IRAN

Journal of the British Institute of Persian Studies

VOLUME V 1967

CONTENTS Page
. Governing Council . Statement of Aims and Activities .
.

.
.

.
.

.
.

.
.

.
.

. .
.

ii iii v I

. . . . . . Director's Report. A Lacquer Mirror-case of 1854, by B. W. Robinson .

.
.

.
.

An Inscription of Darius from Pasargadae, by George G. Cameron . . The Iranian Migration into the Zagros, by T. Cuyler Young, Jr. The Sufi Master as Exemplified in Persian Sufi Literature, by Seyyed
Hossein Nasr . .
.

7
II 35

The Evolution of the Iqta' in Medieval Iran, by Ann K. S. Lambton . Notes on the Baluchi Spoken in Persian Baluchistan, by Brian Spooner
The Sherley Myth, by R. M. Savory Moorey . . . . . .
. . . .

41 51
73 83 99
123

. . .
.

Some Ancient Metal Belts: Their Antecedents and Relatives, by P. R. S.

. . .
.

.
.

Qal'eh-i Yazdigird: A Sasanian Palace Strongholdin PersianKurdistan,

. . . . . . by E. J. Keall Ceramiques Peintes de Tureng Tep6, by J. Deshayes Survey of Excavations in Iran during 1965-66 . . . . . .

133

Published annuallyby

THE BRITISH

INSTITUTE

OF PERSIAN

STUDIES

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

Price: ?2 Ios. od.

NOTES

FOR

CONTRIBUTORS

All contributions must be in typescript, with double spacing and wide margins. Only clear, glossy photographic prints or strong outline drawings should be submitted for consideration as illustrations. The titles of books and periodicals should be underlined, while the titles of articles REFERENCES: in periodicals should be between quotation marks. Both the volume number and the date of publication of a work should be cited in the first reference. Abbreviations should follow the list given on the last page of the Journal. The transliteration into Roman script of names and words in Oriental TRANSLITERATION: languages (other than modern Turkish) should be in accordance with the system employed by learned bodies such as the Royal Asiatic Society. Modern Turkish names and words should be written in the current Turkish orthography. The closing date for the receipt of articles for each issue of Iran is the end of September. Material should be sent either to The Editor, Iran, c/o The British Academy, Burlington Gardens, London, W.I; or to the Director of the British Institute of Persian Studies, P.O. Box 2617, Tehran.

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 C?2 od. entitles the subscriber to receive the Journal. los. Application Forms at back of Journal.

IRAN
Journal
of the British
Institute

of

Persian

Studies

VOLUME

1967

CONTENTS Governing Council Director's Report .

....
. .

Statement of Aims and Activities

.
. . .

Page ii iii v

. A Lacquer Mirror-caseof 1854, by B. W. Robinson . . An Inscription of Darius from Pasargadae, by George G. Cameron .

7 II 35 4'
5r

. The Iranian Migration into the Zagros, by T. Cuyler Young, Jr. The Sufi Master as Exemplified in Persian Sufi Literature, by Seyyed . . Hossein Nasr The Evolution of the Iqtd' in Medieval Iran, by Ann K. S. Lambton . Notes on the Baluchi Spoken in Persian Baluchistan, by Brian Spooner
. . . The Sherley Myth, by R. M. Savory Some Ancient Metal Belts: Their Antecedents and Relatives, by P. R. S. . Moorey ...

..

73 83 99
123

Qal'eh-i Yazdigird: A Sasanian Palace Strongholdin Persian Kurdistan, . . by E. J. Keall .. . . . Ceramiques Peintes de Tureng T1pe, by J. Deshayes Survey of Excavations in Iran during 1965-66 . by annually Pzblished
. . .

133

THE BRITISH

INSTITUTE

OF PERSIAN

STUDIES

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

BRITISH INSTITUTE OF PERSIAN STUDIES GOVERNING COUNCIL President *ProfessorM. E. L. MALLOWAN, C.B.E., M.A., D.Lit., F.B.A., F.S.A. VicePresident ProfessorA. J. ARBERRY, M.A., Litt.D., D.Litt., F.B.A. Members 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. ProfessorW. B. FISHER, B.A., D. de l'Univ., F.R.A.I. BASIL GRAY, Esq., C.B.E., F.B.A. ProfessorA. K. S. LAMBTON, O.B.E., D.Lit., Ph.D. ProfessorSETON 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.B.A., F.S.A. ProfessorR. C. ZAEHNER, M.A., F.B.A. Hon. Treasurer Sir JOHN LE ROUGETEL, K.C.M.G., M.C. Hon. Secretary E. F. GUERITZ, Esq., M.A. JOHN Hon. Editor Mrs. LUKE HERRMANN Hon.Assistant Editor CHRISTOPHER SHACKLE, Esq., B.A. OFFICERS IN IRAN Director DAVID STRONACH, Esq., M.A., F.S.A.
Assistant Director

BRIAN SPOONER, Esq., M.A. c/o The British Academy, Burlington Gardens, LONDON, W.I.
*Denotes Founder Member

P.O. Box 2617,

Tehran, IRAN.

STATEMENT OF AIMS AND ACTIVITIES

i. The Institute has an establishment in Tehran at which British scholars, men of learning versed in the arts, friends of Iran, may reside and meet their Iranian colleagues in order to discuss with them subjects of common interest; the arts, archaeology, history, literature, linguistics, religion, philosophy and cognate subjects.
2.

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

3. The Institute, whilst concerned with Persian culture in the widest sense, is particularly concerned with the development of archaeological techniques, and seeks the co-operation of Iranian scholars and students in applying current methods to the resolution of archaeological and historical problems. 4. Archaeological excavation using modern scientific techniques as ancillary aids is one of the Institute's primary tasks. These activities, which entail a fresh appraisal of previous discoveries, have already yielded new historical, architectural, and archaeological evidence which is adding to our knowledge of the past and of its bearing on the modