. It is tempting to propose a formation dast [P] +-dl (by analogy with changal, danbdl, etc.) in the sense of " arm " of the bow, i.e. the limb (cf. Ar.yad al-qaws " arm of the bow ", i.e. the upper limb), or, alternatively, anddkh (>anddsh) + + -tdr [P] > andastdr"that which throws, shoots ". The loss ofan- can be paralleled (see s.v. kushtuwdn), but the shift kh > sh is difficult to explain. this word most probably Ibranjak, n.64 Spelt ifranjak in the Wddih65and ibranjaq in Arab Archery,66 derives from a diminutive of abranjan[P] " bangle ", for it denotes a roughly semicircular strip, or sliver, of hard material (bone, ivory, hardwood) running round the belly of the bow's handle and filling the gap left in the early stages of manufacture between the upper and lower sections of the horn with which the belly of the composite was lined.67 In Turkish it was known as the felik.68 It was not, as supposed by Elmer,69 and after him Boudot-Lamotte,70 a shock-absorber, but merely a plug.71 The meaning of the term was extended to denote the " belly of the handle ".72 Jarkh [P charkh" wheel "] for qaws al-jarkh " wheel-bow ", n.73 For Taybughd it was not only a crossbow, but also a horse-archer's crossbow with a stirrup at the front of the stock. As the rider rose in his stirrups he pressed down on the stirrup of his crossbow and drew the string with hooks attached to a shoulder-strap.74 Sincejarkh in this sense is corrected to qaws in almost all manuscripts,75 ar-ri'l we may suspect that in the context it is a misnomer.76 It is my belief that the jarkh was basically the same kind of crossbow, but that it differed by having a small wheel on the stock such as we see illustrated in Payne-Gallwey's Crossbow.77Used in conjunction with a cord which hooked on to the bowman's belt, it provided a simple pulley mechanism which it would have been impracticable to operate on horseback. Jild/jild/chuld, n.78 [P chillah " bowstring "]. A bowstring of new, wound silk of high quality, twisted lengthwise. Jinkilbdz, n.79 A mode of gripping the handle of a bow, particularly favoured in flight shooting.80 The thumb was set on the back of the middle finger and the end of the index finger tucked in behind the back of the handle. Taybugha writes: " The Turks (sc. the Mamlfks) call this jinkilbdz. It is the flight archer's grasp because he can use it to increase his draught and bring in the arrowhead right up to the very base of the thumb, so that it leaves the handle of the bow."s8 With the index finger out of the way there would be no danger of its creeping out in front of the arrowhead which at full draw lay behind the handle. The etymology is almost certainly changdl-ibdz [P] " falcon's talons ". When seen in the flesh, the grasp described strongly resembles the grip of certain species of falcon, and in a society familiar with falconry the comparison would be most apt. A nineteenth-century Ottoman writer drawing on medieval sources, TaybughT included, compares the configuration of the flight archer's fist to the head of a harp (feng),82 but we may well suspect that through some misunderstanding of his authorities he has taken chang " talons " as chang " harp ". The suspicion draws strength from the fact that he finds it necessary to explain that a harp is a " Frankish stringed instrument ". n.83 Bow-case [P kamdnddn.] Kaminddn/kaminddna,
83

Add. 23489, fos. 22b ff. Ibid., fo. 83b. Cf. Cambridge MS. Qq. 240(7), fo. Ioob. 64Add. 23489, fos. 90 f., 92b f., 94a; III. Ahmet 2608, fo. 22a; Or. 1358, fo. 4ob; Arabe 2833, fo. I6a; Huntington 208, fo. 16b. 66 Add. 23489, fos. 30b, 36b. 66
P. 15.
67 66

62

76

Writing two centuries earlier, Mardli clearly distinguishes between jarkh and qaws ar-rikdb (Cahen, op. cit., BEO XII,

Arab Archery, 162. p. 70 Op. cit., p. I63. 71 Paterson, op. cit., p. 76. 72 As in Add. 23489, fos. gob, 92b f., 94a. 73 Ibid., fos. 23b, 35b, 74a ff. 74 Ibid., fo. 74b. 75 See Arabe 616o, fo. 77a; Cambridge MS. Qq. 178(8), fo. 57b; Leiden Cod. Or. 74, fo. 6oa. etc.

69

Ibid., p. 76.

See Paterson, JESHO IX, p. 75 f.

Payne-Gallwey, The Crossbow (London 1958), (reprint), fig. 30. 78Aya Sofya MS. 3800, fo. 31b; Arabe 616o, fo. 20a; III. Ahmet 2608, fo. 24b, etc. 7, Add. 23489, fo. 41b. 80 Flight shooting is a sport in which archers compete for maximum distance. It is an art requiring considerable skill. For full details see J. D. Latham, " The Meaning of Mayddnas-sibdq" in JSS XIII (1968), pp. 242-7. s8 Add. 23489, fo. 41b. 82See Hein in Der Islam XV (1926), pp. 62-383 Add. 23489, fo. 86a

p. 132). 7 Sir Ralph

THE

ARCHERS

OF THE

MIDDLE

EAST:

THE

TURCO-IRANIAN

BACKGROUND

103

Kandr,n.84 Edge of the bow dividing back from belly. [P kandr" side ".] n.85 From the context in which the word occurs there is little doubt that it denotes a Kharkumin, " bracing-board", i.e. a device to facilitate the bracing of strong bows. It clearly derives from a Persian compound noun, of which the second element can hardly be anything but kamdn bow ". " The first is possibly kharin the sense of " bridge " as for a stringed instrument. If so, the device would consist of a base board with a central upright support, or bridge. With the loop of the string in place one limb of the bow would be secured to one end of the base and its grip then set on the support (presumably padded). The archer could now put his weight on the other limb and press downwards until it was sufficiently flexed to allow the free loop to be slipped into the nock. etc. Kushtibdn/kustibdn, As kushtuwdn, q.v. " or n.86 Kushtuwdn, kushtuwdndt, [P angushtvdn, angushtvdnehfinger-stall "]. Archer's thumb-guard pl. of leather, or drawing-ring of bone, horn, metal, or other hard substance. It occurs in the Tabsira in the distorted consonantal form 'bjstydn87 'njs(sh)tbdn, angushtvdneh in Boudot-Lamotte's not as (for I have discussed this word and the object it denotes in JRAS (1968), pp. 65-7text88). n. this word seems to indicate (a) a protective covering, pl. Saysara/sisara,89 saydsir,90 In the Ghunya of leather, to safeguardthe loop of the bowstringagainst chafe at the point where the loop probably comes in contact with the bow's nock (I know of at least one such covering on an extant Persian and string91)92 (b) the lip of a leather thumb-guardprotecting the pad of the thumb.93 In the Nihadya the term certainly denotes some part of the loop.94 The etymology is puzzling. Si-sar [P] " thirty heads " comes most readily to mind, but what meaning can it have? Is it in some abstruse way connected with chillah" bowstring " and the numeral chil " forty "? n.95 An elaborate kind of tubular arrow-guidefunctioning in much the same way Shdh-majrd mijrdt), (or as the barrelof a gun. The device was slotted on opposite sides to allow a bowstring to pass through " channel it and along its length. [P shdh" king " and Ar. ".] majrd/mi'rdt vbl. n.96 The act of strikingthe arrow-pass,or daymak/dimak Tadmik, (q.v.) as a result of faulty shooting technique. " n.97 Bow-case. [P tarkash quiver ".] Tirkdsh, aux Ulki, n.98 Registered by Dozy in his Supple'ment dictionnaires arabes, s.v., the word is defined as " espece d'exercise militaire ". The definition could and should have been a little more precise, for Dozy's authority, Reinaud, clearly appreciated that the word was connected with archery and leaves the we impressionthat it had to do with a target of some kind.99 In some recensionsof the Ghunya learn: " The full target (al-hadafat-tamdm) which is called the ulki in Turkish and is known as the long field (al-mayddn at-tawil) has a length of between 130 and I4o arm-spans, though it may be someyards, since the " long field " with which it is said to have been synonymous is given elsewhere as between I3o and 15o arm-spans.101It is almost certainly to be identified with iilkii [T] "target,
".102

times less . . ."100 It was, then, a target in the order of 284-305 yards, and possibly 284-328

goal ", which survives in the modern language in the sense of " ideal, something to aim at

P. P4 80. 95Add. 23489, fo. Io2b. 86Ibid., fos. I9b, 36b-37a. 96 Ibid., fos. 47b (in corrupt form tadmfd), 87b, 90a. 87 BodleianMS. Huntington264, fos. 63a-b. (Following foliation 97 Ibid., fo. 25b. in occidentalnumerals.) 98 Ibid., fo. Iooa. 98 Tabsira, " De l'art militaire chez les arabes au moyen age " in JA (1848), p. 1I7. 9 89See below notes 92, 93 for references. p. 221. 90 Nihdya, 8o. p. 100 III. Ahmet 2608, fo. 146a; Or. 1358, fo. 213b; Arabe 2833, 91For drawing my attention to this I am indebted to Lt. Cdr. fo. 72a. W. F. Paterson,R.N. 101III. Ahmet 2608, fo. 29a; Or. 1358, fo. 50a; Arabe 2833, 92 Add. fo. 21b; cf. Aya Sofya 3314, fo. 21a; Huntington 208, fo. 50a. 23489, fos. 26a, 8ib; Nihdya, 61, 80. pp. 102 A. V. Moran, Tiirkfe-ingilizce Sozliik (Istanbul 1945), s.v. 98Add. 23489, fos. 36b f.
15 Ibid., fo. 8oa.

84Ibid., fo. 28a.

SOME KURDISH PROVERBS By D. N. MacKenzie

By chance the year 1957 saw the publication of no less than three major collections of Kurdish proverbs. Two are in the related dialects of Sanandaj, centre of Ardalan (Kurdistan), and of Sulaiminlya in 'Irdq. The formercollection occurs as an appendix to the second volume of the valuable Kurdish-Persian-Arabic dictionary Kitdb-i farhang-i Mardakh by Shaikh Muhammad Mardikh Kurdistani (chdpkhdna-yi artish) and contains approximately 900 proverbs and idioms. The second, entitled in Kurdish Pand-i pishindn, by Shakh Muhammad-i KhMl (Ma'arif Press, Baghdad) has approximately 13oo entries. The last collection, of about 350 proverbs, is to be found in a book of folk-tales, songs, etc., Filkl6rdKurmdnji& (Haypethrat, Erevan), in the remoter dialect of the Kurds of the Armenian S.S.R., compiled by the indefatigable IjajiRJindi. We are much indebted to these learned editors, for not only are the compilations of interest for their subject matter but they also show very clearly the value of proverbs as material for the dialectologist. In quoting and comparing a few examples, selected mainly from the firsttwo books, my intention is both to bring them to the attention of a wider public and to support a plea that others, in the fortunate position of being able to do so, should record local and dialect forms of proverbs. Unlike many of the proverbs current in standard Persian, which are of literary origin, these local proverbs are often expressed in words which vividly portray rustic life and experience. They are a mine of vocabulary, both everyday and technical, archaic and modern. Moreover, they provide the only type of text which, while being a natural growth rather than a translation of an alien text, is often directly comparable with its counterpart in related dialects. In the following selection I have included only a few of the many proverbswhich are nearly identical in the two dialects. Instead I have concentrated on those which exhibit significant differencesin form. After the two versions of each proverb (marked S. for Sulaimini and A. for Ardalani) I give a literal translation, as close to the original as possible. Unfortunately, most collectors of proverbsassume that their meanings are obvious or well known to the reader, but this is seldom the case when he is a foreigner. Indeed, the meanings of proverbs are often so elusive or ambiguous as to evade even a native collector. The explanations given below are based on the interpretationsprovided by Shakh Muhammad-i Khal (for all that he occasionally mistakes them as texts for a sermon). Many of the proverbs naturally have more or less close parallels in Persian, Arabic, or other neighbouring languages, but the almost endless task of matching them I must leave to the reader. Only in a few cases do I quote a parallel from the other side of the Iranian world, from Pashto. This is not the place for a grammatical study of the dialects. A description of Ardalani forms the introduction to Shaikh Mardiikh's first volume, but being written in the dialect itself it is not readily available to the non-specialist. The dialect of Sulaiminiya I have described in detail in my Kurdish DialectStudies (London 1961) and a comprehensivedictionary has since appeared (Taufiq Wahby and I C. J. Edmonds, A Kurdish-English [Oxford 1966]). Here I shall only draw attention to some Dictionary of the chief differences revealed by these examples. One is the strong Persian influence on the vocabulary of Ardalani, in contrast to the Turkish and Arabic influence seen in Sulaimani. For example, Ard. has kdr (3), zwdn (I2), iapal (I9), mardim, pat
(54) in place of Sul. if, qisa, pis, xalq, pdra, and also Jha (2), dsydw (3), hayzr (I4), qalay (48), where Sul. has the truly Kurdish words Zsk,ds, tir&, Jwdn. Although the transcription of the examples may not always represent the exact pronunciation, and therefore obliges us to be cautious in drawing conclusions from it, one difference is certainly accurately reproduced. This is in the form of the past participle of all verbs, ending in -g in Ard., -w in Sul., e.g. 105

106

JOURNAL

OF PERSIAN

STUDIES

xwirdig: xwirduw (4), gayltig : gailtuw (8), dig: ddw (20), and similarly in the words for " man ", " pydg : pydw (I3), " alive ", zinig : zinda (50), and bead ", marig : muar (51).

Ard. has verbs with infinitives ending in -ydn,e.g. kuyidn(34), fizydn(61), dJsydn (64), xiliskydn (72), which correspond both to intransitive verbs ending in -dn in Sul., e.g. ku~in, fi~dn, xiliskin, and to passive verbs in -rdn, e.g. dJrdin. Compare the pres. stems in *kiny- : kanri (6o) " be dug up ".1 Causatives are formed with -in- in Ard., -in- in Sul., e.g. : birYn (32), and the verb " to break " wirtin: Jkin- (I I, I3). The verb " is " is -a in both dialects, but only in Ard. does the -s- reappear, Ikinkudiw-a-t-awa(the origin of the -t- obscure), if-a. Cf. also Ard. wa-s-a " is like " to Sul. wd-ya(69-71). Both dialects have preserved the suffix, traceable in Old and Middle Iranian, meaning " also ". Ard. has the archaic form -ic, Sul. -(i)s, e.g. _hiz-i': _hi~z-(41), hdwla-yc:dwla-4 (48). The proverbs below exhibit many other differencesbetween the two dialects, too numerous to list. The full collections probably provide examples enough for a complete grammatical description of the dialects. I trust that this selection will serve to give an idea of the interest of the collections, in many respects, and of the value similar collections in less well-known dialects could have. I.
either before or after another vowel, e.g. has (30), ku'ydg-as-awa (34), a-s (41), against Sul. haya,

a-ba. S. agarkawlt p~s(t) udka-ba ba Idn-ixdwan-ix6-y-awa A. kalpis xds bfi-dyj kol-i xdwan-i-aw ba ba. " If the sheepskincoat had been any good it would have been on the shouldersof its owner." The implication is " practise what you preach ". 2. S. dsk-ina-girdw a-baxf. A. dha-yna-girydg a-waxyj. " He is giving away a gazelle not (yet) caught." Counting his chickens before they are hatched. 3. S. d il-i xJ-ya-ki w iaqanadama din-i xi-y a-SkinW. A. isydw kar-ixwa-ya-kd,iaqiaqa sar-i xwa-yt-iri-t-a eF. " The mill is doing its own work and the tappet breaks its own mouth and teeth (A. brings its own head to pain)." The tappet (iaqana,iaqiaqa) is a stick, fixed to the near-horizontalchute at the bottom of the grain hopper, which is bounced up and down by the revolving millstone, causing the grain to trickle down. This is said of a person who talks a great deal about doing a job while someone else is getting on with it. wd 4. S. aw xurmd-ya-y xwirduw-t-a danik-aka-y la girfin-i min-d. to A. aw xurmd xwirdig-t-aminba qilinika-y tj bdzi-mkirdig-a. " That date which you have eaten, here is its stone in my pocket (A. I have (already) played with its stone)." In other words, the trick you are playing will not catch an old hand like me.
5. S.

A. dw la sar minbigird,Ei gaz, i sad gaz. yak " When the water has passed over my head, what (mattersit) one ell or a hundred ells? " Jindi's version is a little further removed from the Persian: dvku dasiri, i buhust-ak, iarbuhust."... ii what one span, what four spans? " 6. S. dw-j bajfga-yak-i diilot hara-bj bi-iwd. ia ba A. iw-tk jnga-yk-i har a-. bi-hwl.
" (When) a water has gone into a watercourse it must just go (on)." That is, it cannot stop itself: old customs cannot easily be changed. S. aw-i la kisa-di bi ba kawiik dar (d)>. ba A. awa la kisa-di be~ kawiik dar t-g. " That which may be in the bowl will come out by the spoon(ful)." " Truth will out ": a threat to " spill the beans " is implied. Another version is: S. harii la di~a-di bi ba ask" dar (d)&.

iw la sar ka tj paff, ii yak gaz, i sad gaz.

7.

1It

appears to be an idiosyncrasy of Shaikh Mardfikh to write the -v- of such passive present stems with the letter gdf. The putative -y- forms are marked with an asterisk below.

SOME

KURDISH

PROVERBS

107

A. harci la diza-dd bi ba kawlik dar t-j.

" Whatever may be in the pot will come out by the ladle (A. spoon)."
8. S. ba bwdr na-gaistuw-a w darpi dd a-kanj. A. ba bigdr na-gayitig vwdl a-kane.

" He has not (yet) reached the ford and he is already taking off his trousers." To meet trouble half-way. Cf. Pashto: " The river has not come yet and he is tucking up his drawers beforehand " (J. Darmesteter,
wahi. ld sind rdyalayna day aw badaye w.runbay

9.

Chantspopulairesdes Afghans II, p. 223, no. 56). S. ba gul-j bahdrnd-yat. A. ba tdq-a gul-j bahdrnd-yat. bi gul-ak-j bihdr nd-e.

" Spring does not come with a (single) flower." Jindi also:

Io.

S. ba huftir-ydnwut, kuf-it baiw-a. wut-i, bdr-i xj-m la sar pilt-i xi-m-a. man-a. A. wut-ydn, wultir, bird-t bag-a. wut-i, bdr-i min har IEil

A Northern Kurdish version (Roger Lescot, Textes kurdesI [Paris I940], p. 199, no. 70) is: min zi mizgzni ddn-a kari, gj, dahsjak ta-ra ba. got, im-&min kim ba, bdr-& mzda ba.

" They said to the camel, 'You have acquired a son.' He said,' My load is on my own back.' " (A. " They said,' Camel, you have acquired a brother.' He said,' My load is still forty maunds.' "

" They gave the donkey the good news, saying, 'You have acquired a foal.' He said, 'My fodder has become less, my load more.' "
I 1.

S. ba kdrakarbi-1jyxdnim, har~ikdsa w kawlik-a a-y-lkind. A. ba kdrakarb-~~i xdnim, harMc kdsa w kawcik-a a-4kinj. S. ba zimdn-i irin (or, qisa-y xJv) mdr la kun (d)&t-adar. A. ba zwdn-i xwav mdr la kund dar t-&.

" If you say to the maid, ' Madam,' she will break all the crockery there is." " By means of a pleasant tongue (words) a snake will come out of its hole." In N. Kurd. also
(Lescot, op. cit., p. 217, no. 201): bi xabar-d xwav mdr i kul-d xwa dar di-kava.

I2.

I3.

S. barz fuifn mil-i pydw a-kin&j. A. barz pafin mil-i pydg a-Skini.

" Flying high breaks a man's neck." This is said to be merely a counsel of moderation, but it suggests rather Proverbsxvi, 18 on pride. " At the time of the grapes (ripening) the gardener's ear is heavy (i.e. he is hard of hearing; A. does not hear)." He does not want to acknowledge the demands, or needs, of the less fortunate.
S. A. "A our bird bird-ya, bdzad (or, _hisdb) Jyd-ya. birdi-mdnbirdi, kisa-mdn Jydi. brother is a brother, business (accounting) is separate." (A. " Our brotherhood brotherhood, purse-a separation.") Jindi has: bird birdti, bdzdr baydnti. "A brother-brotherhood, business-savagery." The sentiment is the same in Pashto: tar wrori xori ba kawu, hisadb myanja. " We shall practise brotherhood and sisterhood (but with) accounting between." S. Japla ba dast-tk li nd-dirt. A. ba tanyd das-i lap *nd-kutyit. " A clap is not hit with (only) one hand."

14.

S. bdxawdnla waxt-i tiri-dd gii-y girdn-a. A. bdxawdnla waxt-i hayar-dgiika-y nd-Zinaw&.

I5.

16.

108

JOURNAL

OF PERSIAN

STUDIES

S. also: dast-jkba tanyd,taqa-ynd-yat. " A single hand, its clap does not come." It takes two to make a quarrel, or a friendship. The N. Kurd. version (Jindi and Lescot) is similar: dast-j (bi) t'ani, dangzi nd-i. la 17. S. day-i dahk1 diirxoi-a. A. day-i dawl la darxwal-a. " The sound of a drum is pleasant from afar." la A. also: sadd-ydawl-a-sifnd dfir-aw xwal-a. " The sound of drum and shawm. .. ." 18. S. di-, lagal to-m-a; diwdr,go-t li b&. A. dir, la tak ti-m-a; diwdi, bi-zinawa. " Tree, my (business) is with you; wall, pay thou heed (A. listen)." There are other versions:
g'-t " Mother-in-law (My daughter), my (business) is with you; daughter-in-law, pay thou heed." Pashto has a very similar form: lure,td ta wdyam; ywagpre wanisa. ngore, I am telling you; daughter-in-law, pay heed to it." " Daughter, ba I9. S. daryd dam-isagpis nd-bi. ba A. daryd dam-isag capatnd-wl. " The sea is not defiled by the mouth of (one) dog." Similarly, in both dialects: jogala darydit nd-kd(t). " A little stream does not make the sea turbid." A lone detractor of a great man is disregarded.
20.

S.

xasa (or, kil-im), lagat ti-m-a; bfki, tj

li bj.

S. ddw-i-ala hawtdw, qul-1tar na-baiw-a. A. ddg-i-ala haft dw, qul-i tar na-wg-a. " He has forded (lit. struck) seven waters, his ankle has not become wet." taf S. also: la hazardw-i ddw-a,qul-a-pi-y na-biw-a.
" He has forded a thousand waters. ...." He has learnt much from experience.

2 I.

S. dit bardnia, la gut ndsik-tir-a. A. dit kuciknia, la gut ndzik-tir-a. " The heart is not a stone, it is more tender than a flower." S. dit bj dit figd-yhaya. A. dit ba dit fega-yhas. " From heart to heart there is a way." Love will find a way. S. dit-i kas ma-jlna; ka Jdn-it, bi-andia ma-ba. A. dit-i kas fil ma-ka;fil-it kird,bi-aneima-wa. " Do not hurt the heart of anybody; if you have hurt (one), do not be heedless " of the probable consequences. KhMlcomments: bj. ka Jir-itwaldnd,a-bj qatlyn-itp& " If you have brandished the sword, you must have a shield about you."
kaybdni bi-y-Skindtaqa-y nd-yat. S. diza-yak A. kadi-kkaywdni bi-y-3kin& taqa-y nd-y&. "A bowl which the lady of the house breaks makes no noise." Similarly: S. dast-&k _hkim bi-y-bifi xn-ii nia. A. das- 4dkim bi-y-wr~xin-i nia. " A hand which the governor cuts off has no blood." S. di Jiti ba dast-tk hal nd-gir&. A. ba das-g di kdtak hat *nd-giry&. "Two water-melons (A. melons) cannot be taken up by one hand."

22.

23.

24.

25.

Do one thing at a

time.

SOME

KURDISH

PROVERBS

109

26. S. g6za ham jdar-&k sdyi nd-gaRft-awa kdni. ba la A. gJza hami~a kdniba sdq nd-Tt-aw. la

" A water-pot does not return sound from the spring every time (A. always)." Tout passe, tout casse.

S. gurg-aw la pist-i maf-4-ya. A. gurg-ala p6s-i me-s-d. " He is a wolf in a sheep's skin." Also: S. la pest-i maf-dgurgia-kdt. A. la pis-i m-ad gurgia-kd. " In a sheep's skin he is acting as a wolf." 28. S. hamamal-jkhanjfir-xjr na-a-md. hanjirba ddr-awa bai-dya A. agargi mal-i mjwa-xwar mjwaba ddr-aw na-a-ma. b7a-jd
27.

like a justification for the inequality of society.

29.

" If every bird were a fig-eater (A. fruit-eater) there would be no figs (fruit) left on the trees." Khal says that this is a tradesman's excuse for not revealing the tricks of his trade, but it also reads

S. harkas-a adw-a-fwdn-i tida-yxalq bi la birs-da-mird. A. har-kaEdw-a-nwaf-i mardim la birsi-da-miri. bi tJSa-y " Whoever expects (to live on) the provisions of (other) people will die of hunger." God helps
them that help themselves.

30. S. harsar-jksawdd-yak-i haya. A. harsar-i sawdd-yk-i has.

" Every head has its preoccupation."

31. S. hatdmdlwastd-bimizgawt _hardm-a. A. td mdtwisd-wi mizgitharam-a.

" So long as the house be standing the mosque is forbidden." Do not give alms until your family is satisfied. Charity begins at home.

32.

S. hilka ba bin-hayil a-bir~ini. A. hilka ba binagila-wirzin&. " He broils eggs under his armpit." He is young and hot-blooded. 33. S. hilka-yimfJlajauiik-i sibaynibad-tir-a. A. hilka-yimfj lajiijala-ysizi xds-tir-a.
" Today's egg is better than tomorrow's chick." A bird in the hand is worth two in the bush.

34 S. jay kuw-a-t-awa, 6ilminba dndar-aa-garf. S. A. jag kuyadg-as-aw, &ilmin-a Jwa-gafl. son " The battle has died down, the snotty one is going about in search of a stick." This is said of any belated assistance. hi 35. S. ka hi x6-t na-ba,kayan-aka x6-t ba.
A. ka hin-i xwa-t nia, kdaddn hin-i xwa-t-a. " The straw was (A. is) not your own, the straw-store was (is)." This is a reproach to somebody who over-indulges at another's table.

36.

S. kar la k6 kawtuw-a w kunda la k6 diadw-a. A. kar la k6 kaftig-a, kunna la kii difydg-a. " Where has the donkey fallen (and) where was the water-skin torn? " The meaning is, " What has what you are saying got to do with what I am talking about? " S. kiw ba kiw na-gdt, adam ba ddam a-gdt. A. kdf ba kef nd-gay baldm binyddamba binyddam a-gay. " (S. If) mountain does not reach mountain, (A. but) man reaches man."

37.

110 38.

JOURNAL

OF PERSIAN

STUDIES

39.

40.

41.

S. ki6-jk ddyk-i madh-i bi-kat, a-bi xdlt bi-y-xwdze. A. kanilk-j ddyk wasp-i bi-ka, a-se xdci bi-y-xwize. " (If) a girl's mother (needs to) praise her, her mother's brother will have to ask for her (in marriage)." Plainly nobody else wants her. Good wine needs no bush. S. kiira, 'i-tawi? di~~adw-i say. A. kiir, wut-ydn,ca-t garak-a? wut-i, di dw-i sdq. " Blind man, (A. they said,) what do you want ? (He said,) Two sound eyes." Another version is: S. kiir td a-miri ba tamd-ycdw-a. A. kiir td aw fj'-a a-miri tamcddr-idii dw-i sdq-a. " The blind man, until (A. the day) he dies, yearns for (A. two sound) eyes." Pashto has a version closer to the Persian: rund la xwddya ca ywdri? dwe starge. " What does the blind man seek of God ? Two eyes." S. ku'ik td na-jiqtj(A. na-jiwj) sayin-a. " A stone is heavy so long as it does not move." S. also, as in Persian: bard la jl-y xo-y sayin-a. " A stone is heavy in its own place." S. kuf-i kuf-i hziz-if mdl-i bi -a? md-i bi EW-a? dzad A. kuf-i dzd mdl-i bi ca-s? ku.-i malt- bi ea-s? " What does a brave lad need wealth for ? What does a cowardly youth need wealth for, either ? " _hz-i{ The first can gain what he wants and the other will lose whatever he has. S. la baredz m-yak bi-kirit-awa walifat-a. A. la xirs mfi-ykbi-kanit-aw, walifat-a. " If (only) one hair be taken from the boar (A. you pluck from the bear), it is a duty." The least harm you can do to an enemy is better than nothing. The pronunciation of Arabic t (as here in watifat) as -1- is an affectation of mullahs, who have been known to inject a fine emphatic Arabic t into such plain Kurdish words as mindt " child ", balEm" but ", a-hj " says ", etc. In the proverb it perhaps lends a mock air of sanctity to the " duty ". S. la csmdn-dastira-yak-i nia. _hawt A. la _haft dsmdnhasdra-yk-inia. " In (all) seven heavens he has no star." The star of his fortune has set. Pashto goes further with: pa zmaka kxe syorayna laral aw pa dsmdnkke storay " not to have a shadow on earth or a star in heaven." S. lagal gurg sai a-kdt, lagal maf sin a-ket. A. la tak gurg- gilt a-xwd w la tak paz-d ~iwan a-kd. " He feasts (A. eats meat) with the wolf and laments with the sheep." A version from Erevan the newspaper Ryd t'aza) is: (from duxa (=di-xwa) l xway-fae ine di-ka. gur-r' eats with the wolf but laments with the owner (of the sheep)." " He S. manjat (or, qaezen)i mard-in ba Eil (or, hawt) sel (d)&t-a jfI. A. qezen-i merd-galba Cil set t-e~t-a kul. " The cauldron of (true) men comes to the boil in forty (or, seven) years." They are slow to anger. la S. ma-raengez gurns-ifal i sipi a-sitame~t-awa. la A. maere-ngez gurns-i fal i bazig a-silame~t-aw. " The snake-bitten shies away from a black and white (A. mottled) rope." S. marg lagal ewal-in jazn-a. A. marg la tak hewmelt-ejain-a. " Death together with comrades (A. a house-mate) is a celebration." Troubles shared are troubles halved.

42.

43.

44.

45.

46.

47.

SOME

KURDISH

PROVERBS

111

48. S. mayman zjrjwdn ba, dwla-1-idardd. A. mayman qayal ba, hdwla-yt-i hdwurd. dar fira " The monkey was very handsome; it produced pocks too." To make the worst of a bad job. ka 49. S. mtwa-yak gai agarna-y-kayt-awa a-gani. A. mjwala takgai, na-y-kanit-aw a-kafr. dd " When a fruit has ripened, if you do not pick it, it will spoil (A. fall off)." Nubile girls should be found a husband as soon as possible. 50. S. mirdabapalaqdza (or,Swan) zindand-bit-awa. A. mirdigbapalaqdaZ zinig-awnd-wit-aw; A. am mirdig-a b-am3iwan-anad-i. " The dead will not come alive again through thrashing about with hands and feet (i.e. kicking up a fuss) " or " with lamentation." A. also, " This dead body will not live (again) with this lamentation." There is no use crying over split milk. hdwmdl 5'1. S. mara-y magarba law bi-kirit-amil. A. marig-ihdwsd-mdl saw bi-y-kayt-a mil. magar " The beads of a house-mate (A. neighbour'shouse) can only be put (A. you can only put) round the neck at night." There is no point in stealing what you cannot use. na 52. S. na mdIl-hayafahmn bi-y-bdt, din-i hayalaytdnbi-y-bdt. A. na mdl-jk-im fahmdn has has bi-y-wd,na din-&k-im saytdnbi-y-wd. " He has (A. I have) neither any property, for Rahman to take it, nor any religion, for Satan to take it." The epitome of misfortune. 53. S. niza la pi-d a-bi ~j-y bi-kayt-awa, bi-y-dizit. infj A. nayzala bar-da-sejfga-y bi-kayt-aw, waxt-abi-y-dizi. aw " First you must make a place for the lance, then steal it." S. also: fimjf bi-kar-awa, bi-y-diza. injd " A lance-(first) make a place, then steal it." 54. S. pdra-yxalq kisa a-difini. A. pal-i mardim a-difc. kisa " (Other) people's money tears the purse." The sooner debts are repaid, the better. 55. S. pi ba qadbafa-yxi-t id kia. A. ba andza-ybaf-i xwa-tpd dd kjia. " Stretch out your legs to the extent of your rug." In Pashto it is " your quilt ": domra ugdawa pve de ce comra brastan wi. 56. S. piSiladam-ina-a-gaift-a dag, a-y-wut,sdr-a. A. kitik dam-ina-a-gaiba dag, a-y-wut,sit-a. " (When) the cat's mouth did not reach the fat sheep's tail it would say, ' It is salty.' " The version known from Aesop's fable is also found in S.: dam-ina-a-gailt-atiri, a-y-wut,tirl-a. f&wi " (When) the fox's mouth did not reach the grapes it would say, 'They are sour.' "
57. . ba darmdna-Jyd kun-dw-kun piilta ga-y a-y-drd-awa. " The cat's dung was suitable as medicine (so) it would hide it in every (possible) hole." A. kitik wut-ydnp&,ga-t bo darmdna-d. kird-i-a xdk-aw. " They said to the cat, 'Your dung is suitable zir medicine.' It put it under the earth." Like a as dog in the manger. S. pydw-i ndbat xizm-i dar a' qarz-i kn baydd a-kdt. A. pydg waxt-& mdtit b&qawm-i dar a qart-i kina a-kdt-aydd. " A penniless man (A. When a man becomes distressed, he) recalls to mind distant relatives and old debts."

58.

112 59.

JOURNAL

OF PERSIAN

STUDIES

5. few xj-y ba kun-awa na-a-~l,hazg-ik-i ba kilk-awaba. A. fiwi na-a-vau kund-dd, ba hazg-jk-ibast-aba xwa-y-aw. " The fox could not get into the hole itself; (what's more) it had a piece of dry brushwood (stuck)
in its tail (A. it had tied ...

60o.

61.

62.

63.

64.

ma'rni-kdr-bibi-cuyd Z~ qul-dxwa, _hazak xwa-vdgire dd. A Pashto version strives for the ridiculous: pa magak ydr naIu nanawatalay, ye pa lakay poreyuta ka9r. caj " The mouse could not enter the cave, it tied a winnowing sieve to its tail." S. fiZsa-y dd-bastrdw kiz-a-bdhat nd-kan(d)rij. ba A. fisa-y bastig-aba siis-a-bd l *nd-kinyj. hat " A root which has taken (well) is not dug up by a gentle breeze." a-kdtba xgr-i bdwk-i. S. fJn-i raiw A. fin-i rfiydg a-kdt-axayr-ibdwk-i. " He makes the spilt oil into a charitable gift (in the name) of his father." nd-biba swdr. S. swdrtd na-gil& A. swdrtd *na-gilyjnd-wit-aswdr. " A horseman does not become a horseman until he falls." .Jt-it? a-li, bo-ma-lwl. S.ta, S. bo "' Madman, why are you mad ? ' He says, 'It suits me.' " A. wut-ydn 3t, bj Ia iti a-kay? wut-i, bo-m*a-.ilyj. ba " They said to the madman, ' Why do you act mad ? ' He said, 'It suits me.' " According to Khal, this is said of someone who comes off well in any demand he makes. S. Ir ka dl}rdnd-6it-awa gwdn. A. sir la tak d6s>y nd-Mit-aw gwdn. " Once the milk has been milked it will not go back into the udder." Similarly: S. tir la kawdndar u nd-gafit-awa -aw). (A. " The arrow (which) has left the bow will not return."
S. M1 ba taff na-bi nd-camit-awa.

to itself)." Jindi has a snake:

65.

" A withy, (if) it be not while (it is still) moist, will not bend." A. tih ba taff na-lamit-aw nd-camit-aw. itir " (If) a withy does not bend while (still) moist, it will not bend any more." You cannot teach an old dog new tricks. ' 66. S. td a-lit ' barsila tirJpi a-gdt. ' A. td a-yz'J hdaa hayiir a-gayi. pj " By the time he says ' unripe grape ' the grapes ripen." Said of someone who is always behind the times. 67. S. tdtu soir-i dinyd-y cajtuw-a. zor A. tdt sol-i dinyd-y faltig-a. fira " He has tasted much of the bitter and the salty of the world." 68. S. and A. tir dgd-yla birsinia.
" The sated is not aware of the hungry." In the North, Jindi: zik-7 tir hd 4 &birli nina. Lescot (op. cit., p. 209, no. I43): ti-rhdy zik-&birli tu na-ya. zik-i " The full belly is not aware of the hungry one." S. wak kar-i ndwjfga wd-ya, la har da li a-xwadt. A. wasa kar-i ndwfjga, la har-tik lI a-xwd. " He is like a donkey in the middle of a stream. He eats from both sides." He is getting the best of both worlds.

69.

SOME

KURDISH

PROVERBS

113

70.

71.

72.

73.

74.

75.

S. wakpaltamdl-ihamdm, sdt-j ba bar har yak-jk-awa-ya. har A. wasafota-y himdm, hd ba barkas-ik-aw. fro-j " He is like the towel of a Turkish bath; every hour (A. day) it is on a (different) person." S. wak sag-ipi-sitdw wd-ya. A. wasasag-ipd-sgzydg. " He is like a dog whose foot has been burnt," he is so restless. S. xiliskdntila-ypalakar-a. A. xiliskydn jira-ypalakdr-a. " To slip is the retribution (A. reward) of the hastener." More haste, less speed. S. xjratdw-iqardygiswdnadawdm-i nia. A. xwaratdw-i bdndawdm-i liv nia. " Sunlight on the edge of the eaves (A. lip of the roof) has no duration." The setting sun seems to be falling off the edge of the roof. This is said to someone who comes to help when the task is almost done. la S. Xwd yazab bi-gir& m rala bdl-i lj a-fwjnT. A. Xwd yazaw a-giri la mir&ca pi a-dd. bdl-i " If (A. when) God becomes angry with the ant He makes wings grow on it (A. gives it wings)," so that the birds can catch it. The Northern version is: kj ajal-j muirie 6angaipar-i w& di-bin. hdt lj " When the ant's hour of death is nigh it gets claws and wings." S. yak-&k dwdi-ydn na-a-dd kaciahwdl-imdl-ikixd-ya-pirsi. i-y A. yak-j fga-ydn na-a-dd ndwazwi, a-y-wut,mdl-i qJxd kdm-a? " They did not allow someone to come into the village, yet he asked about (A. he said, " Which is) the house of the headman." Cf. Lescot (op. cit., p. 193, no. 35): min jih- yaki ii gundnin-bif, di-g6,jfih(i)k-& bi-binmdl-dmdlxwj. " There was no place for someone in a village (but) he was saying, 'Take my bedding to the headman's house.' "
S. zik-jk-I tir la hazdr zik-i birsi ?dk-tir-a.

76.

A. zik- tir la hazdrzik-i birsixds-tir-a. " A full belly is better than a thousand empty bellies." dast 77. . Szuind bi-dayt-a ndlifa a-kdtba sar-a-zil-aka-y-d. A. sifnd bi-dayt-a ndaiw pif ba sar-a-gawra-ka-y-d das a-kd. " If you give a shawm into a novice's hand he will blow into the big end of it."

THE ARABIC ELEMENT IN PERSIAN GRAMMAR: A PRELIMINARY REPORT* By MohammadAli Jazayery

I.

In the seventh century of the Christian era, Iran was joined to the Muslim empire, of which it remained a part for some six centuries, though tight and complete Arab control lasted only throughout the Umayyad period (41-132 A.H. = 661-750 A.D.), during which Iran was ruled by Arab governors, accompanied by large numbers of Arabs who settled in Iran. With the accession to power of the 'Abbdsids (I32-656 A.H. = 750-1258 A.D.), Iranians found their way into the caliphal administration, on which, as governors, secretaries, ministers, etc., they exerted great and far-reaching influence. One of the Iranians in the 'Abbisid service founded a semi-independent dynasty in the opening years of the third century A.H. (about 820 A.D.), followed by other dynasties, who ruled smaller or larger portions of Iran. These dynasties maintained varying degrees of independence from the caliphs, though they continued to be part of the caliphate, often in name only, until the Baghdad caliphate was ended by Hfiligfi in 1258 A.D.
2.1

The linguafranca throughout the Muslim world was Arabic, used in religion, science and scholarship. It was also used in communication among the provincial governments, and between them and the central caliphal administration. In every way, the language spread with Islam, and a knowledge of Arabic came to become necessary for anyone interested in political and social affairs, or in scholarship, especially since the number of Arabs-soldiers and other temporary residents as well as immigrants -grew in both Iran and elsewhere in the caliphate.'
2.2

In Iran, Arabic was the language of administration at the beginning. Soon, however, parts at least of the governmental records, such as those of the taxes, came to be kept in Persian. About 75 A.H., at the suggestion of an Iranian, the records were converted into Arabic.2 Towards the end of the fourth century A.H., an Iranian minister of the Ghaznavid Mahmud made Persian the official language of administration; his next minister brought Arabic back into use,3 though this was temporary, and Persian was restored later. 2.3 Throughout the period of full Arab domination, which lasted until the middle of the third century A.H., the spoken language in Iran was Persian, in its various dialects. Even religious rituals, including the daily prayers, were at first conducted in Persian, at least in some regions.4
* This is a slightly revised version of a paper read before the Islamic Section of the XXVIIth International Congress of Orientalists, held from August the I3th to the 2oth 1967, in Ann Arbor, Michigan. The text published here is for the most part that which was presented orally, and therefore suffers from the restrictions caused by the twenty-minute time limit imposed on all sectional papers. The revisions consist of a reworking of 3-3.1-3.3-3, and the addition of 3.3.4 and 7.1-7.5. dar 1 Z. Safd, Tdrix-e adabiyydt Irdn (Tehran 1332/1953), vol. I, pp. 134-52 3

Safa, vol. I, pp. I17, 159-60. 'A. Eqbal, Tdrix-e 'omumi va Irdn bardye sdl-e cahdrom-e dabirestdnhd (Tehran 1318/1939), pp. 267-8.

4 Safa, vol. I, p. 125; Eqb~l, p. 107.

115

116

JOURNAL

OF PERSIAN

STUDIES

2.4 The use of Persian as a literary medium began soon after the appearance of the semi-independent Tdhirid dynasty (205-259 A.H. = 820-872 A.D.), and especially after the Samanids (261-389 A.H. = 874-999 A.D.) came to power.5 In the fourth century A.H. (tenth century A.D.) some prose and poetry was produced in Persian, though Arabic remained the principal language of culture. In the development of prose, translation, especially from Arabic, played an important part. The situation was more or less the same in the fifth (eleventh) century, except that what might be considered literature proper began to develop.6 In the fifth and sixth (eleventh and twelfth) centuries, Persian literature, both prose and poetry, grew in maturity as well as in quantity; and more literary, historical and religious works appeared than ever before.' The sixth century was particularly rich in scientific works.s Literary and scholarly works continued to be produced during the Mongol (654-750 A.H. = 1256-1349 A.D.) and Timfirid (771-906 A.H. = 1369-1500 A.D.) periods.9 The chief feature of the Safavid period (907-1148 A.H. = 1502-1736 A.D.) was the appearance in Persian of large numbers of religious works, which had heretofore been written in Arabic.10 Today, Iranian writers hardly ever write in Arabic, except for certain types of religious writings, but the language of religious ritual among Iranian Muslims remains Arabic. 3.1 The Persian language that emerged from these cultural developments was influenced by them and also by the Arabic language. The influence of Arabic was at first not very great. However, upon the accession of the 'Abbdsids to the caliphate, Iranians became active participants in Islamic politics as well as Islamic learning and scholarship, attaining mastery in such clearly Arabic-Islamic subjects as commentaries on the Qur'an, the hadith, and Arabic lexicography and grammar. They translated and wrote in Arabic on these subjects, and even composed poetry in it." This situation led to the borrowing of increasingly large numbers of Arabic words into Persian. At the beginning, the borrowed words were mostly connected with new, or greatly changed, cultural concepts, chiefly those in the fields of religion and government administration. To these were added words which poets borrowed to meet the requirements of rhyme or metre. There were also a few Arabic synonyms for already-existing Persian words, some of which eventually disappeared or became archaic, obsolete, poetic, or in other ways specialized.12 3.2 Despite all this, at the beginning of the fifth (eleventh) century, Persian was still relatively free from an Arabic admixture. From the fifth century onwards, however, the proportion of Arabic words in Persian increased. There were a number of reasons for this. Most important of all was, of course, the growth of the influence of Islam, and, with it, of the Arabic language. In the fifth and sixth centuries there was a great increase in the number of Islamic colleges, where Arabic was the medium of instruction in all fields, and where, furthermore, Arabic language and literature constituted a major part of the curriculum. Another prominent factor was that Arabic itself had by now become quite enriched, and thus capable of influencing other languages of the Muslims. This enrichment was due to the development of a considerable body of scientific terminology in Arabic, resulting from compilation of scientific works, in most cases translations from Greek, Syriac and Pahlavi, produced in the period from the second to the fourth centuries.13 In the fifth century, thus, " the gate of the language was laid open to foreign, particularly Arabic, words ".14 But it was in the middle of the sixth (twelfth) century
5 6

Saf, vol. I, pp. 143-4, 179, 304, 315-16. G. Lazard, La languedesplus anciensmonuments la prosepersane de (Paris 1963), pp. 34-5(Tehran 7R. Safaq, Tdrix-e adabiyydt-eIrdn bardyedabirestdnhd n.d.), pp. I05-o6. 8 Safa, vol. II, pp. 325-6. " gafaq, pp. 328-9.

fifty works on religious matters in Persian. Irdn, 2nd edn. (Tehran 1340/1961), Homn'i, Tdrix-eadabiyydt-e p. 46I. 12 M. T. Bahir, Sabksendsi (Tehran I337/1958), vol. I, pp. 259-66. 13 Safd, vol. II, pp. 231, 32714 Bahdr, vol. I, p. 264.
n'

o10afaq, p. 340. As Safaq points out, Majlesi alone wrote about

THE

ARABIC

ELEMENT

IN

PERSIAN

GRAMMAR

117

that there began the great influx of Arabic words, as well as Arabic phrases, sentences and grammatical features-a situation aided by the practice of quoting Arabic proverbs, poems, traditions, Qur'anic passages, etc., in Persian writings.15 During the Mongol and Timiirid periods the barriers between Persian and Arabic were all but removed, at least in the vocabulary.16 3.3.-1 a A full and detailed investigation of the Arabic element in the Persian vocabulary-including based precise statistical study-is yet to be undertaken. Statements on the subject are nearly always on general impressions or, at best, on imprecise and unclearly formulated statistics. There are at least three partial statistical studies, however, all using the same method, which do not suffer from these limitations and will be summarized here. One of these studies covers eight prose texts (four histories and four commentaries on the Qur'dn) written in the tenth to twelfth centuries of the Christian era. Another paper is about two classical lyric poets of the thirteenth and fourteenth centuries, Sa'di and H~ifez respectively. The third study concerns a twentieth century novelist, Buzurg 'Alavi.'7 In all of these papers, a distinction is made between the number of individual lexical items in a text and the number of the occurrences of these items. As an example, one of the samples contained 2376 occurrences but only 450 lexical items, since several items occurred more than once. 3.3.2 The statistics given in the studies referred to indicate that the ratio of the Arabic lexical items was about 25 or 30 per cent in the tenth century texts, about 50 per cent in the twelfth century texts, about 40 per cent in the two classical poets, and about 46 per cent in the twentieth century novelist. The total occurrences of the Arabic words were io per cent in the tenth century texts, 20 or 25 per cent in the twelfth century texts, about 20 per cent in the two poets, and about 20 per cent in the twentieth century novelist.s8 Interpreting these and other figures, Lazard concludes that " the Arabic vocabulary thus became richer and more varied " between the tenth and twelfth centuries,19 and Skalmowski concludes that the ratio of the Arabic lexical items in Persian has remained considerably constant since the classical period, that the high degree of" Arabicization " observed in some periods in Persian literature has not had a lasting effect on the language, and that " the penetration of Arabic seems to have taken place in a single thrust, during the epoch of the formation of the Persian language ".20 3.3.3 The proportion of Arabic words in contemporary Persian is usually estimated at " half ", " about half ", " over half", or in some such phrase, the estimate presumably referring to occurrences rather than to lexical items (though this is not usually specified). Actually, the ratio of Arabic words varies in accordance with a number of variables. For example, it is higher in prose than in poetry.21 Within prose, it is larger in certain technical fields-such as law, philosophy and religion-than in others like
15Safi, vol. II, pp. 327-8, 888-9. 19 Lazard, op. cit., p. 63. 20 Lazard, op. cit., p. 55. Another set of statistics, given by M. T. Bahir, will be summarized here, mostly because it attempts to cover the entire span of Modern Persian literature, though it must be noted that his statistics and the interpretations based on them are neither precise nor clear. According to Bahar, the Arabic words in the Persian prose of the fourth century A.H. (tenth A.D.) constituted no more than 3 to 5 per cent (somewhat higher in poetry), but increased to about io per cent in the first part of the fifth (eleventh), to more than 50 per cent in the middle of the fifth (third quarter of the eleventh), and to as much as 8o per cent in the sixth to eighth (twelfth to fourteenth) centuries. (Bahar, the work cited in footnote 12, vol. I, pp. 266, 275; vol. II, p. 96.) 21 Homd'i, pp. 482-4.

xeBahdr, vol. I, p. 288; gafaq, p. 247.

S17The three studies, in the order of their publication, are:

R. Koppe, " Statistik und Semantik der arabischen Lehnw6rter in der Sprache 'Alawi's ", WissenschaftlicheZeitschrift der Humboldt-Universitat Berlin, ges.-undsprachwiss.Reihe, Jg. IX zu (1959-60), pp. 585-619; W. Skalmowski, " Ein Beitrag zur Statistik der arabischen Lehnw6rter im Neupersischen ", Folia Orientalia III (1961), pp. 171-5; and Gilbert Lazard, " Les emprunts arabes dans la prose persane du Xe au XIIe siecle: apergu statistique ", Revue de l'Lcole Nationale des LanguesOrientales (1965), PP- 53-67. My information on the II first two studies is based on Lazard's article, in which they are summarized.

18 Lazard, op. cit., pp. 61, 5411

118

JOURNAL

OF PERSIAN

STUDIES

the natural and mathematical sciences (where French loanwords are in the majority). Again, there is a higher proportion of Arabic loanwords in the more formal styles than in the less formal ones, and, generally speaking, in writing than in speech. And allowance must be made for individual differences. A cursory examination of a very small sample of Contemporary Persian prose showed that the occurrences of the Arabic words ranged from 20 per cent to 75 per cent of the total, with an overall average of slightly below 45 per cent.22

3.3.4
As may be gathered from the preceding summary, the Arabic element constitutes a sizable portion of the vocabulary of Contemporary Persian. There are literally thousands of Arabic loanwords. Furthermore, the Arabic borrowings represent a number of varied phenomena. The largest group is made up of simple lexical items such as the following:23 vaqt " time "; g'a-z" food "; daqiq " precise "; ma'mul " common "; san'at " industry"; dowlat " government "; mokdleme sobh " morning"; conversation "; fekr " thought "; tahavvol " evolution "; lebds " clothes "; bahs " discussion "; " jadid " new "; mohtaram" respectable "; 'aziz " dear "; tekrdr " repetition "; Joru' " beginning "; jahl " ignorance "; jdhel " ignorant "; and countless others. Another group is composed of compound substantives (nouns and adjectives) containing the Arabic article -al-, such as: jadid-ol-vorud" newly arrived "; haqq-oz-zahme " fee "; sahih-ol-'amal " honest "; fdsed-ol-axldq " of loose morals "; raqiq-ol" qalb " soft-hearted "; taht-ol-hemdye protectorate ". Another major category consists of phrases of various types. Adverbial compounds may be composed entirely of bound morphemes, as in 'aldhdag " therefore "; and ma'zdlek " nevertheless "; or of combinations of free and bound morphemes, such as the following: fe-l-masal " for example "; fe-1" " haqiqe" in truth "; 'ala-r-rasm" as usual "; be-l-qovve potentially "; 'end-al-motdlebe upon demand "; " as far as possible "; in all of which a bound morpheme is joined to a free morpheme hatta-l-emkdn with the help of the article -al-. The adverbial compound may consist of an initial al- and a noun; e.g.
22

These estimates are based on an examination of fourteen short passages, in a number of different fields, which this writer undertook about 1956-57. The ratio of the Arabic words in each passage was as follows: (i) a personal letter written by a person with little formal education, 20oper cent; (2) memoirs, 39 per cent; (3) an article on D.D.T., 40 per cent; (4) a prime minister's report to the Iranian Parliament, 40 per cent; (5) a book on Persian grammar, 41 per cent; (6) a letter written by a university professor with a wide range of education in Iran and Europe, 41 per cent; (7) a play, 42 per cent; (8) a literary journal, 43 per cent; (9) a book on literary history, 46 per cent; (Io) a news item in a newspaper, 47 per cent; (I1) a book on linguistics, 48 per cent; (12) a history of a newspaper, 58 per cent; (13) a newspaper editorial, 59 per cent; (14) an article in a law, 75 per cent. Strictly scientific texts were excluded, as were the writings of authors known to have a noticeable predilection for or against using Arabic words. The sample was far too small, totalling to 961 words. Nevertheless, several observations may be worth making: (a) The ratio of Arabic words takes a considerable jump (19 per cent) between passages (I) and (2) ; again (by 16 per cent) between (I I) and (12); and finally (by 16 per cent) between the last two passages (13) and (14). Otherwise, in about 71 per cent of the passages (2)-(I 1), the rise in the ratio from each passage to the next is gradual, the ratio ranging between 39 and 48 per cent. (b) The overall average for all passages is 44 7 per cent. (c) If we leave out passages (i) and (14), each of which, as we have seen, is separated from the next passage by a large percentage, the average for the remaining twelve per cent, which is above the overall passages is 45"3 average. On the other hand, if we leave out passage (i) as well as the last three passages (I2)-(14) (once again because of the existence of a wide gap at each of these

points), the average for the remaining ten passages is 42 - I per cent, which is belowthe overall average. (d) The three averages arrived at in the manner described fall within a very small range, from 42 - I to per cent. 45"3 The estimate of " slightly below 45 per cent " given in the text is based on the preceding calculation. This figure, it may be noted, is more than twice the figure quoted above for 'Alavi, the twentieth-century novelist. The difference can perhaps be explained by certain facts: the smallness of my sample, the unsophisticated nature of my statistical procedures, and 'Alavi's subject matter as well as his individual style. But perhaps even more important is the fact that the 'Alavi text is a novel and hence close to the informal spoken style, which in general contains a lower ratio of Arabic words.
23

Persian examples are cited in their current Persian pronunciation, not the original Arabic forms, in a transliteration system in which the following relations between sounds and letters are observed: s = th (as in think); j = j (as injar); c = ch (as in church); h = voiceless pharyngeal spirant; x = voiceless postvelar spirant (usually transliterated kh); z = th (as in them); i = zh (as in pleasure); i = sh (as in shop); s = emphatic voiceless dental spirant; z = the voiced counterpart of s; t = emphatic voiced interdental spirant; ' (a raised v = voiced postapostrophe) = voiceless pharyngeal spirant; = voiceless postvelar stop; y = semivowel velar spirant; q (always as in yet); ' (apostrophe) = glottal stop (hamza); a = /fa/ (fatha); e = /e/ (kasra); o = /o/ (zamma); =/a/ -= (alef); u = /u/ (the high back vowel); i = /i/ (the high front vowel); ow =- /ow/ and ey = /ey/ (both diphthongs). The other letters have approximately their English values. The socalled silent h (representing the phoneme /e/ in the stressed final position) is not transliterated; it can be assumed after any stressed final /-e/.

THE

ARABIC

ELEMENT

IN

PERSIAN

GRAMMAR

119

" in short ". Another type of phrase is made up of co-ordinated elements, as in tow'an va karhan al-,araz " willingly or unwillingly "; leylan va nahdran" by night and by day "; beyni va beyn-al-ldh" between God and me ". More complex types of phrases are illustrated by the following examples: be-ahsan-e " vojuh" in the best of ways "; kamdfe-s-sdbeq" as before "; menheys--al-majmu' on the whole "; men-al badv-eel-al-xatm " from start to finish "; and a number of other types. Finally, there are a few complete clauses; e.g. va qes 'ald zdlek " and so forth by analogy "; fa hov al-matlub " so much the better "; tavakkal-to'ala-l-ldh " I trust in God ".24 4.I In the first four centuries or so of the Islamic era, the Arabic element in Persian consisted almost exclusively of individual words, which underwent changes in pronunciation and/or meaning. Sometimes, they were also modified in their grammatical function, as when, for example, Persian plural suffixes were added to the Arabic broken plurals, which were, in these situations, treated as singulars. in the sixth and seventh (twelfth and thirteenth) centuries, and even Gradually, however-especially more in the tenth and eleventh (sixteenth and seventeenth) centuries-writers in Persian came to use, not only simple Arabic forms, but also grammatically complex ones, such as forms containing the feminine suffix, or those showing gender concord. In the past hundred years or so these conditions have continued, although, as we shall see, some reaction has been shown against them in the last few years. It is such grammatically complex forms which we subsume under the broad term " grammatical element ", and to which we shall now turn. 4.2 Writers on linguistic borrowing do not agree on whether or not grammatical interference is possible between languages. At one extreme is the older view that completely denies the possibility of such interference. At the other extreme, it is held that nothing in language is " secure against invasion by foreign material ".25 As a number of scholars have pointed out, such contradictory views are due to the absence of universally accepted concepts and definitions.26 Of particular relevance is the lack of agreement on the lines, if any, separating morphology from syntax, or grammar from lexicon, or, for that matter, free morphemes from bound morphemes. Some linguists have in recent years attempted methodological innovations, or modifications, which they hope will solve the problem.27 In this paper no efforts will be made to discuss the various views, or to make new proposals. Rather, we shall work on certain tentative, and unstated, assumptions. 5. The Arabic element in Contemporary Persian grammar can be described under several headings, as follows: Phenomena I. Introduction Non-Persian Grammatical of This category includes features and devices peculiar to Arabic and non-existent in Persian except in the Arabic words within it, such as the following: (I) The device of discontinuous morphemes, an entirely Semitic device now used, almost without limit, in Persian, an Indo-European language. Thus, from the root k-t-b we have: ketab " book"; " kotob " books "; kadteb scribe "; kottdb " scribes "; maktab " school "; makdteb" schools "; maktub
24

This list of categories is far from exhaustive, but will probably give a very general picture of the situation, and place in perspective the linguistic phenomena described in the following paragraphs as examples of the Arabic grammatical features in Persian. 25 For examples of the various views on the subject, see the following works: 0. Jespersen, Language: Its Nature, Development,and Origin (New York 1923), pp. 213-15; U. Weinreich,

Languagesin Contact: Findings and Problems(New York 1953), p. 29; L. Deroy, L'empruntlinguistique (Paris 1956), pp. 72-1x 0. 26 Weinrich, op. cit., p. 29; E. Haugen, Bilingualism in the Americas: A Bibliography and Research Guide (University, Alabama 1956), p. 63 (Publications of the American Dialect Society, no. 26). 27 Weinreich, op. cit., pp. 29-31; Haugen, op. cit., pp. 6o-62.

120

JOURNAL

OF PERSIAN

STUDIES

" letter "; makdtib " letters "; mokdtebe" correspondence "; from the root q-t-l: qatl " murder "; " murderer "; maqtul " murdered "; qattdl " murderous, deadly "; from the root '-l-m: 'elm qdtel " knowledge, science "; 'olum " sciences "; 'dlem " scientist, learned "; 'olamd " scientists, learned men "; 'alldme " the most learned "; 'alim " very learned "; ma'lum " known, obvious "; este'ldm " enquiring "; ta'lim " teaching "; ta'allom " learning "; mo'allem" teacher "; mota'allem" student ". Discontinuous morphemes appear both in word formation and in the rest of grammar, as illustrated by these examples, and as will be further discussed later. (2) Grammatical " categories ": " (a) the dual, as in taraf" side ": tarafeyn" the two sides "; zowj husband "; zowje " wife ": " husband and wife ". zowjeyn mo'alleme "woman-teacher "; zdni (b) the feminine gender, as in mo'allem "teacher": " adulterer ": zdniye " adulteress "; mohtaram: mohtarame respectable " " (3) Grammatical " relations ". Concord appears in two grammatical environments in which Persian does not otherwise observe it: " (a) concord in number between nouns and adjectives. Thus, in dowlateyn-emo'dhedateyn the the noun and the adjective are both in the dual; and in (two) contracting governments ", 'olamd-yea'ldm " the distinguished learned men ", both components are in the plural. (b) concord in (number and) gender, in such cases as the following: maqdldt-evdrede" articles received "; amvdl-e masruqe " stolen properties ". In these and similar examples, the feminine forms of the adjectives are used with the plural nouns, as required in Arabic, where the feminine singular adjective is usually employed with a sound feminine plural noun or with a broken plural (though in Persian the rule is often applied incorrectly). II. Additionof New Items Within Existing Categories In this group, no new grammatical category is involved. What has taken place is an enlargement of already existing categories through the introduction of new items from Arabic: (I) Two types of Arabic plurals: (a) those containing the suffixes -dt, -in, -un, as in: daraje" degree ": darajdt; mosdfer" traveller ": mosdferin; nahvi " grammarian ": nahviyyun. (b) " broken " plurals, as in: masjed " mosque ": masdjed; madrese" school ": maddres; fekr " thought ": afkdr; amr " command ": avdmer; Sd'er" poet ": Jo'ard. " (2) Arabic elative forms: .Sahir famous ": alhar; fdsed " corrupt ": afsad; sari' " quick ": asra'. (3) A number of Arabic morphemes used as derivational affixes such as zu (in several allomorphs, " and both in the singular and the plural) in zo-l-jaldl " glorious "; gu-haydteyn amphibious "; gav-el" rational beings "; la in ld-'aldj " incurable "; ld-qeyd" nonchalant, unconcerned "; ld-maghab 'oqul
"irregligious "; etc.

(4) Words in the Arabic suffix -an (the tanvin), used as adverbs in Persian. A word of explanation may be in order here. In Contemporary Persian, in the native element, there is a syntactic class of adverbials, but not a morphological class. The adverbials are of two types. By far the largest number of them are adjectives or nouns used in the adverbial position; e.g. xub which means both " good " and " well "; daheste slow, slowly "; tond " fast, quickly "; sobh " morning, in the morning "; etc.; " and the adjectives ending in the suffix -dne, as in dustdne" brotherly, in a brotherly manner "; marddne " manly, like a man ". The other group of adverbs is made up of a very small number of words which function only as adverbs; e.g. hargez " never "; hamile " always ". There is nothing in the form of such words, or in the affixes that can be added to them, that is peculiar to adverbs and that can thus mark them as such. The only morphologically marked adverbs are those ending in the -an form of the Arabic tanvin(marking the accusative case of the noun), as in: sari' " quick ": sari'an " soon, quickly "; " apparently "; etc. Thus, on the morphological level, Persian has no edher " apparent ": adverbs except those ending in the suffix -an, which is borrowed from Arabic. .dheran

THE

ARABIC

ELEMENT

IN

PERSIAN

GRAMMAR

121

in Patterns III. RearrangementstheStructural

Strictly speaking, all the phenomena discussed so far-indeed any interference-result in some rearrangement in the structure of the recipient language. However, what we mean by " rearrangement " here is somewhat more specific. An example is provided by what has happened to the Persian phonological system under the influence of Arabic. For one thing, the number of consonant clusters, and their frequency, has been affected by the introduction of thousands of Arabic loanwords. Similarly, the allophonic relationship seems to have been affected in the case of [q] and [y] (the post-velar stop and spirant respectively). And the introduction of Arabic morphemes containing the voiced pharyngeal spirant [q] (the 'ayn), and those containing the glottal stop [P] (the hamza), both of which have been rendered as glottal stops in Persian, has had repercussions on the system, and on (linguistic) style. Under " rearrangements " mention may also be made of possible effects of Arabic on the word order in Persian sentences. This, however, is very difficult to pinpoint, since the more or less free variation that exists within certain types of sentences, mostly in the informal style, seems to have existed, at least to some extent, quite early in the development of the language, independently of the Arabic influence.28 6.I cases have been cited in the preceding paragraphs to illustrate the Arabic element Perhaps enough in Persian grammar, though the subject has by no means been exhausted. We add one or two general observations. To begin with, of the grammatical features mentioned, all but four are obligatory in Persian grammar, quite aside from the Arabic element. The non-obligatory features are the dual and the feminine forms of nouns and adjectives, and concord in number and gender between nouns and adjectives. All of these can be ignored in Persian. Duality and femininity can, if desired, be expressed by semantic devices. Concord need not-and often is not-observed at all, even in combinations consisting exclusively of Arabic words. The remaining features are obligatory in Persian, but for each Arabic form used there is at least one native Persian equivalent. Thus, the Arabic plurals can be replaced by native Persian plural forms. The Arabic elatives have equivalents with Persian suffixes. And so on. This is somewhat less true of the device of discontinuous morphemes, or the adverbs ending in the Arabic suffix -an. Even here, however, Persian equivalents can often be found, though they may be used less commonly than the Arabic forms. 6.2 Another important question is, whether the Arabic grammatical features can be combined with non-Arabic Persian words. There is, in fact, a handful of cases where Arabic grammatical morphemes are combined with Persian morphemes or words, or where concord is applied to sequences of native " letters Persian nouns and adjectives; e.g. kafi " shoe ": kaffd " shoemaker "; ndmehd-ye received ". The largest number of such forms, however, is considered " substandard ", " uneducated ", vs.ele " incorrect ", etc., even by those scholars who would freely use Arabic forms in their speech and writings. 6.3 The very large number of Arabic words in Persian, considered along with the Arabic grammatical features they contain, would create the impression that a great deal of grammatical borrowing has taken place. On closer examination, however, this impression turns out to be inaccurate. What Persian has done is to take over a large number of Arabic words which happen to contain Arabic grammatical morphemes and devices. The number of instances where Arabic bound morphemes are combined with Persian words is infinitesimal, especially in the case of the discontinuous morphemes.29
28

de Lazard, La Languedes plus anciensmonuments la prosepersane, pp. 464-5. 29 Attention is here called to the situation in English, where French, Greek and Latin affixes are added to native English

words quite freely. (See O. Jespersen, Growthand Structure of the English Language,4th edn. (New York 1923), pp. io6-io,
I24-8.)

122

JOURNAL

OF PERSIAN

STUDIES

The situation, in other words, is similar to that represented by the Latin plurals in English. What we have, then, is an Arabic stratum, or layer, within the Persian grammar, along with the native Persian stratum, each keeping, to a very large degree, its internal identity and independence. It has been suggested that grammatical borrowing can be said to have occurred only when a grammatical morpheme (or device) is taken over, without accompanying free morphemes, from one language into another, and applied to the forms of the borrowing language. That is, if a language borrows a number of foreign words containing a foreign grammatical morpheme, or some other phenomenon, then isolates this phenomenon and applies it to native morphemes, what we have is not grammatical borrowing. Rather, it is a case of extension by analogy, which is constantly at work in language, and is the process through which native speakers learn their own language, i.e. through which they discover, isolate, and apply the rules of their language in new situations. In these terms, also, there are no cases of grammatical borrowing from Arabic into Persian. 7.-1 in Persian, which was first introduced in the early centuries of the Islamic era, The Arabic element has continued to expand. In recent decades, however, two related developments have curbed the increase in the proportion of that element in the language. On the one hand, new Arabic borrowings have been introduced at a slower rate than before; and, on the other, some of the old borrowings have gone out of current use. There seem to be two major reasons for this trend away from Arabic loans. One reason is the simplification of the written style, both literary and otherwise, bringing written Persian closer to the language of common speech, in which the Arabic element is less extensive. Another major reason is the movement for purification. 7.2 The desire for purifying the Persian vocabulary, or, at any rate, the preference, on the part of some writers, for Persian words over Arabic words, has a long history. Ferdowsi, the tenth-century poet, for example, used a rather small number of Arabic words in the 6o,ooo couplets of his great epic. Others, on the other hand, seem to have made a point of using Arabic words and complex Arabic forms to the Malek greatest extent possible. Perhaps the best representatives of this group are the historians and Vassdf (fourteenth century). These two historians are often almost 'At. Joveyni (thirteenth century) completely unintelligible even to a moderately educated twentieth-century Persian speaker, who finds it relatively easy to read such earlier writers as the scientist-philosopher Ebn Sind, the politician-scholar Neztm al-Molk, and the philosopher-traveller Naser-e Xosrow, as well as the poet Ferdowsi.30 7.3 In more recent times, conscious efforts to purify Persian from its Arabic element began in the nineteenth century, became stronger in the constitutional period, and enjoyed full organized governmental support in the nineteen-thirties.31 The most extreme product of these efforts is the style known as Fdrsi-ye Sare, which had its heyday during that decade. The advocates of this style insisted on throwing out all the Arabic words and replacing them with " pure " Persian ones. They based their choice of the Persian substitutes on the flimsiest evidence, such as the absence, in a word, of certain letters of the alphabet which were supposed to occur only in Arabic words, and the presence of certain others supposed to occur in Persian only. Their sources were usually the all-too-often unreliable Persian dictionaries. The result was quite frequently unintelligible. In a passage in Frsi-ye Sare, it was not uncommon to find nine out of every ten words to be entirely unknown to the readers; and the authors had to gloss them in parenthesis or in footnotes.32
30

NegarA (a pseudonym), Ardyes va pirdyed-e zabdn-efdrsi (Tehran 1327/1958), p. 27. 81 'Aziz Dowlatdbhdi, Tdrix-e tahavvol-enasr-e mo'dser (Tabriz 1333/1954), chs. ii and iii.

32

Perhaps the most notorious supporter of Fdrsi-yeSare was the newspaper Irdn-e bdstdn,published in the nineteen-thirties in Tehran. This paper published articles in Fdrsi-yeSare (Dowlatdbddi, pp. 15-16).

THE

ARABIC

ELEMENT

IN

PERSIAN

GRAMMAR

123

Among the leaders of the opposition to purification have been S. Hasan Taqizade and the late Mohammad Qazvini.33 These scholarsdemonstratein their own writings a strong partiality for Arabic words, including quite a few which are not part of the common stock of ContemporaryPersian vocabulary. In addition, they use many Arabic grammatical formsof the type describedearlierin this paper. Sareand with Taqizide and The late Ahmad Kasravi disagreed both with the advocates of Fdrsi-ye The major features of his style can be summed up as follows. He retained the commonest Qazvini.34 Arabic words, especially when these were simple bases. Of the less common or the complex ones, he replaced many with Persian words which were easily intelligible and often more common than the Arabic words they replaced. Still others he replaced with words made up with the help of Persian affixes,or with Persian compounds. By using these methods he was able to avoid a great many Arabic forms without becoming unintelligible; if anything, he often became more intelligible than both of the extremist groups. Two other features made it possible for Kasravi'sstyle to avoid Arabic forms and yet remain natural. One was his use of Persian phrases and collocations, many of which he had picked up in the course of reading Classical Persian texts. The other was his realization that it was not necessary to find a Persianreplacement for every Arabic form used in Persian. No one else seems to have realized this fact, with the result that the advocates of Fdrsi-ye Sarenot infrequently used unknown, often incorrect, forms supposed to be pure Persian when they could have used expressionsconsisting of Persian components already familiar. In his later years, Kasravi did introduce a handful of unknown words of the type which he had originally consideredunacceptable, but these constitute a very small number and account for a very low proportion of the actual occurrences in running text. On the whole, he continued to retain a natural style.35 7.4 or An important development in the purificationmovement was the founding of the Farhangestdn, the Iranian Academy, in 1935. The constitution of this governmental organization listed twelve objectives, but in practice its activities were practically limited to finding, or coining, Persiansubstitutesfor foreign words. On the whole, it was not a successfulventure, and became inactive in 1941.36
33Taqizide has perhaps been the most persistent in his views, to which his prestige as an elder politician-diplomat-scholar has lent great support. His major presentation of these views was in the first lecture (on a scholarly subject) that he gave in Tehran after returning from an extended stay in Europe. Entitled " Lozum-ehefz-efdrsi-yefasih " (= " The Necessity of Preserving the Eloquent Persian "), the lecture was delivered at the Teachers College on 5th Esfand 3326 (February 22nd 1948), and later published in the journal radgdr IV (1948) and also separately (Tehran 1327/1948). The book by Negard cited in footnote 30 above was a reply to this lecture. 34 Ahmad Kasravi (1890-I946) was a distinguished Iranian historian, linguist, jurist, and an advocate of, and prolific writer on, extensive and far-reaching social reforms in all aspects of culture. His views on language are closely related to this last aspect of his activities, and are to be found in a large number of articles as well as the book Zaban-e pak (Tehran 1322/1943). A collection of some of the articles he published during 1933-37 can be found in Ahmad Kasravi, Zaban-efdrsi va rdh-erasa va tavdnd dn, garddnidan-e ed. YahyA Zokd' (Tehran 1334/1955). 35 It should be emphasized that Kasravi opposed the Fdrsi-yeSare style, and he did so in no uncertain terms. He found that style repellent. He considered it as " nothing but a disgrace to the Persian language " (Zdban-efdrsi, p. 2), and believed that it would " destroy the language even more " (ibid., p. 3). The term " Fdrsi-yesare " is usually rendered as " Pure Persian " in English; even the Persian-speaking author Dowlatib~di (footnote 31 above) discusses Kasravi in the chapter on the Fdrsi-yeSare. The English rendering is not quite adequate, and DowlatbAbdi's classification of Kasravi with the writers in Fdrsi-yeSare misses the point, though his description to some extent makes up for his classification. The word sare is hardly ever used in Contemporary Persian except in connection with the language reform, and there it has come to designate the extreme style described above. The word Kasravi most often uses in this content is pirdstan,which he uses in the sense of " to prune ", " to improve by taking something away ". Moreover, he uses the term to refer to much more than the vocabulary. Again, Kasravi's method is entirely different from that used by the supporters of Fdrsi-ye Sare, who, as we have seen, would throw out all Arabic words, and would replace them with entirely unknown words supposed to be " sare Persian ". 36 On the Farhangestdn, R. Lescot, " La reforme du vocabulaire see en Iran ", Rivuedesitudesislamiques(1939, Cahier I), pp. 75-96. see For a list of the words adopted by the Farhangestdn, Farhangestdn-eIrdn, Vdie-ha-yenow ke ta pdydn-esdl-e 319 dar Farhangestdn-e Irdn pagirofte bode ast (Tehran 1319/1941). For the views of a prominent Iranian scholar on language reform and the Farhangestan, see H. Mass6, " La lettre a l'Acad6mie Iranienne de S.A. Mohammed Ali Foroughi ", Revuedes itudes islamiques(1939, Cahier I), pp. 17-74. It is interesting to note that the Iranian Academy was founded only three years after the establishment of the Turkish Linguistic Society. There is, in fact, some evidence that Iranian language reform owes something to the developments along the same lines in Turkey. The first book in the Fdrsi-yesare was written by an Iranian resident in Turkey, and published in Istanbul, in 1300 A.H. (1883 A.D.). (See Dowlatibadi, pp. I 1-12.) On the Turkish situation, see C. E. Bosworth, " Language Reform and Nationalism in Modern Turkey ", The Muslim World LV (1965), pp. 58-65, I 17-24, where the Iranian language reform movement, and the are Farhangestdn, briefly mentioned (pp. I22-3). His very brief statement on Kasravi needs to be revised, but his evaluation of the activities of the Farhangestdn,in comparison with the Turkish Linguistic Society, though brief, is to the point.

124

JOURNAL

OF PERSIAN

STUDIES

7.5 Advocates of language purification have generally been concerned with the vocabulary alone. It is true that writers in the Fdrsi-yeSare, avoiding Arabic words, generally did not use grammatically complex Arabic forms, but this was only an incidental consequence of their original concern, which looks at language primarily (if not exclusively) as a collection of individual words. In recent decades, a few scholars, such as Parviz N. Xanlari, have now and then prescribed against using such Arabic phenomena as the broken plurals in Persian. On the whole, however, they have not offered many specific details. The only notable exception was Kasravi, who followed two general principles in his discussionof language. One was that a language should maintain its independence; and the other was that a language should be entirely regular and consistent in the application of its rules.37 Both of these concepts dictated against the use of Arabic grammatical patterns in Persian. He set a good example himself, for he did not as a rule use Arabic patterns in his own writings from 1933 onwards.

7 A. Kasravi,
17-I9.

Zabdn-e pik (footnote

34 above),

pp. 3-5,

I2-13,

THE GENESIS OF SAFAWID RELIGIOUS PAINTING ByJ. M. Rogers*

The existence of religious wall paintings dating from the late Safawid period has been known to historians for some considerable time and, indeed, formed the subject of a paper read to the Third International Congressof Iranian Art and Archaeology at Leningrad in 1935,1when the question of its origins was also opened. Our knowledge of the monuments, notably the Imamzdda Shah Zayd at Isfahan and a second Im~mzada at Amul has unfortunately increased little in the thirty years or so since this first report, and it is far from clear how many more of these wall paintings remain to be discovered. Whatever the state of affairs, the fundamental question of the style and content of the paintings must await a complete survey of the extant monuments, as well as more detailed publication of those already known.2 The purpose of the present study is, therefore, severely circumscribed: questions of iconography are of necessity rather neglected, and the main emphasis is upon the intellectual process by which such an exceptional genre of painting for Muslim society apparently came to be accepted by the orthodox. The pictorial sources upon which it drew are still mysterious,but it is, meanwhile, possible to throw light upon the reasonswhy in the late Safawid period there seems to have been a breakdown of the traditional Muslim prejudices, Shi'i as much as Sunni, which, it is still customary to believe, so much inflamed the Muslim 'ulamd'at every period against painting in a religious context. The argument falls into three parts, firstly, a discussionof the nature of religiouspainting in general; secondly, an examination, in rather more detail than has previously been attempted, of a number of literary sources for the existence of religious painting in Islam in the pre-Safawid period; and, finally, an attempt to suggest that the development of the religious paintings of Safawid Persia finds its origins in the culture of Iran and Transoxania of the fourteenth-fifteenthcenturies, which are characterized by a strikinggrowth of mystical or devotional literature and which, on some occasions, is accompanied by manuscript painting which, itself, has a fair title to be called " religious ". It should, however, be clearly stated initially that the problem which is under discussionis not one which may readily be answered by reference to current Shi'i practice in Iran or Iraq. Sociological or anthropological investigation, while doubtless interesting in what it may reveal of what the orthodox Shi'i now believes to have been the case, will only mislead us if we allow it to do duty for an historical enquiry. More precisely, it may well be the case that Diba3 is correct in describing the pictures of 'Ali which now appear in many public places in Iran as icons, in something approachingthe liturgical sense of the term; but it would be grossly improper to assume from this without careful investigation that when these paintings first appeared they played a r61le analogous to icons of the period (or of any other in Byzantine society. Moreover, it is widely believed, even by most orthodox Muslims at the period)
* The present article is a revised version of a communication originally delivered to the Fifth International Congress of Iranian Art and Archaeology, Tehran, April 1968. I am indebted to the Research Fund of the American University in Cairo for assistance towards the writing of the article. 1 Mehdi Diba, " Quelques aspects de l'Iconographie Musulmane en Iran ", Proceedings of the 3rd International Congress of Iranian Art and Archaeology (Moscow/Leningrad 1939), PP. 53-6. 2 The best bibliography to date, which also includes zoomorphic decoration employed in religious architecture in Persia, is to be found in Monneret de Villard, Introduzione allo Studio dell'ArcheologiaIslamica (Venice/Rome 1966), p. 263. This Ornament may be further supplemented by Denike, Arkhitekturny Srednei Azii (Moscow 1939), p. 2 and fig. 197, and Stchoukine, La Peinturedes ManuscritsTimourides(Paris 1954), p. 3 both on the paintings in the tomb of Shirin Biqa Aqi (1385) in the Shah-i Zinda at Samarqand; Pugachenkova, Puti Razvitiya Arkhitektury Yuzhnogo Turkmenistana Pory Rabovladeniya i Feodalizma(Moscow 1958), pp. 429-30, and Rempel', ArkhitekUzbekistana(Tashkent 1961), p. 362, both for the turnyornament dragons on the mosque at Anau (1456) and for the Shir Dir Madrasa at Samarqand (1619-36) with its panels of tigers and a sun, doubtless copying those on Timur's palace at Shahr-i Syabs (138-I1405) which Clavijo describes.
3

Op. cit., p. 53-

125

126

JOURNAL

OF PERSIAN

STUDIES

present day, that painting in Islam has grown up in defiance of certain explicit pronouncements in the Qur'an and the hadith literature; whereas, dispassionate consideration of the Qur'dn shows that it contains nothing at all on this question, and that the hadith is very far from being clear.4 Current opinion, therefore, is not to be accepted unquestioningly. To proceed, however, with the question of what is to count as Muslim religious painting. Arnold's Painting in Islam5 takes it for granted that this is simply a matter of picturing prophets, saints or any scene related to some episode in the Qur'dn. It is, of course, undeniable that many scenes first described in the Qur'an have been illustrated at some time or another, notably, the story of Joseph (Silra XII which, exceptionally, is entirely narrative), Solomon, his jinns and the Queen of Sheba (Siira XXVI, verses 16-44), and a number of episodes from the Exodus in which Moses plays a central part (Stiras VII, XX and XXVI). It is also true that many scenes are illustrated with great fidelity to the details given in the Qur'dn, for example, the incident where Zulaykhd, inflamed with love for Joseph, pursues him and rips the shirt from his back (Saira XII, verses 25 ff.). But against this pictorial tradition, there is the equally undeniable fact that there has been at all periods a fundamental prejudice in Islam against illustrating the Qur'dn itself. What, then is the status of these paintings ? The answer appears to lie in the fact that they were held to illustrate not the Qur'an itself but rather the secular romances roughly based upon it, like Firdawsi's Yiasufu Zulaykhdand the many other versions which followed it, or historical works and panegyrics, like Rashid al-Din's Jdmi' al-tawdrikh and al-Birfini's al-Athdr albdqiya, which were not held to contain religious paintings at all. The point is, I think, clear from the Il-Khanid illustrations to the latter two manuscripts, where, although illustrations of scenes from the Qur'an and from the lives of Muhammad and other Prophets are frequent, they have no devotional purpose and are, rather, intended to emphasize the historical connection of the il-Khans with the earlier Islamic dynasties and, by fitting them into the general development of history as seen in the Qur'an, to minimize the strangeness and lack of antecedents of the r6gime and, as it were, legitimize it. In the case of the secular romances it is rather more difficult to say. Even supposing that Firdawsi's original purpose in writing a poem on the subject of Joseph may have been secular entertainment rather than religious instruction, the ability of subsequent Sfifi commentators to find allegory in almost anything makes it now very difficult to judge on internal evidence alone whether subsequent romances on this theme were secular or religious in intention. However, if we may judge by the illustrations to these later romances, it is perfectly clear that for the most part the painters envisaged their task as the entertaining illustration of a story which was in their eyes itself primarily conceived as an entertainment. It follows, therefore, that Arnold's definition of religious painting is inadequate, and that whether painting is religious or not depends not upon the subjects chosen, nor the expression of religious feeling, nor even upon the feelings which inspire the painter at this work, but upon its function. The relevant functions which religious painting may fulfil are various-the explanation of a text, the teaching of a lesson, allegorical, spiritual or moral, and perhaps (though probably only as a by-product of the foregoing) the inculcation of reverence towards revealed religion in the beholder. This will be clear from consideration of any work of the Italian Renaissance which pretends to the title of religious art. But, in contrast, none of the illustrations of the works referred to above seems to have had these functions. They do not explain the Persian text and are, for various reasons, incomprehensible without it, so that they cannot be said to teach any lesson which the text itself does not teach. Is this true of Islamic painting as a whole ? It is certainly true that it would not be profitable to illustrate the Qur'in. It follows from Islam's denial of the Incarnation that Christ, like the other Prophets, was never more than a man, a good man, doubtless, but still far from embodying in Himself the absolute prescriptions of divine law contained in the Qur'in. Hence in contrast to Christianity where the central conception of Christ as God sanctifies the practice and teaching of religion through His example as a man, in Islam the notion that one may best acquire sanctity by imitating the Prophets and the saints is necessarily secondary. By virtue of the
* For example, al-Nawawi (al-Minhiaj sharhMuslim b. al-Hfajjdj fi VIII, p. 398), whose general exposition [Cairo 1304-05], as late as the thirteenth century, is one of the first Muslim attempts to consider the various pronouncements of the 'ulama' on the subject as a whole up to the time when he wrote, and who is unable to resolve the clear contradictions which the hadithcontain. 5 Oxford 1928, pp. 90- 16.

THE

GENESIS

OF SAFAWID

RELIGIOUS

PAINTING

127

Incarnation the New Testament is both the chronicle of the acts of the man, Christ, and the revelation of God to His faithful, the Old Testament being envisaged as the foreshadowing of this, to the extent that the Prophets and Kings of the Jews are seen as types of Christ. This is, indeed, the basis of the argument used by the Church Fathers to justify the illustration of Holy Writ, since the life of Christ in pictures could be seen as a convenient way of demonstrating to those who could not read the moral and spiritual lessons which they had to take to heart. The narrative episodes of the Qur'dn, however, do not ever seem to have been intended to form the connected narrative of God's revelation of Himself to man, and in general seem to take second place to the rules which normally fill out the Stiras in which they occur; indeed, the relationship of narrative episodes to legislation is often unclear and may on some occasions not even be illustrative. And M uhammad, unlike Christ, was never seen as more than a man so that, originally at least, his life was not viewed by the faithful as having essential didactic importance; the concern of exegetes was rather to discourage the orthodox from polytheism by emphasizing the complete difference between the reverence or respect due to him as a Prophet, or as the greatest of the Prophets, and the worship due to God alone. In the light of these considerations we may come to recognize the impossibility, rather than the impermissibility, of illustrating the Qur'an. The narrative elements form only a part of God's revelation, and the life of Muhammad is irrelevant to them in a doctrinal sense; and as for illustrating the legislative elements, as well illustrate the Code Napoldon. The clear impossibility, therefore, of must have been well understood in the early centuries of usefully illustrating the Qur'dn-which Islam-may ultimately in the late eighth-ninth centuries have come to be rather misunderstood as a prohibitionagainst doing so. But there is, of course, little point in prohibiting what cannot adequately be done anyway. This somewhat arid conclusion did not find much sympathy in Iran, where the status of 'Ali, if difficult to define in terms of Christian theology, has traditionally been higher than that accorded to Muhammad by the early Sunni theologians, even to the point of his being considered the perfect man. Furthermore, the early development of Sfifi mysticism in Iran laid stress on personal sanctification rather than a rigid legalism. Also, in regard to the sixteenth-seventeenth centuries, with which we are here concerned, it is clearly the case with Safawid religious painting that it satisfies some of the conditions for religious art which were established above, in that it aims to encourage the faithful to emulate both the spirit of devotion even unto death of the early Shi'i martyrs and the reverence for them of their teachers. The figures are splendid, even if not always well painted, and their faces are veiled, with a nimbus where appropriate. The function which these late-Safawid paintings fulfil is, however, so far from being new to the spirit of Shi'ism that the chief grounds for surprise at its appearance in Persia at this time should be that it appears only in the handful of monuments which we know so far and at such a late date. In concluding I shall attempt to sketch out its relation to earlier developments in Muslim painting; but it is perhaps as well to admit at this point that the question of its excessively late appearance in Iran is one which it seems at the moment impossible to answer fully. Although the monuments are deficient there are, however, various literary testimonials to the existence of painting in Islam which might properly be described as religious. Although they are, I think, conclusive proof that the Safawid wall paintings known to us cannot have been the earliest examples of the genre, they are, nevertheless, rather difficult to evaluate. For one thing, there are so few of them. We are almost entirely dependent upon the reports of observant Western travellers, but many of these, for example Chardin (who actually devotes a whole chapter of his Voyages6to the state of painting in Persia and yet does not mention the wall paintings of the Imimzida Shah Zayd at Isfah7n, which he almost certainly saw), utterly fail to mention them. Yet again, it is difficult to know exactly
6 Voyagesdu Chevalierde Chardin,ed. Langlhs (Paris 18I11), V, especially pp. 202-03, where he actually states that painting is forbidden to the extent that the Shi'i 'ulamd'will not tolerate portraits of human beings or even pictures of animals. But since he adduces this fact as an explanation not of the fact that there is no such painting in Persia but of the badnessof what

there is (at least, in his view), he cannot be regardedas conclusive proofeitherway. SFor Chardin'svisit to Isfahdnsee Y. Godard, " L'Imamzade Zayd d'Isfahan: Un Edifice decor6 de Peinturesreligieuses Musulmanes Athar-e'Iran, II, part 2 (Paris I937), ",
pp. 341-8.

128

JOURNAL

OF PERSIAN

STUDIES

what to make of this lacuna: it certainly cannot be taken as proof of the non-existence of the genre, and factors like the widespread Shi'i unwillingness to allow non-Muslims into their shrines could perfectly well explain the absence of references by so many Western travellers to any paintings which they might have contained. Even so, it is still surprising that they do not mention rumours of them. The second difficulty arising from the sources is that they are very far from proving the existence of the same phenomenon. This is perhaps less serious than it at first appears, since it may merely reflect what we might conclude from the paucity of the extant monuments, namely the largely unformulated and probably unofficial character of religious art in the pre-Safawid period. Ibn Battilta will, therefore, be admissible for his account of the " picture of the Prophet 'Ali " which he saw in the Crimea; Don Juan of Persia for his information that by the end of the sixteenth century there was an established tradition of funerary portraiture of Shi'i martyrs (shuhadd') in some Persian mosques; and the Qa?di Ahmad for his romantic attribution of all sorts of (figural) paintings to the tomb of Bihzdd (and, as we shall see, the description is important whether it was true or not). Ibn Battflta, who in the course of one of his journeys arrived at Kerch on the northern shores of the Black Sea, recounts a very curious discovery: " I saw a church (kinisa)so we made towards it. In it I found a monk (rdhib)and on one of the walls of the church (ft ahadhaytdn al-kinisa-inside, I think) I saw the figure (lra) of an Arab man wearing a turban, girt with a sword and carrying a spear in his hand (firat rajul'arabi'alayhi'imdma sayfan),and in front of mutaqallid him was a lamp which was alight. I said to the monk 'What is this figure ? ' and when he replied 'This is the figure of the Prophet 'Ali ' I was filled with astonishment. .. ."8 Nevertheless, he goes on to say, they spent the night there and even tried to cook some chickens which, unfortunately, had gone bad on the voyage. There is, of course, no need to emphasize the extraordinary nature of this report, though the grounds of Ibn Battita's astonishment are not entirely clear. He is, however, in general a very careful reporter of what he sees. Thus he tries to find Muslim equivalents for non-Islamic institutions (in describing Mosul, for example, he remarks that the mdnistir of the Greeks or Christians (rmi?yyin) corresponds to the Muslim zdwaya), and he frequently gives the spelling of words which might be strange to his public. We must bear this rather schoolmasterish approach in mind when we consider the accuracy of his account. First of all, kinisa. Ibn Battfta is especially precise about the words he uses for places of for worship--jdmi' and masjid for Muslims, zdwiuya the Dervish monasteries where he usually lodged, and dayrfor most of the Christian monasteries; the word, ribdt, with very few exceptions, being used for caravansarays.9 As his account suggests, the building must have been visible for some distance, and the use of the word kinisa thus suggests a Christian church. Why, then, the " portrait of the Prophet 'Ali "? And how would they, as obvious Muslims, have been allowed to spend the night inside it and even cook a meal? Allowing for the possibility that the monk was just too scared to refuse, it would appear that Ibn Battiita's use of the word kinisa is meant to reflect his own puzzlement as to what the place really was. Secondly, although Ibn Batttita is no less precise about the Shi'i shrines of Mesopotamia which he visited and the special rites of the pilgrims visiting them, he makes no mention of any images there, and it would appear that if at the present time the cult of images is established there this is as much the work of Persian Shi'ism in comparatively recent times as of any tradition associated with the shrines of Iraq. Moreover, the Golden Horde, which ruled the hinterland of Kerch at the time of his visit, was certainly Sunni. Kerch was, however, a port with a large variety of foreign colonies in it, and it is more than possible that it may have contained small pockets of Persian Shi' merchants, who would naturally have had their separate places of worship. There is, therefore, on religious grounds no reason why we should rule out Ibn Batfita's account a priori.
8 Gibb, The Travels of Ibn

II (London 1964), P 467; Batta.ta Defr6mery and Sanguinetti, Voyages d'Ibn Battuta II (Paris 1877), pp. 355-6. The italicized words above must be regarded as a series of exclamations at the apparent sense of the text.

* Gibb,

p. 349; Defr6mery and Sanguinetti, II, p. 137. In one case, however (the Ribit Yfinus near Nineveh, which seems from the collection of lavatories and water channels (matdhir), masjids and mihrHbswhich it contains to have been a Muslim buylt kathira. place of worship), he says of it ribd.tfhi

THE

GENESIS

OF SAFAWID

RELIGIOUS

PAINTING

129

al-nabi 'Ali? Gibb, the most recent translator of Ibn Thirdly, we have the problem, why has suggested that he merely misunderstood his informants, that the edifice was indeed a si.rat Battiita,10 church and that the painting was actually an icon of Elias (Greek genitive, Elia). This interpretation raises the question of why the church should contain an icon of Elias at all, but it is very difficult to see how Ibn Battita could have come to make such a mistake in the first place. As is well known, the letter 'aynis extremely persistent in Arabic phonology and is never dropped in favour of alifmadda, as the 'All-Elias transformation would demand. Although Greek monks might hardly be expected to recognize the difference between the two, Ibn Battita himself would certainly have done so, and with his usual pedantry about orthography and divergent pronunciations, would certainly have enquired further. The possibility that the monk was scared and gave the answer which he thought would please most, merely leads to the problem why he gave that particular answer, and not something more innocuous. A Muslim might hardly be expected to take such an answer for granted, and the fact that Ibn Batitta seems to have done so suggests that he at least considered the answer to be well-founded. An examination of alternatives to Elias in Byzantine iconography is of no help either. The Byzantine calendar abounds in warrior saints, but it is difficult to think of any wearing what could be described as a turban. Some Byzantine officials at this period wore costumes superficially resembling Muslim dress, like the turban-like bonnet (the skiadion,or sunshade) and the long khil'a-likell robe of the Grand Logothete or Comptroller of the Imperial Treasury. But there is no evidence that even Byzantine generals, let alone soldiers, wore similar uniforms; that of the Grand Logothete is the only one known to me from Byzantine art.12 As for the suggestion that the kinisa may have been Italian, with a consequent Italian origin for the painting which Ibn Battita describes, though it might appear probable on the grounds that there was a large Genoese colony in fourteenth-century Kerch, we may dismiss it out of hand on grounds of style. Fidelity to details of costume, particularly Byzantine or Islamic, is rare in Italian art before the time of Gentile Bellini, who visited Istanbul and Damascus in 1480-81; and it is equally difficult to think of an identification for the figure in terms of an Italian saint. It follows that the objections to Ibn Battfita's admittedly surprising account do not hold and that, on internal grounds, he may well be accurate. Our suspicions may well remain, however, since we have no other even remotely contemporary evidence for such religious paintings in Islam, except, indirectly, an inscription from the Jdmi' al-QOadi in Aleppo, an undated Maml-ik building, written in a fourteenth-fifteenth century Mamlkik
naskhi":13

" Cursed be he who concerns himself with images representing living creatures near this mosque, or who erects a statue of a living creature for people to assemble in front of, or who exposes it for sale.. " There follows a quotation of al-Bukhdri's hadithto the effect that the painter at the Last Judgement will be given the impossible task of breathing life into the images he has made. There are many possible reasons for such an emphatic proscription, and the inscription may therefore merely be ammunition against Christian churches in the vicinity. However, in Mamlfik Syria there can have been few Christian communities confident enough to expose themselves to this form of attack; and, in any case, should the Christians have been scandalizing good Muslims, the remedy would not be to place a comminatory inscription on the wall of a mosque, but to destroy the images and, doubtless, punish the perpetrators. It could also be seen as a purely general reminder to the orthodox not to diverge from the path of virtue, though it is quite unclear why it should have been thought necessary just at this period to make it. Indeed, reflection suggests that it must have had a specific
10

1xKhil'a is used in a fairly strict sense (see Mayer, Mamluk

Costume [Geneva 1953]) to mean a long robe opening down the front, fastened with a belt and very much like a modern dressing gown. The garment is still worn in the Bukhara Oasis, where it is known as a khalat. 12 Underwood, The Kariye Djami I (London 1967), frontispiece and pp. 42-3. Theodore Metochites, the final restorer of the Church of St. Saviour in Chora at Constantinople at the time he occupied the position of Grand Logothete (1321-22), is

Gibb, op. cit., p. 467.

13

represented in the magnificent series of mosaics which he commissioned as being in court uniform. The date is conveniently near that of Ibn Battfita's travels, and it is especially unfortunate that Ibn Battuta during his long stay in Constantinople was never able to enter the churches and inspect their decoration because (as he tells us) of the requirement that visitors should prostrate themselves before the crucifix in the narthex. Herzfeld, " Inscriptions et Monuments d'Alep " in Corpus ArabicarumII, Syrie du Nord (Cairo 1956), pp. Inscriptionum
321-2.

130

JOURNAL

OF PERSIAN

STUDIES

purpose. An analogous case can easily be constructed. Murder is known to be forbidden to all Christians, yet we should never expect to see the commandment " Thou shalt not kill ", with a rehearsal of the special punishments for murderers, as an inscription on a church, unless some particularly flagrant case in the vicinity, or in the church itself, had just taken place and it was thought necessary to give the faithful a good lesson. Although we cannot place any great weight on the occurrence of this inscription it certainly suggests that at the time that it was carved, there had been attempts to introduce paintings of living creatures, or even statues, into the mosque. Just who made the attempt and why is quite impossible to say. Fortunately, the other testimonials which follow are less obscure. Don Juan of Pcrsia, originally a Shi'i Muslim called Uruch Beg, left Persia in 1599 as one of the four secretaries of the Embassy sent to Europe by Shah 'Abbds under the guidance of Sir Antony Sherley. In his Relaciones,written in Spanish after his conversion to Catholicism, he states that his father, 'Ali Beg Baydt, was killed at Tabriz in the Turkish siege of I58o and that Muhammad Khudibanda, the ruler of Tabriz, afterwards ordered a picture to be painted representing 'Ali Beg standing over the seven Turkish commanders whom he had accounted for. The picture, Don Juan says,14 may still be seen placed over the door of the mosque in Tabriz dedicated to the Amir Haydtr, whom he reports to have been held to be a saint (sic: the building is not now identifiable, and it is difficult to know whether he meant wali or not). The account is particularly interesting, since it demonstrates the existence of a genre of which no examples seem to exist at the present time, namely, of funerary portraiture; and because it shows the special consideration accorded to Shi'is who had fallen in battle and thus become martyrs. Although this is the only precise literary testimonial known to me, the emphasis which it places upon martyrdom is a very characteristic aspect of the militant Shi'ism of Safawid Persia. A third source, a treatise on painting written by the Ahmad c. i6o6, and which deals with the Qa.di great painters and calligraphers known to him, is of special interest in that it presents the attitude of a thoroughly sophisticated member of the Shi'i 'ulamd'towards painting in a religious context.15 Painting, in the view of the Qadi, is a product of the qalam, itself sanctified by being the instrument by which the Qur'an was written down. It is made even more respectable, however, because of the " miracleeven quoting a working pictures drawn by the pen of Muhammad and 'All ibn Abi Talib ", the set of verses in mutaqdribmetre to the effect that one of Muhammad's additional superiorities over the Q?.di Chinese (!) is that he excelled them in painting images. These protestations might strike one as somewhat dubious, since their exaggeration suggests an attempt to persuade unconvinced orthodox Shi'is that portrait painting, being recommended by the highest authorities, was therefore entirely above reproach. There is, in any case, no need to take his claims at their face value-which would in any case be small-and we must see his account more as an apologia for a genre which was widely cultivated but which up to that time had found little theological backing. However, the importance of the move from seeing painting as something at best permissible under certain circumstances to seeing it as something sanctified, as in the case of the father of Don Juan of Persia cited above, cannot be underestimated. The Qadi Ahmad is also interesting in that he uses as his criterion of the goodness of a painting its likeness to life, that is, precisely that quality which had been the ground of its condemnation by the early hadith, since it was precisely in this respect that the painter was held to usurp the creative activity of God. Just how great a revolution this was by the time the Qadi wrote is difficult to say, though it is certainly expressed, in a non-theological context, by 'Abd al-Razzaq Samarqandi in his account of the mission sent by Shah Rukh to China in the I420s. In his lavish descriptions of the Chinese temples seen by the embassy, we find no word of disapproval, and only praise, for the extreme life-like qualities of the statues and painted figures which they contained.'" Once again, it does not matter whether Samarqandi convinces us that his observations are true or accurate; the important fact is that he should use the criterion of vivacity and likeness to life as criteria of the quality of the painting and sculpture which he saw.
14 Ed. Le Strange, Don Juan of Persia, a Shi'ah Catholic,1560-1604 (London 1926). and Ahmadb. 15 Minorsky, Calligraphers Painters: A Treatiseof Qad.i
Mir Munshi (Washington, Freer Gallery 1959)-

16

Translated by Quatrembre in Notices et Extraits XIV (I843), passim. Samarqandi's judgement is, of course, odd.

THE

GENESIS

OF SAFAWID

RELIGIOUS

PAINTING

131

The most interesting observation from our point of view, however, is that after lavish praise Qadi's of Bihzdd he says: " He died at Herat and was buried in an enclosure (rawda, or tomb) full of paintings and ornaments (naqsh-u nigdr)." Minorsky in his commentary" writes that he dare not translate naqsh-u nigdr as " painted images ", since " even for Bihzad, such embellishment for a resting place would be extravagant ". But the extravagance is precisely the point. Bihzdd was still for the Safawids the ideal to which the painter should aspire, and the Qadi is not so much giving a historical account as ascribing to him the sort of praise he felt to be his due. The most important factor is that being buried in a tomb decorated with painted images is thought of as the most appropriate reward for the greatest Persian painter. The Qadi's account is therefore useful as testimony to the question of funerary painting in Persia. Moreover, it is possible that the Qadi was not being entirely fanciful. The tomb of Timur's younger sister, Shirin Biqa Aqa, in the Shdh-i Zinda at Samarqand (1385)18 contains a series of wall paintings with magpies in a landscape of trees, and a number of Chinese phoenixes painted in gold on some panels of the ceramic revetment of the tomb. The motifs are the only occurrence that we have, so that their meaning is unclear. However, it is possible that some reference is intended to the well-known hadiththat the souls of those who die for the faith are as green birds in Paradise, which does occur as a funerary inscription elsewhere.19 Equally, as Denike is inclined to suggest, there may be a revival of the traditional Iranian customs of tomb-painting which Mas'fidi describes, but which do not seem to have persisted into the Islamic High Middle Ages. The actual reason does not concern us here; all that is otherwise rather important is that the tomb in the Shdh-i Zinda gives some factual content to the highly-stated claims. Qa.di's The above cases make no pretence to being a complete list of all the references to religious painting in the pre-Safawid period, and it is certainly possible that further researches would bring to light evidence at least as important. It is not clear, however, that they would go any further towards showing exactly what the reasons for the development of religious painting were, and it is perhaps useless to expect a single answer to this question. The most that we can hope to do is to identify the attitude which brought the genre into being and, so far as possible, trace its origins. The third and last part of this article will be devoted to this. The question arises first as to where the attitude which lies behind the development of religious painting first arose. As we have seen, the paintings themselves seem to be restricted to Safawid Persia, and the fact that at the present day paintings appear to be canonized in the Shi'i shrines of Iraq is almost certainly ascribable to the influence of Persian Shi'ism. Although Ibn Battata's account may lead us to suppose that in some Sunni communities Muslim religious painting was tolerated, we should still, I think, have to postulate a Persian colony in Kerch to which the " picture of the Prophet 'Ali " belonged. But at the same time the Shi'i 'ulamd' have not been markedly less hostile to religious painting than the Sunni ones, and it will not do to assert without qualification that religious painting is a necessary characteristic of any Shi'i society. It is true that Shi'ism has been a missionary religion far longer and to a much greater degree than Sunnism. Because of the deep suspicion with which Shi'i Muslims, and particularly the Safawids, were viewed by orthodox Muslims, they were obliged constantly to fight against extermination, and the propagation of the faith depended in large measure upon a highly-trained and informed dlite which could make new converts as orthodoxy killed off the old. Since the propagation of the faith has always occupied an important place in the minds of the 'ulamd' (although they have often depended upon
17
18

Minorsky, op. cit., p. i8o and note 631. Ornament Srednei Azii (Moscow 1939), p. 202 Denike, Arkhitekturny and fig. 97; Stchoukine, La Peinture Manuscrits des Timourides (Paris 1954),P-3. They agree that the blue objects in the landscape, which had previously been identified as dragons, are in fact streams.

19 For example, in the mausoleum of the Vizir, Fakhr al-Din 'Ali in Konya (c. 1285). That the inscription should be used in his case points, I think, to the reburial of his two sons there, who had been killed in 1277 in the Jimri episode.

132

JOURNAL

OF PERSIAN

STUDIES

personal contact to make proselytes, for example, the Shi'i babasof the Janissary regiments in sixteenthcentury Turkey), there has been constant awareness of the problem of reaching the non-literate elements of the population. Hence the emphasis has been largely upon teaching by example rather than by precept. This essential feature of Persian Shi'ism stems, as I have suggested above, from the fact that the perfection of 'All has always been a central pre-occupation of Shi'i theology, which gives at least the starting point for the teaching of religion by urging the faithful to follow his example. This central characteristic is well shown by the early appearance in an Iranian context of lives of, and devotion to, Stifi saints, intended not just as edification for the particular community to which the saint happened to belong, but for the faithful at large. An even more important phenomenon is the the Muharram passion plays which commemorate the sufferings of Hasan and growth of the ta'zzya,20 in the Shi'i world as a whole and not just in Persia. Once again, they are presented as an IHusayn, individual example for the faithful to follow, even unto death, in their pursuit of virtue. This is not, of course, to assert that religion by example rather than precept is in no way a feature of Sunnism, but only that in Shi'ism its role is much more central and irreplaceable. The contrast between traditional Sunni views ofjihdd, for example, which is a general duty incumbent upon all Muslims but which is very often seen as having no particular reference to individual cases, is very striking.21 The passion plays have as their purpose not just the re-enactment of certain historical events but a moral and didactic significance as well, the awakening of the mass of the faithful to the urgency of their duty as mujtahidin. The ta'zfya exhibits, in theatrical terms, exactly the requirement which must be stipulated as necessary for the emergence of religious painting. It has often been asserted that these and related aspects of Shi'i piety are due to the influence of Christianity. In the case of religious art this is in one sense undeniable, since many modern paintings of 'Ali are taken almost directly from Western paintings of Christ, and Arnold has shown that as early as the Edinburgh Jdmi' al-tawdrikh,we find the il-Khanid painters using Byzantine originals rather than following the details of the Qur'anic narrative, e.g. in the case of the Annunciation.22 But one must distinguish very carefully between the appropriation of a Christian iconographic tradition and contamination by certain Christian dogmas, and in this latter case, there is not, I think, any evidence that Shi'i Muslims at any period were less than sincere in their overt hostility to Christianity. Hence it does not follow that the precise religious beliefs, of which the cult of images in Safawid Persia was an expression, have any more than faint analogies to the religious dogmas which in Christianity justified and encouraged the use of images. Paintings of Muhammad (we cannot call them portraits for there was, as far as we know, no tradition upon which the painters could draw) appear first in the il-Khdnid manuscripts of the early fourteenth century, sc. the 1307 Jdmi' al-tawdrikhand the Athdr al-biqaya of the same year.23 However, so far from being a symptom of the development of a religious art, these occurrences are part of the tremendous concentration upon the history of Persia which makes itself felt as the il-Khdnids became settled, and which, after the conversion of Ghdzdn Khan in 1295, became a concentration upon the history of Islam as well. The Mongols, by identifying themselves with these two traditions, aimed to achieve a sort of literary respectability. This desire and its consummation are very clear if we compare the Tarikh-iJahdn-gushdy,written by Juwayni (who died in 1283) and Rashid al-Din's Jdmi' al-tawdrikh, written only twenty years later. Juwayni, a Muslim and a Persian, attempts to reconcile his own beliefs with the necessity of not criticizing the r6gime by showing the Mongol invasion of Persia and the Near
20

21

Y. Godard suggests (loc. cit.) that the ta'zfya can be no earlier than the late seventeenth century, since observant travellers like Chardin and Olearius fail to mention it. We have already seen how inconclusive " observant travellers " may be; and in fact, Tavernier writing in 1667 describes something very comparable to a ta'z4ya. Even so, it is interesting that these passion plays appear to gain in popularity more or less at the same time that the religious paintings under discussion first appear. The Seljuqs and the later Ottomans are, for example, fond of titulature associated with the jilhdd (e.g. qdtil al-kafara wa'land of calling themselves ghdzfs; but it is not always mushrikin)

easy to see their lives, or their policies, as the pursuit of martyrdom in the cause of the Faith. 22 Arnold, Painting in Islam (Oxford 1928), p. Ioo and pl. XXIV. 23 The late Bishr Fdris argued that the frontispiece of a volume of the Kitdb al-Aghdniin the National Library in Cairo represents Muhammad engaged in a mubdhala, ritual exchange of or curses, with the Christians of Najrin. However, this theory has been conclusively refuted by Rice (" The Aghani Miniatures and Religious Painting in Islam ", BurlingtonMagazine [April I953]), who points out that the personage shown, if any historical character is intended, must be the Atabeg Lu'lu' of Mosul.

THE

GENESIS

OF SAFAWID

RELIGIOUS

PAINTING

133

East and its accompanying excessesas the punishment of God upon Islam for the deficienciesof Muslims and their failure to observe the precepts of the Qur'an. Twenty years later the position has absolutely changed. Rashid al-Din is concerned to emphasize the historicalcontinuity of the period preceding and following the Mongol invasion. The dynasty has become now the newest of the Islamic dynasties, fully integrated also into the scheme of Persian historiography, and the inconvenient fact that Hiulegiiwas responsiblefor the destruction of the Caliphate is nicely glossed over. and This important change of viewpoint reflectsitself in the illustrationsof both the Jdmi' al-tawdrikh which contain illustrationsof a range of historical subjects from the Fall onwards, the Athdral-biqaya, including both Qur'anic and non-Qur'anic material, as well as a selection of events subsequent to the Hijra. The illustrationsof Muhammad and his Companions have importance, however, as history and not as didactic illustration of a theological or moral lesson. Thus, although Muhammad is presented as the last and greatest of the Prophets,he is not distinguishedpictorially in any way from other important and the religious figures,like Buddha, or even from the rulersof the Jdhilfyya. Thus he has no nimbus,24 of enlarging the size of the device customary in Persian painting of the fourteenth-fifteenth centuries figuresin a painting in accordance with their relative importance, if applied in Il-Khdnid illustration to Muhammad, is also applied to the most important figuresin many other compositionsin which he is not represented. Nevertheless, these early paintings of Muhammad and the other Prophets are of importance, partly because of the enormous impetus which Il-Khdnid patronage gave to the diffusionof illustratedmanuscripts (which in the later fourteenth century gradually turned from solid history to romance), and partly because of the manner of representingMuhammad and other Prophets, without veiling, nimbus or even enlargement, a manner which persisted even when the intention was to inculcate reverence in the beholder. The most important developments, however, seem to come in the Timurid period in the remains of the Central Asian empire of the Chaghatay Turks. This may well have been due to the Timurid invasion itself which, though scarcely less destructivethan that of Chinggis Khan, had entirely different consequencesin the cultural field. Timur was not only a Muslim to begin with, but a respectablePersian scholar, so that the long period of adaptation which the il-Khinids underwent in Persia became unnecessary, and his immediate descendants were remarkable for their erudition, refinement and civilization. So when their taste turned towards China,25 this was a caprice characteristic of a wellestablished dynasty and not the effortsof parvenus to acquire a past for themselves. The admiration of 'Abd-al-Razzdq Samarqandi for the vivacity and beauty of Chinese Buddhist sculpture and painting, to which we have already alluded, is the mark of a man thoroughly sure of his cultural and a fortiori religious position.26 More pertinently, the cult of saints itself owes much to Timur and his descendants, not only in the but as great increasein the foundation of khdnqdhs memorial constructions,27 also in a spate of devotional but narrative and religious literature: not the mystical, highly obscure, work of much Persian some of which were also allegoricalin intention. The taste for religiousnarrativeand biographyis S.ifism, works, not, of course, an invention of the Timurid period, sira-literaturehaving been cultivated in many areas of Islam from its earliest centuries. However, in the Timurid period we may note two important developments,firstlyin the unusual quantity and variety of works translated or re-written, and secondly
24

The earliest occurrence of the typical flaming nimbus of 26 Quatremeire, Noticeset Extraits XIV (1843). The impossibility of a Persian expressing such sentiments in the thirteenth Muslim painting seems to be on the Angel Gabriel in the Annunciation miniature of the Edinburgh al-Athdr al-bdqiya century is not just a matter of religious prejudice, but is see Stchoukine, La Peinture Iraniennesous les derniers part of the deep suspicion which the still uniranicized (I307), et 'Abbasides les Il-Khans (Bruges 1936), pl. LXIV. But there is Mongols, with their tolerance of most religions and favouring of Buddhism, awakened among the 'ulamd', whose high apparently no case of a Prophetbeing represented with such a offices had previously depended upon their being orthodox nimbus before the Topkapi Saray Mi 'raj-ndma (Album, Muslims. Hazine 2154), possibly to be dated to the first half of the 27 See Bartold, " O Pogrebenii Timura " in Sochinenmya (2) II fourteenth century. (Moscow 1964), P. 433, for an illuminating discussion of 25Amongst the copious literature on the subject one should single in the spread of khdnqdhs Transoxania as memorial construcout Grube's discussion of some Timurid paintings in the Hazine tions. Albums of the Topkapi Saray (Kunstdes Orients V/i [ 1968]) and other collections, which clearly copy Chinese prototypes.
12

134

JOURNAL OF PERSIAN STUDIES

the large proportion of them which were illustrated.28 Some of these works seem to be direct translations or recensions of earlier works, for example, Farid al-Din Mantiq al-tayr and his Tadhkirat al-awliyd', and here the important fact is that they are now illustrated, gradually coming to be designed 'Att.r's in a way which might enhance the didactic importance of the text. Others are works of pure devotion, like the Khawardn-ndma,written in 1426 by Muhammad ibn IHusayn and containing the exploits of 'Ali, decorated in a manuscript of c. 1480 with 155 miniatures which are meant to bring out his closeness to God and his pre-eminence among human beings; Nafahdt al-uns (I480), a series of biographies Jrmi's of Sffi saints, and the same author's Shawdhidal-nubuwwa, a biographical account of how Muhammad revealed himself to be a Prophet, both in his lifetime and through his Companions and followers down to the early Sifis; and HIusayn al-Kaishifi's Rawdat al-shuhadd',a book of Shi'i martyrs. These are all works of the type which might properly be called " devotional ". Further examples could be given, but the most interesting development is not so much the content of these new texts but the approach. In this respect the Uighur text of the Tadhkiratal-awliyd', bound in with the Mi'rdj-ndmawritten at Herat in 1436 by Malik Bakhshi, which is now in the Bibliotheque Nationale, is of particular importance.29 The Uighur text of the Tadhkirat al-awliyd is a fairly exact translation of the earlier work of the same name by Farid al-Din The translator's preface, however, is strikingly new. introduction is concerned to say that the lives of the saints are useful primarily to supplement the 'Att.r. 'Att.r's teachings of hadith, since they are the heirs of the Prophets and in this way have religious authority: in other words, they provide a sort of ijtihdd for the faithful. Malik Bakhshi is much more explicit. He says that to teach religion is a work of merit. To understand the Qur'an is a work of much labour, but since the lives of the saints conform exactly to its precepts, reading their lives is a short cut to sanctity. Their sayings show the faithful which are the true dervishes (sc. the proper examples to imitate), they strengthen resolution, making men of women and lions of men, bringing them to the stage where they The above brief paraphrase is sufficient to show the very remarkable similarity of the sentiments put forward as a justification for reverence for the saints, to the principles held to justify the use of painting first in the Byzantine and then in the Western Churches, sc. painting as a short cut to sanctity, inspiring men with devotion to duty and the love of God and leading even the illiterate to an understanding of the Scriptures. It does not matter, of course, whether Malik Bakhshi's preface to the Tadkhiratal-awliyd' is the first expression of these views in Islam; but it is noteworthy that the type of literature which these views encourage, namely devotional or didactic, accompanies the first appearance of painting which seems intended to be didactic and which is, anyway, intended to inspire reverence in the beholder. There is, it must be admitted, little reason for bringing the analogy with the birth of Christian art even closer by saying that these early Muslim didactic paintings were intended to bring the nature of faith and morals nearer to those who could not read, since the texts with illustrations could hardly have come into the hands of the illiterate at all. However, I think it may fairly be said that these developments in Timurid manuscript painting, in works illustrated in or intended for a Sultan's library, are an important prior condition for the appearance of those paintings in the Safawid shrines of Persia which were intended to make plain the doctrines of Shi'ism to all and to inspire the faithful with enthusiasm and reverence. As I have said above, it is difficult to draw a dividing line. The rather jejune tradition of inserting standard pictures of men to decorate lives of the Prophets certainly persisted into seventeenth-century Istanbul, and there are many other cases where it is clear that the illustrator had no knowledge of the text he was dealing with. However, certain manuscripts, notably the famous Topkapi Kitdb siyar
28

are the 61itewhose soulsarethe veryessence suffering God'scause.30 in of

It is, of course, impossible to draw a rigid distinction. Mir 'All Shir Nevd'i, for example, who also wrote a Tadhkirat al-awliyd', the text of which was to serve as the basis for Ottoman versions like the Tdrikh-i Alti Parmdq, also wrote works of purely secular biography like the Tadhkirat al-shu'ard', lives of the poets from whom he had learned. It is probable that even in the sixteenth century many biographical works of Sfifi mystics were not primarily devotional in intention, but the general tendency is clear. Phenomena which should also be taken into account are the rediscovery of the bogus grave of 'All near Balkh, which was refurbished by Ijusayn Bayqard in I480-8I

(Bartold, " Mir 'Ali Shir i Politicheskaya Zhizn' " in SochineniyaII (2) [Moscow 1964), ch. IV, p. 235) and which became the modern shrine of Mazar-i Sherif. This led to the " discovery " of further bogus tombs at Herdt and other places, which eventually were denounced. It illustrates the highlycharged atmosphere of religious enthusiasm which characterizes the later Timurid period. 29 Bibliotheque Nationale, MS Suppl6ment Turc 190. Translated by Pavet de Courteille, Le Mimorial des Saints (Paris 1889). 30 Ibid., p. 2 of translation; pp. 3-4 of text.

THE

GENESIS

OF SAFAWID

RELIGIOUS

PAINTING

135

al-nabi,31illustrated in Istanbul in 1595-96, a series of texts of the Mi'rdj-nama, of which the 1436 Herat mentioned above, all show manuscript is the most important for our purposes, and the Khawardn-ndma this approach very clearly. The Herat Mi'rdj-namais worthy of closer attention. The subject, the Ascension of Muhammad, was of interest to Muslim theologians long before the fifteenth century, since we have a work by Abu'l-Qasim al-Qushayri (died 074 A.D.) collating and commenting upon the various versions of the Mi 'rdjknown to him. However, the textual question is for once less important than the illustrations of the Herat manuscript of 1436, since in the Library of the Topkapi Saray there is a series of early Ascension miniatures cut from the text which they illustrated and pasted at random into an album (Hazine 2154), and with an attribution written in a later hand (possibly that of DUst Muhammad, whose treatise on Persian painters antedates the Qadi Alhmad's by some fifty years) to the legendary Persian painter, Ahmad Mfisa.32 Since it is to Duist Muhammad's treatise that we are indebted for the knowledge that Ahmad Mfisa actually it wrote a Mi'rMj-ndma, would appear that the late attribution-signature is not independent evidence for the authorship of the miniatures. The paintings from the album, Hazine 2154, show a certain stiffness, and although they can fairly easily be related to later texts known to us, they do not suggest that the illustrator felt it important to be faithful to the text. As Ettinghausen has remarked, comparison of the Topkapi illustrations with the Herit manuscript suggests that they must have been of two different texts; but it is scarcely surprising to discover that the iconographic tradition is of greater importance, since much the same point about fidelity to the text might be made in the case of the Herat manuscript. The most important feature of the Herat Mi'rdj-ndma is that each of the fifty-eight miniatures which decorate it is accompanied by an explanation in Anatolian Turkish (one written in a sixteenth century naskhi, and one in a hand nearer to nasta'liq which may be later), in Arabic, and sometimes in Persian as well (in shikasta). This last hand suggests that the commentaries were added in Istanbul, since shikastadoes not appear to have been used in fifteenth-century Transoxania. The additions are not surprising: the text in Uighur cannot have been accessible to many at the Ottoman court, and we can readily understand that a commentary should have been thought necessary. What is surprising is that it should be the illustrations which bear the comments, not the text, even though the content of the additions is, as already remarked, closer to the text than what is depicted.33 The obvious conclusion is
31The text is a translation of Wdqidi by MustafA b. Yfisuf Erzurfimi written in 1388 and re-copied in Istanbul in the Library of Murad III under the supervision in 1595-96al-.Darir of a naqqdsh-bashi, head painter. Stchoukine (La Peinture or I' illustres(i) : deSulayman OsmanII d'aprisles Manuscrits Turque out that in the first [Paris 1966], p. 85) points (152o-1622) of the six volumes Muhammad is only once represented with his face veiled, which suggests that in later sixteenth-century Turkey there was still a certain latitude about this. The paintings of the Kitdbsiyaral-nabfare of especial importance in that they seem to derive from a careful reading of the text. Stchoukine (ibid.) has remarked upon their " archaism ", but if we compare them with the Mi'rdj-ndma illustrations from the Album, Hazine 2154, which have been attributed to Ahmad Mfisd, we shall find that this is not an accurate description. The paintings certainly show a very striking feeling for colour, which is doubtless an intentional way of awakening awe at the events illustrated. But although in this respect they appear to be in the spirit of certain Italian primitives, there is no reason to suppose that the painters in the Library of Murad III were looking back to what was, even then, a remote past. 32 Ettinghausen, Persian AscensionMiniatures of the 14th Century (Accademia Nazionale dei Lincei, Rome 1957), PP- 360-83. Two Western texts, both thirteenth century, have been delle published by Cerulli in II Libro della Scala e la Questione Fonti Arabo-Spagnole della Divina Commedia (Vatican 1949), and the Uighur text of Malik Bakhshi by Pavet de Courteille (Paris I882). The texts are, of course, quite different. In general, the Uighur text is briefer and more confused, conflating various different episodes which appear in the Western texts, and considerably altering the order. The one interesting feature, however, which Western and Eastern texts have in common is that they ascribe to God human characteristics, a voice, position (being seated on a throne), though not movement and activity. The movement is not new, of course, and under the Seljuqs accusations of tashbihor anthropomorphism had been rife, but it is particularly interesting in the Turkish context. The paintings themselves do not get any further to God than the outermost of the seven veils before the Throne (f. 45 verso of the Uighur manuscript, published in Blochet, Les Enluminuresdes AManuscrits Turcs, Arabes et Persans de la BibliothiqueNationale [Paris 1926], pl. XXXVII), though the motive for this may well have been reverence, rather than the fear of falling into idolatry which is customarily supposed to have afflicted the 'ulama' with regard to painting. 33 It has not been possible for me to compare all the commentaries with the Uighur text. One example should be sufficient, however, to show that it was the text, rather than the illustrations, of the Herdt manuscript which was attached as commentary. The Uighur text describing an angel (Pavet de Courteille, text p. 9, translation p. 6) with seventy heads, seventy tongues in each head, each reciting seventy different forms of tasbih, is exactly followed by the commentary in Anatolian Turkish (verher balindayetmij dili olup ve hem bir dili yetmis diirliitesbihokurdu...) (compare Esin, " An Angel in the Miscellany, Album H.2152, of Topkapi " in Beitrdgezur Kunst desAsiens[Istanbul 19631, p. 279). The Western texts translated by Cerulli (op. cit., pp. 61-3) elaborate somewhat on the possibilities of multiplication, replacing seventy by seventy thousand.The illustration of the Herit manuscript (f. 17 recto) (Blochet, Les Enluminures,pl. XXXVI) has, rather tamely, only thirty-three heads.

136

JOURNAL

OF PERSIAN

STUDIES

that by the time the manuscript came to be appreciated at the Ottoman court, the paintings it contained were viewed as interesting or important in themselves and not just, as so often before, as decoration for a text. One of the prior conditions for the existence of religious art is, therefore, fulfilled, namely that the art should stand first, if not independently, so that it can teach or moralize without benefit of text, or at least with only the sort of minimal titling which we find in the Ottoman additions to the Mi'rdj-ndma.Also, this particular case is only one among a whole class of manuscripts in sixteenth-seventeenth-centuryIstanbul. We still know very little about the origins of this genre of painting (which, one may suggest, is properly called religious) in Turkey, where, in any case, equivalents of modern Shi'i " icons " and of the Safawid wall paintings do not appear ever to have existed. Its development may have been affected, at least in its iconography and possibly in its literary source material as well, by works of medieval and Renaissance piety imported from Europe by the extensive foreign colonies resident in Istanbul; a proportion of these must almost certainly have come into general circulation outside the capital.34 But it is more probable that its origins are to be found in the later Timurid kingdoms of Transoxania. The transmissionof manuscripts was enormously facilitated by the sack of Herat in 1507 by the Safawid Shah IsmT'il,who carried off the Royal Library, together with Bihzdd and others of its staff, to Tabriz; and by the subsequent sack of Tabriz itself by the Ottoman Selim I Yavuz, who likewise carried off the Library of the Safawids, and with the painters from it laid the foundations of the Ottoman painting of reached the later sixteenth century. This is almost certainly the route by which the Herat Mi 'rdj-ndma Istanbul, and demonstrates the extent to which the Ottomans had come to be the direct heirs of the Central Asian later Timurid schools. As for the contents of works acquired in this way, I do not think that the Ottomans cared much a whether works illustrated were Shi'i or Sunni in sympathy. The fact that Fudiili's Hadiqatal-su'add', translation of IHusaynal-Kdshifi's Rawdatal-shuhadd', strongly Shi'i Book of Martyrs, was illustrated a in the Library of Murad III in 1596, more or less at the same time as the orthodox Sunni Kitdbsiyar al-nabi,certainly does not show that Murad himself was at all heterodox. Although we know that the or Janissary regimentscame under strong Shi'i influence from their Bektashibdbds, imams, the extent of in Anatolia is a very vexed question, and it has been so far impossible to determine Shi'i penetration how far the Shi'i propagandizing in sixteenth-century Anatolia, for which there is concrete evidence, was actually literary. There obviously were pockets of Shi'i influence in Istanbul, and doubtless within the Ottoman court itself. But the transmission of texts and traditions with their illustrations seems adequately explained by the successivesacking of the two great Persianlibrariesrather than by recourse to our tenuous knowledge of the politico-religious history of Anatolia. The Ottoman taste for didactic, moralizing illustration is thereforeof relevance to our subject firstly because it was a development of tendencies already present in the literature and in certain manuscript paintings of the later Timurid kingdoms; and secondly because after its adoption by the Ottomans, it developed in Istanbul in ways comparableto painting in Persia,in particularin its concentrationupon the devotional aspectswhich are already significantin the Kitdbsiyaral-nabiof 1595-96. It is no accident
34The effects of the importation of Italian works make themselves felt at a much earlier period, though we have by no means so much documentation on the results of the Genoese and Venetian penetration of the Middle East in the il-Khanid period. Leaving aside the extraordinarily Venetian St. George and the Dragon, which occurs, apparently, quite by the way in the Pierpont Morgan Mandfi' al-hayawdn (1295) and which, Mr. Basil Gray kindly informs me, is probably a modern addition, there are a number of paintings which demonstrate that Italian painting was known in Il-Khdnid Persia. One of the most striking of these is a miniature in the fourteenthcentury Tabriz style from a series of albums acquired in Istanbul by the Prussian Ambassador, Diez (1784-91) (Ipsiroglu, Saray Alben [Wiesbaden I964], pl. XIX), showing an apparition of the Virgin and Child to two kneeling figures, the Virgin being surrounded by the flaming mandorla later used in Islamic painting for the Prophets of Islam. The setting is of pointed Gothic arches, with the stone coursing carefully drawn in, and with a gallery above containing two (angel) figures playing an organ. We have here a straightforward transference of elements from Italian primitive painting or book-illustration into a setting, doubtless Christian, characteristic of north-west Persia. Further evidence is suggested by the report of the Dominican, John of Montecorvino (edited by Yule, Cathayand the Way ThitherIII [London 19141, p. 53), when he writes that he has had six pictures made (sc. by local craftsmen) illustrating the Old and New Testaments for the instruction of the ignorant, with explanations carved in Latin, Persian and Tarsic (i.e. Uighur), that all may understand. The analogy to the Ottoman addition to the paintings of the Mi'raj-nama,is, of course, most striking. John of Montecorvino's pictures, if executed at Peking as his account suggests, must nevertheless have been intended for a Mongol-Persian milieu; otherwise the inscriptions would have been in other languages.

THE

GENESIS

OF SAFAWID

RELIGIOUS

PAINTING

137

that the flaming nimbus of the Prophets, given to them in the first place out of piety, becomes more splendid and sometimes even replaces the human figure altogether. Faces are more and more frequently veiled, and clothing and gesture become grander as the movement advances, in Ottoman book painting concurrently with the wall paintings of Safawid Persia.35 Painting, as we have said, had tended to develop the sense of its didactic possibilitiessome considerable time after the increasein moralizing and devotional literatureof the fourteenth-sixteenthcenturies. This delay is notjust a matter of the greater conservatismor lack of imagination shown by the painters of the time in Persia and Transoxania. Rather, lack of originality, rather than its presence, was the quality most admired in painters, and the greatestpraise that writerson painting like the and Duist QadiAhmad can give,36 even to is not that he is sensitiveor observantor novel, but that he copies Muhammad Bihz.d, well, and that in his copying he merelyhappened to excel everyone else. This general prejudice,together with a certain taste, perhaps imitating the Chinese, among the Timurids for genre-painting rather than narrative,explains why it is only towardsthe end of the sixteenthcentury in both Istanbul and Persiathat we see the full development of original narrativepainting, and why the taste for religious wall paintings makes itself apparent at more or less the identical period (sc. the late seventeenth century, if we accept only the Imamzddas, but some hundred years earlier if we regard Don Juan's account as relevant). A manuscript of the Mantiqal-tayr,recently acquired by the Metropolitan Museum,37will perhaps serve as a final example. The copy in question was written in Herdt at the court of Husayn Bayqardin 1483. At a slightly later date it was decorated with genre scenes unrelated to the events describedin the text in the style of Bihzdd; but before the paintings in the manuscript could all be completed it seems to have been carriedoff by Shah Ismd'il in the sack of Hertt in 1507. In 16o09it came to the attention of Shah 'Abbas, who was looking for a suitable present to make to the family shrine at Ardebil. The incorporated into the manuscript states that he had the margins re-drawn and gilded and the waqfiyya four pages which still remained blank filled with fine miniatures, which, it is scarcely surprising to discover, are all narrative or explanatory, in contrast to the fifteenth-centurypaintings. One of them, a portrait group of the birds themselves (the si murgh,who wander in search of the Simurgh), so is accurately painted that it is possible to give them their Linnaean names; one is of what is apparently the Martyrdom of St. John the Baptist; and the remaining two are of incidents from the well-known tale of the Shaykh of San'in, who denied his faith for love of a cruel Christian maiden, but who eventually returns to the light after she has forced him to keep swine and do various other improper things. In the light of earlier arguments,this change in approach (which could, after all, be merely symptomatic of a general, non-religious, taste for narrative art rather than decorative genre scenes) is worthy of remark, especially since we must bear in mind the possibility that Shah 'Abbas was influenced in commissioningthe new miniatures by the consideration that the manuscript was to be presented to the
Ardebil Shrine. The date of the waqfiyya, 16o9, thus represents a terminus ante quem,since the transition

clearly comes well before the great Ottoman religious miniaturesof the late sixteenth century. In fact a translationof the Mantiqal-tayr, by Mir 'Ali Shir Neva'i, copied at Bukharain 1558,38is also decorated

35It would appear that when in fifteenth-century painting 36 Dfist Muhammad, it is true, credits Ahmad Mfisd with having Muhammad is shown veiled, as in the possibly fourteenth-century drawn the veil from the face of painting and inventing " the Mi'raj-ndmaillustrations from the Topkapi Saray attributed to style of painting current at the present day " (sc. c. 1550) Ahmad Mfisk, this is the result of later overpainting. In the next (Wilkinson and Gray, Persian Miniature Painting [London/ hundred years, however, there is a great increase in the use of Oxford 1933], pp. I83-4), but this is the only occasion known the veil, only two clear cases being known where Muhammad is to me on which a painter is given credit for his originality. Nor shown unveiled: a line drawing in the Topkapi Saray (illustrated does the judgement, it might be said, reflect much credit upon in Wilkinson and Gray, Persian Miniature Painting [London/ Dfist Muhammad's powers of discrimination. Oxford 19331, pl. XCV), and in a manuscript of Nizdmi (illus" The Seventeenth Century Miniatures in the Lantrated in Guest, ShirazPaintingin thei 6thCentury, Freer Gallery of 37 Grube, Museumof" Art guage of the Birds ", Bulletin of the Metropolitan Art Oriental Studies, I [Washington 1949], no. FGA o8.278), (May 1967). where the Mi 'raj serves as the frontispiece to the Iskandar-ncma. Mrs. Guest, to whose work I am indebted for these examples, 38 Bibliotheque Nationale, MS Suppl6ment Turc 996. Illustrated makes the point, with which I fully agree, that the principal in Blochet (Les Enluminures, pls. LV-LVI, commentary, reason for the veiling of the face was pietism. In the later period pp. xo5-o7). The illustrations in this particular manuscript the point applies to historical works as well, since, for example in a neglect the main " plot " of the birds' search and confine themselves to the moral anecdotes; and this again suggests a manuscript of Mirkhwind's Rawdatal-Safa', dated 595 (Arnold, Paintingin Islam,pl. XXI) Muhammad is also shown with veiled change of emphasis, from too literal allegory to more straightface. forward moral teaching.

138

JOURNAL

OF PERSIAN

STUDIES

with illustrationsof the moralizing and didactic sort, again from the life of the Shaykh of San'an, which suggeststhat the transitionwhich the I6o9 additions of Shah 'Abbdsindicate can be pushed back to the middle of the sixteenth century. It may, of course, be questioned whether this clearly identifiable phenomenon of the development of didactic, moralizing and even devotional painting in books,which we can attributeto the predominantly Sunni society of Transoxania of the fifteenth-sixteenth centuries, has any necessaryconnection with the development in Safawid Persia of Shi'i wall painting. For in contrast to paintings decorating the walls of a shrine, book-painting-even given the tremendous impetus to the diffusion of manuscriptsby the dynasties which followed the two Mongol invasions-can never have been a public art. It is, indeed, only as a result of the creation of Royal Librariesin the fifteenth century that we have any real knowledge of the painting of these, or earlier, periods.39Exactly the same objection might be brought against the present hypothesis as that which was brought against the view that Shi'i religious painting might have developed as a result of Christian influence: that similar phenomena may be the expression of radically opposed religious beliefs. However, I think that the evidence sketched above is too strong for this particular objection to hold. In a society where even the Shi'i 'ulamd' had expressedthemselvesso violently against the representation of living things (and there is no reason to disbelieve Chardin's testimony on this point40)we must look not only for changes in taste but for changes in the approach to the teaching of religion. And to the traditional emphasis which Shi'ism has placed upon the good life as a matter of imitating the martyrs and other saints, the development in the later Timurid courts of a literature, and then painting, concerned with the teaching of morals and religion by example, comes as the ideal complement. In these special circumstances, the temporal discrepancy between the developments in bookpainting and the firstdated religiouswall paintings known to us (at the Imdmztda Shah Zayd in Isfahan, dated to 1685-86 by a restoration inscription) is not too serious a problem. The literary testimonials advanced above allow us to posit the existence of important religious paintings, of the type of the funerary portrait of Don Juan's father, at least a century previously. This makes it therefore more plausible to suggest that the Safawids themselves had some responsibility; with their expansionist campaigns of the early sixteenth century and their insistence upon the exalted religious status of their the rulers,41 didactic approach shown in the wall paintings was the ideal way of teaching those who could not otherwise be reached by reading or preaching. The foregoing argument has been necessarilysomewhat diffuse. I have attempted to deal with the phenomenon of Safawid religiouspainting in monumentsof the late seventeenth century not by studying the iconography of the paintings, but by endeavouring to trace the growth of certain attitudes towards the teaching of religion in the post-Mongol period in Persiaand Central Asia. These were characterized by the development of a taste for didactic and devotional works, expounding the virtues of the saints, prophets and martyrs of Islam to the faithful and encouraging them in the path of heroic virtue. The texts, though often written by Shi'i Muslims, seem to have been read by Sunnis as well (as, for example, in the Library of the Ottoman Sultan Murad III), and the precise religious sympathies of the audience do not appear to have had very much to do with the question. In l1-KhdnidPersia historical works presented Islam and Muhammad predominantly in their historical aspects, in order to underline the continuity of the Mongols with the 'Abbasid Caliphate, which had come to an end as a result of the activities of Hiilegii. After the Timurid invasion, when the conquerors had no need to assert their
relationship either to Islam or to the Iranian tradition, taste turned towards the heroic, rather than the purely historical, aspects of Islam, and the inculcation of heroism by recounting the life and works of men of pre-eminent virtue. These were often of course, but given the Siifi custom of taqfya and Si.fis,
8 With the exception of scientific works, of medicine, botany,

astronomy and engineering (for example, al-Jazari's Kitdbft with its diagrams of automata), ma'rifat al-hiyal al-handasiyya illustrated manuscripts in Islam have rarely found their way outside royal libraries. There is, of course, nothing exceptional in this. Where illustrations are not necessary for the understanding of a text, as they would be in the majority of scientific works, they indicate by their presence little more than a taste

for the greatest luxury possible. Exactly the same point can be made about Byzantine book-illustration; see Wessel's illuminating article " Buch-illustration " in Reallexikon zur Kunst, fascicule V (Stuttgart 1965). Byzantinischen 40 Loc. cit. 41 For example, the habit of Shdh Ismd'il's troops, noted by the Venetian traveller Contarini, of crying as they went into battle " Ismd'il wali Allah ".

THE

GENESIS

OF SAFAWID

RELIGIOUS

PAINTING

139

the consequent concentrationof Sfi commentatorsupon often obscure allegory, it is ratherstrikingthat so much of the Timurid and post-Timurid devotional literature should be comparatively straightforward. Painting followed in the course of this development, though more slowly, because the tradition (establishedin so much Byzantine painting) of book illustrationsas sheer decoration to increase the costliness of a work, died hard, and because innovation in much of the painting of this period came in spite of the painters, who saw their chief job to be that of copiers. This already favourablemilieu for the development of religiouspainting was aided still furtherby the Safawid conquest of Persia and the declaration of Shi'ism as the state religion. Faced on all sides by enemies whose orthodoxy gave them a pretext for making war on Persia, it became even more urgent for the Safawids to adopt new ways of encouraging their armies and the populace to remain strong in their religious beliefs; here, the requirementsof politics and religion exactly coincided. One may regard the seventeenth-centurywall paintings at Isfahdn of 'All and of the martyrdomof Hasan and IHusaynas a recognition, some little time after the event, of the potentialities of large pictures as a means of instructing the faithful. It has been important to lay some stresson the above points, for studies of the iconography of the paintings may lead us to different conclusions, particularly in regard to the sources of the subject matter. From the Zand period onwards, Persianpainting has certainly been highly indebted to Italian sources, particularly to works of piety popularized doubtless by the Franciscan missions. But, as I have emphasized, this does not license us to say that the religious forms, of which these paintings are an expression, were themselves influenced by Christian doctrine. Thus although we may agree that the reasons for the development of a Shi'I religious iconography may have been much the same as those which, in the early centuriesof the Church, were used to canonize religious art, it is scarcely plausible to suggest that the Shi'i 'ulamd'could at the instance of some Christian missionary have changed their views on the advisability of religious painting. Having said this, one must note that the iconographic sources of these Safawid wall paintings are so far unstudied. The problem, which I have deliberately avoided in this article, of how it could become possible for artistsin the often violently anti-Christiansociety of Safawid Persia to reconcile themselves with borrowed Christian subjects and styles, remains unsolved.

EXCAVATIONS AT BABA JAN, 1968: Third Preliminary Report By Clare Goff

The third season of excavations at Bibd Jan began on July Ist 1968, and continued for ten weeks. As before, the work was sponsored by the Institute of Archaeology of the University of London, and directed by the writer. The team consistedof Mr. Stephen Mitchell and Mr. Robert Hurst (seniorfield assistants), Miss Judy Pullar, Miss Noel Siver and Mr. Jeffrey Frye (field assistants), and Mr. Julian Baldick who took charge of the reconstruction and registration of the tiles from the " Painted Chamber ". Mr. Hans Smits was with us for the first part of the season as photographer, and several people were able to visit us for shorterperiods, notably Mr. Nigel Palmer, Mr. Peter Kramer, who made a second visit with the Institute Land-Rover to help transportthe tiles back to Tehran, and Miss Kora Kramer. I am particularly grateful to Miss Pullar and Mr. Baldick who stayed on at the end of the season to help with the drawing and packing-up, and with the somewhat eventful drive back to England. Our thanks are also due to the Iranian Archaeological Service, particularly Mr. Pourmand the Director General, and Mr. Khorramabadi, Deputy Director; our representative Miss Farkhondeh Etessam, whose tact and sympathy when dealing with the villagers went a long way towards smoothing over an unstable local situation; the Department of Education in Khorramabad, particularly Mr. Mohammed Reza Jazaeri, who as always kept open house for our team; the Sharififamily of Nilrabad and Mr. Kiumas Sultani and engineers of the Khawa road-building team for the hire of a bulldozer at the end of the season. I also feel that it is high time that we acknowledged the help of the two Nfirdbad doctors, who for three seasons have regularly treated a constant stream of minor ailments incurred by members of the team and by our workmen. Finally, I am deeply indebted to the following people and institutions for their help with equipment and funds: the British Academy; the Stein Arnold Fellowship;' the Ashmolean, Birmingham and Bolton Museums; and Mr. James Bomford,without whose most unexpected contributionofC?5ooat the end of the season, we should have been unable to complete our work. The personal expenses of the Director and of Mr. Stephen Mitchell were met by grants from the Gerald Avery Wainwright Near Eastern Fund, and those of Mr. Robert Hurst by a grant from the Margery Fund of the Institute of Archaeology. Mr. Jeffrey Frye paid his own expenses, and was also able to contribute something towards the excavations. I should also like to thank Mr. David Stronach, Director of the British Institute of Persian Studies, for allowing me to park a Land-Rover and caravan in the front courtyard of the Institute for two years, and for the loan of a Land-Rover, tents and other equipment for the excavations. THE EXCAVATIONS TheSequence A preliminary report on this season's excavations has already appeared in IranVII and should be read in conjunction with this account.2 As described in this report, the main problem before us was to finish the excavation of the " Painted Chamber " with its beautifully decorated tiles and to determine
its relationship to the large fortified building lying immediately to the west below the summit of the
I would like to take this opportunity of thanking Sir Mortimer Wheeler, Sir Max Mallowan and Professor Seton Lloyd for their encouragement over the last three seasons, and practical help in obtaining funds. 2 Goff, " Bdba Jan " in the " Survey of Excavations ", Iran VII
1

(1969), pp. I69-70. Preliminary reports on the two earlier seasons can be found in Goff, " Excavations at Baba Jan, 1967 ", Iran VII, pp. 115-3o and Goff Meade, " Lfiristan in the First Half of the First Millennium B.C.", Iran VI (1968), pp. 105-34-

141

142

JOURNAL

OF PERSIAN

STUDIES

East Mound. We also hoped to be able to define more exactly the sequence of pottery starting with the earliest " genre Lfiristan" levels when these monumental buildings were first constructed, through a series of complicated rebuildings when the " Painted Chamber " and its adjoining Courtyard appear to have been occupied by squatters up to the latest sixth century levels overlying them. By the end of the season, when we at last understood the configuration of the mound, it was clear that we could drasticallysimplifythe numerationof our levels, reducing the number of major phasesfrom five to three. The revised nomenclatureis set out in tabular form in Fig. I, below. A few words of explanation should perhaps be added.
EAST MOUND. CENTRAL MOUND GRAE5. POTTERY

PEQIOD

i?

S b
1(iii)

S 5TONE

ENCLOSURES.
GAP ? STONE /UT TABLES. 1ORkS/OPS

M F/INE SLIPPED. RED

DEQENEQATE "G-L. " "/VUSI//-IA/

WhEEL MADE &UFFI LATE/PON IT

b (Iv)
(va)

/G/-FOOO UILD/N

?ARLY If /R.ON

/ib (vb)

ii//

FOi2TQE5 PAINTED

I1

MANOR. " '/ION

STONEViLLAGE GAP
i/TCAIEN

MANOR.

LATE EX ?ENRELU21/STAIV"

Z
MIXED. IRON I
MIDDLE

C-AMBER. ? 4a TQ.C. CRzz

4b-c

/M. G/YAN /V-/// qOD/N .

5a-d
(o 6b

STONEVILLAGE GAP

B8ONZE

MIXED.

7o-b

FLOOR5ONLY,
UNEXCAVA TED.

CLUB R/MS.

BUFFSTRAW-TEMP PROTOLITERATE ,

+ NEO. C/A LCOL.

Fig. I. Sequence buildinglevelsat BdbdJdn. of

At the end of the second season it was clear that there were two major buildings on the East Mound: the " Painted Chamber " and the Fort, but only a small section of each had been excavated, and it was only after a visit to the very similar complex at Nfsh-i Jan that we began to suspect that the Fort walls might extend from the very top of the tepe to virgin soil, and that the two buildings might be contemporary, or nearly so.3 But since the sequence was still far from clear it seemed safer to stick to our original level numbers-i for the Late Iron III phase; ii for the Fort itself which lay directly under these levels on the summit of the mound; iii, iv and va for the squatting levels above and within the " Painted Chamber "; and vb for the " Painted Chamber " itself. In our new scheme the Late Iron III levels ia and ib are grouped in Phase I; all the squatting phases, which include high floors and alterations within the Fort, comprise Phase II; and the " Painted Chamber " and Fort become Phase III. Since " pottery very similar to the the lowest floors of both these buildings contain painted " genre Liristmn first three levels on the summit of the Central Mound, the Manors and village are also included in Phase III ;4 Phase IV comprises two levels below them which produced Bronze Age pottery similar to
3 The original advocate of this theory was the site-supervisor,

Miss Kay Wright, who was able to draw on her experience of a similar stratigraphical situation from Jerusalem. For references

to Nfish-i Jan, see Stronach, " Excavations at Tepe Nfish-i Jan, 1967 ", Iran VII (1969), pp. 1-20. 4 Goff Meade, Iran VI, pp. 12-14.

--

PAT UPPE.

TLEVEL

rnLALL..

II ADD

OF FO

LEVEL II RECONST FLOO 'METEE//

I M ET ZET.E5. --

LEUCTIOV.
RECONST7. UCT

" Fig. 2. East Mound: " PaintedChamber andFort.

144

JOURNAL

OF PERSIAN

STUDIES

that from Godin Tepe III;5 and Phase V, of which only a tiny area was dug, is probably Protoliterate.6 The actual building levels on the East Mound are still designated by lower case numerals, revised in accordance with our new scheme, and those on the Central Mound by their original Arabic numerals. The " Painted " Chamber The most spectacular construction on the East Mound was undoubtedly the " Painted Chamber " (Figs. 2 and 4; Pls. Ia and Ib). This was a large, irregularly-shapedhall, with its central area measurm. An extension to the west 4-8o m. long gave access through an ing approximately 10o40 12"50 elaborately-recesseddoorway (Pl. Ic) to a small side chamber. On grounds of symmetry one can

Fig. 3. Details of " Painted Chamber". (a) Blind window. (b) Fireplace and door. 5 Personal communication from T. C. Young, Jr., who generously allowed me to see all his material from Godin Tepe while I was visiting the site. See also T. Cuyler Young, Jr., Excavationsat Godin Tepe: First ProgressReport,Occasional Paper I7, Art and Archaeology, Royal Ontario Museum, figs. 17-19. 6 These levels have produced a burnished buff straw-tempered ware with club rims. Similar pottery fragments have been collected by me from mounds in south-west Lfiristin where they are associated with red and black polychrome wares otherwise found in the " couche interm6diaire " at Susa. See L. le Breton, " The Early Periods at Susa, Mesopotamian Relations ", Iraq XIX (I957), P. 99, fig. 14, and personal communication from R. H. Dyson.

" " panels; 9S " Nine Square filled panels; 9Sh " Nine Squares with hollowpanels; 9C " Nine Squares with Chequered mixed tiles direction panels; C FourChequered panels. Undifferentiated all " Standard"design. Arrowsindicate offall. Plan d field drawings Kay WrightandS. Mitchell,assistedbyJudith Pullar andJ. Baldick. by thewriter of from theoriginal

Key: X " St. Andrew's Cross ";

Fig. 4. Plan of tiles, " Painted Chamber". W " Wheel "; R " Reverse "; i6S Sixteen squares with Central Cross; 9Sf " Nine Square

146

JOURNAL

OF PERSIAN

STUDIES

perhaps postulate a similar arrangementto the east, destroyedbeyond recovery by the Level I occupation directly above it and by modern quarrying (Pl. Ib). This tentative reconstructionis supported by evidence from Ntish-i Jan, where the Central Building contains a rather similarly-shaped room, entered on its long axis, with its side walls stepping in towards the rear.7 The main entrance to the Chamber is placed off-centre towards its western end, and is directly opposite a modern-lookingfireplace set into the thicknessof the back wall (Pl. IIb).8 Further along, in the very centre of the back wall, is a second recessed doorway beyond which we have still to excavate. The way in which a low seat is set into the side of the doorway (Fig. 3b) is particularlyinterestingsince some similar arrangement was postulated for the main doorway of the Level I Manor on the Central had been completely destroyed.9 Another feature of the Mound, although the mud-bricksuperstructure room which recalls Ntish-i Jan is the elaborately-recessedblind window in the west wall,'o which probably originally had its counterpart to the east. Whereas the main chamber, recess and side chamber (room 20) were originally painted white, the west wall of the Chamber had two further coats of red-paintedplaster superimposed,and furthertraces of red plaster were found on the blind window, fireplace, recessed doorway and at the foot of the southern wall. The floor, by contrast, seems to have remained white, but was probably covered with gaily-patterned rugs in which red would again have been the dominant colour. The arrangementsfor lighting the building are obscure. If we are to judge by the Central Building at Nfish-i Jan, there would have been no windows at all,1" surely such a colourfulinteriorwas meant but to be well lit, and certainly there would have been plenty of room in the upper part of the walls for large windows overlooking the Courtyard. The Tiles The greatest problem in the " Painted Chamber " is presented by the tiles. These, as described in an earlier report,12were found scattered all over the floor of the Chamber, to which they had fallen from some higher elevation (Fig. 4 and P1. IId). Our present season's excavations brings the total number of complete or nearly complete examplesto 176 and there must have originally been at least 2oo. When we first encountered the tiles at the end of our second season, we suspected that the " Painted Chamber " was a long rectangularroom, rather similarin plan to the East and West Long Rooms of the Manors. The tiles, which seemed to slope down from the north to the south, could then be assumed to have slid down from a frieze near the top of the back wall where they could have been employed either in a purely decorative scheme, or perhaps as panels between a series of small windows forming a clerestoryoverlooking the surroundingroofs. Their enormous weight and size made it seem extremely unlikely that they had come from the ceiling where they would have been difficult to fix in position and where any loosenessof constructionmight have proved fatal to people in the room below. Unfortunately this convenient theory is no longer tenable. In the firstplace the room was far bigger than we originally anticipated, and when extending our excavations at the beginning of the season we discovered, not the back wall as imagined, but two large column bases made of stone with plaster edging. Faint traces of red paint around the sides of these bases enabled us to deduce that each had originally supported a wooden shaft 90 cm. in diameter painted red like the walls of the room. Despite the fact that these two columns were up to five metres from any wall, they were surroundedby tiles and tile fragmentswhich would seem to have fallen from the area immediately above, whatever may have been the case with the
tiles closer to the edges of the room. When we plotted the position of each tile together with the direction in which it had fallen, a second pattern emerged. The tiles had fallen roughly in three parallel rows, running across the room from north to south. The majority lay between the columns and the side walls, but there was also a considerable number in a line down the centre of the room itself. This immediately suggested that the main
7 8

Stronach, op. cit., pp. 9-ro and fig. 3. On analogy with a very similar structure excavated in the Fort (P1. IVa) there would have been a chimney leading up from the back of the fireplace within the thickness of the wall, but the " Painted Chamber " example was less well preserved.

9 Goff, IranVII, p. 120o. Further excavations in this area in 1969 indicated that this " seat " was a secondary feature. 10 Stronach, op. cit., pls. Ic and Id and p. Io. ix Stronach, op. cit., pp. 9-I I. 12 Goff, Iran VII, pp. I27-9.

EXCAVATIONS

AT

BABA

JAN,

1968

147

supportsfor the roof had run in a north-south direction between the columns and the longer walls. On analogy with modern roofs, these heavy cross members-burnt traces of which were found lying on the floor of the Chamber-would have been joined by lighter timbers, and one assumesthat in any catastrophe, such as a fire, these lighter pieces would have been the first to give way, spilling the tiles which they somehow supported into the gaps below. The direction in which the tiles have fallen lends weight to this theory since the vast majority appear to have fallen from towards the centre of the room outwards; their uppermost edges are those furthest away from the walls, and their painted faces look towards the walls themselves. Had they slid down from a high frieze or clerestory, they would surely have fallen in the opposite direction, their decorated faces would have looked towards the centre of the room, and their highest edges would have been those nearest to the walls. The next question is, how were they attached ? The tiles did not lie directly on the floor, but occurred throughout 50-70 cm. of mud-brick tumble below the lowest of the Phase II floors. Many of them still had considerable lumps of clayey brick adhering to their backs. Yet one cannot conceive of someone simply building a mud ceiling and then fastening such heavy tiles into it in the hope that they would stay there. As noted above, wood must have played a large part in the basic construction, whatever variety of coffering was superimposedupon it. It has been suggested, as a way round this problem, that the tiles fell not from the ceiling of the Chamber itself, but from the floor of the room above, which would impose no constructionalproblems. This may be the true explanation, but one would expect their faces to show some trace of wear, where people had walked over them. Moreover, there is a striking tendency for all the tiles found lying flat on the ground to face downwards, suggested that they had faced that way originally. Again, the decorative red and white scheme of the tiles fits in very well with that found in the rest of the Chamber. It seems almost certain that the enormous painted Hall was the important ceremonial room, not its upper storey, and that the tiles were an integral part of its appointments.13 As for the purpose of the room, it may have been either a throne room or a temple. Comparative material comes from IHasanlfiIV, the Achaemenid sites in Fars, Dum Surkh and Ntish-i Jan. In a recent review of the architectureof the Burned Buildings of IHasanli IV,14T. C. Young, Jr., discussed the possibility that Burned Building II could have had a religious function, only to reject it in favour of the more obvious interpretation of a throne room, which in itself was only a more elaborately constructed version of the normal, large, private house. None of the other buildings so far discovered at IHasanlfi could be religious in purpose, although there could be room for a temple in the unexcavated of the mound, and D. B. Stronach has argued most convincingly that the Zenddn-i Suleimdn portions at Pasargadae, and Ka'ba-yi Zardusht at Naqsh-i Rustam were modelled on the standard Urartian temple.15 Cyrus, it is suggested, adopted this alien form because before his reign (if we may believe Herodotus) the Persians did not build shrines, but offered sacrifice to " Zeus " on the highest peaks of the mountains.16 This view, however, did not take into account the evidence from Dum Surkh, and was written before the discovery of the Central Building at Nilsh-i Jan, which, as we have seen, appears to be closely related to the " Painted Chamber ". Dum Surkh"7-in default of the final publication of the evidencewe can perhaps discount. It was supposed to show Assyrianinfluence, and, since Kfih-i Dasht is only on the fringes of the Iron II (genre Lfiristdn)culture province of Lfiristdnin the eighth century,18it may still have been dominated by Kassites or Elamites. But the Central Building is almost certainly Median
and pre-Cyrus, and the fact that it was carefully and laboriously filled in with shale soon after it was built could suggest some form of ritual destruction.19
13

I am extremely grateful to Professor Seton Lloyd for bringing his professional architectural knowledge to bear upon the problem. We eventually decided that a conclusion was impossible, since the necessary evidence was lacking. Any other constructive suggestions as to how the tiles could have been used will be welcomed. It should be noted in passing that the evidence for a flat timber roof is considerably stronger than that for a vault. 14T. Cuyler Young, Jr., " Thoughts on the Architecture of

Uasanlfi IV ", IranicaAntiquaVI (1966), p. 59" Urartian and Achaemenid Tower Temples ", 15 Stronach, XXVI (1967), pp. 278-88. JNES 16 i, Ibid., p. 286. Herodotus p. 131. 17 E. F. Schmidt, " The Second Holmes Expedition to LfiristAn ", BAIIAA V (3) (1938), pp. 205-16. wares and their relation18 For the distribution of"genre Lfiristan" ship to the Iranian invasions, cf. Goff Meade, op. cit., p. 130 f. 19 Stronach, Iran VII, pp. Io-II.

148
ROOM 3

JOURNAL OF PERSIAN STUDIES

IO4m. ROOM 4

,0

.. . ....."... .
/2/m. FINAL N /Olin" D"115 .... UCTI. EL A WLL.

103 M.

/I
II

WAL

102<m
"

~
-A51 C1A/COAL
BRIGT

lolm.

~-~~-

A:54 4U 5U SCCE531VE C5~T-REPLA5TERING5

Ch

/0/ na1 MONIVV 5TRATA

Dt57AUCT/OA/

100 m
/4/1CRIEH 7

loomOO

FL/aO

/OWN

/// VI/Il N(I/NT PLASTFP//

qqm

Ch
98qsm

-?

EM1

TR

q
N f I9

UN,(')FLOOO

X C A V A T E Z

Fig. 5. SectionA-A through Key: Pl Plaster;

It is also perhaps significant that in the earlier stages of Phase II at Bdbd Jan, when, judging by the pottery, relations with the area around Malayer must have been extremely close, the " Painted Chamber " was remodelled in a way that reduced the roof span, and made the ground plan even more like that of Nfish-i Jdn20-this at a time when squatters' workshops were already huddling along the edge of the east wall of the Courtyard, obscuring the fagade, and destroying the aura of ceremonial magnificence. Only in the final stages of Phase II was the original purpose of the Chamber lost, when a row of stables, backing on to the Fort, was built across the western end of Courtyard and Chamber. Whatever its exact function, however, it seems that Fort and " Painted Chamber " together comprised the residence of a local ruler. The ceremonial portions fell gradually into disrepair with the decline in importance of the site in Phase II, but the fortified end was kept open for military reasons. How this ties in with the Manors on the Central Tepe, and with what we know of the rest of the area, will be discussed at length at the end of this article.

The Tile Designs

It remains to say a few words about the individual tiles themselves. Most fall into the categories enumerated last season (Goff, Iran VII, pp. 128-9, Fig. 9 and P1. IV). The vast majority still depict the " Standard " design. There are several new variations on the " Reverse " theme, with a central red cross and we also have complete examples of the " Nine Square Chequers " (P1. IIIc) and " Nine Squares " with unfilled central panels (P1. IIId). The most interesting, however, are several new designs: an entirely new form of " St. Andrew's Cross " (P1. IIIa); a sixteen square " Chequer Board " with a central cross; and a group of four " Wheel " designs (P1. IIIb). Furthermore, it is now possible to see how these rarer, more unusual tiles were arranged in groups against a background of tiles of the " Standard " design (Fig. 4). All four " Wheel " tiles fell in close proximity to each other, and unreconstructable fragments of a fifth were found in 1967, suggesting a cross with four arms around a centre. Even more striking is the arrangement in the very centre of the room, where the unique " St. Andrew's Cross " appears to have been surrounded by four more orthodox examples, the whole forming a " Kassite Cross " just opposite the main doorway.
20Ibid., p. I f.

EXCAVATIONS
ROOM 5

AT

BABX JAN,

1968
CO U TYA R D

149

MODZ*RN PIT

IEREBU/LD. LEVEL

/
r

y E 'D RRICR TUMBLE

E1 -:
77....-......

oLEVEL

~TEQEACINTEL ,E'/ ER

1
~/0/
. . ," '

?Z] p1:PETERS 'OUT WWALL i O

OLAPSED iFOR WAL r

A5-1 RL;RBLE
Ic.

m~

?B~
WALcL.

a
CLAY YELLOWl 1M.
NORTIT WALL

/V--
P.~

i" / " .-

|???
LESE?EEE

h
WIT/-I CII/ FLEE/N

/
PL

mp TOPVS ALL-,lFDD7

V
7,,,

l OF 51TIClnRD. OF COUR COURTYARD. ~YA

WEST WALLOF

ElLDE2q
+-LI/MIT OF EXCAVATIONS,%S TEP, bUTTRESS Ri WITU C1I

--A CL.ir ~b 4?4

~;k
PL bia FLOOR

RID

C
-' P?T/WD

//

rI~T-

FLOOR

Alo

P05/T/() OF DR~ f~ol I-51

FL

LDOR2I

Fort and westendof Courtyard. Ch Charcoal.

The Fort

The Fort appears to resemble the Manors on the summit of the Central Mound, but the rooms are slightly smaller and the walls more massive. The military nature of the building is deduced both from its appearance and on analogy with Naish-iJin. Owing to its impressivestate of preservation,we were able to do little more than cut a section through the centre of the building and clear rooms I and 3 and a small section of 4 down to their lowest floor levels. The remaining walls shown in the published plan (Fig. 2) were in most cases not taken down below the highest of the Phase II floors. Since it is already apparent that several modifications to the existing structure took place during this later period of occupation, certain features of the plan may have to be revised in the light of next season's excavations. This is particularly the case with regard to room 4. At present this has been reconstructed, on analogy with the Manors of the Central Tepe, as a large, squarish Hall, which in Phase II may have had a row of post holes across the centre.21 The uppermost stratum of the central area of the Hall was occupied by an enormous, closely-packedmass of articulated mud-bricksstanding on end (P1.IIa). From its original appearance in section (Fig. 5) I interpretthis mass as a fallen mud-brick wall which had collapsed in toto and still remained in an astonishingly good state of preservation. To the west of room 4 was a long room I I X 5 m. with three large buttressesdown each side. This was entered by a door still preserved intact, which appears to have been roofed with some form of corbelling, though the actual details are not very clear. At the northern end was another fireplacevery like that from the " Painted Chamber ", with a plastered chimney running up into the thicknessof the wall. On either side of the fireplacewere two narrow slots leading through into a cavity within the wall, and similar slots were found between the buttresseson the western wall, which presumablyled through to the outer face. These slots could have admitted little light, and do not seem to have been constructional, being too high for a floor and too low for a second storey. They may have supportedscaffolding but are not matched by other slots on the face of the inner, eastern wall. Like the " windows " at Niish-i Jin,22 their main purpose would seem to have been to admit air.
21

Excavations in 1969 have indeed revealed post holes but arranged in a different pattern. The so-called post hole of
13A

22

1968 is the beginning of a possible drainage channel. Ibid., pls. IVd and Vd, and pp. 12-13.

150

JOURNAL

OF

PERSIAN

STUDIES

The existence of a second long buttressedroom flanking the eastern end of room 4 is inferredfrom a short stretch of buttressedwalling, only the top of which has so far been excavated. It is immediately apparent, when reviewing this area, that the fortresswas built on a different alignment to the " Painted Chamber " and the far side of the room has still to be determined. At the end of the last season the allimportant junction was still obscured by baulks. Baulks also obscure the complicated rebuilding that must have gone on in Phase II, when the arch over the door between rooms 4 and 5 appears to have collapsed, and when the gap was blocked at a slightly later phase by a cross wall superimposedon the rubble (Fig. 5). TheSpiralStaircase Ramp or The only other room in the Fort which we had time to excavate completely was room I, which contained a spiral staircase or ramp. The principle of the construction is shown in the sections and elevations of Fig. 6. Room I is entered at its south-westcornerfrom room 2. As in room 5 the top of the door appears to have collapsed and only the foundations are visible. This seems to be because the westernmostwall collapsed into the centre of the room, destroying everything above a certain height.

MOCUY

COLLAPSE.DN

FLOO

h EA-----?M

FWOODEN

AMP Q

OO U-"Ex.

N7P.

LAN7DING

/I

EFALCT OLENN A
AOA CLIDN

AWALL.
NIC E OVEI

MAC

I;t
EXITINCr I-IGWT CENTPAL PIEIPI

.....P_ /
5 ECTION A-A ELEVATION CC

/ /

S..
LANDINq
IZCOACLIN L, 1 1

C 5P "TAII
m
_.2
' ...

5 E C T 10 N

B-B

E L E V A T 10

C"C

and Fig. 6. Sections elevations roomI. of

EXCAVATIONS

AT

BABA

JAN,

1968

151

From the door, a gently-slopingpise ramp leads up to a mud-bricklanding. The ramp restson huge scattered stone foundations. The landing has a sharp drop of 1 25 m. on its northern side, but, as we shall see, probably had a wooden continuation. The central pier was built independently of the landing, but bonded into a higher landing on the west. It supporteda demi-archwhich must have linked up to the north wall (P1.IIc). This arch was in a particularly bad state of preservation, and collapsed before it could be properly recorded. It appears however to have been some sort of corbelled construction. It seems likely that the two landings were linked by a wooden ramp or flight of steps running up to the west beneath the arch. In the first place, the lower part of the walls in the area beneath the arch had been baked red, presumablyin some violent conflagrationconsuming a great deal of timber or other combustible material. Secondly, when excavating under the arch, the sections revealed an area of burnt charcoal and clinker reaching up as far as the low east landing with collapsed yellow mud-brick on top. As the section receded further back under the arch, so the height of the charcoal-filled area increased until it reached approximately the height of the upper landing. Although I examined the successivesections with extreme care, I could not find anything to suggest that the mud-bricksformed a structure, or were anything more than tumble from the roof or walls above which had fallen on to the burnt remains of a wooden substructure-presumably either a ramp or steps. Masses of charcoal also adhered to the north wall, particularlyits north-westcorner,suggestingburnt postswhich had supported the upper part of the structure. If the arch had always been at the same height, there would have been very little head room at its western end. However, there is a diagonal line of breakage across the north face of the central pier, where the arch cracked off and collapsed, sloping up from the lower landing to the higher one. Below this line of breakage there is no sign of the arch beginning to spring outwards, although at the eastern end the springing started low down, below the level of the landing. Thus the roof of the arch also appears to have sloped upwards from east to west, parallel to the slope of the ramp beneath it, giving nearly two metres of head room throughout its length. The furthercourse of the ramp is uncertain, but there is a slight diagonal ledge on the southern side of the central pier, suggesting a higher continuation above the original low pis6 ramp, which was destroyed when the upper vaulting and wall collapsed. PhaseII The details of the later history of the Fort still need a further season's elucidation. The building appears to have been burnt, possibly at the end of Phase III, or at an early stage of iic, and then substantially reconstructed. Much of this reconstruction seems to have been of a rather shoddy nature, although room 4 was recoated with white plaster, and there appear to have been elaborate reconstructions of the upper part of the walls and buttressesin room 3. All the rooms east of room 4 have several higher floor levels, probably correspondingto Levels iib and iia in the Courtyardand " Painted Chamber ", although it is difficult to link up the floors within and without the Fort on the basis of anything except absolute levels. After a second conflagration when the stables of Level iib were completely destroyed with their horse and donkey occupants, there was a short occupation in Level iia before the Fort was finally abandoned. However, the shell was still sufficiently well-preservedin the sixth century B.C. to serve as a foundationfor the rathershoddy stone constructionswhich the inhabitants of Phase I built above it.
The Pottery The pottery from Phases I and III has been adequately described in an earlier report when it was classified under the headings of BibB Jan A and B.23 TWOfurther seasons' excavations have added to our repertoire of shapes, without otherwise causing us to modify our original remarks as to ware and affinities. Within Phase III we now have a great variety of beautiful complete or semi-complete " genre Liristan "wares, some of the best of which came from the floor of room 3. The" Rosette "or" Wheel " Unusual fabrics are design, recalling patterns on the tiles, is particularly common (cf. Fig. 7: Io, II).
23 Goff Meade,

Iran VI, pp. I 19-26.

152

JOURNAL

OF PERSIAN

STUDIES

rare, but a small bowl (Fig. 7: 9) in greyish black gritty ware with a very high burnish on the exterior and grooved decoration on the base is presumably imported from a neighbouring " grey ware " culture province.24 In Phase I, small bowls with flanged or nail headed rims still predominate (Fig. 8: 8, 9) but we also paralleled at Hasanlil in Level IIIa; small " chamber pots " with paired handles (Io); goblets (2, 3) and small globular drinking vessels (4, 5), which also occur in the Achaemenid level at Niish-i Jan.25 A small proportion of the Phase I pottery consists of a very fine red ware with a highly-burnished surface and one such fragment was decorated with a geometric design of opposed triangles in purple paint (6). " Bird handles " occur both in common and red ware recalling similar examples from Godin II and Nish-i Jdn.26 An unusual dish (7) was made of a hard, rathergritty and metallic buff ware with a design in red carelessly splashed on the rim. Similar pottery turns up in Tehran antique shops and is said to come from the region of Ardabil.27It would also appear to be related to pottery from an upper level of Yanik Tepe near Tabriz.28 The assemblage as a whole can be most closely paralleled at Godin Tepe, Level II, where practically every shape illustratedin Young's recent publication has its BdbdJdn counterpart,and where there are also parallels for our fine red slipped and common wares. The Godin Tepe assemblage would appear to have a far greater proportion of fine wares-as one would expect from a building of obvious political importance.29 The main contribution of the last two seasons has been the recognition of a new pottery phase correspondingto Phase II. In my first preliminaryreport, I drew attention to the fact that much of the pottery from room I in Level va (now iic) appeared to be plain and rather dark in colour, and this was " tradition.30 attributed to a degeneration in potting techniques impinging on a basic " genre Liiristmn We now know that a new type of ware also makes its appearance alongside the older wares, and that this is very similar to that from the Fort at Tepe Nfsh-i Jan, both in fabric and in its most characteristic shapes. The ware is hard and well levigated, with numeroustiny grit and mica inclusionswhich glitter when the sherd is held up to the light. This mica tempering also occurs in Phase I, but never in the earliest " genre Liristan " period. Unlike the PhaseIII wares, the pottery is usually wheel-turned. The colour varies from peach-pink through to reddish or greyish-brown,but on many of the sherds the core is fired much darker than the exterior, which appears as buff, or pale cream giving a " false slipped " effect, particularlyon the outside of the pot which often presentsa banded effect, also observed on the Naish-i Jan examples.31 The outer surface is usually carefully smoothed and often burnished. A few fragmentsof the same mica temperedware with a reddishcore, may have a thin red slip on the exterior, though again it is often rather difficult to distinguish between the effects of slip and of differential firing. Shapes also reflect the Nfish-i Jan styles. Jars with straight trefoil spouts (Fig. 7: I; compare Stronach, op. cit., IranVII, fig. 7: 3); small jars with paired handles (Fig. 7: 3; compare ibid., fig. 6: 7 and pl. XIa) and bowls with inturned rims ornamentedwith knobs and horizontal handles (Fig. 7: 5, 6; compare ibid., fig. 6: 3, 4 and 5). Grooved and nail-headed rims are also fairly common (Fig. 7: 2, 3 and 4). There are still insufficient sealed deposits to estimate exactly the proportion of the " Nish-i Jan
ware " to " genre Liristan ", but judging by sherd counts, and by deposits such as that in room I, Level iic, plain forms of the latter undoubtedly predominate. This suggests that there may have been
24The ware-but not the shape-is like some of the pottery from Necropole B at Sialk. 25 Personal communication from D. Stronach. 26 at Stronach, Iran VII, pl. XIIa. T. C. Young, Jr., Excavations Godin Tepe, fig. 44: 3. " 27 Goff, New Evidence of Cultural Development in Lfiristdn in the Late Second and Early First Millennia ", Ph.D. thesis presented to the University of London, July 2966, pl. 149: 2-4. The vessels are not identical but have the same "feel" to them. 28 C. A. Burney, " Excavations at Yanik Tepe, Azerbaijan, 1961 ", Iraq XXIV (1962), pp. 147-9, pl. XLV 33; cf. also

have numerous larger examples (I I, 12). Other frequently-occurring shapes are trefoil rimmed jars (I),

Dyson, " Problems of Protohistoric Iran as seen from IHasanli ", JNES XXIV (I965), p. 205, n. 3. 29 T. C. Young, Jr., Excavations Godin Tepe,figs. 42-44 and pp. at 30-31. For the architecture cf. idem, figs. 36-41 and pp. 23-30. I am indebted to Dr. Young for allowing me to go through his unpublished material on the site. 30 Goff Meade, Iran VI, p. 119. 31 Stronach, Iran VII, pl. XIa. I am extremely grateful to Mr. Stronach for going through the Nfish-i Jan and Bba Jan pottery with me at the end of the season, and discussing the various comparisons listed here.

EXCAVATIONS

AT

BABA

JAN,

1968

153

Fig. 7. Pottery from PhasesII andIII.

154

JOURNAL

OF PERSIAN

STUDIES

Fig. 8. Potteryfrom Phase L

EXCAVATIONS

AT

BABA

JAN,

1968

155

no basic population change at the beginning of Phase II, but that trade relations, and hence economic and political ties with the Malayer region, were close. While it is obviously unwise to generalize too dogmatically from the evidence of only a few sites, the discoveryof wares most closely paralleled at Ntish-iJan below those which would appear to equate with Godin II suggeststhat the Iron Age sequence for the Central Zagros runs: Bdba Jan III; Niish-i Jan; Godin II; with a gradual progressionof architecturalforms away from the single or double colonnaded halls of IHasanlGi and the Bdbd Jan Manors towards the Imperial tradition of Fort and Apadana. IV This sequence would also tie in well with Stronach'sradio-carbondate of 723+ 220 B.C. for the Central Building, and his attribution of the Fort to the seventh century B.C. (Stronach, op. cit., p. I6), but would Though one could possibly place Ntish-i Jan and BdbSJan II within the eighth century B.C. it is quite inconceivable that they are as early as the ninth. In fact, as discussedin an earlier report, Baba Jan I -which may be marginallylater than Godin II-has affinitieswith Pasargadae,and lies most probably within the sixth century B.C. (Goff Meade, IranVI, pp. 121-2). Detailed comparison of the material from all three sites, together with a more extensive seriesof radio-carbondates should eventually provide an exact sequence. Conclusions At some time in about the ninth century, a group of Iranian tribesmen,pushing down from the north reached the broad plain of Delfan, where an adequate water supply and large expanses of arable land made permanent settlement all the year round a practical proposition, and a large stone-walled village soon spread all over the summit of the deserted Bronze Age mound and out into the surroundingfields. Helped by its favourable strategic position athwart the main migration and trade routes, the site grew in importance, and not long after it had been occupied, the section of the village immediately on top of the mound was pulled down and the first of the Manors built in its place, presumably at the command of the local chief. How the settlement on the top of the Central Mound is connected with the new buildings to the east we have no means of knowing. Certainly the pottery from the eastern buildings shows subtle differencesin design-differences which continue in a debased form into the " squatting " phases. Moreover the " Painted Chamber " with its sophisticated decorative schemes suggests a rich and well-established community. One may therefore infer that the sequence ran Village, Manor 2,
Fort, " Painted Chamber ". make Young's date of 824?+0Io3 B.C. for the earliest stages of Level II suspect (Young, op. cit., p. 31).

And what of Manor I ? In last season'sreport I pointed out that the latest building on the Central Mound looked unfinished. Not only was there almost no mud-bricksuperstructureat the northern end -extraordinary if the walls were as high and massiveas those on the East Mound-but this is in contrast to the well-preservedremnants of walling above the Level 2 foundations. More extraordinarystill, this grandiose construction,occupying the best strategic position in the area, was completely deserted at the end of Phase III along with its surroundingvillage. Hardly a scrap of " Niish-i Jan ware " came from either the Manors or the surrounding houses. This seems inexplicable unless we can suppose that the redesigned Manor was never finished and that the sudden appearance of a new threat from outside led to the withdrawal of the existing population within the walls of the only building strong enough to withstand the new danger. The " workshops" cluttering the Courtyardmay thus representthe houses of the erstwhiledwellers of the Central Mound, who had retreatedthere for protection, converting what
was once a chief's residence into a motte and bailey castle. At first there seems to have been an attempt to preserve the ceremonial importance of the Chamber, but presently this too was abandoned as being too expensive a luxury in the new and insecure conditions prevailing, and the site was buried beneath a row of donkey sheds and stables. This phase would correspond to the seventh century-the time of renewed Assyrian incursions into the Zagros, and possibly the " Scythian interregnum " while the tightening of trade and political relations with the Malayer area, in the heart of historical Media, can best be interpreted as a stage in the gradual unification of the Medes and other Iranian tribes to counter these new threats to their independence.

156

JOURNAL

OF

PERSIAN

STUDIES

CATALOGUE
Fig. 7. Pottery from PhasesII andIII . " Nfish-i Jdn ware ". Pale orange-buff, gritty with mica inclusions. Wheel marked surface. Cream exterior with slight burnish. Height: 17 cm.; Diameter 16 cm. East Mound. Room 20, kiln. Upper iic? 2. " Nish-i Jdn ware ". Warm buff with small mica and grit inclusions. Smoothed surface. Diameter: 26 cm. Fort. II. 3. " Nish-i Jan ware ". Light brown with paler exterior. Mica tempering. Diameter: 25 cm. "Painted Chamber ". iic/iiib fill. " ware ". Creamy buff with grit and mica inclusions. Diameter: 22 cm. Fort. Room 3. 4. Nfish-i Jan II fill. 5. " Nfish-i Jdn ware ". Pinkish with tiny grit and mica inclusions. Creamy surface with streaky burnish.
Diameter:

6. 7. 8. 9.
10.

I1.

" Nfish-i Jan ware ". Light brown with mica and grit temper. Wheel made. Outside creamy with hard burnish. Three bumps on rim. Height: cm.; Diameter: I3.8 cm. Fort. Upper II. " 7"2 Jan ware ". Pinkish brown, wheel made, hard, with white grits and mica inclusions. Outside Nfish-i creamy, with " sandpaper " finish. Diameter: c. 9 cm. Room 20, kiln, II. " Nfish-i Jan ware ". Hard greyish buff with fine grit temper. Wheel made. Off-white surface. Height: cm. Room 20, kiln, II. BJ/68/279. Io cm.; Diameter: 8.8 Core grey to buff, well levigated with small brown grits. Surface mottled with very high burnish and hard sheen. Grooved decoration. Restored from fragments. Height: cm.; Diameter: 7 cm. Fort. 5"3 13" Room 3. III floor. BJ/68/237. " Genre Lfiristan ". Core grey with white and grey grit tempering. Cream slip has fired green on one side and spout. Paint, dark green. Height: cm.; Diameter: 8-4 cm. Grave I (from between 17"5 mounds). BJ/67/I34" Genre Lfiristan ". Core dirty buff with small black grits. Surface dirty buff, burnt in conflagration; smoothed hard. Paint, blackish brown, vitrified. Restored from fragments. Height: 30 cm.; Diameter: cm. Fort. Room 3. III floor. BJ/68/238. Io.I Fig. 8. Pottery from Levelsia and b All pottery wheel made.

14'5 cm. Fort. Upper II.

i.

2. 3. 4. 5. 6.

7. 8. 9. Io. I1.
12.

Pale orange/buff common ware. Height: c. 12 cm.; Diameter: cm. Grave 6. BJ/66/9. Fine cream common ware. Height: 7"2 cm.; Diameter: 7-7 cm. Level ib. BJ/68/242. lo02 Fine light reddish-brown common ware. Height: I0-3 cm.; Diameter: -4 cm. Level i. BJ/68/I oo. Fine pale orange common ware with large scattered grits; surface burnished; friable. Height: 7" 9 cm.; Diameter: I I 8 cm. Level ib. BJ/68/74. 15" Side of jar. Fine brick red ware, well levigated with wheel marks on interior. Outside: hard burnish. Vertical pattern burnish on neck. Paint: dark red, matt, applied after burnishing. Erosion below Level ib (from grave?). BJ/68/I i. Hard, rather gritty creamy peach ware. Surface: slightly eroded. Paint: reddish brown, splashed on in dollops. Diameter: 24 cm. Level i. Fine, soft grey common ware. Slight burnish on outside, inside smoothed. Height: cm.; Diameter: 5"5 c. 20 cm. Level i. BJ/68/277. Fine buff common ware, rather "floury". Scraped surface. Level i. Height: cm.; Diameter: 5"5 cm. BJ/68/89. 20"2 Warm buff common ware, rather " floury " surface. Height: cm. Level ia. cm.; Diameter: I5.8 I8.4 BJ/68/39. Medium coarse ware, burnished surface. Diameter: 33 cm. Level i. BJ/66/I99. Buff common ware with brown and white grit tempering. Slight burnish. Height: 20 cm.; Diameter: 42 cm. Level i. BJ/67/194.

Peach common ware. Height:

30-3 cm.;

Diameter:

9-2 cm. BJ/67/189.

" west. PI. Ia. " PaintedChamber looking

Pl. Ib. " Painted Chamber" looking east.

Pl. Ic. Recesseddoorway to room 20.

Pl. Ha. Collapsed wall or vault in room 4.

Pl. IIb. " Painted Chamber" : fireplace.

Pl. IIc. Stairwell: east landing and arch.

Pl. IId. Collapsed tiles on floor of" Painted Chamber".

c Pl. III. Tiles from " Painted Chamber".

a. Diagonal. b. Wheel. c. " Nine Squares " with Chequeredcentres. d. " Nine Squares " with unfilled centres.

b
P1. IV. Details of Fort: room 3. a. Fireplace. b. East wall and door. c. Door.

EXCAVATIONS AT HAFTAVAN TEPE 1968: First Preliminary Report By Charles Burney

Archaeological activity in Azerbaijan has greatly increased over the past decade, since the opening of the excavations at Hasanli in I956-57. These were closely followed by work at the neighbouring prehistoric sites of IHajjiFirtiz, Dalma Tepe and Pisdeli Tepe, significant for the sixth, fifth and fourth millennia B.C. respectively.' The same expedition likewise conducted surveys of sites in the Ushnu valley, and of course thoroughly investigated the Solduz plain, of which the mound of I;asanloi is the largest site. In 1948 T. Burton-Brownhad already carried out pioneer excavations at Geoy Tepe, near Reza'iyeh, the significance of which should not be underestimated.2 One of the periods especially illuminated by the work at Geoy Tepe was the third millennium B.C., commonly termed the Early Bronze Age, in which three sub-periods of Geoy Tepe K were distinguished, since provided with a around Lake Urmia in 1958-59, in which evidence of occupation from the fourth until the first millenB.C. was found in abundance. It was discovered that, during the third millennium B.C., the Urmia region belonged to one vast cultural province including eastern Anatolia within the highland zone and also Trans-Caucasia, though then it was not fully appreciated how numerous were the sites of this cultural zone nor how wide its geographical bounds. Yet no site belonging to this culture, apart from Geoy Tepe and Karaz near Erzurum, had then been excavated outside the U.S.S.R. Iasanlii seemed unlikely to provide evidence of this period. The excavations at Yanik Tepe, south-west of Tabriz, were thereforeundertakenin 1960-62, with resultsin terms of pottery and architecturebeyond all original expectations.3 A series of radio-carbon dates generously provided by the Radio-Carbon Laboratory of the University Museum, Philadelphia, has shown that the Early Bronze Age occupation at Yanik Tepe lasted from c. 2600 B.c. until c. 1750 B.C. or even later, the final levels being badly destroyed by later disturbance on the top of the mound. This occupation can now be fitted into the wider picture provided by the work of many Soviet archaeologistsin Armenia, Georgia and Azerbaijan. It cannot in any circumstancesbe studied solely in the context of north-westernIran, though the results from Godin Tepe have revealed a southward movement to the Hamadan region in the time of Godin Tepe IV.4 The history of occupation at Yanik Tepe came to an end with the close of the second period of the Early BronzeAge at that site. Later material comes almost entirelyfrom graves and from pits containing painted and plain pottery dating to c. 600 B.C. The second millennium B.C. is thus, except for its first quarter, hardly illumined at all by Yanik Tepe. In recent seasons of field work the Philadelphia expedition has obtained a firm sequence of cultural periods from IHasanlti,where Period V, characterized by grey ware, is now termed Iron I and may even date back as early as c. 1350 B.C., though a slightly lower date may seem more plausible. Contemporary with IHasanliiVI are levels at Dinkha Tepe,5 in the Ushnu valley, which contain a proportionof Khibir painted ware, giving clear indications of links with upper Mesopotamia and north Syria, possibly to be associatedat least in part with the tin
nium
1

radio-carbon date of 2574k 146 B.c. for K3. The writer carried out a survey of prehistoric mounds

For a recent brief summary, see R. H. Dyson, " A Decade in Iran ", ExpeditionII, no. 2 (1969), PP. 39-47. For a general in discussion, see R. H. Dyson, in R. W. Ehrich, Chronologies Old World Archaeology (Chicago 1965), pp. 215-56. See also Iran VII (1969), pp. 179-80 (HIjjji Firfiz); and Antiquity XXXIV (1960), pp. 19-28 (Pisdeli Tepe). 2 T. Burton-Brown, Excavations in Azerbaijanz948 (London i95 x). s C. A. Burney, Iraq XXIII (I961), pp. 138-53; Iraq XXIV XXVI (1964), pp. 54-6L. (1962), pp. 134-53; Iraq

4 T. Cuyler Young, Progress Report (Royal Ontario Museum 1969); for the later periods at Godin Tepe, see Young, " The Chronology of the late Third and Second Millennia in Central Western Iran as seen from Godin Tepe ", A.J.A. 73 (1969), pp. 287-91. 5 R. H. Dyson, " The Archaeological Evidence of the Second Millennium B.c. on the Persian Plateau", C.A.H. I, fascicule 66 (1968), pp. 20-23. See also Expedition Ii (1969), pp. 39-47.

157

158

JOURNAL

OF

PERSIAN

STUDIES

trade. These new discoveries are of the utmost interest for the districtsjust south and south-west of Lake Urmia; but they hardly apply to the same extent, if at all, to the rest of the Urmia region, in which the second millennium B.c. remains a little-known period. Side by side with the growing knowledge of the prehistoric periods of north-westernIran has gone an ever-increasingactivity in the work of discovery and excavation of remains from the period of the kingdom of Urartu, in Turkey, the Armenian S.S.R. and Iran. Though investigation of Urartian sites now within the frontiers of Iran has developed only very recently, this of course does not mean that much will not be found and recorded in the near future. There can surely be little doubt that Urartu at its zenith included within its frontiers more or less the whole of the Urmia region. Attempts to restrictits limits to the north-westdistricts near Lake Urmia seem quite unjustifiablein the light of the known evidence from Ta?tepe, Agrab Tepe and now also from Qalatgar, a typical Urartian fortress on a spur commanding the fertile Ushnu valley below. This fortress doubtless protected the southeastern frontier of Urartu against Assyrian penetration; an inscription of Menua (c. 810-781 B.C.) found at its foot makes its attribution to the time of Urartian expansion into those parts fairly secure. There is certainly no probability that this marks only a raid or temporary territorialgain. Moreover, the small fort of Agrab Tepe,8 situated close to HIasanloi probably built as a frontier post, has walls and in a typically Urartian manner and contains Urartian pottery, including the form of jar constructed shaped like an inverted bell which is known also from Giriktepe (Patnos), whence examples can now be seen in the Archaeological Museum at Ankara. The excavations currently in progress at Bastam, between Khoi and Maku, will reveal the plan of a major Urartian stronghold; and Dr. Kleiss has also found a number of Urartian strongholds extending northwards from the district of Shihpair towards Maku, a region naturally included firmly within the frontiersof Urartu.7 Excavations at Nish-i Jan and at Godin Tepe have provided the beginnings of an understanding of Median material civilization in the Hamadan region,8the hub of the Median kingdom in the closing years of Assyria and Urartu and in the little known period immediately preceding the rise of the Achaemenian empire under Cyrus the Great. The Urmia region formed the province of Media Atropatene, which seems to have declared its effective independence of the Macedonian empire after the victories of Alexander the Great, who never found time to campaign there. But of the earlier history of this part of Media, of precisely how and when it was lost to Urartu and annexed to the then parvenu kingdom of the Medes, little or nothing is yet known. Any indications of Median civilization in this region would be of the very greatest interest, all the more so since the Urmia region must have played a very significant role in the transmission of Urartian cultural traditions, particularly in building techniques but also in metal-working, to the nascent Achaemenian state. The problem confronting the archaeologist in the Urmia region is to try to disentangle the distinctively Urartian and Median elements in pottery and in other aspects of the material culture from the long-lasting traditions indigenous to the region and doubtless originating even before the arrival of the first Iranian tribes bringing their distinctive grey pottery with them from their earlier homeland to the east or north. If the Median period was anyhow relatively brief and remains tantalizingly obscure, the material culture of the Achaemenian period in the Urmia region and in the surroundingdistrictsis likewise little understood. No attempt to understandeven the pottery alone on the evidence of any one site is likely to bear close examination: Hasanlfi IIIA must be set beside the material from Takht-i Suleiman. Parthian and Sasanian civilization around Lake Urmia likewise require investigation. Among the
best-known monuments is the Sasanian rock relief near the modern town of Shihptir, whose name (formerly Dilman) is derived from this carving. The Parthian period remains elusive. The evidence of the precise measurements of mud bricks, normally square and thin, is one of the chief aids to the provisional dating of building levels from the Achaemenian period until as late as early Islamic times. The above outline of problems awaiting solution in the Urmia region may help to demonstrate the reasons for the choice of Haftavan Tepe for a campaign of excavations which is planned to last several
o R. H. Dyson, " Problems of Protohistoric Iran as seen from IHasanli ", J.N.E.S. XXIV (1965), PP. I93-217 and especially pl. XLIII.
7

XVIII (1968); IranVI (1968), W. Kleiss, Istanbuler Mitteilungen pp. 166-7; Iran VII (1969), p. 188. 8 David Stronach, " Excavations at Tepe Nfish-i Jain 1967 ", Iran VII (1969), pp. 1-20. See also note 4 above.

EXCAVATIONS

AT

HAFTAVAN

TEPE,

1968

159

seasons. Any archaeologistapproaching a site for the first time in a region relatively little known has rightly to be able to justify his choice and to state in intelligible terms his main objectives and hopes. If these are not all achieved, it should not be for want of consideration or of planning the location of trenches on the site to gain the maximum of information for the minimum of outlay in time, money and effort.

SNakhichevan.

SU.
Ozhulf a

S.S.R.

Marand

TEY K UR -Yanik
N Rez [eh

Tepe

I RAQ

,
me'f,

DinhTep Hasant

Sahabad
Fig. I. Map showingthelocation HaftavdnTepe. of

The size of the mound of Haftavan, alternatively called Butan Tepe, Toban Tepe or Tam Tepe, is shown by the site plan (Fig. 3), while the general situation is indicated on the sketch map (Fig. I). This situation makes it likely to have been open to cultural contacts and influencesfrom both north and south, from Hasanlii and its surroundings and from the less well-known sites of Trans-Caucasia. East-west connections are on the whole likely to prove less prominent, though excavations could disprove this suggestion. Any site in north-westernIran which can illustrate links with Trans-Caucasia would be worth investigation for that reason alone. The large area of Haftavan Tepe, one of the three

160

JOURNAL

OF PERSIAN

STUDIES

or four biggest mounds near Lake Urmia, suggests its importance both for its situation and as the centre of a small but well-watered and extremely fertile plain. Surface examination of the mound, carried out by the writer in 1958 and 1959, with a subsequent visit in 1962, suggested that the site was mainly of the second millennium and the early first millennium Clarification of the sequence of pottery and of the general cultural development in the second B.C. millennium B.c. was therefore a clear possibility from the surface evidence; but the dating of much of the pottery thus collected to either the early first millennium B.C. or earlier was not at all apparent. Excavation alone could provide the solutions to the problems raised by surveys. Proportionately little that could be directly related to IHasanlti or to Urartian pottery was to be found. That the citadel crowning the mound was of later date than the larger area forming the " lower town " was immediately realized, but how much later was uncertain. Fears of deep levels of later date than the periods of special interest, that is of post-Achaemenian date, were diminished by the evident lack of Islamic surface material, even though surveys elsewhere in the Near East have sometimes failed to reveal sufficiently clearly the presence on the top of the mound of a considerable depth of late occupation obscuring the levels of the periods under investigation.

N/i

POWDERY MIXED

N/2
N/2-POWDERY
W13
FOUIDFWNUMS

luCKY
SAME WALI
OF

W lay
WIRE /DRAR SHY I

Wdn1ERN//, 7/f~PNS ~ ~l ~
ME SET

~N3D~YR~un F--RnR--EASYL oE-wm&E mAREA N

AREA

HAR lUK

rtn~is E14El/hB

side of Trench Fig. 2. Longsection(north B-E-N).

At the time of the writer's first visit to Haftavdn Tepe the citadel was badly scarred by a great trench driven straight through its centre from east to west. A similar trench was found to extend along a roughly parallel line a short distance to the south, cutting into the south slope of the citadel. These trenches had obviously been dug with a purpose some years previously: subsequent enquiries, in 1968, elicited details including the date of the digging of these trenches, about 1933-34. No report is known to have been published, but plainly discoveries of sufficient interest to give an incentive for such a large effort must have been made at about that time or slightly earlier. The main trench seems not to have been continued down below the level of the foot of the citadel. In the years between 1962 and 1968 extensive damage was done to the lower slopes of the mound, especially to the south and west, only the north side escaping untouched. Many of these scoops, brought about by villagers carting off soil to their fields, are only shallow, such as that at the foot of the east end of the citadel; but others are deep and have reduced the area of the mound which will repay excavation. Fortunately there is a great expanse left untouched, the largest such area being the gentle slope of the lower town at its east end, towards the spring, which still provides good drinking water, and

EXCAVATIONS

AT

HAFTAVAN

TEPE,

1968

161

which may well have been the chief source of supply for the prehistoric community. There is no difficulty in finding water in the fertile plain, the average depth of the water table in recent years being approximately 5 to 7 m., although there has now been a rise in the level of the water table, resulting in the overflowing of wells in the higher parts of the plain. Salinization of the water is apparently no problem, owing to a strong flow of sweet water, washing salts from marly hills down into Lake Rezd'iyeh. Nevertheless, salts in the levels of Haftavan Tepe are very apparent, perhaps especially in Area J, down near the spring, and these salts cause damage to the pottery; but no such hard salt crust as occurs on many mounds near Lake Rezd'iyeh is to be seen on the surface of Haftavan Tepe, which has a scattered covering of vegetation. As the site plan (Fig. 3) shows, the site is a large one, with a length from east to west of at least 550 m. and of up to 4oo m. from north to south, the site plan being divided into 20 m. squares for purposes of surveying. Datum point was at the level of the spring by the mill. The highest point of the mound was 27 m. above datum, suggesting a total depth of deposit of about 25 m., an estimate which will be proved or disproved by a sondage in Area J currently being carried out in the second (1969) season.

E1/4 IBURNING WALL ,-ST-RMU

S IME 11 - BRMWWITH

OF E ____MsT.= M11..,..

HAlW B T Muo YEOWH ES OAY ATI1 9UNWH?, HARD YELNVo KEAS OR EMR
VEHY

El

l R/LOOSEurxY) FILL /

... FILL RY E/1 LOOSE PeTTRY MXEO] YEL M

B/2~

E/1

r~l~l._3_r~si.E/2
-77
I BURNT WIBTHBRJIM

_E RDYELLMI VRY MICKY

RnIIAon E) isuiR
r

--E/[omrr~nE/5

OCEATO----N
I
'0.1% L /OFXED WE FACE WAL SMC

EIB
IEADS
am TO
WAa

CGPATIaA
16

-.

E/2tobottom oftopwrse sones of

som E/Aoto ofthus AL

M &%ais

BRmIM [ IO

V f YEE15 sow1M.. fTSN AEwE STFASNL IL -(F ASHY WMI

B_-Znfln---"'

WALL NAC-YTENRIS FACE EADS STRAW ETC. SOFTTECAYED

WITH

BRIM

BM4U

STONES

MEDOF SUY nu

SI TEWO

CCV AREA E -5 1
HAFTAVAN

AREA B

2
TEPE

19
-1968
SElN

AMEA

NE&B

NORTH

Several features of the mound were discernible before the start of excavations. The summit of the citadel was shaped like a shallow saucer, with a rim and dipping slightly down towards the centre, where the great trench had torn the citadel in two. These contours thus suggested a citadel wall, an impression strengthened by the uniform and fairly steep gradient of the north slope, where the best hope of finding the citadel defences undamaged clearly lay. The marshy nature of the ground at the foot of the site on three sides seemed to indicate that the main approach to Haftavan Tepe must have been from the west, the side facing the village of Haftavan and the only direction from which a vehicle can now reach the mound. Irrigation now adds to the effects of streams flowing down towards the lake to make much of the surrounding plain not easily passable: this accounts for the difficult access to Haftavan Tepe, even though in a beeline it lies not at all far from the main Shdhpir-Rezd'iyeh road, from which it is clearly visible, standing up from the plain to an impressive height. The relatively small area of the citadel compared with that of the site as a whole suggested a late period or periods of occupation, when the greater part of the site was no longer occupied. This would give opportunities for distribution of trenches with the aim of revealing levels of different periods not far beneath the present

.
. -

A-1
___ ___ __ _____ _ _ _ _ _ _ _ _ _ _ _ _ ___

___

1 1 11. 1 1116 1
18

19
21 22 23

20 24 2

26

-__

__
______

FOF

---

if
R5----------

-_-_

-_

X--

__

6 8 9 10i 12 131 15 1i 17 18 19

L
20 212
Fig. 3. Site plan (contoursat half metre intervals).

'_

24 5

OF
ROVEN level5

7' PACKED7
FeLOO ROH levl6 FLOOR

PIT

Wevel

/PIIlO

HAFTAVAN
AWs

TEPE-1968
LMS

C3

K2

K1

3W

/4,5&6/

5&6

Fig. 4. Plan of the Urartianlevelon the citadel(AreasK andC3).

164

JOURNAL

OF PERSIAN

STUDIES

surface of the mound. Perhaps the most important period in the long history of the settlement would be found towards the spring, where reports of discovery of graves provided an additional reason for location of one of the areas of excavation at the east end of the site (Area J). The main effort, however, was concentrated on the summit of the citadel (Areas C, G and K), with a long trench down the north side (H and F), and on the west side, where a long 5 m. wide trench (B-E-N) was excavated from close to the west end of the old trench across the citadel. Later in the 1968 season extensions (M and P) were made on either side of E. The strategy of the first season was inevitably determined by the surface contours of the mound, supplemented by the evidence of strata visible in the innumerable cuts into the slopes of the Citadel and of the lower town. The first season of excavations at Tepe was sponsored by the University of Manchester, the expedition being further financed by grants from the British which also provided generous support,Haftav.n Academy, the Calouste Gulbenkian Foundation and other institutions. To the Archaeological Service of Iran a special debt of thanks is owed for the permit allowing the excavations to be carried out; and Mr. Ahad Darbani, the representative appointed to work with the expedition, is owed a personal gratitude for his help and advice at all times throughout the season. The authorities in Shdhpiir most kindly allowed the village school at Haftavan to be used as the headquarters of the expedition, thus enabling us to stay in a particularly delightful village. The school, built by the Armenian community, is situated within a compound together with the church. The members of the expedition staff totalled ten, comprising Charles Burney (director), Brigit Burney (photographer), George Learmonth (conservator), Malcolm Stephenson (architect), Duncan Noble, Regnar Kearton, Edgar Peltenburg, Timothy Dickinson and Peter Burney (archaeological assistants) and Krissa Kearton (housekeeper). Excavations began on July 9th 1968 and the camp was closed on September 13th. The number of workmen employed on the excavations was at first small, since there were none that had had any previous experience of such work; but, as the season drew on, the number was increased to a maximum of about 120, being halved for the last few days of the excavations. It was decided as a matter of policy not to confine the locations of trenches or areas of excavation to squares aligned exactly with the grid. Such an alignment, though facilitating the work of the surveyor, would necessarily fail to take into account either the condition of the site at the beginning of the excavations or the orientation and probable extent of buildings as discovered in the early trenches. Certain areas were, however, laid out on an east-west or north-south alignment. Letters of the alphabet were assigned to each new area as it was opened up, trenches being either 5 or 4 m. wide and squares being I o x I o m. The same letter was assigned to four quadrants making a whole square of 20 x 20 m., Area J being the only such completely opened up in the 1968 season. Three quadrants (CI, C2, C3) of one area were opened up on the citadel. Such a method of laying out and describing areas of excavation is eminently suited to a campaign based largely on the desirability of uncovering extensive plans rather than of concentration on the stratigraphy of certain limited areas. At Haftavdn Tepe it has always been hoped that a full stratified sequence can be obtained by taking advantage of the presence of the main town quarters on different parts of the mound at different periods. The Early Bronze Age having been investigated at Yanik Tepe, where levels of that time are to be found on the surface of most of the site, there was no intention in the first season to dig in strata of the third millennium B.c. In fact, however, a small trench just below the bottom of the great ditch, in Trench E (Fig. 2), did yield pottery of very late Early Bronze Age appearance, consistent with a date even as late as the second quarter of the second millennium B.C. Not till the 1969 season was an area (R) opened up at the foot of the east end of the citadel, where it has been much destroyed, which revealed a building level of unmistakably Early Bronze date, probably c. 2000 B.C. Both the pottery and a fine stone axe or hammer-axe confirm this dating. The architecture likewise is utterly different from that found in the later levels of Haftavin Tepe. It is noteworthy that the 17 m. contour crosses Trench R, suggesting that well over half the present height of the citadel existed already by the end of the third millennium B.c. The fill is much harder than most of the later levels, where mud brick tends

EXCAVATIONS

AT

HAFTAVAN

TEPE,

1968

165

to be very poorly preserved, possibly in part the outcome of the action of frost; but more probably it was the unsuitable mud used which caused the bricks of later periods to disintegrate so badly. The next phase distinguishedin the 1968 season is representedby the massive stone building shown in axonometric projection (Fig. 6 and Pl. IIb-d): its discoverywas indeed a happy chance, since it lay within the width of the long trench down the west side of the mound (B-E-N). The meagre pottery from this level was insufficient to provide any secure dating evidence in a region where no ceramic sequence is yet known for the second millennium B.C. A date falling between c. I7o0 and c. 1400 B.C. seems suggested by the decadent Early Bronze level just beneath and by the occurrenceof grey wares of V IHasanlGi type (c. I350-c. I000 B.C.) in an overlying level to the west of the main building. There, in E I and N, a plan of rather unimpressive houses contemporarywith the tall building was excavated. The construction of these walls and of the tall building alike was of stones of widely differing size set in hard mud. Little of the original mud-brick superstructureremained in situ, and the faces of these walls, often more mud than stones, were not always easy to trace. Similar construction,hardly fitted to endure many of the wet springs and early summers now so characteristicof western Azerbaijan, is to be seen in neighbouring villages today. There a row of stones is quite often set high up in a terre pis6 wall. Natural lines of cleavage in the soil of the mound, percolated by salts, add many a false trail in the search for an elusive wall face. Yet even brickworkin this region is all too liable to have been very poorly preserved unless hardened by burning: such burning made the task of the excavators of the IV buildings of the IHasanlGi citadel relatively easy.9 A provisional conclusion, reached before proper study of all the relevant pottery, is that during the period equivalent to ;Hasanli V the site was only very sparsely settled, with the familiar grey ware imported in comparatively small quantities from the south. A large pit (JX 3A) near the spring may be assigned to the Iron I (Hasanli V) period, together with one burial found immediately overlying the fill of this pit and stratified beneath all the other burials uncovered in Area J. In the long trench (Level E I/5) dark grey wares predominate, with " button " bases and a " worm " inside the rims of bowls being characteristic.10 The stratified sequence being obtained from the Iox 0om. sounding (JX) near the mill should without doubt stretch back before c. 1500 B.C.,and has already (1969 season) revealed connections with Trans-Caucasia in its wide variety of painted pottery, even though this is much outnumbered by plain wares of predominantlyred tone. Correlationswith the tall stone building in the long trench will be among the results of the sounding in JX. The evidence available to date would seem to indicate that the town reached its greatest prosperity V and the subsequent regional in the early years of the first millennium B.C.,after the end of diversificationdiscerned by T. Cuyler Young as a result of his extensive field work in western Iran.11 .Hasanloi IV Pottery of the period contemporarywith IHasanlii no longer shows the same affinitieswith that site; but Urartian elements are not at all discernible, and all the evidence points to a period earlier than the main building level so far reached on the citadel. Moreover, building remains of the early first millennium B.C. (tenth and ninth centuriesB.C.)have been found both in Area J near the mill and extensively over the western slopes of the mound. The contours of the site suggest an extensive town quarter over the wide area of the eastern slope of the mound; and the opening of a new area (S) immediately west of J should provide some more evidence of the layout of the town. It will, however, be several seasons before a very extensive plan can be uncovered. Levels I and 2 in J (Pl. IIIa-b) are preservedonly to the top of their stone footings; but these are enough to show their sound construction and the wide areas enclosed by them. The writer owes to ProfessorSeton Lloyd the suggestion that possibly here may be the remains of a bazaar or market near the edge of the town. Large enclosures rather than houses seem to be indicated. The date of Level J2 must surely lie in the tenth century B.c., with some phase in the ninth or eighth century being the likeliest date for the burials found scattered over Area J. One burial contained a bridge-spoutedjug of dull red ware, burnishedoutside, with coarse grit temper
H. Dyson, J.N.E.S. 24 (1965), for a list of his articles on the citadel of Hasanlfi IV in various numbers of I.L.N., Expedition, etc. See also T. Cuyler Young, " Thoughts on the Architecture of Hasanlfi IV ", Iranica 7 (1967). 10R. H. Dyson, J.N.E.S. 24 (1965), pp. 193-217, especially
O R. 14A

pp. 195-7 and fig. 2; see also T. Cuyler Young, " A Comparative Ceramic Chronology for Western Iran ", Iran III (1965), pp. 53-86. T. xx Cuyler Young, " The Iranian Migration into the Zagros ", Iran V (1967), pp. 1-34-

166

JOURNAL

OF PERSIAN

STUDIES

and hand-made (Fig. 8, No. 2): pots were only rarely placed with these burials, so that this is one of more significance than its quality deserves as dating evidence. It could point to a date in the eighth century B.C. for these burials in J. A cremationjar found in J2 (Level 2) is of light red ware with only the slightest burnish outside, and is wheel-made (Fig. 8, No. 3); this appeared at first sight as perhaps of Urartian type, but this is not certain. It contained burnt human bones and rodent remains. S-shaped earrings of bronze occur in burials found both in J and in A, the area on the south-west slope of the mound where excavationsfirstbegan and where remainsof a building level were uncoveredimmediately beneath the surface. If the impression of an eighth century date for the pottery from Areas A and D is correct, then the burials from A and J could fall as late as c. 700 B.C., though a higher date seems more probable. Discoveries in a burnt level on the citadel in the 1969 season indicate continuity in bronze jewellery, comprising tassels and S-shaped earrings, accompanied by changes in beads, yellow glass becoming especiallypopular. A date in the seventh century B.C.for this burnt level seems probable, so that fashions in jewellery seem to have lasted with relatively minor modificationsfor three centuries or more.
110 pFSFNT nFM01 -sSURFArF

IXED DEBRIS K/1 POWDERY

K/1.
STEONE

MK/2
K/3
HARD
YEL

ISH]

BRICKU

BRIS

PIr/

TEARUERKIK

BIC KBRIC3 K/2

FEEUEN F

'K

on

0 '.5

HAFTAVAN TEPE-19T18
AREA

WEST

SECTION

acrossAreaK on thecitadel. Fig. 5. Section

distinguished by the presence of long bronze chains, each one worn as a personal ornament, a feature not without parallel at IIasanlGi. The other jewellery includes a large number of bronze tassels, too numerous to be described as earrings. Doubtless they were worn round the bronze headband which was also distinctive of these burials (Fig. 7 and P1. IVa-d). These must be earlier than the burials in A and J, though a date as late as the end of the ninth century B.C.is by no means impossible. The popularity of carnelian beads is strikingly illustrated by the discovery of a hoard, doubtless belonging to several necklaces, in a jug in (P1. Va). A less expected occurrence not far removed A[ in date is of painted pottery in Area P, where sherds, including some painted outside in light red, dark red and black on a buff ground, were found in some quantity and have been supplemented by material from P2 and Q in the 1969 season. Even though disturbance caused by the digging of the adjoining ditch, so prominent a feature of the long section (Fig. 2), must be reckoned with, there seems to be growing evidence of continuing use of painted pottery well into the firstmillennium B.c. It was certainly not restricted to levels preceding the grey ware of IIasanlfi V type. Classification of painted wares should prove to be one of the most rewarding results of work at Haftavan Tepe.

B.C.) are two burials found at a high level near the west end of the long trench (N): these are Sooo-800oo

Probably to be ascribed to the opening years of the period contemporary with HasanlGiIV (c.

EXCAVATIONS

AT

HAFTAVAN

TEPE,

1968

167

E. AXONOMETRIC-AREA
stone in E. of projection themajor building Trench Fig. 6. Axonometric The plan of the main building level reached on the citadel (Fig. 4 and P1. VIa-c) may serve to indicate how in the first season a tantalizing glimpse of a major period in the history of the site was obtained. Even though the settlement was probably by then limited to the area of the citadel, these walls must surely show that a citadel or possibly a fortress of some importance was then standing on the summit, commanding the surrounding plain in typically Urartian style. Though not all the pottery is of Urartian appearance, a date in late Urartian times still seems much more probable than an attribution to the early sixth century B.C. Such a dating would of course imply a connection with the Medes; but nothing that can be called distinctively Median has yet been found at Haftavdn Tepe. Since,

168

JOURNAL

OF

PERSIAN

STUDIES

TEPE 1968 HAFTAVAN - BURIALS

SCALE 0 01 15
0.2 03

metres 05

INFANT A/3 No.1 BURIAL

DOER IS SKULL FACING
0M SMALL RI BRONZE G---L

BRONZE BEAD i ,

OBJECTS SKULL DETAILS BENEATH OF [1:101

OF OUTLINE ABOVE SKULL

" 03

BRONZE BRONZE PENOANT \ BRACELET BRONZE 0 RING FINGER LINE OF . K/LARGE BRONZE CORNELAN, & BEADS FRIT

RING BRONZE

81 RING BRONZE

OF PIT DGEBURAL

FEMALE ADULT N/1 No.5

PENDANTS BRONZE BRONZE PEANT OF" CLUSTER

IRONNS

BRONZE CIRCLET

EANT CUTE OFBRONZE E ; STEMS )

-

CORNELAN BEADS
ON ONG BONZE FRAGMENTS

PENDANTS BRONZE

10IRON

PIN - --CORNEUIAN & NECOLACE FRIT BLUENECDSAE

POT FIT

FEMALE ADULT Cl No.1

BRONZE RING INGER

SBRORBN7ZE

RIN

MIRROR,

RING BRONZE OKEN

B.C.; middle-c. tenthcentury B.C.; bottom-earlySasanian periods: top-c. eighthcentury period. Fig. 7. Burialsof different

EXCAVATIONS

AT

HAFTAVAN

TEPE,

1968

169

however, no evidence has been published to provide guide lines in the quest for relics of the Medes in the Urmia region, it is quite impossible to do more than guess at parallels until intensive study has been carried out. One parallel with Nfish-i Jan has been found several times on the citadel in the later levels, overlying the major building shown in Fig. 4: this is a coarse, shallow platter. Some finer pottery forms, largely bowls, also appear comparable with Median material from the Hamadan region. Pottery contemporary with the main level includes unmistakably Urartian wares, among them the now well-known red polished ware, once termed by the writer " Toprakkaleware ".12 Since, however, it is found throughout the length and breadth of Urartu, albeit often only in small quantities, such a description seems too restricted. The difficulty in defining the period of Urartian domination at Haftavdn Tepe and elsewhere around Lake Urmia lies largely in the persistence,side by side with the imported Urartian elements, in the material culture of indigenous traditions in pottery and metalworking. No such cultural uniformity as characterizes Urartian sites in the central homeland of the kingdom of Van is to be expected for this same period at Haftavdn Tepe. Without excavations no significant comparisonscan be attempted with the Urartian fortressescommanding the plain in which Haftavdn Tepe stands. Preliminaryconclusions based on the first half of the 1969 season suggested the possibility that the citadel in the late Urartian period or immediately thereafter (seventh or early sixth century B.C.) comprised a row of buildings sited round the perimeter and facing in towards an open square in the centre; but the latest discoveriesmake this less certain. Associatedwith the building level shown in Fig. 4 was a rather thin layer of burnt debris, also found in Area G3 (1969 season) and in the area of the citadel wall of later date. At least three buildings belonging to this main level have now been in part uncovered; and there is evidence of a secondary phase, with additions of inferiormasonry. Clearly there was a violent destruction and burning of the citadel. The relatively modest depth of the burnt debris and the indications of nearly complete disappearancefrom many of the walls of the mud brick superstructuresuggest a phase of abandonment of the citadel, when some stone robbing occurred, paving stones being removed from the contemporary surface in K. The later levels on the citadel of HaftavAnTepe are, so far as excavated to date, in a very ruinous condition, mainly owing to the digging of innumerable pits all over the area. These pits, not unique to this site but found elsewhereat comparable periods in westernIran, suggest a very retrograde phase. They have moreover made any intelligible plans of the remains of the later building levels almost unobtainable. Only the citadel perimeter wall was untouched, till seriouslydestroyed in recent years. Part of this wall was found (in Area G) near the surface, exposed by a modern cut into the east end of the citadel (P1.VIIa). The bricksare of thin square form, 42 x 42 x Io cm., a size possiblymore suggestive of Parthian than of Achaemenian date. The pottery, however, points to the earlier of these two IIIA. A heavy bowl of wheel-made dull red ware (Fig. 8, periods, having clear parallels with IHasanlGi No. 12) deservesspecial mention; and another bowl is distinguishedby the shaving of its outer surface (Fig. 8, No. I I and P1. VId). Found cut down into the Achaemenian building levels on the citadel were graves containing a fairly standard equipment. At first, however, when the grave shown in Fig. 7 (C2, Grave I) was uncovered, its date was far from certain: the design on the bronze finger-ring,perhaps representingthe sacred tree in stylized form, appeared to indicate an Achaemenian attribution. It was with the increasing occurrences of glass that a later period came to be plainly demonstrated. A silver coin of
Shahpir II (P1. VIIIa) seemed to clinch a dating to Sasanian rather than to Parthian times, most probably early in the Sasanian period. The finest object in the burial illustrated in Fig. 7 is the long bronze hairpin, found against the skull of the woman who had presumably worn it in her lifetime: this pin is cm. long, with a decadent version of the pomegranate type of head which appeared at 15"5 the time of discovery to lend support to an Achaemenian date. What made this most improbable was a miniature glass mirror found with the same female burial: this mirror, presumably for cosmetic uses, was of clear glass with a backing of bituminous material. It would seem that silvering must have been used, for the mirror is as effective as one manufactured today. Such a mirror turned out to be one of the items of equipment normally occurring with these Sasanian burials. Another standard item was a
12 C. A. Burney, A.S. VII (I957),

p. 42.

10

15cm.

[except 3,10 figs &121

56

RED EEP

10

12

13&

EXCAVATIONS

AT

HAFTAVAN

TEPE,

1968

171

miniature glass bottle, presumably for scent or perfume. One burial in C2 (Grave 4) was particularly rich in beads, having two necklaces, one of 33 beads of glass, with white patina and green inside, the other of two pendants and I 14 beads, mostly of glass, the colours including black, white, blue, yellow, green and olive green (Pl. VIIc). A mirror and glass bottle were also found with this same burial. Black stone beads of thin square form with green and orange glass spacers were found in another Sasanian grave (Pl. VIIIc). Excavationsin the areas of the citadel opened up in the 1969 season suggest that this Sasanian cemetery was restrictedlargely or even entirely to the north-westpart of the citadel, especially in C2. Traces of two Sasanian burials right on top of the late citadel perimeter wall, added to the evidence of the section in Area K (Fig. 5), provide further indication, beyond the pottery, that that wall must be pre-Sasanian or very early Sasanian. An early Sasaniandate, already suggestedfor the cemetery on the citadel summit, is furtherindicated the absence of evidence for an Islamic date for the one building level found to seal these graves.13 by This contained a stone-paved street (P1.VIIId) acrossArea C, with remains of stone walls, petering out towards the centre of the citadel, with every indication of very severe stone-robbing. This was the final stage in the long history of settlement at Haftavan Tepe, when there may have been a stone-paved approach road up into the citadel from the west side. No such approach has yet been traced in earlier periods. Haftavan Tepe may not yet be known beyond the small number of those archaeologists familiar with the Urmia region. It may not always have been a great town, nor was it necessarilyoccupied with unbroken continuity throughout the millennia from its earliest level to the last buildings on the citadel. Yet assuredlyit will, through the current campaign of excavations, steadily be revealed as a site of the first significance for the archaeology of north-westernIran and neighbouring regions. Such evidence can only be obtained by sustained effort, the reason for the decision to continue work for several seasons. In the next preliminary report more of the documentation for some of the conclusions here tentatively put forward will be provided.
x3The practice of inhumation in the Sasanian period in this part of Iran has implications which cannot be pursued here.

Fig. 8.

(i)
(2)

(3) (4) (5) (6) (7) (8) (9) (to) (I I)

(12)

(13)

Greyware (Jr, Grave7). jug Drab redware (Jr, Grave jug 3). Red burnished warecremation (Jr, Level2). jar Two typicalbronzetassels. and Bronzependant earring(J3, Gravei). Bowl of rather coarse ware (Ai, LevelI). red Lightgreywarecup (Ji, Grave7). ware Polychrome sherd(PI, LevelI). ware Polychrome sherd(PI, LevelI). Two-handled of coarse ware (H2, Level2). red-brown jug Shavedwarebowl, buffoutside,light redinside,wheelmade pit in Level2d). (CI, Largebowl of dull redwheelmadeware (Ci, Level2). Ribbed glass bottle, pale green,basebroken (CI, LevelI). off

north-east. P1. Ia. Viewof plain of Shdhpir,looking

seen Pl. Ib. The mound from the

Pl. Ic. Viewsouthwards from themound.

in Pl. Id. Stoneandpise wall construction H

P1. IIa. Trench N, looking east.

Pl. IIb. Trench E-N, withstone

E: thestonebuilding excavation. Pl. IId. Trench during

E: thes Pl. IIc. Trench

Pl. Ila. AreaJ, Level2.

Pl. IIIb. AreaJ, Level2 andLevel3 (in soun

P1. IIIc. Ji, Grave7: a burialof theIHasanllV period.

Pl. IIId. A laterburialgroupin AreaJ (c. ei

N: burialwith bronze chain,etc. P1. IVa. Trench

Pl. IVb. Trench N: contents thesa of

N: contents thesecond burialwith bronze chain,etc. Pl. IVc. Trench of

Pl. IVd. Trench N: blue frit beadnec

Pl. Va. AreaA: jug with contents carnelian frit beads. and of

an Pl. Vb. AreaA: contents childburial(bronzes of

centuries Pl. Vc. AreaP, LevelI (c. ninth-eighth B.C.).

ware P Pl. Vd. Polychrome from Trench (cf. Fi

with latercitadelwall above. P1. Via. AreaK: Urartianbuildinglevel (foreground)

Pl. VIb. AreaK: Urartian buildinglevel(foregr

Pl. VIc. Area C3: Urartian building level, with later walls in foreground.

Pl. VId. Shaved ware bowl (cJ: Fig.

citadelwall. PI. VIIa. The latermud-brick

Pl. VIIb. Pits in AreaC on the

Pl. VIIc. A Sasanian burial group.

Pl. VIId. A Sasanian miniaturegl

Pl. VIlla. Silver coin of ShjhpiurII (309-379 A.D.).

Pl.

V1t.

A Sasanian burial g

Pl. VIIIc. Sasanian black stone beads with glass spacers.

Pl. VIIId. The stone-pavedstreet on the citadel, ove

"FURTHER NOTES ON THE SHAFT-HOLE PICK-AXE FROM KHURAB, MAKRAN"

Iran VII (1969), pp. 163-8

CORRIGENDA TO THE TECHNICAL APPENDIX Through a series of unfortunate circumstances, I was unable to see the galleys of the Khurdb pick-axe article before it went to press, and a number of serious errorsappear in the publication which were not in the original manuscript of the technical report. Note should be made of the following corrections. Page andline Correction
164, Spectrographic Si

Analysis Table Sn
166, line II

<o - ooi

per cent

Not detected (detectable to 20 p.p.m.) 166, line 7 The last metal to solidify would do so at theplace where the metal was hottest ... leading, suggestingthat the densities, which representbubbles in the metal, have some material substance and are contained as inclusions in the casting. They are actually indications of voids, the absence of matter in the bronze.)
Small, almost roundareas of high film density ... (The term " bodies " is mis-

166, lines 30-31

(The abrupt change in the thicknessof the metal is fundamental to the explanation of a hot tear in this instance.)
167, line 30 The corrosion within the groove ...

... where the mould changes abruptlyfrom a relativelythin to a relativelythicksize.

The reproduction of the radiograph is so poor that none of the details cited in the text and in the caption to this plate are observable, yet they are extremely important and constitute the major evidence for the casting technique and the arguments I have made for reconstructionof that technique. More important, however, is the unfortunate choice of this particular radiograph for publication. This film was taken under special conditions (the pick-axe was buried in very fine copper shot to remove secondary radiation in order to achieve a uniformity of film density and a clarity of outline of the object). For precisely this reason, however, it is very misleading as far as the structureof the casting is concerned, for it shows an apparently constant thicknessof metal throughout the object and completely obscures the existence of the socket cavity. It reproducespoorly the details of porosity, cleavage, hot tears, dross, and shrinkage all of which are treated in considerable detail in the text. (Printsof the x-radiographthat should have appeared as P1. IIa are available and will be furnished on request.) in and P1.IIb The typesetterinterchanged the words groove protuberance the plate labels. The uppermost arrow at the left should be labeled protuberance, the lower and arrow at the left should be labeled groove. P1. IIa
HEATHER LECHTMAN

Laboratory for Research on Archaeological Materials Massachusetts Institute of Technology Room 8-432 Cambridge, Mass. o2139 U.S.A. 173

SURVEY OF EXCAVATIONS IN IRAN DURING 1968-69

I. EXCAVATION REPORTS

BdbdJan The fourth season's excavations at BibA Jdn Tepe started on July I4th and continued for two months. As it was intended that this should be the final season at the site, our main effortswere directed towards completing the plan of the large building complex comprising the Fort and " Painted Chamber ", discovered last season on the East Mound. A full report of the 1968 season appears in the main section of this journal and should be read in conjunction with this much briefer notice. Our main discoveriestook place in the central area of the so-called Fort. Here, the central room 4, is shown to have had a flat roof supportedprobably by four columns arrangedin the form of an irregular square. Set in the west wall was a large fireplace with an elaborately fluted plaster surround. In the

'

.,/

U.

S.

S.
-

R.

*TABRIZ o Tepe .HoFbavin *r SHosolO SKaluroz ,Tureng

S.,astam

CASPAN
SEA .,

....
Tepe

Takht-i-Suloiman "(

* Gheykaryeh TEHRAN

MESHED i-.. M l 7
I

Tpe iTepe *'.

iGar

*HAMADAN
Jan

Darch

*Bb1a

c4

ChoSadsefd

. ,,Khorromobad SISFAHAN *
S~ su

Sulaiman

c*cS Mish HaFtTepe l

ishpOr hp G Prsepolis Po SHIRAZ

Shahr-iSokhte Sh-ohdad
ZAHDAN ZAHEDANt .

...

..
A RA BI 8IA

>

"TepeYahya Sir&F

0
1. I f 1

50o 1

Kilometres

175

176

JOURNAL

OF PERSIAN

STUDIES

north wall, a central doorway, flanked on either side by no less than five reveals, gave access to a long room to the north. Measurementstaken on the collapsed wall debris found in the centre of the room allowed us to estimate that the total height of the walls must have been at least Io m. There is, however, insufficientevidence for us to calculate whether the columns also reached to this height or whether there was a second storey. Enormous beams found lying on the lower Level III floor indicate how the room was roofed in its earliest phase, but there is some evidence to suggest that in the reoccupation period II, this central area was unroofed and formed an open courtyard. The transition from the rectangular central rooms of the Manors on the summit of the Central Mound, to a square columned hall, may reflect the beginnings of the new architectural tradition that was to culminate in the thirty-columned halls of Godin Tepe and Pasargadae, and the Apadana of Persepolis. Certainly, as with the Manors, one is left with the impression of builders feeling their way with a new form of experimental architecture with which they were not yet familiar. The " Painted Chamber "-which this season'swork has shown to be either contemporarywith, or slightly later than, the Fort, is already more assured and obviously planned with considerable forethought. West and east of the central room were two long rectangular rooms, 3 and 5, the first of which was completely excavated in 1968. Room 5 proved to be nearly identical, except for the presence of a small stepped plinth in the north-east corner. Tucked behind the plinth was a beautifully modelled and painted statuette of the lower half of a leg encased in a laced boot, and with three anklets. One assumes that this either formed part of some larger statue which once stood on the plinth or else reflects the influence of contemporary " boot vases " found in Eastern Anatolia. The only other room completely excavated was room 9, south of the courtyardflankingthe " Painted Chamber ". This proved to be a long narrow living room, again with an elaborate fireplace to the north. The lower sections of the walls had been whitewashed, and the floor was covered with fragments Luristdn of beautifulgenre pottery. Carefulstratigraphicaldigging through the reoccupation levels of this room, and those of the courtyardimmediately to the east, confirmedlast season'spottery sequence, and also gave us our second bronze-a two-faced Janus head, affixed to the top of a screw. It was Lurist.n found in a deposit dating to the latest phase ofII, stratigraphically just above an earlier level which had and Niish-i Jan pottery, and two elbow fibulae. produced mixed genreLuristdn GOFF CLARE Bastam Das Deutsche Archaologische Institut in Teheran fuihrtein der Zeit vom I. Juli bis 31. August eine Grabung an dem urartaischen Platz Bastam bei Qara Zia ed-Din in Nordwest-Azerbaidjandurch. Es war die erste lIngere Kampagne in der 1967 aufgefundenenAnlage (vgl. IranVI (1968), S. 166 und VII (1969), S. 188). Hauptaufgabewar die Erstellungeines Planes der Festung und der anschliessenden Siedlung, die Klarung der damit verbundenentopographischenFragen und die Gewinnung datierbarer Keramik von verschiedenen Stellen der Anlage. Der durch zum Teil noch bis in 5 m hoch erhaltenen Mauern und Felsabtreppungen als Spuren verschwundener Mauern befestigte Festungsbergist rund 850 m lang, etwa 400 m breit und erhebt sich 18o m hoch tiber der Ebene. Bastam ist damit mit Abstand die gr6sste der bisher auf iranischem Boden festgestelltenurartaischenFestungen und folgt in der Gr6ssenach Vankalesi und Toprakkalebei Van. Die am Fusse des Festungsbergessich erstreckende m Siedlung ist rund 6oo00 lang und etwa 300 m breit. Ihre Gebude sind grossenteilsnoch als Schuttwalle im Gelande sichtbar. Eine besondere Stadtmauer scheint nicht angelegt worden zu sein. Auf Festung und Siedlung bezieht sich die Angabe der im Archaologischen Museum in Teheran befindlichen " Bastam-Inschrift" die ausser dem Bau eines Haldi-Tempels den Namen der Stadt, Rusahinili, erwahnt. Die gesamte Anlage ist in der Zeit Rusa II im 7. Jahrhundert v. Chr. gegriindet worden und die Siedlung hat nach den Ergebnissen der diesjahrigen Grabung maximal 60-80 Jahre bestanden, das v. heisst, sie ging beim Ende des Reiches Urartu um 600oo Chr. zu Grunde. Fuir diese Annahme spricht das Vorhandensein von nur einem Stratum im Stadtgebiet und zusitzlich die Keramik, bei der es sich im Stadtgebiet vorwiegend um einfache Gebrauchskeramik spiturartaischer Zeit handelt. Die Siedlung ist mit der Festung durch ein umfangreiches Kammertor verbunden, das in dieser Kampagne zusammen mit flankierenden Bauten in Form von Schnittgrabungen untersucht wurde. Dabei konnten drei

SURVEY

OF EXCAVATIONS

177

aufeinanderfolgende Bauphasen deutlich unterschieden werden. Ein ailterer, langrechteckiger Hallenbau mit Risaliten auf seiner Aussenseite bildet die il1teste Phase. Er ist aus gr6sseren B16cken in sorgfiltig aufgeschichteter Bruchsteintechnik errichtet. Gegen diesen Bau 1iuft eine Risalitmauer, die ihre Fortsetzung in der grossen, noch erhaltenen Aufwegmauer auf die H6he des Festungsberges findet. Als dritte Phase wurde die eigentliche Toranlage mit zwei starken, etwa quadratischen Tiirmen, an diese Risalitmauer angebaut. Auf der Aussenseite des langrechteckigen, ilteren Hallenbaus wurden einige, durch herabgerutschte B16cke der Mauern stark beschiidigte Skelette freigelegt und in Fundzusammenhang mit diesen Skeletten Bronzenadeln, Bronzeringe mit Siegelsteinen, Karneolringe und

"

SHOTLU
A.(AE.

SANGAR

" .... H ....''

U..S..R.

\BASTArI

S
. :D3ULFA

rr

-,r

J~

rz ALE "'I
KIZKALE.....
K140

," .
Y

U . EVOGHL

j
TURKEY
,,

MA MRAND

WAZIRI ." SHAHPUR KALE

? s-"
l(ARNARO*

SOFIAN

.
?

K . '... L.... GAUHiARC(iNt. .V

. -.' :...:'... .. .: . ..

".

G9

eine Eisenfibel. Da einer der Bronzeringe ein typisches armenisches Kreuz zeigt, scheint die Annahme gerechtfertigt zu sein, dass es sich bei diesen Toten um die Opfer einer Katastrophe aus friihchristlicharmenischer Zeit handeln k6nnte. In die urartaiische Zeit passen diese Funde stilistisch nicht. Ausser dem n6rdlichen Tor zur Festung wurde auch das Siidtor in seinen aiusseren Teilen freigelegt. Hier ergaben sich zwei urartaiische Bauphasen, darunter als Anbau ein Pfeilerraum, der bisher nur teilweise freigelegt werden konnte. In armenischer Zeit wurde der Raum des Siidtores durch Hiitten eingenommen, entsprechende Keramik und eine Bronzemiinze wurden in der Grabung gefunden. Auf der H6he des Festungsberges ergab in der Nihe der grossen Steinterrasse, die vorliufig als wahrscheinlicher Platz des von Rusa II in seiner Steininschrift erwaihnten Haldi-Tempels gelten kann, ein Schnitt zwei Riiume. Der untere, am Hang gelegene, angeschnittene Pithosraum erbrachte Keilinschriften und Massangaben an den grossen Vorratsgef'assen. Der hangaufwirts anschliessende

178

JOURNAL OF PERSIAN STUDIES

Raum erbrachte grosse Mengen der typischen urartaiischenrotpolierten Keramik des 8.-7. Jahrhunderts v. Chr. (sog. Toprakkale-Ware), wie sie auch in der gleichzeitig mit unserer Grabung laufenden Grabung in Haftvan-Tepe bei Shahpur (Charles Burney) gefunden wurde. Schliesslich wurde ein zu Fiissen des Festungsbergesgelegener Tumulus angegraben, der aus Flusskiesbdndern und Schichten von Material eines alteren Tepe aufgehauft ist. Die Bestattung konnten wir in dieser Kampagne noch nicht erreichen. Die bemalte Keramik aus den Schichten des Tepematerials geht bis in den Anfang des 2. Jahrtausends v. Chr. zuriick. Ausser der eigentlichen Grabung wurde die Surveytitigkeit in Azerbaidjan weiter fortgefuhrtund auch in diesem Jahr wieder einige urartdiische Anlagen neu gefunden. Sie werden durch entsprechende durch die typische Mauertechnik oder durch die ffir die Urartter bezeichnenden FelsabtKeramik, reppungen als urartiische Anlagen ausgewiesen. Die in den Jahren 1967, 1968 und 1969 festgestellten urartiischen Anlagen sind auf der Obersichtskarteangegeben. Ausser diesen urartiischen Neufunden sind eine Reihe weiterer vor-und friihgeschichtlicher, mittelalterlich armenischer und islamischer Befunde gemacht worden, die, wie die Survey-Ergebnisse der Jahre 1967 und 1968 in den Archdologischen aus Mitteilungen Iran,N.F. II (1969), S. 7 ff und wie die Ergebnisse der Bastamgrabungin den AMI III, vorgelegt werden sollen.
KLEISS WOLFRAM

Bishdpar The historic city of Shdhpfir, known as Bishaipir, was built in 266 A.D., in the twenty-fifth year of the reign of ShdhpUrI. We owe this discovery to the bilingual inscription in Pahlavi/Arsacid and Sasanian written on one of the columns at the so-called " place of offerings ". The French Archaeological Mission carried out excavations at Bishaipiron five occasions between the years 1933 and 1940, revealing a number of handsome monuments. Unfortunately their work could not be continued, and after a break of several years the Archaeological Service sent an Iranian archaeological mission to Bishapiir in 1968. The Archaeological Mission began its work near the northern wall of the city, exposing a long stretch of the perimeter wall more than I50 m. in length. In its first phase of construction the outer face of the wall appears to have been supplied with a virtually continuous fagade of rounded towers, each 7 -5 m. in diameter with an interval of only 40 cm. between them. The tallest examples still stand to a height of 3 m. During a second period of construction, again within the Sasanian period, many of the buildings of the town were transformed-including the outer face of the wall. Thus in each seriesof three towers, one tower was removed. Apart from these transformations,another monument, characteristicof the Sasanian period, was built beside the River Shahptir, only Io m. from the perimeter wall. In this building-the function of which is still uncertain-a central iwan is to be found with two wings and two corridors. Engaged columns on rectangular bases also enliven the fagade. The remains of the Islamic buildings recovered so far fall into three distinct periods of which the most splendid date from the late eleventh and early twelfth centuries A.D.
ALI AKBAR SARFARAZ

Ganj Dareh Tepe Excavations were resumed at this small mound in Kermanshih District for nine weeks from June to late August 1969. The research was sponsored by the Department d'Anthropologie, UniversitC de Montrial and financed by the Canada Council, with the collaboration of the Archaeological Service of Iran. This was the second full season of work here, following a brief sondage in 1965 and two months of excavation in 1967 (see Iran V and VI). These had shown that (a) early phases of the Neolithic were represented extending to the eighth and perhaps ninth millennia, (b) there existed a village level constructed of brick with the buildings unusually intact due to its partial destruction by fire, and (c) evidence of an early stage of pottery making was present.

SURVEY

OF EXCAVATIONS

179

In 1969 we concentrated on exposing in the central part of the mound larger areas of the last four levels (tentatively called A, B, C and D) revealed in 1967. The basal deposits of the site (Level E) were not investigated in 1969; thus we are not yet certain as to its relation to subsequent levels nor whether solid architecture is present. It is this level, or part of it, which may date to the ninth millennium if a single radiocarbon determination from the 1965 sondage (GaK-8o7: 8450?I50 B.C.) is B.C., accepted. In the " burned village " level (D) about 60 square metres have now been exposed. This has revealed a more complex situation than suspected earlier. It is made up largely of small rooms, alcoves or cubicles, most of which seem far too restricted to have been used for living purposes. Most are rectangular but some are round and silo-like and were perhaps storage places, though no traces of stored materials were preserved. Some of the walls have round or oval " portholes " which are often sealed by means of flat clay discs. In several rooms there were clay containers resembling storagejars, as well as small domed " bins ", all built against the walls. These were immovable containers, but there was also a very large portable " storage jar " about 8o m. high (already reported in 1967 but in not removed until this year) standing on a plasteredsurfaceo. a cornerof one room. All these containers " bins " are made of plaster-likemud, apparently sun-bakedoriginally and preservedby the action and of the fire which destroyed the village. The structuresin the village are of the cluster type, built against each other with no paths or lanes between them and no clearly defined doorways. Most are built of straw-temperedplano-convex bricks, some nearly I m. in length, laid between layers of mud mortar. The walls, which in some cases are still standing to heights of nearly 2 m., are faced inside and outside with mud plaster. In addition to the brick walls there is an unusual form of walling (unknown to the writer elsewhere in Iran) in which alternating layers of mud and fine plaster were built up in strips and then coated with plaster on both faces. There are no hearths or other signs of domestic occupation in the rooms though some stone tools and animal bones occur. If the sectors so far excavated in this " burned village " represent a storage area then possibly the living areas are to be found on the flanks of the mound. It is not yet clear how many building phases are represented in this level. Several (unpublished) radiocarbon determinationsrecently obtained at the University of Pennsylvania suggest that it was occupied in the second half of the eighth millennium B.C. Level C was not well representedin this season's excavations and little can be said of it at present. A larger area of Level B was exposed including several rectangularrooms with rich occupation deposits and hearths and a few pieces of crude pottery. The walls are made of brick as well as chineh.Another recent Pennsylvania radiocarbon determination suggests this level dates to about 7000 B.C. The last I -5 m. of deposits at the site are grouped as Level A although it probably correspondsto more than a single occupation. It is badly leached and disturbedby roots and animal burrows,and very little remains intact. Bricks are visible in some profiles but are badly decomposed and very few wall sections could be traced either by following them into the profilesor by scrapinghorizontally. However, there is a well-made domed structure preserved, possibly an oven or kiln, with several strata of ashy materials. We recovered none of the curious sherds with crescent-shapedor " fingernail" impressions
found in 1967.

The flint tools seem to be very similar typologically from base to top of this site. The assemblages are characterized by backed bladelets, truncated blades, heavy flake scrapers,end-scrapers,many use-

retouched pieces and some fine bladelet cores. Some blades have extensive zones of polish and may be elements of reaping knives. For the first time we recovered a very few geometric microliths (trapezes, segments of circles). No obsidian was used. No ground stone axes or celts occur and only a few fragments of polished stone vessels, but there are many rubbing stones, pestles and large limestone mortars. Clay animal figurines and cones are abundant, especially in Level D, and several more female figurines were found including the " stalky headed " type known at Tepe Sarab. Up to now very little in the way ofvegetal material has been recovered although flotation techniques have been employed in both seasons. Study of the faunal remains by Dexter Perkins suggests that goats at least were domesticated here. This interpretation may be further strengthened by discovery in 1969 of hoof prints, probably goat or sheep, on a number of the piano-convex bricks from Level D. Apparently the animals had walked over the bricks while they were being sun-dried.

180

JOURNAL

OF PERSIAN

STUDIES

For the first time we have recoveredprehistorichuman skeletal remains at Ganj Dareh. A complete skeleton of an adult, flexed and without grave goods, occurred near the top of Level D rubble and thus is of later age. In the " burned village " itself some very burned bones from several alcoves represent three children and an adult. These may have been burials which were later calcined by the fire. Perhaps the most significant general result of the work so far at Ganj Dareh is that at last we have a Zagros site which falls, at least in part, in the eighth millennium B.C. The rich occupation deposits of Level B and the well-preservedarchitectureand early pottery of Level D promise to throw much light on a millennium for which hitherto no dated materials were known in the Zagros. Another season of excavation at Ganj Dareh is planned, especially to investigate the antecedents of these in Level E.
PHILIP E. L. SMITH

Gheytaryeh The Gheytaryeh district of north Tehran, where the excavations took place, is situated in a chain of hills at the foot of the Elburz range, between Chizar, Farmanieh, Darrous and Gholhak. Here local building operations had revealed the existence of an important Iron Age cemetery probably dating from I200 to 800 B.C. The remains of this civilization, known by its grey-black pottery, have already been discovered at Darrous and Chizar in the area around Tehran, and also at Khorvin and Chendar in the Karaj-Qazvin area. Pottery discoveredin the valleys of Gilan (Amlash-Daylamdn), in the Iranian Talish, at IHasanlii IV, in Tepe Giyan I, in cemeteries A and B at Sialk and at Tepe (Ddmghdn), dating from the late second millennium and the beginning of the first millennium B.C.,can all be compared to the pottery .Hisr found at Gheytaryeh. One could say, therefore, that the Gheytaryeh pottery is one sample of the arts that came to be practised by the immigrants who, according to archaeologists'theories, entered Iran towards the end of the second millennium B.C. by routes running both east and west of the Caspian Sea. The dead from Gheytaryeh were buried haphazardly, with some on their left side and others on their right, but usually with the face directed towards the sunrise. The graves themselves consist of earth pits. Usually there were between five and twenty pottery vessels next to each skeleton, apart from other small objects in terracotta and bronze. The size of the ancient cemetery-it occupies an area close to 15,ooo square metres-is proof of a long period of civilization in this region and its environs. Unfortunately the numerousmodern buildings that now exist at Gheytaryeh may make it difficult to locate the settlement that once went with the cemetery.
KAMBAKHSH FARD

GodinTepe Full scale excavations were again conducted by the Joint Canadian-Iranian Expedition of the University of Toronto, the Royal Ontario Museum and the Archaeological Service of Iran at Godin Tepe in western Iran from mid-June to late September 1969. The excavations were supported financially by the Royal Ontario Museum, the Harvie Foundation, the Babylonian Collection of Yale University and the Ford Foundation. Excavations concentrated on (I) continuing clearance of the Period II or Median structure;
(2) extending downward the " Deep Sounding " over an area of some 700oosquare metres further into the Period III or second millennium B.c. deposits; and (3) opening sizable areas of the upper Period III building levels which appear at the surface of the mound in most places thus far tested. The Period II structure: The 1967 excavations suggested that the whole of the Upper Citadel was formed from the remains of a large fortified manor house dating to Period II. The 1969 excavations proved the point. We have now cleared a single structure, albeit built in units over a period of time, covering some 3600 square metres. The fortification wall laid bare over some 65 m. in 1967 was found to continue for another 40 m. further east along the steeply eroded north face of the Upper Citadel. This newly excavated segment of wall is marked by six pilasters or bastions, a large tower toward its eastern end, and arrow slots in and between each pilaster and at suitable points in the fagade of the

Pl. Ia. Terrasse de Masjid-i Solaiman.

Pl. Ib. Masjid-i Solaiman, le grand escalier.

Pl. Ic. Masjid-i Solaiman, fagade du templed'Hiracles.

P1. IIb. Masjid-i Solaiman, Statuette de danseur, bronze. Face et dos.

en P1. Ha. Masjid-i Solaiman, support bronze.

P1. IRd. Masjid-i Solaiman, animal en bronze.

P1. IIc. Masjid-i Solaiman, torse de guerrier. Pierre.

P1. Ile. Masjid-i Solaiman, lampe en bronze.

tWte P1. IIIa. Masjid-i Solaiman, d'homme.Pierre.

d'orant. Pierre. Pl. IIIb. Masjid-i Solaiman,bas-relief

Pl. IIIc. Masjid-i Solaiman, tWte femme. Pierre. de

Pl. IIId. Masjid-i Solaiman, tWte d'homme. Pierre.

Pl. IVa. Masjid-i Solaiman, installation de l'jpoque achiminide.

Pl. IVb. Masjid-i Solaiman, poterieachmie'nide.

Pl. V. Mazir-i Sipahsadlr,Tabas.

PB Pl. VI. Takht-iSuleimdn, Pfeilerhalle von Osten.

Pl. VIIa. Takht-iSuleimdn, E vonSiiden. Hof

Pl. VIIb. Takht-iSuleimdn, Bruchsteinfu

P1. VIIIb. Kz7ficinscription in the Macjid-i Sar-i K Pl. VIIIa. Remains of the mihrdb of the Masjid-i Imdm Hasan, Ardistdn.

and detailsof thedome drum Pl. IXa. Mausoleum Shaikh'Abdas-Saldm, showing of the tomb and adjoining secondary chamber. pattern, themihrab

Pl. IXb. Minaretof theMasjid-iJami', Gurgdn, showin

and Pl. Xa. Detail of theminaret the Masjid-i Plmenar,Sabzavdr, patterns inscription. showingbrick of

Pl. Xb. Squinch at

Ribd.t

Mdhi showing cannulargroove.

Pl. XIb. Djaffarabad, 1968-69, four de

Pl. XIa. Bronze figurine from Shahr-i Sokhte-29

cm. in height.

d'en P1. XIc. Djaffarabad, i969, tombe

Pl. XIIa. Suse, acropole.

salle hypostyle. Pl. XIIb. Suse: apadana,

du Pl. XIIc. Suse: apadana,la salle hypostyle palais achdmenide.

SURVEY

OF EXCAVATIONS

181

tower. We have yet to find the eastern limit of either the wall or of the structure immediately inside or south of the wall. Here lies a double bank of six great magazines, each store room entered from a long corridor or hallway running down the east-west axis of the building. The six magazines off the north corridor are each approximately 2-75 m. wide and 8-75 m. long. The walls separating the rooms are approximately 1-25 m. thick, while the wall separating the north and south banks of magazines, which was once the outside wall of the fortress, is 2-50 m. thick. The south magazines are approximately 7-25 m. long and 3 o m. wide. The south wall of the south magazines was apparently the south wall of the entire fortified complex in its final stage. Between these magazines and the great columned hall uncovered in 1967 we found a series of small rooms, difficult to describe without reference to a detailed plan, as well as the remains of two rather fine stairways which once gave access to the second storey that covered this whole wing of the building. Finally, to the west of the great columned hall we discovered yet another, smaller hall with its roof supported originally by a double row of four columns. The western wall of this hall is the western wall of the complex. Stratigraphic and structural evidence indicates that the original structure included the two columned halls and an open courtyard or collection of small rooms immediately east of the great hall-approximately what we excavated in 1967. To this was added first the north bank of magazines (note that the wall separating the north and south magazines contained arrow slots). A third stage of construction is represented by the south bank of magazines. Finally, since our resident local Median prince now had a building shaped roughly like a " C ", he filled in the area between the south magazines and the great hall with the rooms and stairways mentioned above. Eventually, the entire structure lost its original function and was taken over by people who blocked off many of the doorways and passages (the entire east wing apparently went out of use) and who cut up the monumental spaces with small, rather jerry-built curtain walls. This phase was in turn followed by a " squatter " occupation of peoples who may be said to have simply camped in the much-decayed ruins of a once grand building. Preliminary field analysis of the pottery from this structure and consultation with colleagues working on sites in the same general time range suggest perhaps a late eighth century, but more likely a seventh-early sixth century date for Period II. On the whole, any significant Achaemenid occupation would seem unlikely and we appear to be quite firmly in the Median time range. Whether eventually we shall be able to date the several stages in the construction of the building on the basis of the pottery found seems doubtful. Small finds other than pottery were rare, and all data indicate that the building was peacefully abandoned. The " Deep Sounding": An area of some 700 square metres has been devoted to a stratigraphic sounding that we hope will eventually test the remains of Periods II to V (mid-fourth millennium B.c.) through a depth of deposit of about 15 m. In 1969 this project carried us some 3 m. deeper into the Period III levels, and the sounding now stands at a total depth of something over 6 m. We completed the clearance of Level 111:2, begun in 1967, and defined Levels 111:3 and 111:4. This, the lowest level reached thus far, represents the architectural remains of several small village houses fronting on a packed earth street. The walls were preserved to an unusual height. A great many pots were found smashed in situ on the stone-paved floors and courtyards which are a hallmark of the level, and small finds in general were unusually numerous. The skeleton of a man crushed by falling roof debris was found on the stone floor of one room. All of this evidence suggests some kind of violent destruction, and an earthquake is perhaps the best explanation (cf. the destruction of Level III:2). There may be evidence in both the architecture and the small finds for specialized craft functions in the area excavated. A guess date for this level is c. seventeenth-sixteenth centuries B.c. Attempts to open up large areas of the upper Period III construction level elsewhere on the Citadel were not so successful. Unfortunately, the area chosen turned out to have been comparatively open and few structures were recovered. Work here did demonstrate, however, that there had been a limited Period II occupation, presumably contemporary with the fortified manor house on the Upper Citadel, dug into the upper Period III deposit. The remains suggest a few flimsy structures. All other areas on the Citadel and Outer Town thus far tested have shown Period III materials immediately below the present surface.
16A

182

JOURNAL

OF

PERSIAN

STUDIES

Excavations in 1971 should see the end of work on the Median fortification. We shall also continue with the " Deep Sounding ", and may return to do further work in the Period III cemetery on the south side of the Outer Town mound. T. CUYLER YOUNG, JR. HaftavdnTepe The second season of excavations at Haftavdn Tepe began in the second week of July 1969 and continued till mid-September, the camp being closed on September 18th. There was a staff of fourteen and a work force averaging eighty in number. Work was continued in several of the areas opened in the first season, while new trenches were opened up both on the citadel and below. The intention of investigating as many periods as possible with the minimum of clearing of overlying levels was fairly well achieved as before, though the later levels on the citadel proved to be as severely damaged as in the previously excavated areas. The earliest level to be found was in a new trench (R) at the foot of the north-eastside of the citadel, where recent quarryingof soil made it possible to examine the earlier phases of the citadel itself. Rather unexpectedly, since at a height of 17 m. above datum at the spring, Early Bronze II pottery was immediately found, together with a well-preserved and compact building plan, of very different character from any of the later levels at this site. The graphite burnish on some of the pottery gives a clear parallel with the Early BronzeII period at Yanik Tepe and also with pottery from sites in Georgia: a date around 2000 B.C. is indicated. The east end of the Early Bronze mound seems to have been near the lower end of Trench R, while the west end was probably not far east of the major stone building in Trench E found in 1968. Thus the present shape of the citadel was brought about by that of the underlying Early Bronze mound. The outer parts of the site are therefore almost certainly entirely of the second millennium B.C.and later. A ceramic sequence is being obtained from a sounding (JX) of 10 I0 m. within the area (J) o in 1968 at the east end of the site, near the spring. Seven levels have now been found here, the opened lowermost being very well preservedand clearly not far above virgin soil, although there is at least one more building level as yet unexcavated. The greater part of the pottery from all these levels is unpainted; and grey ware of Iron I (Ijasanlfi V) type occurs commonly in only one level, the fill of a very large pit. This is one of the indications that the site was perhaps only very sparselyinhabited during the Iron I period (c. 1350-1ooo B.c.). In the earlier levels of this sounding, perhaps dating back to c. 1600 B.C., there is a small proportion of painted sherds exhibiting a wide variety and in part demonstrating connections with Trans-Caucasia to the north. At the west end of the site a major extension to the 1968 excavations was begun, but the great depth of deposit made it impossible to complete this to the level of the great stone building till the next season. In these areas (P2 and Q) several building levels were distinguished, though the architecture was in general poorly preserved. These levels appear to date back not beforec. 1000 B.C.,as the contents and stratigraphicposition of two burials, each of a young girl, indicate. For the first time the positions on head and neck of several types of personal ornament found in the 1968 season became comprehensible. These ornaments include earrings, discs, head-band and tassels of bronze and a variety of beads of glazed frit, shells, etc. Underneath the head, but separated from it by a line of burning in the
soil, was a cylinder-seal of local design, not so far attributable on stylistic evidence to any clearly-known class, though possibly with Mitannian affinities. The principal achievement of the 1969 season has been the uncovering of a large part of the plan of the citadel buildings belonging to the Urartian period. Pottery, including several vessels of the characteristic fine red polished ware, and other finds make this dating beyond any doubt, though a more precise date within the period c. 850-c. 6oo00 B.c. would be hard to venture: on the whole this building level, planned to one over-all scheme for the citadel and on one alignment, seems most probably to be of the eighth or early seventh century B.c. At first a line of six column-bases was thought to indicate the fagade of one major building, though this was situated very close to another; but finally it was realized that these columns belonged not to any external portico but supported an upper floor,

SURVEY

OF

EXCAVATIONS

183

in whose debris was found some of the fine red polished ware, also in similar context in a small room forming a corner tower close to the north edge of the citadel. This was part of the building rather than of any defences. Next to it was a small gateway. This building level was constructed directly on top of the debris of a fire, and was in its turn burnt: in this burning were found fragments of bronzes and of beads, including many of yellow glass, scarcely found in the 1968 season and therefore probably later than the contents of burials in the outer parts of the mound. There seems to have been fairly considerable stone robbing. There is evidence of the rebuilding of at least the brickwork of the perimeter wall along the very short stretch of the north side of the citadel so far excavated. The level immediately overlying the Urartian buildings may well be Median (i.e. early sixth century but is so badly destroyed by pits that no very clear plan emerges. There seems little hope of B.c.), finding areas of the citadel where the post-Urartian levels may be less damaged: those parts next to the perimeter seem the least destroyed. But the Urartian level will assuredly repay more extensive excavation. Overlying the Urartian perimeter wall was one constructed of a few stones and mud packing, all that has survived. Above this was a wall built without use of any stones but with large flat square mud-bricks averaging 40 x 40 x 10 cm. or slightly more: these bricks and an approximately semicircular tower forming part of these defences seem to suggest a Sasanian date; but the occurrence immediately over the remains of this wall of two Sasanian burials comparable with those found elsewhere on the citadel shows that these defences must be placed very early in the Sasanian period. Otherwise they could be Parthian, though there is no parallel to support such a dating. Fewer Sasanian burials were found than in the 1968 season, showing that they were probably mostly confined to the north-west part of the citadel. Only when the pottery from the post-Urartian levels has been thoroughly analysed will any certainty on their dating be possible. The two seasons of work at Haftavan Tepe have demonstrated that it was perhaps only partially occupied in the second millennium B.C., after a long earlier duration. In the tenth and ninth centuries B.C., however, it became a large town, which, contrary to the writer's previous opinion, continued as a major town under Urartian rule. The distinctive Urartian pottery was possibly used only in the citadel, and there in the better rooms alone: elsewhere the local wares, made in the tradition flourishing before the establishment of Urartu in the ninth century B.C., were still in use. Several more seasons of excavations are planned, the next one for 1971. C. A. BURNEY

Among other excavations conducted by the Iranian Archaeological Service during the summer of 1969, work was again resumed at the important eighth to sixth century cemetery of Kaluraz, which is situated 55 km. south of Rasht.

Kaluraz

La seconde campagne de fouilles sur la terrasse sacree de Sar-Masjid, a Masjid-i Solaiman, dans les montagnes des Bakhtiari, a dur6 six semaines, au printemps de 1968. Sa modeste equipe ne comprenait que son chef et Madame T. Ghirshman, t qui vint s'ajouter M. Naji Fathali, le technicien du " Survey Operations of I.O.E. and P.C.", aimablement mis t notre disposition par le Consortium du PWtroleIranien, et qui assuma tout le travail des relev6s d'architecture mise au jour lors des fouilles. Ahmad Ettemadi, le chef d'Cquipe qui a travaillk sur mes chantiers depuis 1934, vint nous rejoindre aprbs avoir termind sa collaboration avec le P. Steve, t Suse. Le repr6sentant du Service archdologique de l'Empire Iranien 6tait M. Gandjavi, en qui la Mission trouva un organisateur dynamique et un actif fouilleur. Au nom de la Mission je lui exprime ici notre gratitude pour sa loyale et sincere collaboration.*
* J'6xprime ma profonde gratitude &Mr. C. A. E. O'Brien, Chairman and General Managing Director, Iranian Oil Operating Companies, ainsi qu'd Mr. G. Birks, Field General Director, pour leur large hospitalit6.
16B

Masjid-iSolaiman

184

JOURNAL

OF PERSIAN

STUDIES

On sait par notre court rapport sur la campagne pr6cedente 'a Masjid-i Solaiman, publie dans cette meme revue (voir Iran VI (1968), p. 165), que le plus grave problkme qui s'6tait pose t moi pour une exploration methodique de la terrasse sacree, etait que pres de la moiti6 de la surface de celle-ci 6tait couverte d'un cimetiere oii les inhumations ne se pratiquaient plus depuis pres d'un siecle. Mais l'obstacle pour nos travaux 6tait d'une importance majeure: c'est ainsi que le degagement du temple d'Heraclks, decouvert en 1967, fut stop6 devant les tombes et n'a 6t6 reconnu que sur a peine la moitie de sa superficie. Grace a l'intervention de la plus haute instance du pouvoir en Iran, cette difficult6 fut aplanie et nous avons eu la possibilit6 de poursuivre nos recherches, sans pour cela porter atteinte aux sentiments traditionnels des habitants du grand centre petrolier qu'est Masjid-i Solaiman. la La Pl. Ia presente l'aspect de la terrasse "a fin de cette seconde campagne. On y distingue, au un grand batiment dont la destination reste encore 'adeterminer apris que son degagepremier plan, sa ment aura 6t6 termin6, et, a% gauche, le temple d'H6raclks entierement reconnu. La P1. Ib donne droite. La P1. Ic montre l'idee du grand escalier de 25 metres de largeur qui, sur la P1. Ia se trouve a% des dallages de pierre sur le sol, qui l'tat actuel de la faqade de temple d'Heraclks avec les restes indiquent, peut-etre, des voies processionnelles, ainsi que des gradins ou banquettes faits avec de puissants blocs de pierre appuyds contre le mur de la faqade perc6e de deux portes donnant dans l'antecella. Un certain nombre de petits bronzes provenaient de ce sanctuaire, offrandes votives sans doute, parmi lesquels un danseur (P1. IIb) qui, tout comme on peut le voir de nos jours, tient en dansant sa main gauche derriere la nuque; une lampe (P1. IIe); un petit animal (P1. IId); et un pied ou support en forme de g6nie f6minin ailk, qui se termine en bas par une patte griffue (P1. IIa). Des sculptures de l'Cpoque parthe sont venues au jour lors du degagement du temple ou de ses approches. Parmi elles, il faut mentionner un torse d'homme arm6 d'un poignard enfonc6 dans sa tunique sur la hanche droite - oeuvre en ronde bosse mais d'une platitude que sauve l'interessant decor du tissu fait de losanges encadrant un motif 'aquatre petales (Pl. IIc). La partie inf6rieure d'un autre bas-relief presente un noble Parthe dans la pose d'orant, son bras droit (dont on ne distingue que le coude) lev6 en signe de v6neration, et qui tient dans sa main gauche un objet indetermin6 (fruit?). Lui aussi porte sur la hanche droite un poignard dont on ne voit que la poignee, le reste 6tant gliss6 dans une poche faite dans la tunique. Celle-ci, ainsi que le pantalon long gliss6 dans la chaussure que serre un ruban atboucle, sont faits d'un riche tissu orne de losanges et de bandes de motifs en rinceaux au dessin lin6aire (P1. IIIb). La presentation sur les sculptures parthes de ces cofteux tissus, ne souffre pas le moindre pli susceptible de nuire 'al'appreciation de la valeur du vetement, et contraste avec d'autres repr6sentations d'orants dont les vetements d'un tissu plus modeste ne sont pas depourvus d'un certain mouvement de vie. Parmi les tetes mises au jour, il faut citer celle d'un homme barbu, aux cheveux trait6s par des figure glabre, aux yeux rangees de bouclettes en bossettes (Pl. IIId); une autre tete, d'adolescent, "a la bouche petite aux lvres serrees, porte quelques lointains reflets de l'art " grecogrands ouverts, bouddhique" son contemporain (P1. IIIa); et, enfin, une petite tate de femme en ronde bosse (P1. IIIc). Les recherches de cette annie ont enrichi nos connaissances, parmi tant d'autres, de deux r6sultats auxquels nous attachons une importance primordiale. J'attribuais la cr6ation de cette terrasse de Masjid-i Solaiman i l'6poque ach~minide, et ceci depuis ma premikre visite sur les lieux, il y a exactement vingt ans. Plusieurs raisons m'y invitaient: la conception mime de la terrasse; I'appareillage de son coffrage; l'drection de son podium, quoique remanik et moins bien conserve que celui de Bard-6 Nechandeh voisin. La campagne de 1968 nous a apportC des preuves supplkmentaires susceptibles d'6tayer cette hypoth~se et de maintenir une date qui correspondrait a l'6poque oi les tribus perses vinrent s'installer dans cette partie des monts Zagros. Derrikre le temple d'H6racls et un niveau plus bas que son sol, entre ce sanctuaire et la montagne contre laquelle est adossde la terrasse,. nous avons mis aujour des batiments (P1. IVa) oii fut dfcouverte une poterie dotte de versoirs trbs caractfristiques, connue en Iran et attestfe aussi bien i Suse que dans le nord-ouest du pays comme 6tant celle de l'Cpoque achfmfnide (P1. IVb).

SURVEY OF EXCAVATIONS

185

La seconde acquisition touche plut6t le domaine de l'histoire des religions. Je m'explique: aussi bien &Masjid-i Solaiman qu'a Bard-a Nechandeh, nous nous sommes trouv6s en pr6sence d'une monstrueuse annihilation de tout ce qui pouvait exister dans et autour des sanctuaires d'images de fiddlesou de statues de culte, sculpt6esen pierre ou couldes en bronze. Devant nous, sortaient de terre des " membra disjecta " des images des humains et s'6talaient au sol des tetes sans corps, des corps sans tate ni jambes, des jambes sans corps. Ma premiere pens6e fut de croire a des destructionsdues aux fanatiques conqudrantsmusulmans, puisqu'on sait l'aversion qu'avait l'Islam a cette 6poque pour la representation humaine: je suivais, comme la suite l'a d6montr6, une fausse piste. La fouille se poursuivait et, parallklementaux trouvailles de sculpture s'accumulaient les monnaies quelques rares emissionsparthes d6couvertes. Mais a c6t6 d'une bonne centaine de pikces 61ymaiennes, et sassanides, aucune monnaie islamique ne vint grossir le lot numismatique sur aucune des deux terrasses. Le plus surprenantpour nous a 6t6 r6v616 l'6tude des emissionssassanidesdont aucune ne par datait de plus bas que le IVe siecle, et cela ' l'encontre de ce qu'on trouve sur des sites de cette 6poque oi les monnaies des Khosroes abondent. Serions-nousdone en pr6sencedes suites d'une action de l'Eglise zoroastriennequi, avec les rdformes de Kartir, devint celle de l'Etat sassanide,caract6risde sa violence contre tout ce qui 6tait consid6r6, par par les reformateurs,comme directement non-zoroastrien,et les images des sanctuairesen premier lieu ? Cette hypothese ne s'accorderait-ellepas avec les id6es qu'exprime Nyberg dans son etude r6cemment aus Mitteilungen Iran,ressuscit6esgrace au Dr Heinz Luschey (H.F. 1968, parue dans les Archaeologische pp. 39-48)? R. GHIRSHMAN in Palaeolithic Excavations the ZagrosArea An expedition was organized from the University Department of Archaeology and Anthropology, Cambridge, during the Long Vacation of 1969, with the primary object of exploring reported prehistoric cave sites in the Hazar-i Masjid (Kopet Dagh) range in Eastern Iran and regions to the south and east. The problem at issue was to throw light on the expansion of the Upper Palaeolithic blade industries in South-West Asia, at the expense of the earlier Middle Palaeolithic flake assemblage. In the event exploration from this point of view of regions south and east of the Hazar-i-Masjid was frustratedby an outbreak of cholera in the area planned for study, while the sites made available by the Iranian authorities in the Hazar-i-Masjid all proved to contain only medieval and later occupations. The expedition accordingly turned westwards to examine certain sites in the Kiih-i-Dasht area of Luristan, already briefly visited earlier in the season. Our first task was to examine in greater detail a series of painted rock shelters in the escarpment which crowns the Ktih-i-Sarunmassif, a few kilometressouth-east of the town of Kfih-i-Dasht and at an altitude of some 2000 m. above sea level. In addition to the problem of dating the wall paintings, the sites in question and others near them (kindly shown to the writer by Mr. Sarfaraz of the Antiquities Service of the Iran Bastan Museum, during our earlier visit) gave promise of deposits of Palaeolithic date, possibly capable of contributing to the original purpose of the expedition. In all, four sites were sounded and three were dug in some detail. The first, on the west side of the mountain, known as Mirmalas, yielded only deposits of Neolithic date. These were some 2 m. thick, rested on rock bottom and were clearly stratified and undisturbed. They contained large slabs fallen from the wall, whose parent position could be clearly identified from their shape. Since a number of paintings-including the large hunting scene for which the site is chiefly remarkable-had been painted afterthe fall of the slabs, we had here clear demonstrationof the Neolithic or post-Neolithic age of these examples at least. It is hoped to describe these resultsin greater detail, together with illustrationsof the art, in a later report. Six other decorated shelters in the area on the eastern side of the mountain, all with paintings showing points of stylistic resemblance to Mirmalas, were also studied but failed to produce further direct evidence of date. Several of these however contained deposits of quite different lithological character to Mirmalas and it was these last that provided contributions to the second problem.

186

JOURNAL

OF PERSIAN

STUDIES

was the first to be examined BarddSpid I (the Luri version of the place name, otherwise Sang-i-Safid I) on the Eastern escarpment. It consisted of three openings into a single complex of small caves, affording an overlapping stratigraphical sequence in three segments. The latest segment of this upper portion was not unlike Mirmalas and yielded some coarse pottery of Neolithic date. The second, containing relatively fine-grained deposits, passed downwards gradually into a thermoclastic scree, of a type elsewhere known to result from very cold or glacial climatic conditions. This yielded traces of flintwork of Upper Palaeolithic character and carbon in its upper portion, but below, deeper in the same scree, was a small series of artifacts of unquestionably Mousterian typology. Carbon samples were obtained which, it is hoped, may be sufficient to establish a dated framework for the Barde Spid I finds. A similar sedimentological profile was obtained in the decorated site of Humian II some two miles to the south-east in the same escarpment. Here the only datable object was a flint blade of Mesolithic or later type in situ immediately above the scree. Finally in the important decorated site of Humian I, an extensive formation of cemented scree was found to contain a succession of rich Mousterian horizons throughout at least 3 m. of depth. An upper horizon was studied in detail from the point of view of horizontal scatter while the earlier horizons were exposed in trial soundings. As far as the investigation has proceeded, the Humian I succession would seem to be of special interest as offering a high montane aspect of the Mousterian already known from such lower level sites in the same general region as Kunji, Yafteh, Bisitun, etc. It also offers possibilities of studying the evolutionary pattern of the Mousterian in the area. The realization of these and other objectives must now await the laboratory work in progress. It already seems likely however that the results from the four sites, all from a closely similar topographical and geological context, afford a documented succession of changing sedimentary processes and hence some evidence of climatic change over the last 40,000-50,000 years. If preliminary indicaan interval roughly in the order of tions are fulfilled, the human settlements at Humian I should cover 0,000ooo years. They probably indicate specialized summer encampments designed to exploit such animals of the high mountain environment as ibex, wild sheep, etc. A preliminary examination of the bones at least seems consistent with this suggestion. The raw material of the industry could have been obtained in part locally but in part is likely to have been brought from natural out-crops at a lower level. A brief survey of surface finds suggests the chert presence of hunters with the same basic stone-working tradition both in the neighbourhood of microoutcrops and elsewhere near water supplies on the mountain plateau. It remains to be seen if scopic examination of the charcoal collected will provide any evidence of the contemporary vegetation but traces of invertebrate fauna were extremely scarce.

C. B. M. McBURNEY

Persepolis

On an aerial photograph taken by Dr. E. Schmidt in 1940 one can see, away from the ruins scattered on the terrace at Persepolis, a certain quantity of hillocks a regular distance apart and linked by a line in links high relief which seems to protect the eastern wing of the terrace. These hillocks and the line that them are less easily seen by the tourist who strolls among the ancient ruins. Dr. Schmidt assumed from his aerial views that these hillocks were the remains of towers that had once defended the eastern flank of the royal palaces. When the Archaeological Service of the Ministry of Culture and Arts charged me, in the course of the year 1968, to proceed with investigations and excavations at Persepolis I thought it would be worthwhile to excavate these remains with a view to identifying the so-called fortifications. After some preliminary studies I chose the highest mound, which dominated the terrace and could be considered as the highest of the ancient remains at the site. From the beginning of the second day of the excavations we could easily distinguish, at a depth of 30 cm., the remains of mud-brick walls that were still almost totally intact. We stopped using picks and, in co-operation with our more experienced workmen, we continued uncovering the walls, this time with knives and brushes. The result of our excavations was the discovery of a number of corridors, rooms, a vaulted fortified tower and a small central court-all of which helped to make up part of a single fortified point in the continuous defences of the Persepolitan terrace.

SURVEY

OF EXCAVATIONS

187

The foundations of these fortifications are of unquarried stones on which are built walls of large mud-bricks, 32 x 32 x 13 cm. in size. Some parts of the walls are made of a mixture of small stones embedded in clay, on which are placed mud-bricks. This is also true of the floors of the rooms and corridors. The slope on which the foundations are laid was only rarely modified or smoothed over. The enclosing wall, like that of the terrace itself, possesses niches which retain traces of arches. These niches would have facilitated surveillance of the royal guards. In the northern wing of the court is a large drain, I -80 m. high, carved in the rock and covered with great flagstones. This drain must have served to let rain water flow away. The door which leads to one of the guard rooms is covered by a cradle vault in which either ordinary bricks or possibly long mud-brick units were used to compose the underside of the vault. The discovery of cradle vaults at Persepolis is of great importance in the reconstruction of the history of Achaemenian architecture. Achaemenian art, in fact, has always appeared to us as a specifically imperial art, designed to glorify the power of the King of Kings. The architectural elements in the palaces of Persepolis and Siisa consist of stone columns supporting a horizontal roof. At the very time that the grandiose palaces of the two principal capitals of Darius and Xerxes were being built, the majority of people lived in mud-brick houses roofed with cradle vaults. The discovery of cradle vaults at Persepolis, the home of high stone columns, supports the hypothesis that the vault was used in Iran concurrently with the flat roofs used for the great palaces. Mr. A. Sami, ex-Director of the Scientific Mission to Persepolis, has also discovered on the terrace itself some cradle vaults which he has published in Vol. II of ArchaeologicalReports under the title " Excavations in one of the original rooms in the sub-soil of the palace of Hadich " (A.H. I330, pp. 51-2). He found there five fairly narrow rooms and two long corridors. " The roofing of the corridors ", he says, " like that of the rooms, is composed of vaults in mud-brick. On this roof are other buildings on the same level as the eastern rooms. In the corner of one of these rooms was found the base of a cradle vault. Until now one could only assume that vaults existed in the Achaemenian period without being able to give a concrete example." Unfortunately the remains of the cradle discovered by Mr. Sami have disappeared and vaults in modern baked brick have replaced it. We do not even have photographs showing these remains of the birth of a cradle vault in mud-brick. The discovery of identical cradle vaults in a tower of the Persepolitan fortifications becomes even more precious therefore for those who wish to study Achaemenian architecture. Seal impressionsin baked earth: North-west of the fortified tower, at the bottom of a staircase in corridor B, were the broken remains of a pottery vase, which must originally have contained the fiftytwo seal impressions that were found nearby. These impressions were formerly of crude clay, but all had been baked in a fire that had raged through the room in which they lay. The designs represented in these impressions are most interesting. Some of them show an Achaemenian general (whom one can identify as the King of Kings in person) fighting the enemy. Most of them show a conquered Greek soldier on his knees. Behind the general, three officers take part in the fighting. The general wears Achaemenian costume and with his left hand holds the arm of his enemy, whom he bears to the ground with his lance. The costume of the Greek soldiers is minutely represented. On one impression one sees a fight between a Persian officer and a Scythian general. The Persian grasps his enemy by both hands and forces him to the ground in hand-to-hand combat. The artist has shown all the details on this little impression; he even depicts the weakness of the falling general. The fluting of the tiara worn by the Iranian warrior is visible and the proportions of the figures are well kept.
AKBAR TADJVIDI

Shahddd In the spring of 1968, with the help of the Geographical Institute of the University of Tehran and the collaboration of Mr. Ali Hakemi, certain archaeological researches were made on the border of the Lit Desert by the Archaeological Service of Iran. The area concerned was situated to the east of Shihdad

188

JOURNAL

OF PERSIAN

STUDIES

(the ancient Khabis) in a locality known as Agous. In the course of a ten-day excavation the mission's team succeeded in discovering a number of tombs of the third and fourth millennia B.c. The grave goods were found to include painted pottery, vessels of stone and copper, and jewellery of silver, agate and lapis lazuli. Of particular interest were several grey stone vases engraved with animal designs including a lion, a bull and other animals. These stone vesselswere each beautifully shaped and had been fashionedwith great elegance. Similar vases have also been discovered at Sfisa, in Baluchistan and at Kirman. One unusual feature of the site was the absence of superimposedstrata; the remains representing different civilizations succeed one another on the infinite horizon of the desert. Every two or three hundred years, each successive culture appears to have established itself elsewhere, usually shifting progressivelyfurtherwest towards more mountainous districts. The violent winds and floods of the Lfit may have encouraged this pattern of movement,just as they may also be held responsiblefor the severe erosion of the graves and houses of each community.
ALI HAKEMI

Shahr-i Sokhte The researcheson the earliest cultures of Sistan have been continued this year, by enlarging excavations on the main Bronze Age site of the region: Shahr-i Sokhte. The Italian Archaeological Mission has been operating for two and a half months (September-November 1969) on an area of over 1200 square metres. The main aims this year were to explore a wider surface of the dense inhabitation area concluded in the (Periods I-III), to study in all details one more building beside the Houseof Staircases, season (Period II-III), to increase documentation on the Late Chalcolithic layers (Period I), and past to start excavations on the highest point of the mound. With the last aim we intended to uncover a main, well-preservedbuilding, which could provide us with extensive information on the large-scale architecture. Our hypothesis turned out to be correct as at the end of the season we had completely isolated a building, extending for at least 500 square metres, with massive mud-brick walls, rising to a height of over 3 m. Three sides of the building's perimeter had unfortunately been washed off by erosion, but we were able to isolate the fourth one and the south-east corner. This large building represents a new period in the cultural sequence (Shahr-i Sokhte IV), completely undocumented in past campaigns. Period IV marksthe introduction of the fast wheel in pottery manufacture; the pottery types are characterized by articulated shapes with straight corners, buff or brick-redin colour, rarely painted, but frequently red slipped or burnished in parallel lines; a much smaller percentage are of black-burnishedgrayware and buff jars decorated in Late Bamptir style, carrying painted and moulded elements. Period IV appears to be datable to the first quarter of the second millennium and the general evidence shows, notwithstanding the sharp changes, a remarkable cultural continuity in the pottery, as well as in the metallurgical traditions. The plan of the building is a series of square and rectangular rooms, peripheral corridors, straight-sideddoors with well-marked sills and two staircases, leading to an upper terrace-store,whose steps have the rims strengthenedby timber frames: a technique destined to last for long in the architecture of Iran. The end of the palace was certainly a violent one as the whole of the building shows evident traces of an extensive fire: collapsed beams, wall-plastersand floors
fully burnt, broken jars still in situ and finally the dramatic discovery of a human skeleton lying on the floor of the most northern, isolated room. The bones were completely burnt a few hours after death, as is proved by the carbonization of the brain in the exploded skull. The body fell among Period IV buff pottery, and the right hand was still holding a stone pestle. The Palace shows only a single building phase, because directly underneath the burnt floor level, the setting of the foundations cut the deposits of Period II, characterized by buff painted ware. The discovery of a fourth period of occupation, to which can be related the Burnt Palace, completes the cultural sequence of Shahr-i Sokhte. It can now be divided in the following way: Period I: Late Chalcolithic, with bichrome buff and red/black on grey ware, painted with geometric designs. The general picture is very close to that of the Geoksyr Culture (Namazga III of the South Turkmenian Sequence). Among the most exciting finds were three sealings of cylindrical seals of a type apparently very

SURVEY

OF EXCAVATIONS

189

close to late Uruk or Jemdet Nasr. Periods andIII: Early and Middle Bronze Age: they show a slow II but continuous development of the earlier Chalcolithic models into highly standardized pottery types. Bichrome wares disappear and the pottery classes are limited to black on buff and black on grey wares, with the exception of a polychrome pottery, usually limited to a relatively small number of specimens. Stamp-seals of a compartmented type are widely employed, and frequently occur: clay anthropomorphic and theriomorphic figurines, various types of metal, timber and stone tools and a wonderful bronze figurine, 29 cm. high, representing a woman carrying a jar on her head, with striking Mesopotamian features (P1. XIa). It was discovered on the surface, but is most likely to relate to these periods, which in terms of absolute chronology could mean the second half of the third millennium. This period has connections with Namazga IV and Bampiir I-IV. PeriodIV: Late Bronze Age. As already stated, this period sees the end of the painted tradition, the introduction of the fast wheel, new types of stamp seals and anthropomorphicfigurines and, apparently, a great increase of architectural skill. On the basis of the pottery found in the destructionlevel (Level 4), we might associatethis period with Bampfir V-VI and even with the Umm an-NMrCulture of the Oman peninsula. TapeRud-iBiyaban: This mound lies approximately 25 km. south of Shahr-i Sokhte in the middle of the ancient Delta system of the Helmand. It is covered with pottery slugs and wasters, and on the surface at least forty-two large and small pottery kilns can be recognized. In a preliminary excavation four of these were cleaned out. The pottery associated with them and covering the mound can be recognized as of Shahr-i Sokhte III and IV types. Kilns Nos. I and 3 have one chamber, vitrified walls, vaulted roofs and their bases are lined with a vertical row of bricks. An identical technique appears to have been employed for kilns 2 and 4 with the difference that these last two have two fire chambers divided by a vitrified earth pillar and average m. in length, twice as large as the other two. 5.50 Excavation of this unique Bronze Age pottery mill will be resumed on a much bigger scale in the following season and will be used for palaeomagnetic studies for the region on the basis of the wall vitrification.
MAURIZIO TosI

The fourth season of the Institute's excavations at Sirdfbegan in October 1969 and is still in progress as I write. Aided by generous grants from the British Museum, the British Academy, the Royal Ontario Museum, the Pilkington Glass Museum and others, we are investigating three areas: the Great Mosque (Site B), the residential suburb (Site F) and the pottery (Site D). At the same time we are adding contours to the large-scale map of Sirdf and recording additional buildings, including a small mosque situated in one of the cemeteries. Site B: In 1968 we began to excavate the structurespreserved beneath the platform of the Great Mosque (see above, " Third Interim Report ", p. 7). This year we are continuing the work by removing the earth and rubble make-up for the platform, but leaving the piers and foundations of the mosque intact. As the excavation proceeds, it becomes progressivelymore probable that the remains are those of a single large building, measuring at least 70 x 55 m. We have now exposed some 1300 square metres of the building, revealing an irregular plan consisting of numerous interconnecting rooms and small cobbled yards. In the area excavated, we have fifty-two rooms and four yards, two of which contain
isolated piers or plinths (cp. " Third Interim Report ", fig. 3, bay 2F). Beneath the north angle of the mosque, the early walls are unusually thick, suggesting that they supported at least one upper storey, and a room beneath bays 13D-E possesses an apse. Otherwise, the building contains few distinctive features and its identity remains unknown. We still know little about its date. However, finds from the late eighth or early ninth century make-up for the mosque include several earlier objects, notably an intaglio bearing a Pahlavi inscription, and it is possible that the building, which is considerably larger than the Great Mosque, is pre-Islamic. Site F: Work has continued at Site F with the object of investigating the early history of the suburb (" Third Interim Report ", pp. 8-15). In House W, which we excavated in 1967-68 (" Second Interim Report ", fig. 4), removal of the existing floors has revealed the floors and footings of an earlier courtyard house. The early house closely resembles the standing building, although its north wall is

Sirdf

190

JOURNAL

OF PERSIAN

STUDIES

almost 4 m. farther south and consequently the overall length is only 23 m. The house has two entrances the more impressive of which is in the south wall. It appears possible now that the original main entrance to the standing house was also in the south side, beneath the loggia, and that the elaborate north entrance is a later addition. Outside House W, investigation of the mosque has shown that it contains two main periods of building. In the first period a small ablution area occupied the angle between the mosque and the main street. This was demolished in Period 2, when the mosque was enlarged. In House S, excavated last season, we have discovered that the semi-circular pilasters belong to Periods I and 2 (not Periods 2 and 3, as originally believed) and that the building was reconstructed using rectangular pilasters in Period 3. In House R we are recovering the remains of an earlier, and no less irregular, structure. Site D: In 1966 we discovered the site of a pottery near the western edge of Sirgf, in a quarter constructed when the city reached its greatest extent. An exploratory excavation revealed that the site, which was being eroded by the sea, contained numerous kilns and pits filled with wasters (" First Interim Report ", pp. I2-14). This season we are attempting to excavate the entire pottery. It is emerging as a complex of at least five walled enclosures, bounded by the sea on the south side and by streets to the north and west. A sixth enclosure may exist on the east side. Excluding this hypothetical enclosure, which awaits investigation, the surviving parts of the pottery occupy an area of some 1250 square metres. In the north-west angle of the complex is a courtyard house measuring 16 5 x 15 8 m., at first, the which resembles the houses at Site F. Although it may have had a purely domestic function house was later converted into workshops. A potter's wheel was installed in one of the entrance passages, the rooms in one angle were demolished to make way for kilns and two further kilns were built in the yard. To the south of the house lie two enclosures, one of which is almost empty and presumably served for storing clay, fuel and finished pots. In contrast, the second enclosure is crowded with kilns, a well, a cistern and a workshop. Additional kilns exist in the enclosures to the east and south-east of the house and at present we have excavated a total of twenty-six. The excavation has already yielded a vast number of coarse pottery wasters. It is clear that glazed pottery was also made on the site and the persistent occurrence of glass waste and slag continues to suggest that a glass furnace exists in the vicinity.
DAVID WHITEHOUSE

Suseet Susiane

' Les recherches entreprises Suse et en Susiane au cours de l'hiver 1968-1969 relkvent d'un programme de cinq ans qui a pour but d'orienter clairement et de prdparer les recherches a venir. Les essais de synthese prdsentes rdcemment par M. R. Ghirshman et par M. P. Amiet ont permis de cerner les lacunes de notre connaissance et d'en souligner les insuffisances. Des problkmes se trouvent posds. Reste a envisager les moyens de les resoudre de la manie're la plus efficace. Pour cela nous avons dtcidd d'entreprendre, a Suse meme, une exploration extensive, mettant a profit la topographie actuelle du site, au moyen d'operations de contr6le stratigraphique suffisamment 'tendues afin d'Ctablir la continuit6 de la sequence archdologique et culturelle. Simultandment, hors de Suse, nous avons commenc6 un inventaire des nombreux sites archdologiques de la Susiane et de l'ensemble du Khuzistan, recherche realisde avec l'aide du C.N.R.S., en 'troite collaboration avec les chercheurs iraniens et amdricains dans ce secteur. La connaissance du terrain et des ressources offertes nous permettra plus tard de faire pour nos chantiers les meilleurs choix. Des problkmes poses &Suse trouveront peut-8tre hors de Suse leur plus rapide solution. Ainsi esparons-nous dans quelques anndes 8tre 5 m~me de prdsenter un programme cohdrent r~pondant aux questions les plus urgentes. A la campagne 1968-1969 ont participd activement M. I. Nafici, reprdsentant du Service des Antiquitds et des &tudiants de l'Universit6 de Tthdran. Mlle. G. Dollfus et Mlle. M. Lechevallier du C.N.R.S. ont dirig6 la fouille de Djaffarabad et en donnent le compte-rendu ci-dessous. A. Lebrun m'a assistt sur le chantier de l'Acropole. Les autres membres de la mission 6taient F. Vallat, apigraphiste, A. Assaf, prdhistorien, A. Labrousse, architecte, D. Ladiray et Mme. Terry Haass, dessinateurs, Mlle. M. Mignon, photographe. Nous avons eu le plaisir d'accueillir &Suse plusieurs collkgues iraniens,

SURVEY OF EXCAVATIONS

191

anglais et amdricains, notamment ceux des universitis de Chicago, de Michigan et de Houston qui prennent part a la prospection archdologique du Khuzistan. Djafarabad: L'objectif de cette premiere campagne 6tait de reconnaitre la stratigraphie du site. Dans cette perspective, une coupe fut etablie d'Est en Ouest sur une longueur de 20 metres et une hauteur de 6 metres, un peu en retrait de la paroi Nord Est du sondage de 1934. En outre, une tranchde de 12 metres de long, large de 2 metres, ouverte sur la bordure occidentale du tipd, dans le prolongement du sondage 1934, permit d'atteindre le sol vierge. Cinq niveaux d'architecture ont dtd reconnus qui ont &t6numdrotds du sommet jusqu'a la base de I a 5. L'examen du matdriel a permis de discerner deux phases distinctes, l'une correspondant aux niveaux 5 et 4, l'autre aux niveaux 3 a i.

Niveaux5-4
Le niveau5, premier niveau architectural reconnu au-dessus du sol vierge se trouve a 7 mitres sous le sommet (13 m); ii est caractdris6 par des sols de terre brilde ou cendreuse, des petits foyers et des murs larges de 0,70 m en briques crues conservis souvent sur 13 a 15 assises. A ce niveau appartient une plateforme large de 2 metres dont le d6gagement n'est pas termind, paisse de deux lits de briques o0 alternent rdgulierement briques brunes et briques vertes (0,40 x o, I o, I i i). Dans le niveau4, l'orientation des murs reste semblable a celle du niveau 5. Des pi ces rectangulaires ont dtd reconnues. En outre, une banquette en briques, en relation avec un foyer construit, a 6td degagde. Dans ce secteur, un feu intense a cuit les sols, deux assises des murs, et ddposd une 6paisse couche de cendres. Cette couche d'incendie, qui a peut-&tre provoque un abandon du site a ce niveau, a etd reparde en coupe en d'autres points du chantier lors du nettoyage des parois du sondage de 1934. Aucune evolution n'a etd remarqude dans le materiel entre les deux niveaux 5 et 4. En ce qui concerne la poterie, a l'exception d'un grand bassin ovale dont la pate est a degraissant mineral, la c6ramique non peinte, mal cuite et friable, prdsente une tres forte proportion de ddgraissant v6g6tal. A l'intdrieur, paroi et fond sont toujours enduits de bitume. Les types les mieux attestds sont un grand pot a carene situde en bas de la panse (morphologiquement diffdrent du " milkbowl " restitu6 par L. Le Breton dont nous n'avons retrouv6 aucun exemplaire) et un bol " renflement cardn6. Les types de formes et de decors de la ceramique peinte sont ceux d6crits par Le Breton comme " Susiane a": " coupes a pied au d6cor incise ou fenestr6, bols ovales a bec verseur, petits bols bord rentrant. Les " d6cors en reserve sont internes et externes et se trouvent sur de grands bols lkvre 6versde et sur des coupes. Sur les bols ovales h engobe chamois, les lignes croisdes, brunes, formant losange a l'intdrieur desquelles peut exister le motif de l'oiseau en vol sont frequents. Deux types non signalks par L. Le Breton sont d'une part des jarres a paule arrondie, haut col et decor de peinture brune, d'autre part des petits bols h6misphdriques a peinture rouge fugitive. L'outillage en pierre consiste essentiellement " en '16ments de faucilles a fine denticulation et en houes fabriqudes partir de galets fendus et retouch6s sur la face dorsale. Une premiere 6tude de la faune montre qu'a c6td du mouton et de la chevre est present un petit bovide qui serait domestiqu6 (J. Wheeler). Niveaux 3-I De mime orientation que les murs des deux niveaux inffrieurs, ceux du niveau3 sont moins larges qu'aux autres niveaux et n'ont souvent qu'une brique d'6paisseur. Plusieurs fours ont dtd creusis a partir des sols de ces niveaux, fours culinaires, mais aussi peut-etre pour l'un d'eux, four B poterie (P1. XIb). Celui-ci posside une double paroi en briques recouvertes d'un enduit. II a 1,60 m de diamitre et est conserve sur I,40o m de hauteur. Dans le haut du remplissage, de grandes briques cuites (o,48 x 0,20 x o, Io), de texture grossibre, provenaient vraisemblablement de la sole. Au niveau 2 appartiennent des murs conserves sur I,50 m de haut. Seuls de petits C1lments ont Ct6 mis au jour en avant de la coupe et nous espbrons que de plus amples digagements permettront de restituer le plan des habitations. Une tombe d'enfant a 6td digagde (P1. XIc). Le mobilier fundraire consistait en trois vases et un sceau ddpos6 t la t~te. Bien que la partie supdrieure du niveau I ait Ctdtrbs perturbde par de nombreuses tombes rdcentes,

192

JOURNAL

OF PERSIAN

STUDIES

des murs et des sols ont dtd reconnus. Des structures domestiques y sont assocides: foyer entourd de champ et dont le fond est constitud aussi par un dallage de briques. Dans I'angle de deux murs se trouvait une tombe d'enfant. Le squelette reposait sur le dos, les jambes replides; le corps 6tait
murets de briques crues, four, bassin ovale de I,6o m x o,90 m construit en briques crues posees de

faucilles et les houes dejanot6s dans les niveaux infirieurs, un outillage lourd en pierre constitudpar des meules, des pilons, des grosses masses polies, des haches, est specifique de ces niveaux. Les figurines animales (oiseau, chien, bovidd, etc....), les fragments de grands cachets en terre cuite, les sceaux en pierre perces longitudinalement, decor6s de caprid6s ou de motifs g6omitriques, complktent cet
assemblage. Dans le niveau 2 des graines ont etd recueillies en abondance parmi lesquelles J. Renfrew, lors d'un premier examen, a pu identifier du bl6, de l'orge a 6 rangs vetue, des lentilles et des pois. A la fin de cette premibre campagne, il semble donc d'une part que la stratigraphie indique une occupation continue; d'autre part, que le matdriel se rapporte aux deux phases d6finies jusqu'dapresent comme " Susiane a " et comme " Suse A ". Il sera donc important de preciser si ces deux phases se succedent sans interruption ou s'il existe une phase de transition. A ce jour, l'examen de la c6ramique ne laisse pas apparaitre de groupe interm6diaire entre la c6ramique des niveaux inf6rieurs et celle des niveaux superieurs. Suse: Acropole. Avant de proc6der a Suse m6me aux ddgagements horizontaux indispensables pour eclairer cette plus ancienne p6riode, il nous a paru n6cessaire d'en 6tablir solidement le fil directeur, l'admirable travail de Le Breton reposant souvent-et la responsabilit6 de cette situation ne lui incombe

La couche de surface est 6paisse de I,8o m. Des tessons sassanides y sont mel1s a quelques fragments d' " 6cuelles grossieres a bord biseaut6 ". ' Les niveaux 3 I ont livr6 des gobelets, coupes, jarres miniatures en pate fine chamois ' decor peint qui sont les types bien connus de " Suse A ", de mime que le sont aussi les jarres et les bassins en pate chamois ou rouge "asurface brunie parfois lustr6e recueillie en abondance. Outre les elements de

par un rang de briques posees de champ, comme c'6tait aussi le cas pour celle du niveau 2.

' envelopp6 dans une natte et un vase peint 6tait placd la tete. La tombe, rectangulaire, est constitute

Du sol vierge, retrouv6 au fond de la tranch6e de Mecquenem, a la base de la galerie F de Morgan qui traverse la partie haute du tdmoin, nous reconnfimes sur une hauteur totale de 15 metres 19 strates principales, numdrot6es de 8 a 27 du haut vers le bas. Ces 15 m de debris representent les vestiges de l'occupation de Suse en ce point depuis sa fondation dans la premiere moitid du 4e mill6naire jusqu' a la fin de la p6riode caract6risde par la poterie peinte dite du " style II "; les niveaux superieurs du timoin correspondent a la fin du 3e millknaire av. J.C. Ainsi les coupes de Djaffarabad et de l'Acropole, dans le prolongement l'une de l'autre, nous permettent d6ja un coup d'oeil sur l'ensemble de la periode du 5e millenaire a la fin du 3e mill6naire av. J.C. En consideration de l'aspect de la coupe, les divisions principales s'etablissent entre les niveaux 23 et 22 et entre les niveaux 17 et 16. Dans les niveaux infdrieurs (2 7-23) a poterie peinte de style " Suse I " on peut distinguer une phase ancienne (27-26) correspondant aux niveaux sup~rieurs de Djaffarabad et une phase r6cente (25--23) caract~risde par l'apparition de formes nouvelles et notamment en 25 de bords a lvre roulke et en 24 de cols dejarre a bord pined de fagon a former a l'ext6rieur un filet qui souligne la 1kvre; en 23 apparaissent de petits bols chamois a lvre coupde en biseau; des fragments d'Xcuelles grossibres se rencontrent dans ce meme niveau, il conviendra de pr6ciser leur appartenance au niveau 23. L'architecture des niveaux 25-23 pr6sente un d6veloppement remarquable. Des murs puissants dans le niveau I21 paraissent devoir 6tre mis en relation avec les grandes constructions signalkes par Mecquenem sur cet horizon auquel R. Dyson attribue encore - a juste titre semble-t-il - la construction de la terrasse qui occupe la partie centrale de l'Acropole.

pas-sur une analyse typologique d'un mat6riel arch6ologique sans contexte stratigraphiquesuffisant. Avec M. P. Amiet, conservateurdu D6partement des Antiquites Orientales du Musde du Louvre, nous avons fait choix pour un premier contr6le stratigraphiqued'un secteur de l'Acropole qui paraissait offrir le maximum d'avantages, au pied du tamoin laisse par de Morgan sur la face Nord du " sondage 2 " de Mecquenem et a proximit6 du sondage effectud par M. R. Dyson en 1954 (P1. XIIa).

SURVEY OF EXCAVATIONS

193

Le niveau 22 est marqud par un changement d'orientation des constructions qui diminuent d'importance. Les murs sont en brique crue de plus petites dimensions que celle des niveaux sous-jacents. Dans le niveau 19, les constructions ont disparu. Il en est de meme dans les niveaux 18 et 17. Les niveaux 22-17 sont caractdrises par des fosses profondes souvent de plusieurs metres et remplies d' " ecuelles grossi'res a bord biseautd " et d'autres debris caramiques (vases a goulot, th6ieres, vases a anse oreillette; le d6cor peint disparait pour faire place a un d6cor incisd ou a r6serve d'engobe). De nombreux petits foyers sont creus6s dans les sols d'occupation qui portent aussi des traces de piquets. Le niveau 18 est caract6ris6 par l'apparition d'une poterie a engobe rouge lie-de-vin ainsi que par des vases &anse torsadde et des jarres-bouteille a goulot tordu vers le bas. ' Une premiere 6tude typologique des " 6cuelles grossieres " en fonction de la stratigraphie a permis M. A. Le Brun de reconnaitre a c6te du type " classique " present dans tous les niveaux et qui pourrait avoir 6tdnon pas tournd mais mould sur une forme: dans les niveaux de base une varidtd 1a vre arrondie soulignde a l'extirieur par un filet irrdgulier; dans les niveaux superieurs une varikt6 a bord renfl6 16gerement rentrant a lvre amincie. Dans le niveau 17 toutes les dcuelles tendent a devenir plus hautes et plus 6troites; elles sont faites alors sur une tournette rapide et presentent dans le fond la marque laissde par la ficelle (string-cut) qui a permis de les d6tacher du support. Ces distinctions morphologiques et stratigraphiques sont importantes pour l'analyse du materiel provenant des rdcoltes de surface sur les sites de Susiane contemporains de cette p6riode. Le niveau 16 est marqud par la rdapparition de constructions importantes dont l'ordonnance est conservde dans les niveaux superieurs (15-I4). M^me observation en divers points de l'Acropole. Du niveau 15, correspondant a un rdaminagement des constructions du niveau prec6dent, proviennent des tablettes proto-l1amites d'une 6criture ddja plus dvolude que celle des plus anciens documents ' de Suse. Cest ablettes, copides par M. F. Vallat, prdsentent l'une un motif decoratif nouveau Suse, sur les deux faces, une empreinte de cylindre montrant des animaux manipulant des vases. Ces l'autre, a des vases et a une figurine en tablettes sont associkes dans le niveau 15 a un cylindre des tdatite brfilde, albatre et a une poterie bichrome rouge et noire a d6cor instable encore en usage dans le niveau 14. Le niveau 15 a livrd galement de nombreux objets en m6tal, armes et outils, et en particulier un hamegon en 6vidente relation avec les nombreux ossements de poissons qui jonchent a ce niveau le sol de plusieurs chambres. Cette operation a l'Acropole sera poursuivie au cours des deux prochaines campagnes. Elle apparait prometteuse en ce sens qu'elle permettra de reclasser une bonne partie du materiel archeologique provenant de la tranchde de Mecquenem. Elle permettra ensuite, par l'6tablissement d'une stratigraphie fine, de pr6ciser la relation de chaque niveau avec ceux des sites de Basse M6sopotamie et le sens l'6criture fait son des dchanges sur l'horizon des pdriodes d'Ourouk et de Djemdet Nasr au moment ofi et ofi s'affirme le caractere urbain des agglomdrations. apparition de l'dvolution La fouille qui porte sur 200 mi, est suffisamment 6tendue pour donner une idle dont elle traverse les vestiges. L'imprestechnologique, economique et culturelle au cours des periodes sion qui pour l'instant pr6vaut, impression confirmde par des observations en divers points du tell, est que le processus d'urbanisation qui s'amorce, semble-t-il avec les manifestations architecturales caracterisant la fin de la premiere p6riode de Suse (niv. 25-23) marque peut-etre ensuite un ralentissement sur l'horizon de la pdriode mdsopotamienne d'Ourouk. Les installations legeres qui remplacent partout les constructions de la ptriode pr~cddente, si elles n'indiquent pas ndcessairement un ralentissement du progrbs technologique, paraissent impliquer le passage i une Cconomie a pridominante t mixte des agriculteurs dleveurs qui ont fondC la premikre agglomdrapastorale, succddant l'dconomie tion de Suse. La structure sociale et le mode de vie de ces populations auraient constitud des conditions moins favorables au ddveloppement du processus d'urbanisation. Suse: Apadana. Sur le tell de l'Apadana, oi furent entrepris avec la participation du Service Iranien des Antiquitds des travaux de prdsentation des vestiges encore visibles, furent faites quelques observations. A0,75 m de profondeur, fut d~gag6 Ainsi B l'Ouest de la cour occidentale du Palais achdminide, un mur de briques cuites lides au bitume dont l'orientation fait avec celle des murs du Palais un angle de Io?. Ce mur se trouve pris dans la terrasse artificielle qui supporte le Palais; il est i la hauteur du

194

JOURNAL OF PERSIAN STUDIES

sommet de la butte dlamite qui constitue le noyau de la terrasse artificielle qui supporte le palais ach6minide. Ce mur est au meme niveau que les fondations de colonnes en pierre mises au jour par Mecquenem dans la cour Est du Palais ainsi qu' avec des 616ments de construction en brique cuite et bitume (MDP XXX, pp. 17 et 55) B l'angle S.O. de la terrasse. Si cette relation etait v6rifide, nous aurions la un ensemble d'616ments architecturaux appartenant a une construction d'une orientation lgerement diff6rente de celle du Palais et fondee t 2,50 m au dessous. On serait tent6 d'y voir des vestiges d'un palais antdrieur a celui que nous connaissons a l'Apadana. Cependant, 1'6tude stratigraphique de la terrasse dans le secteur du mur que nous avons ddgag6 montre que ce mur n'appartient pas " une construction ruinde qui se serait trouvie englobde dans le massif de la terrasse que l'on aurait alors surdlevee. Ce mur repr6sente simplement a ce niveau le debut de r6alisation d'une construction dont le projet initial fut rapidement abandonnd, tandis que l'on ddcidait de surdlever la terrasse de 2,50 m environ. Cette op6ration s'effectua dans un laps de temps tres court puisque les gros murs de brique crue ou de terre pisee qui formant caissons quadrillent en le contenant le gravier de le terrasse, reprennent exactement dans la partie haute, au dessus du mur 701, le plan des murs de la partie infdrieure. Le coffrage de ces murs, lorsqu' il existe, ne montre aucune solution de continuit6 la terrasse artificielle constitue un tout. Cette terrasse mesure en ce point Io m de haut. Elle s'appuie a 1'Est contre la butte dlamite centrale dont le flanc a 6td entam6 puis rectifie par un mur de briques crues. Elle repose a 8 metres seulement au dessus des eaux du Chaour, sur des couches creusdes de tombes 61amites de la premiere moitid du IIe mill6naire av. J.C. La salle hypostyle a et6 nettoyde. Dans l'angle N.O., les fondations de briques cuites du mur, large de 5 metres, s6parant la salle des portiques latdraux, ont dtd degagdes. L'existence de ce mur, mise en doute par certaines reconstructions, avait ddja dte 6tablie par M. Ghirshman. Les vestiges actuels ne permettent pas de reconnaitre l'emplacement des portes (P1. XIIb et c). Les fragments sculpt6s dpars a travers le palais ont Wt6 regroupas en un musde lapidaire; des mesures de conservation provisoires ont 6td prises pour certaines pieces.

J. PERROT

Takht-i Suleimdn(Pls. VI-VII) Die Ausgrabungen des Deutschen Archiologischen Instituts auf dem Takht-i Suleiman in Aserbeidschan wurden mit Unterstiitzung der Deutschen Forschungsgemeinschaft in der Zeit vom 5. 7bis 12. 9. 1969 fortgesetzt. Ziel der Arbeiten war eine weitere Klarung der zeitlichen Abfolge von iuBeren und inneren Umfassungsmauern, die Vervollstandigung des GrundriBplanes der Lehmziegelanlage westlich neben dem Feuertempel, die Untersuchung der achimenidischen Siedlung unter dem Lehmziegelgebiude und die Feststellung sasanidischer Bebauung in der siidlichen Halfte des Takht-Plateaus. AuBerdem waren in einigen Rtumen und H6fen n6rdlich des Feuertempels und am groBen Westiwan Nachuntersuchungen notwendig. Ein im Jahre 1968 begonnener Suchschnitt hinter dem Turm 18 der Umfassungsmauern: wurde bis zur nordwestlichen Eckbastion der, das quadratische Tempelareal diuBeren Ringmauer n6rdlich des Sees umfassenden, inneren Mauer erweitert. Er brachte die alte, m6glicherweise vorsasanidische Lehmziegelmauer zutage, die vor der Erbauung der steinernen Ringmauer die Hiigelkuppe umschloB. (s. Iran VII (1969), S. 192.) Die Lehmziegelmauer hat an dieser Stelle die ungew6hnliche Stirke von mehr als 15 m, und ihre klar erkennbare Innenkante weicht in ihrem Verlauf leicht von dem der steinernen Ringmauer ab. Beide Beobachtungen sind vielleicht dadurch zu erklaren, daB die Sondage einen Befestigungsturm der alten Mauer schneidet. Die Mauerkonstruktion besteht aus einem mehrere Schichten hohen Sockel aus unvermorteltem Bruchstein und einem noch etwa 7 m x x hoch anstehenden Lehmziegelaufbau (durchschnittl. Ziegelformat 45 45 1Ocm). Eine starke, jedoch nur wenig Keramik enthaltende Aschenschicht am FuBe der Mauer ist von Lehmziegelschutt uiberlagert. Offenbar nach einer langeren Verfallszeit, vermutlich in der spitsasanidischen Periode, wurde die steinerne Ringmauer vor die AuBenseite der Lehmziegelmauer gesetzt und, nicht notwenigerweise gleichzeitig, die innere Umfassungsmauer angelegt.

SURVEY

OF EXCAVATIONS

195

24 I \\ s

23
\ 27

',

*
13 oil

\ 2a'
?~~50.70\

,
20 \3

T.

139 0,+46 0
id I

100 3000. il + i V3 r36

+1

200

'
1I

3 \3

\"~~
3;4

1~?
?

\.?l

,' i

\ 38 2

', ",

, \t,

IO4 I

,
"" % "

%% i'

/i

-.

50

AIII

o
C

0
Peid

aoM

-I$IIk

goiln\h

-IbkN'?

196

JOURNAL

OF

PERSIAN

STUDIES

Areal westlich vom Feuertempel: Die Untersuchungen im Geldnde zwischen dem Feuerund dem westlichen Fliigel der inneren Umfassungsmauer lieBen zumindest vier vorislamische tempel Bebauungsperioden erkennen, von denen jede wiederum mehrfache Verdnderungen aufweist. Als friiheste bisher auf dem Takht-Plateau festgestellte Kulturschicht wurde eine im Vorjahr gefundene, ach~menidische Siedlung weiter freigelegt. Ihre unregelmiBigen Raume und H6fe werden durch Feldsteinwinde gebildet, die ein sehr sorgfaltig gesetztes Fischgratenmauerwerk auf flach gelegten Fundamentschichten zeigen. Einige Raume sind mit groBen, gut zusammengepaBten Steinplatten gepflastert. Wiederum fanden sich neben Mauern oder in Mauerwinkeln in den Felsboden gehackte Grabgruben, auBerdem Erdbestattungen und ein Kindergrab in einem Mauerzug. Die Skelette weisen Anzeichen nachtraglicher Umbettung der Toten auf. An Grabbeigaben wurden Arm- und Ohrringe aus Bronze, dreifliigelige Pfeilspitzen und Glasperlen geborgen. - Wdhrend auf das achimenidische Stratum eine mehrere Jahrhunderte dauernde Unterbrechung in der Siedlungsgeschichte des Platzes folgt, zeigen die dariiberliegenden Bebauungsperioden eine bis ans Ende der zoroastrischen Benutzungsphase reichende Kontinuitit. Die erste nachachamenidische Anlage ist ein ausgedehnter Gebaudekomplex mit Lehmziegelwinden (durchschnittl. Lehmziegelformat 45 x 45 x 10 cm) auf Feldsteinfundamenten, dessen Grundplan hiufig abegindert, in seiner Disposition jedoch nie aufgegeben wurde. Als Kern dieser Anlage laBt sich ein Viersaulenraum mit quergelagertem Vorraum und schmalen Seitenkammern an der Nordseite eines vermutlich ungedeckten Vorhofes erschlieBen. Vier Raume eines magazinartigen Traktes, der einer groBangelegten Erweiterung auf der Westseite des Vorhofes anzugeharen scheint, wurden freigelegt. Ihre FuBb6den haben Ziegelpflaster, ein Ofen sowie Metall- und Schlackenriickst~nde lassen auf BronzegieBerei schlieBen. - Von einer schwer erfaBbaren Umbauperiode wurden Lehmziegelwinde ohne eigenes Steinfundament, Becken mit Ziegelwandungen und mit Gipsm6rtel verputzte Podien freigelegt. Offenbar wurden groBe Teile der ersten Anlage, die in einer Brandkatastrophe untergegangen zu sein scheint, durch neue Lehmziegelbauten ersetzt, wahrend die Kerngebaude nur unwesentlich verindert wurden. - Der bedeutungsvollste Wandel begann mit der Errichtung der groBen Pfeilerhallen aus Haustein und Ziegeln (PA und PB), welche vor den zunichst weiterbestehenden Viersaulenraum und seine Seitenkammern gesetzt wurden. Ein sicheres Baudatum iiBt sich noch nicht geben, es kann jedoch nicht nach der friihen Regierungszeit Xusr6 I liegen. In der Folge wurde auch der Viersiulenraum durch Steinbauten ersetzt und iiber den westlich angrenzenden Lehmziegelgebiuden gew6lbte Steinbauten errichtet. Die gr6Btenteils gut erhaltenen Steingebaude wurden nahezu vollstindig freigelegt. Dabei fand sich in dem schreinartigen Kuppelraum (PD), welcher den Vierstiitzensaal des Lehmziegelgebiiudes ersetzt und den AbschluB der Saalfolge PA-PB bildet, ein dreistufiger, quadratischer Sockel von 1,65 m Seitenlange mit der Standfliche eines siulenf6rmigen, gerundeten Schaftes, wahrscheinlich der Unterbau eines groBen Feueraltares. Sowohl der Altarsockel als auch weitere Fragmente von Feuerschalen und kleinen Feueraltiren, die in und neben den Pfeilerhallen gefunden wurden, weisen dem Gebiudekomplex eine sakrale Funktion zu. Der Befund in den westlich angrenzenden Raumen und H6fen IdBt vermuten, daB gemeinschaftliche Mahlzeiten zu den Zeremonien geh6rten. Die Pfeilerhallen waren bereits im 9. Jh. n. Chr. weitgehend zerst6rt, wahrend die sie im Norden abschlieBenden Raume (PC u. PD) noch im 12. Jh. eine Wiederherstellung erfuhren. Areal n6rdlich vom Feuertempel: Das kleine Torgebiude (N) des groBen Vorhofes auf der Nordseite des Feuertempels wurde freigelegt. Es ist in islamischer Zeit durch Zumauerung des Haupteinganges und der westlichen Seitentiir in ein Wohngebaude umgewandelt worden. Der ursprtinglich quadratische Innenraum wurde durch eine Querwand geteilt, die FuBbiden mit Steinplatten belegt. - In einem neben dem nbrdlichen Eingangsiwan des Feuertempels gelegenen Raum (K) wurden zwei FuBb5den aus Sandsteinplatten und Ziegeln aufgedeckt. Ein Tiirdurchbruch zum Innenraum des Feuertempels ist m6glicherweise schon in sasanidischer Zeit angelegt worden. - In dem zum Feuertempelkomplex gehorenden Arkadenhof E wurden die Reste einer in islamischer Zeit errichteten Quermauer entfernt. Darunter lagen in der Hofmitte und vor den 6stlichen und westlichen Arkadenmittelpfeilern Fundamentierungen aus vermirteltem Bruchstein. Der istliche Arkadenpfeiler wird in FuBbodenhohe von einer 50 cm hohen, rechteckigen Offnung durchbrochen, deren Zweck unbekannt ist.

SURVEY

OF

EXCAVATIONS

197

Eine im letzten Jahr angeschnittene Eckbastion unmittelbar siidlich des Seerandbebauung: Westiwankomplexes lieB vermuten, daB der See ebenso wie der Tempelbezirk von einer bastionenbewehrten, inneren Mauer umschlossen war. Bestitigung brachte ein Suchschnitt auf der Westseite des Sees. Es wurde eine Mauer freigelegt, die sowohl in ihrer Konstruktionsweise als auch in ihrer Flucht dem westlichen Flfigel der inneren Umfassungsmauer n6rdlich vom See entspricht; sie ist zweischalig, so daB im Mauerinneren ein Korridor entsteht. Im FuBboden des Korridors lag eine sasanidische Silbermfinze (Xusr6 I, 1.-5. Reg.-Jahr). - In einem Suchschnitt auf der Siidseite des Sees lieB sich keine ungest6rte sasanidische Bebauung feststellen. Friihislamische Siedlungsreste reichen hier bis auf den Kalksteinuntergrund der Hiigelkuppe hinab. In den stark gest6rten Schuttschichten, die Reste der ffir die sasanidischen Anlagen charakteristischen Lehmauffiillung enthalten, wurde eine Silbermiinze Xusr6 II gefunden. Westiwan: Der im 13. Jh. n. Chr. errichtete rechteckige Mittelsaal zwischen den beiden oktogonalen Kuppelbauten auf der Riickseite des Westiwans wurde freigelegt. Es ergaben sich Aufschlfisse fiber die Raumiiberdachung dieses zu den grdBten Raumgebilden des il-khanidischen Palastes geh6renden Saales, fiber die Gliederung der die Westfassade der Anlage bestimmenden AuBenwand und fiber Wand- und FuBbodendekoration. Bei einer nachtriglichen Veranderung noch in il-khanidischer Zeit ist dem Saal ein nach Westen vorspringender Bauteil angegliedert worden, dessen GrundriB die Form eines halbierten Achtecks hat. Fund e: Zu den wichtigsten Funden aus vorislamischer Zeit geh6ren drei Tonbullen spitsasanidischen Datums, mit welchen die Offnungen von GefaiBen versiegelt waren, ein Siegelstein aus Chalzedon mit der Gravierung eines Vogels, ein mandelf6rmiges Gipsstuckelement mit der Darstellung einer doppelten Palmette und ein Bronzeanhinger in Gestalt einer Ente. In islamischen Schichten wurde qualitatvolle Keramik vor allem des 9.-13. Jh. n. Chr. gefunden. Neben den Grabungsarbeiten wurden umfangreiche Restaurierungen vorgenommen. Dank der Hilfe der iranischen Antikenverwaltung konnten besonders gefihrdete Teile des Feuertempels und des groBen Westiwans gesichert werden.
DIETRICH HUFF

The second season of excavation at Tepe Yahyd, supported by the National Science Foundation, Ford Foundation, and the Peabody Museum of Harvard University, was carried out between June 6th and August 24th 1969. This year's excavations concentrated almost exclusively on the horizontal exposure of Periods I, VI-VIII (see Iran VII, 1969). Period I: The top of the mound is capped by a substantial citadel identifiable as Partho-Sasanian. Ceramics consist dominantly of large storage jars with incised and raised motifs. Finer wares of plain buff slip or plain orange-red ware remain in the minority, as does a greenish-blue glazed ware. Of considerable importance was the recovery of a substantial amount of painted fine ware, a black paint on yellowish-buff. This ware was found in considerable quantity in a number of different areas all contextually datable to this period. This type of ware, we believe, has previously been identified as " Londo ware ", but incorrectly attributed to the early centuries of the first millennium B.C. It is characterized by hanging spiral curls and naturalistic motifs. Its clear stratigraphic association with Sasanian pottery and objects, i.e. a rock crystal Gayomard-type seal would appear to mark it as an early centuries A.D. ware or perhaps the last century B.c. A single plain red ware sherd, most probably from a large storage jar, contains an incised inscription tentatively identifiable as Pahlevi. A considerable quantity of iron in the form of a shaft-hole axe, daggers and trilobate arrow heads was also recovered. Decorative belt buckles and ornaments of bronze also appear in limited quantity. A considerable amount of glass, including some whole glass cosmetic vessels, were found. Architecturally this period incorporates large-scale construction of brickextending platforms from an outside containing wall. The several rooms exposed all contained numerous large storage vessels with potter's marks broken in antiquity. In the centre of this complex, excavation revealed the presence of a well-a considerable feat for securing water from the bottom of a

Tepe rayad

198

JOURNAL

OF PERSIAN

STUDIES

carbon dates will be submitted for both this period as well as all others. In future seasons, work will continue on this period while the completed excavation will allow us next to expose the Achaemenian settlement lying directly below and identified as such in the 1968 excavation. PeriodVI: Excavation of this third millennium settlement proved most rewarding. A large quantity of carved steatite bowls and objects were recovered. The incised motifs incorporated curvilinear and geometric designs; triangles, hachures, rectilinear designs; architectural scenes, faqades of doorway (hut-pot), and windows; as well as animal figures. These types of steatite incised bowls have been found on several sites in Mesopotamia and dated to E.D. III. At Tepe Yahyd we have examples of almost all previouslyknown motifs and several new ones. The presenceof unfinishedbeads, stamp seals, bowls, etc., of steatite indicates that these objects were locally made. The above objects of unfinished variety are also found in finished form, including a number of steatite stamp seals. A double-axe (shafthole variety) had carved on each side of the wing an eagle-a common motif on D and on Sfisa pottery cylinder seals of the late and middle third millennium. The presence of a very fine cylinder seal from this period also arguesto the mid- to late third millennium. This seal appearsto have its closest parallels with Akkadian type seals incorporating a vegetation god (?) with palm fronds extending from the shoulders and hands folded before the waist. Two almost identical carvings of the same personage wearing a horned helmet with hair tied in a bun are depicted in excellent workmanship. A single etched carnelian bead has also been found of the type found in the Indus and Mesopotamia; while a few fragmentsof ivory may have served for beads or inlay. The presence of this material is of considerable importance in shedding new light on the commercial relations between south-eastern Iran with the areas to the West as well as to the East. Analysis of the pottery necessitates further work. Ready parallels, however, can be seen in the painted wares of Bampuir(later levels) as well as the site of Shahr-i-Sokhta, while more distant parallels may be suggested at Mundigak, as well as at Sialk IV. The 1969 season provided a single C-I4 date for the lowermost of three architectural phases of this period; a somewhat high date of 3246?465 B.C. (WSU 876). The lower range here of 2780 B.C. could, however, be acceptable. An extensive series of dates has been submitted for the three phases of this period. Architecturallywe have exposed domestic buildings, a single kiln, and what appears to be a very substantial wall around the latest settlement. The latest occupation of this period would appear to date to the end of the third millennium, and sees the virtual disappearanceof painted pottery. Plain wares, many incorporating elaborate potter's marks, with three or four such marks on a single vessel, bear a striking similarity to the proto-Elamite signs evident at Sialk IV. With the extensive evidence for parallels both ceramic and in other materials to the West, on the one hand, and the East on the other, one cannot help thinking of the larger question of third millennium trade communications, the role played by early Elam, and the identification of elusive Magan. Certain it is that this occupation at Yahya during the third millennium has direct bearing on all of the above. PeriodVII: Continued work in this period clarified the stratigraphic order of the building levels and added a very considerableunderstandingto the architecturalcomplexity in the constructionof the fourthmillennium settlement. Pottery is dominantly a painted black on red ware with a continuation of
some of the earlier chaff-tempered coarse wares. Metal chisels, awls and pins, obsidian blades, marble bracelets and bowl fragments, turquoise beads, finely carved mother-of-pearl pendants and animal figures, and clay stamp seals were found in an extensive exposure of domestic architecture. There is a complete absence of any fortification wall about the settlement at this time. A single C-I4 date of 3630+410o B.c. (WSU 872) would be acceptable but for the consistently high ? reported for our dates by this laboratory. (The cause for this has yet to be satisfactorily explained by the laboratory.) The wonderfully preserved architecture, all of a domestic variety, the extensive horizontal exposure and clear stratigraphic association of the materials throughout this and the earlier Period VIII promise to add a considerable understanding of the late fourth millennium. Period VIII: Three major superimposed architectural levels, each with identifiable phases of rebuild, have been uncovered for this period. There is an obvious cultural continuity in these earliest

tectural configuration of this apparent citadel construction. It would appear from the evidence at hand that its history would extend from the last centuries B.c. to the early centuries A.D. A series of radio-

20 m. high mound. Considerable work remains in unravelling the complex stratigraphy and archi-

SURVEY

OF EXCAVATIONS

199

levels, observable in both ceramic continuity and architectural construction and orientation. Walls are constructed of variably sized thumb-impressed brick (chaff tempered) and bound in walls by a chaff tempered mud mortar. Rooms are of a consistently small size, hardly ever exceeding I -5 In. square. Often there is no visible access or doorway, while the inside of the rooms and floors are rarely plastered, as is common in Period VII. Evidence of reed impressions in mud brick-bats found on the floors of some of the rooms suggest the manner of roofing. Material found in the dozen odd rooms of the earliest level do not suggest either the function or the activity carried out in these small rooms. This small size would seem to eliminate them as living areas while there is no evidence for them being used for storage. A coarse chaff tempered " soft ware " is consistently present in each level. However, there may prove to be a significant difference after quantification of the ceramics is carried out. Surface treatment: plain red ware, red slipped ware, burnished red slipped, plain buff with red wash and/or fugitive red painted meandering lines, are associated in the same level. However, different percentages of association can be noted through the different levels together with obvious increases and decreases in the percentage of certain types through time. In the second building level and apparently restricted only to this level, we noted numerous examples of a bichrome painted ware: black and red painted geometric motifs on a fine buff slip. This pottery is without apparent parallels. A wide variety of different coloured flint was used for the production of tools. This material still awaits further analysis. The earliest excavated level appears to be contained by a wall, which was further clarified this season, necessitating a correction from last year's report which viewed the wall as considerably larger than its almost 2 m. width determined this year. A single C-I4 date of 4302 ?- 18o B.C. (GX-1509) can be reported from the 1968 excavations for the fourth level of construction, VIIID. Virgin soil was reached at the base of the mound beneath the lowermost excavated architecture (VIIIE), which we suspect dates from no earlier than the middle of the fifth millennium. The clear association of a wide variety of different type materials in good stratigraphic context and associated with architectural configurations promise to add considerably to our understanding of Iran in the first and third to fifth millennia.
C. C. LAMBERG-KARLOVSKY

II. SURVEY SECTION

TepeAbdulHosein
During our excavations at Bdbd Jan in 1969 we visited Professor Ghirshman's mound of Tepe Jamshidi, which lies in the southernmost part of the plain of Khawa close to the pass at the southern end of the Nehdvand valley. Our attention was drawn to a second mound lying about half a kilometre away from the main tepe above a huge spring which must form one of the main sources of the River Khawa. The sides of the mound had been extensively quarried away by the villagers revealing the lower strata. These consisted of deep bands of charcoal and occupation debris containing flints, but not a scrap of recognizable pottery until at least the top third of the mound was reached. The flint industry (see Figure), as represented by a random surface collection from both top and sides of the mound, falls within the Ganj Dareh, Sarab and Asiab range. Finely prepared pencil-shaped blade cores are smaller than those generally present at the sites mentioned. The small blade artifacts include backed blades, obliquely truncated blades, notched blades, and a single blade broken with secondary retouch and signs of polish along one edge. A single denticulated flake tool was found. Geometric microliths were not represented. No obsidian was found and the stone is mostly chert of good quality. The many mountain streams in the vicinity of the site are a possible source for the stone. An eroded side of the mound showed a mud-brick wall and a series of well-defined charcoal lenses in section. Fragments of unidentifiable bone were also visible in the section. Protected on one side by a mountain range and lying in an open and well-watered valley the area is ideal for sheep and goat herding. Over 40 m. in diameter, and approximately 6 m. high, the mound is at least as large as Ganj Dareh or Tepe Sarab.

200

JOURNAL

OF PERSIAN

STUDIES

The pottery, which appears to be later than the main deposit described above, comprisesa coarse, heavy, straw-temperedware, with red slip. Simple bowl rims predominate. Similar pottery has been picked up on numerous tepes between Harsin and Khorramabhd.
CLARE GOFF and JUDITH PULLAR

SouthWestKhurdsdn From June to August 1969 B. J. Spooner, V. C. Pigott and J. Bockstoce of the Department of Anthropology and the University Museum of the University of Pennsylvania conducted an archaeoof logical surface survey in the south-west of Khurdsin. The survey covered the areas of the shahristdns Kashmar and Tabas from the city of Kashmar in the north to the village of Naiband in the south, and was undertaken as part of a long-term researchprogramme concerned with the cultural ecology of the desert and semi-desert regions of eastern Persia from prehistoric to modern times. A series of eight mounds was recorded and sherded in the relatively fertile plain west of the city of Kashmar. The pottery was largely Islamic but preliminary analysis suggests that some of it probably dates back to the first millennium B.C. No other sites likely to yield evidence of pre-Islamic occupation were located anywhere else in the region, but numerous forts and mines were noted which would be likely to repay further investigation by an Islamic historian/archaeologist.
BRIAN SPOONER

Second Millennium Sitesin Khazistdn The pottery sequence for this survey was developed from studies of unpublished pottery from Haft Tepe and the study collection from ProfessorGhirshman'sexcavation of the Ville Royale at Suisa.

SURVEY

OF EXCAVATIONS

201

This was then employed in field surveysin four areas: (i) Deh Lurdn; (2) central Khizistdn; (3) Ram Hormuz; and (4) the region between the Duveyrij and Karkheh rivers; the last of these was only briefly investigated. Dek Lurdn: In this area six possible Elamite sites were discoveredout of a total of 293 sites surveyed. The most important, Tepe Mussian, is the largest early second millennium site outside of SUsa. Preliminary ceramic analysisindicates that the centre of settlement in the Deh Lurdn area had shifted from Tepe Mussian to nearby Tepe Goughan by the last half of the second millennium. Study of the second millennium Elamite pottery recently excavated at Tepe Farrukhdbadshould provide a more precise dating of the Deh Lurin sites.

S%% Luran SDeh

4%%

% %
% %

% %
%%

SHorffwz

Amms SH

m
yed,
0

z
30(ai

Sce

10 20 0 io 20

Central Khazistdn: In the second area surveyed, a total of fifty-four sites, observed by Adams to to the second millennium, were re-surveyed: several new Elamite sites were found. Slightly to belong the south of the area of Adams' survey a major Elamite site was found on the Karun river near the Golestan suburb of Ahwiz. A comparative study of pottery from Haft Tepe and Srisaindicates that Haft Tepe may include a phase not well representedin the Ville Royale, chantier A, at Sisa. Despite the possibilitiesof regional variation and functional differences of town quarters investigated, it may be speculated that at this there is no major settletime the centre of settlement shifted from Srisato Haft Tepe. Outside of SUisa ment (and few minor ones) which belong to the early part of the second millennium, but by the last half of the millennium the number and size of the sites expand. A high percentage of these major late second millennium sites lie on the banks or in the flood plains of the principal rivers. This expansion may well have been based on the control and use of the major rivers of Khfizistan.
17

202

JOURNAL

OF PERSIAN

STUDIES

RdmHormuz: The third area, in and around the town of Ram Hormuz, was surveyed with Dr. Henry T. Wright of the University of Michigan. Seven of forty sites surveyed were dated to the second millennium. They were characterized by a non-compact, often black cored, buff slipped buff ware whose colour ranged from reddish to brownish to greenish buff. The ware was often decorated with simple geometric designs in black-brown or red-brown paint placed on the vessel shoulder and/or rim. Final publication of the pottery from Tell-i-Ghazir excavated by Dr. D. E. McCown may give a clearer picture of the time range of this ware. It is uncommon in central Khizistan or Deh Luran (the Khtizistdn survey produced only seven sherds on four sites and none were found in the Deh Luran survey), but a few sherdshave been found in the recent Haft Tepe excavations, where the ware has been tentatively dated to the middle of the second millennium. A few painted sherds from Sfisa which are said to resemble those of Giydn III(i) (ArtsAsiatiques XIII (1966), p. Io) have not been examined by this investigator; thus it is difficult to compare them with the Haft Tepe and Ram Hormuz painted sherds. To judge from the illustrations (ibid., fig. 24) there are differences. In any case the distribution of this ceramic should prove interesting. Surface finds in the area included an inscribed brick and a tablet fragment from the major second millennium site in the region, Tepe Bormi. The Duveyrij Karkheh and Rivers: The fourth area between the Duveyrij and Karkheh rivers was only briefly investigated. Preliminaryinvestigation here shows a number of promising prehistoricsites as well as several early to middle second millennium sites. More intensive work is planned, possibly in late 1970.
ELIZABETH CARTER

The Deh Lurdn Region The primary purpose of the reconnaissance and associated limited excavations was to provide a basic fund of information for an intensive study of the prehistoricand early historic water control and irrigation systems. Other than the few sites recorded through the pioneering reconnaissanceundertaken by Gautier and Lampre (J. E. Gautier and G. Lampre, " Fouilles de Moussian ", MDP VIII (1905), pp. 59-148) and the limited survey by Hole and Flannery (Frank Hole, " Archeological Survey and Excavation in Iran, 1961 ", Science CXXXVII, no. 3529 (1962), PP. 524-6), the locations and relationof sites representingthe various periods of occupation on the Deh Lurin plain were unknown. ships The area surveyed surroundsthe presently inhabited small town of Deh Luran and extended into the lower foothills of the Kih-i-Siyah to the north, those of the Ktih-e-Qutbeh (Jebel Hamrin) to the south, and a short distance west of the Mehmeh river and east of the Duveyrij river (see map accompanying preceding report). This forms a roughly rectangular-shapedzone some 26 x 45 km. in maximum dimensions and approximately Iooo square kilometres in area. Although most of the final analysis remains, it is evident that the 1968-69 survey of the Deh Lurin plain has largely succeeded in fulfilling the objectives of the research proposed. The following represents only a brief list of our discoveries. (I) A total of 293 archaeologicalsites and features,including water control and irrigationfeatures, was recorded. This representsan estimated 8o per cent sample of the total number visible on the surface of the plain. This percentage was derived through the intensive coverage of about 70 per cent of the Iooo square kilometres originally designated for study. (2) The sites and features recorded provide evidence of continuous human occupation from at least as early as the Bus Mordeh phase (c. 7500 B.c.) well into the Islamic period (c. A.D. 1300). (3) New techniques enabled us to define the periods of occupation of individual sites and, in many cases, to define the period represented at one particular point on the site and to observe the direction(s) in which the site expanded or " grew ". The survey techniques employed also served to define or predict functionally specific activity loci, which consisted of ceramic kiln areas, flint knapping areas, glass manufacturing areas, smelting areas, burial areas, habitation areas of apparent differing social rank, possible areas of political and religious activities, etc. (4) The reconnaissance produced detailed information concerning settlement patterns and community patterns of each of the cultural phases of the region. Such information pertaining to the Sasanian and Islamic periods is complete because of alluviation, which has been great enough to

SURVEY

OF EXCAVATIONS

203

preserve the architecturalremains of the inhabitants but has not as yet accumulated to the point where it masks the structures. While perhaps more apparent than real, due to the quantities of alluvium covering earlier occupations of the plain, the Sasanian and Islamic periods seem to have witnessed a dramatic population expansion. This is reflected not only through a sharp increase in the number of sites, but also in the constructionof sites and irrigation featuresin areas that had apparently previously not been inhabited, or only very sparsely so. (5) Much information was derived relating to the form and function of water control and irrigation systems. The Sasanian and Islamic systemsare the most fully exposed. However, the possibilityfor recovering nearly comparable data for earlier systems seems quite likely pending future selective tests. While actual features and the techniques utilized remain unknown, there is good floral evidence (Frank Hole, Kent V. Flannery, James A. Neely, op. cit., pp. 406-09) that some form of irrigation was in use as early as the Sabz phase (c. 5500-5000 B.c.). The earliest irrigation featuresyet discovered on the Deh Luran plain apparently date to the early part of the Uruk period. These consist of dams built across small, now intermittent drainages and canals leading from these dams to terraced fields paralleling the drainage channels. In succeeding periods the use of waters from the Mehmeh and Duveyrij rivers permitted an increase in the areas irrigated and a concomitant growth of the canal systems. The first construction of qanat systems apparently took place during the Achaemenian period, or perhaps earlier. Qandt systems are present in both the " classic " form, taking water from the groundwater table or from aquifers, and in a modified version where the system is used to tap water from the deeply incised rivers. The latter form served to filter out much of the alluvium carried by the rivers as the water seeped through the river banks into the systems. The Sasanian and Islamic periods may have witnessed the greatest use of irrigation in the history of the region. Both large and small canals and qandt systems were built. (6) The survey has presented an opportunity tentatively to date and to define the typological nature and cultural context of " utility ware " ceramics. (7) Based primarily on the ceramic assemblage, but also including architectural, settlement and community pattern information, it seems possible to define an early and a late phase of the Uruk period in the Deh Lurdn region, which correspond with cultural developments in parts of Mesopotamia (Henry T. Wright-personal communication). (8) The distribution of Achaemenian sites across the Deh Luran plain in a general north-westsouth-east direction has led to the tentative location of the Achaemenian " Royal Road ". The identification is pending the study of historical documents and aerial photographs. (9) An extensive study of the soils from throughout the Deh Luran plain by the Drs. Kirkby, the indicates that, unlike our previous findings (ibid., pp. 19-20), project's geomorphologists-geographers, a relatively small portion of the region was adverselyaffected through the use of salt-bearingwaters only for irrigation. These findings correspond well with those of the survey-an apparently continuous and intensive occupation of the plain as well as the intensive and widespread use of irrigation throughout the later prehistoric and the historic periods until about A.D. 1300. JAMES A. NEELY
Islamic Monumentsin Kirmadn KhurdsdnProvinces and In the summer of 1969 surveys were undertaken in Kirmin and in Khurashn with the intention of examining known Islamic monuments and locating hitherto unpublished ones, both in the provinces and the routes giving access to them. A number of monuments were visited and studied which have already been recorded in various publications, but from a further study of plans and photographs it is hoped that additional information can be obtained which may be published at some future date. Certain monuments were, however, noted which have either not been previously recorded, or merely noted in passing reference. In Athdr-e' irdn I (1936), M. Andr6 Godard noted the portal surmounted by two minarets attached to the little mosque of Imam Hasan in Ardistin. When Godard saw this monument only one of the two minarets remained: this is now considerably shorter. He described this portal as being that of a Seljuq

204

JOURNAL

OF PERSIAN

STUDIES

madrasa, now disappeared. The small mosque alongside however, which contains the remains of a stucco mihrdb, may well also be part of the original madrasa. Although now very ruinous, sufficient remains of the mihrdb (P1. VIIIa) to show that it must have been important. As yet it has not been possible to decipher the remains of the inscriptions,but certain of the patterns, notably that in the centre of the remaining central arched panel, can be paralleled on the outer decoration of the portal, and also in the nearby Seljuq Masjid-iJami', as well as in the Khwaja Atabek in Kirman. The Ardistin Masjid-i Jdmi' also contains three mihrdbs having similar foliate patterns to the remaining outer bands of the Imam IHasanmihrdb. The brick pattern in the adjacent alcove would appear to confirm a Seljuq date for the building. In the same article Godard mentions the Masjid-i Pamenar at Zavare, but dismisses the Mongol stucco work there as being of no importance. Donald N. Wilber also mentions this stucco in the Architecture IslamicIran: the II KhanidPeriod,although he was unable to visit the mosque. Further of examination reveals that the stucco in question consists of six mihrdbs and surrounding decoration extending as far as the soffits of the arches, much of which is of extremely fine work. A detailed study of the stucco in the mosque, and the possible Seljuq work which it overlays, is planned for the future. Some 3 km. to the east of Nayin is the town of Muhammadiyya in which Godard noted two small mosques of the type which he calls Iwdn-e Kdrkha and of one of which he published a plan in L'Artde l'Iran. This mosque, which Godard calls the Masjid-i Muhammadiyya and which is also known as the Masjid-i Sar-i Kiicha, contains a very fine painted Kaificfrieze which runs round the entire interior of the mosque (P1.VIIIb). Godard dated these mosques to the fourth and fifth centuries of the Hijra and it would seem possible that the inscription dates from the same time. It is hoped to undertake further study of this inscription in the near future. Some 6 km. to the south-west of Mahan on the road to Chtipar is a small hitherto unrecorded monument which would appear to be a Timiirid mausoleum (P1. IXa). Local tradition ascribes the tomb to Shaikh 'Abd as-Salam, the teacher of the Shaikh Nfr ad-Din Ni'mat Allah who died in 1431 and whose tomb is nearby at Mahan. Although the mausoleum is in a somewhat ruined condition, the dome is intact and much of the original tilework still remains. The dome is of baked brick with patterns formed by light and dark bricks, the top of the dome being in light buff coloured bricks below which a series of light buff lozenges depend against a band of darkerbricks. The high drum is ornamented with inlaid tiles each tile being of a single colour and inlaid separately. Below the dome is a band of inlaid Kiffic lettering, below which the drum is encircled with lozenges, each lozenge containing a single word in square Kafic. The tomb chamber is situated immediately below the dome and has a domed plaster ceiling. Adjoining the central tomb chamber is a further single chamber containing another tomb. These two rooms would appear to be all that remains of the original mausoleum, the entrance portico and adjoining room being of more recent date. On the outer wall of the secondarytomb chamber is a simple form of mihrdb. The mausoleum is surroundedby an enclosing wall, one side of which still retains the arcading which probably originally encircled the enclosure. The deserted town of Bam east of Kirmin, was studied and plans made, including detailed plans of the Masjid-i Arg, the Citadel, the main gate of the town, and one of the large private houses near the mosque. Particular notes were made of a large Safavid mihrdb in the Masjid-i Arg which had a finely
traced and intricate star pattern, as well as an inscription frieze. In Gurgin, formerly AstarIbAd, the Masjid-i Jami' contains a minaret which is of considerable interest (P1. IXb). Local tradition insists that the minaret is older than the mosque and is in fact Seljuq. The form of the minaret is typical of the Mazandaran area, but certain ornamental details would tend to confirm its Seljuq dating. The inscription is framed above and below by a band of ornament which is typically Seljuq; the brick measurements, 20 cm. long by 3 cm. thick, as measured in the thickness of the wall inside the window, conform to Seljuq standard sizes; and the plain brickwork of the lower section with the spacing left between the horizontally placed bricks is typical of Seljuq workmanship, although the patterned area has a pattern which cannot immediately be matched in other Seljuq minarets. A recent Iranian publication, Fihrist-i bindhd-yitdrikhi va amdkin-i bdstdni-yiIrdn, refers to the minaret as Seljuq, but the form of the letters of the inscription is unique and has not yet been read. Rabino makes

no mention of this inscriptionin his work Mdzandardn Astardbdd, and although he spent some time in the

SURVEY

OF EXCAVATIONS

205

mosque and deciphered all the other inscriptions. The brick pattern is in many respects similar to that of the nearby Imdmzdda Nilr, to which a fourteenth century date is assigned. At Sabzavar, known in earlier times as Bayhaq, the Masjid-i Pdmenar has a minaret which can be assigned to the Seljuq period. The inscription, which is composed of cut brick, is almost completely illegible (P1.Xa). The brickworkpatterns of the minaret are typically Seljuq and can be paralleled in many instances, although the interstices between the bricks have recently been filled with mortar so that the pattern is less distinguishable. The pattern bands above and below the inscription represent an unusual form of cut brick and if the mortar can be cleaned away from some of them, may give us a new Seljuq brick design. An unusual early squinch form was noted at the Ghaznavid caravanserai of Ribat Mahi, some km. east of Mashhad (P1. Xb). This caravanserai, originally called Ribat Sdhi, was first noted by 80 IrdnIV (I949), although he did not visit it. The first photographs appeared in an Godard in Athdr-6 article by Derek Hill in the LL.N. of August I3th 1966, and again in the second edition of Islamic and Architecture its Decoration D. Hill and 0. Grabar (1967). by The simple semi-dome of the squinch at Ribdt Mdhi is bisected by a cannular groove which runs from the apex of the main squinch arch to the corner angle of the dome chamber below the zone of transition. It may be significant that the brickwork of the arch and semi-dome of the squinch is corbelled and that the lobes of the squinch are purposely separated. It would therefore appear that, in this context, the cannular groove had a specific function rather than being a decorative element, since the line of separationlies along the line of maximum stressof the squinch. One might suggest that the separation of the lobes subsequently evolved into the trilobed squinch. This evolution would be possible in that Ribat Mahi was probably built in 1020-21 and thereforeantedates the Duvdzdah Imim at Yazd by nearly twenty years, the squinch at the Duvdzdah Imam always being consideredthe earliest dated form of the trilobed squinch, and as such the precursorof all subsequent Seljuq squinches.
ANTONY HUTT

IslamicMonuments Northern in Iran In August and September 1968 a preliminary survey of the imdmzidas of Gilan and western Mdzandardnwas carried out. About forty shrines and towers were visited, of which only a very small proportion were of architectural interest. In three imdmzddas, two of them hitherto unpublished, extensive frescoes were recorded, with subjects similar to those on the tomb of Shah Zaid at Isfahdn. South of Rasht the overgrown Qal'a-yi Rfidkhdn, mentioned by Rabino, was visited, as well as the remains of a fine brick bridge, and a well-preservedminaret with decorated brick patterns. Both the latter monuments, which are in the neighbourhood of Minar-i Bdzar, are of uncertain date, though it is unlikely that they are earlier than the Safavid period. The funeral towers of Amul which still remain standing were planned and recorded, including one on the Mahmildabdd road which is apparently unknown. This, like another unfamiliar tower recorded at Kaman Kola just north of Amul, can probably be dated in the fifteenth century. A much earlier dating, possibly in the eleventh century, seems required for the delightful circular red brick tomb tower locally known as that of Darvish Muhammad, near Pul-i Safid, though its shattered sarcophagus is probably roughly contemporary with that of the nearby imdmzdda at Silt Kola, dated 1429.
In the course of October 1969 some little-known buildings to the south of the Caspian provinces were visited, among them the possibly eleventh century " Pir " tomb, a domed square at Takistin near Qazvin. This was thoroughly recorded with plans and sectional drawings. The same was done for the mausoleum locally known as the Mazar-i Sipahsalar (P1. V), some 20 km. from Tabas, which demonstrates the widespread popularity of the large open tomb with an intermediate storey below the dome. Externally the building bears a remarkable resemblance to the Mongol mosque at Aziran, the eleventh century tomb at Sangbast (before its recent restoration), and, more distantly, to the Saljuq Jabal-i Sang.

It is also closely linked to open mausolea in the area to the north and east of Mashhad, though its
material-largely mud-brick-and its limited internal decoration suggest that it is probably later than the Mongol period to which these open mausolea have been assigned, perhaps fifteenth century. The complete lack of tilework would be quite exceptional for a building of this scale in later periods.

206

JOURNAL

OF PERSIAN

STUDIES

Nearby, the Masjid-i Jum'a of Firdaus was recorded. Though badly damaged by the 1968 earthquake, it has retained most of its single iwdn, whose arch still stands. The elegantly incised brickwork, and lavish use of red plaster, rather than faience, to achieve a wider palette suggest a date around 1200. This is supported by striking resemblancesto the mosque at Zuwzan, though at Firdaus the occasional pieces of faience (there is no mosaic faience as at Zuwzan) are still only used for the grace notes of the composition, rather in the manner of the Gunbad-i Surkh at Mardgha.

R. HILLENBRAND

Iran in IslamicTradeRoutes Southern From September 1968 till May 1969 a survey was undertaken in the course of research into the trade patterns of southern Iran in the medieval period. The area investigated included much of Kirman, of Fdrs except the Kilrah of Arrajdn,and a reconnaisanceof Sistan. Importance was attached to the identification of sites along historically attested trade routes. A total of 370 mounds was visited, mainly of the Islamic period. In addition material from 21o sites discovered by Mr. W. M. Sumner in the Kur river valley was classified and sites with medieval pottery were revisited. A special attempt has been made to identify pottery kilns and production centres, which have been located at over a dozen sites, those of the greatest importance in the medieval period being at Sirjdn, near Istakhr and near Hormuz. The products of the identified kilns and of other centres were distributed in strikingly clearly defined regions which overlap very little. The Persian Gulf coast and the garmsir(the " warm land " area predominantly of date cultivation) formed an essentially homogeneous unit characterizedby wares of types common in Mesopotamia and while pottery from the inland centres was exceptionally rare. found stratified at SirMf, In the interior three principal pottery distribution areas can so far be distinguished. The ceramics of the Kur river basin and the adjacent plains were dominated by wares made in the kilns near Istakhr. The central plains of Kirmdn and eastern Fdrs were served by kilns centred at Sirjan. Desert Kirman and Persian Sistdn form an independent region with characteristicpottery types. In the absence of a stratigraphic chronology, such as the excavations at Sirdf provide for the coastal regions, it is not possible to make any fine chronological distinctions in the pottery, but it does appear that the distribution of ceramics was determined principally by natural lines of communication rather than by fluctuating political boundaries. Indeed, in southern Iran long distance trade overland in pottery seems to have been unknown except during the period of florescenceof the " Saljuq " lustre ware kilns of the central plateau; though great cities such as Sirjdn, Kirman, Istakhr and Jiruft enjoyed imported luxury wares at all periods. On the coast, however, pottery and glass were transported long distances in large quantities. Even pottery of low value such as the coarse ware bowls from the Sirdfkilns' is found as far distant as Minib (over 500 km. away). Far Eastern pottery and porcelain were used in great quantities especially in the period after the introduction of" Chekiang celadon ",2 and on several sites comprise over 90 per cent of all glazed ceramic. Yet outside the garmsir very few Far Eastern sherds have been found at any site except Sirjin. In the period of the prosperityof Sirdf before the introduction of " Chekiang celadon " the three sites of Sirjan, Istakhr and Kirman city are the only inland mounds with even a few Oriental sherds. The forty-two sites in the interior with fragments of later celadon or blue and white only
yielded a total (excluding Sirjan) of ninety-one sherds. In comparison, over 10,000 Far Eastern sherds found in the garmsfrtestify to the importance of maritime contact with the East from the earliest Islamic period. An investigation of oyster middens on the Persian Gulf coast dated by associated pottery and glass seems to indicate that the pearl beds which played such a major role in the economic structure of the region have moved little since Sasanian or at the latest early Islamic times. Accounts of several caravansarai and other buildings discovered in the course of this work are being prepared for publication. The prehistoric material is currently being studied by Mr. W. M. Sumner.
1 David Whitehouse, " Excavations at Siraf; First Interim
Report ", Iran VI, p. 16A a This first appears at Sirdf in period 4: ibid., p. 17; " Excavations at Sirdf; Second Interim Report ", Iran VII, pp. 59-6o.

SURVEY

OF EXCAVATIONS

207

It is hoped that during the next year the survey can be completed along the entire coastline of Kirman and Fdrs, as well as extended inland especially around the great medieval city of Sirjan, whose site was located during this year's survey.3
ANDREW WILLIAMSON

ADDENDUM TurengTipde ' Les fouilles de 1969 Tureng T6p6 ont eu un double objectif: d'une part poursuivrel'exploration du tip6 principal (" Mound A " de Wulsin), d'autre part dlargirle sondage entrepris precedemment dans la region SW du site, qui avait r6v616l'existence de couches plus recentes qu'Hissar IIIC. Ce dernier objectif nous a amends h'fouiller une surface de 00oo carr's m2 proximit6 des EI et E2. Plusieurs couches de la p6riode Tureng T6p6 III C2 ont et' atteintes et nous nous sommes arrAtis au moment oi furent rejoints les niveaux mieux connus de la periode Hissar III C (= Tureng Tep6 III CI). Divers edifices de brique crue ont 6t6 decouverts, dont l'un 6tait fonde'sur un lit de grosses pierres et de galets. La ceramique 6tait assez abondante et confirme les conclusions auxquelles nous avions pu aboutir precidemment, en ce qui concerne notamment le r61lede la plaine de Gorgan ta l'origine des cultures taceramique grise qui apparaissentdans le nord-ouest de l'Iran durant la seconde moiti6 du IIe mill6naire. En particulier un fragment d'un bec a jabot analogue ta ceux de Khorvin et d'autres sites a ft6 pour la premiere fois decouvert dans cette region de l'Iran. Des 6chantillons de charbon de bois ont pu etre prdleves. Par ailleurs nous avons relev6 dans les briques crues de ces sondages d'assez nombreux tessons de petite taille, tadecor noir sur fond d'argile rouge lissde, analogues a ceux qui avaient d6jit et ramassis dans le carr6 E2. II apparait d6sormais que cette cdramique pr6sente d' troites analogies avec celle d'Ismailabad-Cheshmeh Ali. Malheureusement la couche d'habitat correspondante ne pourrait etre atteinte qu'au prix d'un effort considerable d'assechement de la nappe d'eau souterraine. Le tepe'principal a 6tdl'objet de plusieursoperationsdistinctes. Nous avons en premierlieu continue la fouille du plus profond des sondages de 1967 et atteint finalement la base des couches sassanides: il s'avere desormais que celles-ci correspondent exclusivement at la derniere p6riode de l'histoire de qui l'empire. Il est probable que les forteresses se sont succeddau sommet du tipe ne sont pas ant'rieures ' F l'entreprisede fortificationde la frontiereseptentrionale laquelle se consacra Chosrois ler (531-79). La fouille de ces forteresseselles-memesa 6td poursuiviedans le quart NW du tepe; des murs de brique crue, temoignant de nombreux remaniements, mesuraient plus de 5 m d'epaisseur. Le materiel etait malheureusementtres pauvre. Dans ce meme secteur NW la forteressela plus recente avait 6t6, apres une longue periode d'abandon entamde par une fosse carrde,de quelque Io m de c6te, dans laquelle un remblai d'argile tres fine avait 6te tass6 pour servir de soubassement't un edifice islamique. Celui-ci, entierement stuque, comportait quatre iwan qui se faisaient vis a vis de chaque c6to d'une cour pourvue en son milieu d'un idicule
carrC t colonnettes de stuc. De nombreux remaniements ont ensuite alt6r6 la physionomie de cet edifice, que l'on peut probablement dater du Xe, atla rigueur du XIe si&clede notre are. La c~ramique y 6tait rare; toutefois un fragment 6maill6 vert fonc6 t mouchetures bleues a 6t6 dtcouvert entre deux couches de stuc. Or de nombreux vases de cette cat6gorie ont ?td trouv6s dans des fosses qui, aprbs l'abandon de cet edifice, l'ont travers6 en plusieurs endroits. Dans ces fosses une abondante et a 6t6 relevte; plusieurs vases comportaient une inscription en caractdres pseudomagnifique cdramique coufiques qui devraient dater du Xe ou du XIe si&cle. D'ailleurs la c~ramique dite de Gorgan, t decor si ce n'est peut-etre par un unique tesson. Il est bleu noir sur fond bleu, n'y 6tait pas reprdsent&e, que tous ces vases proviennent d'un autre batiment, t peine plus recent que l'edifice gt iwan, probable et qui se situerait dans la moiti6 est du tip.
Miles in Persia (London 1902), pp. 430 ff. a The site is different from that mentioned by Sir Percy M. Sykes, Ten Thousand

208

JOURNAL

OF PERSIAN

STUDIES

Enfin plusieurs sondages ont 6t6 entreprissur la pente sud du tep6 de fagon 'a en 6tablir la stratiImm6diatement en dessousse situe le grand batiment brflle dont les ruines sont visibles tout autour du tepe; nous avons pu toutefois constater qu'apres sa destruction par le feu cet edifice avait ete reconstruit au moins une fois. La ceramique permet de dater ces constructionsde l'Age du Fer: elle ressemble a celle qui a 6t6 trouvee au sommet du " Mound C " et qui est probablement de peu anterieure a 1'poque achemenide. Il n'est donc pas exclu que ces batiments gigantesques soient eux aussi des forteresses,que l'on pourrait songer a dater de la periode made. L'un des objectifs de la prochaine campagne sera de preciser la chronologie de ces niveaux intermediaires entre l'Age du Bronze et l'6poque sassanide. ' Pour le moment on peut 6tablir comme suit la succession des cultures qui ont laiss6 des vestiges Tureng Tepe: IA: culture de Jeitun; IB: culture d'Ismailabad; IIA: ceramique grise de type Hissar II et ceramique noire sur rouge de type Shah Tep6 III; IIB: ceramique grise de type Hissar II; IIIA-III CI: civilisation d'Hissar III. ' III C2: derniere phase de l'occupation a l'Age du Bronze, posterieure Hissar IIIC (XVIIe-XVIe
graphie complete. Sous les niveaux sassanides, nous avons atteint quelques couches caracterisees par une ceramique extremement sonore, rougeatre "ataches noires, qui pourrait dater de l'6poque parthe.

siecles ; ?) IV: Age du Fer (VIIe-VIe siecles?); V: epoqueparthe(?); VI: fin de l'empiresassanide(VIe-VIIe siecles); VII: batimentislamiqueau sommetdu grandtep6 (Xe-XIe siecles); VIII: vestigesislamiques sud-ouest site (XIIIe si&cle). au du
JEAN DESHAYES

ABBREVIATIONS AASOR AfO AJA AJSL AJ AK AMI ANET AOr Arch Anz AS BA Besch BASOR Belleten BGA Bib Or BSA BSOAS CAH CIA DAFA EI ESA IAE ILN Iranica JA JAOS JEA JHS JNES JRAI JRAS KF LAAA MAOG MDOG MDP MJ OIC OIP OS PZ RA RCAS REI SAA SAOC Sov Arkh SS Survey TT WO WVDOG ZA ZDMG Annual of American Schools of Oriental Research Archiv fidrOrientforschung American Journal of Archaeology American Journal of Semitic Languages and Literatures Antiquaries' Journal Antike Kunst E. E. Herzfeld, Archaeologische Mitteilungen aus Iran Pritchard, Ancient Near Eastern Texts Archiv Orientalny Archiologischer Anzeiger Anatolian Studies Bulletin van de Vereeniging ... de Antieke Beschaving, Hague Bulletin of American Schools of Oriental Research TUirkTarih Kurumu: Belleten Bibliotheca Geographorum Arabicorum Bibliotheca Orientalis Annual of the British School at Athens Bulletin of the School of Oriental and African Studies Cambridge Ancient History Corpus Inscriptionum Arabicorum D6l6gation Archeologique frangaise en Afghanistan, memoires Encyclopaedia of Islam Eurasia Septentrionalis Antiqua E. E. Herzfeld, Iran in the Ancient East ('94') Illustrated London News Iranica Antiqua Journal Asiatique Journal of American Oriental Society Journal of Egyptian Archaeology Journal of Hellenic Studies Journal of Near Eastern Studies Journal of the Royal Anthropological Institute Journal of the Royal Asiatic Society Kleinasiatische Forschungen Annals of Archaeology and Anthropology, Liverpool Mitteilungen der altorientalischen Gesellschaft Mitteilungen der deutschen Orientgesellschaft Memoires de la DClCgationen Perse Museum Journal, Philadelphia Oriental Institute, Chicago, Communications Oriental Institute, Publications Orientalia Suecana Praehistorische Zeitschrift Revue d'Assyriologie Royal Central Asian Journal Revue des etudes Islamiques Soviet Anthropology and Archaeology Oriental Institute, Studies in Ancient Oriental Civilisation Sovetskaya Arkheologiya Schmidt, H., Heinrich Schliemanns Sammlung trojanischer Altertumer A Survey of Persian Art from Prehistoric Times to the Present, ed. A. U. Pope,
Oxford, 1938

Tarih, Arkeologya ve Etnografya Dergisi TUirk Die Welt des Orients Wissenschaftliche Ver6ffentlichungen der Deutschen Orientgesellschaft Zeitschrift flir Assyriologie Zeitschrift der Deutschen MorgenlaindischenGesellschaft
209

NOTES ON TRANSLITERATIONFOR CONTRIBUTORSTO 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 transliteratedaccording to the table in Kent, Old Persian.Grammar, Texts,Lexicon, p. 12; that for Manichaean Middle Persian and Parthian, the transliterationsystem given in Andreasvol. III, p. 66, should be used; whilst for Pahlavi, the table of Henning, Mitteliranische Manichaica, alphabets given in Nyberg, A Manualof Pahlavi,new edition, p. 129, may be used as a reference for transcription.
II. ISLAMIC AND MODERN PERSIAN

The system used for the Cambridge Historyof Islamshould be used here as far as possible. Consonants (a) Arabic Z 3q k Sb , s ) sh J 1 St th m , s , C j n h .t .d w ( kh .h

b d
Sdh

'

-a (in construct gh state: ; f -at) J r o (b) Persian additional and variant forms. The variant forms should generally be used for Iranian names and for Arabic words used in Persian.
p j 3 z zh ?g v s_ .; Sch The Persian " silent h" should be transliterateda, e.g. ndma. (c) Vowels Arabic and Persian. Short: a
u

Long: Ior 3
J

Notes

?u
~"s

Doubled , iyy (finalform:i) Diphthongs?' au

ai

I. The izdfa should be representedby -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 Abfi and Dhai (Zfi) should be omitted before the definite article, e.g. Abu '1- Abbds (but Abi '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; contributorsshould 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 orthography. 4. Where classical Greek and Latin renderings of Old and Middle Persian names exist, these familiar forms should be used for preference.
210

BRITISH INSTITUTE PERSIAN OF STUDIES APPLICATION INDIVIDUAL FOR MEMBERSHIP

NAMEE ............................................................................................................................................................... DESCRIPTIO N ...................................................................................................... (Title, Decorations,Degrees) ADDRESS

........................................................................................................................................................

OCCUPATION ........................................
Interest in or connection with Iran.......................................................

orderfor this sum.

I wish to apply for Membership of the British Institute of Persian Studies as an individualsubscribing member paying*?1/?4 per annumand *enclose cheque/Banker's *1 wish to make a covenant for a period of seven years. Signature......................................................... Date..................................................
* Delete whicheveris inapplicable.

Paymentmay be madein Dollarsor Rialsas follows: U.S. $10 or Rials720. Price of BackNumbers:?5 ($12 or Rials900). For the convenienceof Subscribersin Europe,paymentmay be made direct to the British Bankof the MiddleEast,7 KingWilliam Street, London,E.C.4.PleasespecifyAccount No. 654-674-01.

APPLICATION MEMBERSHIP A CORPORATE BY BODY FOR

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'sorder for this sum is enclosed. Signature......................................................... Date................................................
* Delete whicheveris inapplicable.

211

MEMBERSHIP-SUBSCRIPTIONS Under paragraph i i of the Articles of Association of the Institute, the Council has prescribed the following scale of subscriptionfor members. i. Individuals (a) Those subscribing ?4 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 ?i per annum. They will have rights as in (a) above, except that they will not receive the Journal. 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 (paragraph9 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.

212

OF FORM COVENANT

sum of

..................................................................................................................................................................... with the BritishInstitute Persian of Studiesthat for a periodof seven herebycovenant shallbe the shorter yearsfromthe date of this Deed or duringmy lifetime(whichever period)I will payto the said Societyfrom my generalfundof taxed incomesuch an Rateyieldsthe net amountannually afterdeductionof incometax at the Standard as

..............................................................................................................................................................................

Thefirst payment to be madeon the.................................................... is Dated this ............. ...19 ....... ............... ..... dayof ........................................................... Signed,sealedanddeliveredby the said: Inthe presenceof:

NOTICE for of It is hopedthat as manyas possibleof those applying Membership the Institutewill undertake a seven-year contributionto the resourcesof the Institute. a covenant,thus making furthersubstantial

ORDER BANKER'S To

..... BANK .................................................................................................................................................................... S ADDRE S ..................................................................................................................................................................... Pleasepayto the BritishBankof the MiddleEast,7 KingWilliamStreet, London No. of Account 654-674-01, for Institute Persian Studies, E.C.4, the creditof the British of the sumof ? : : now and a similarsum on Ist January each year until furthernotice/forthe next six years.* Signed........................................... Date ....................................
* Delete whicheverdoes not apply.

213

S.

. .4o

... -

.-U..

.250

I .Mitesz I

-I
ilometres

-i

Tabriz
RBAI J A N
SL
,.

4,-R
. .
.

U.

sS. o
,,

S.
,

R.

isht. .'ezayeh.

I,
o

??
O4Meshed
.

S::

.orGoran
.Damghani

<*:

Qazvini.L

.*

NDARAN

O VA

mTEHRAN
HImflWaI

KHURASAN

i"

V(

*Kermawshak
"QUiT
.Kashan.

SrA
SIS,

TA

Ni

..PsaryadaeKermaz.
Shi ra 7. .

SrO

AUKE

AD M

AR
J.A.

PRINTED IN GREAT BRITAIN BY HEADLEY BROTHERS LTD 109 KINGSWAY LONDON WC2 AND ASHFORD KENT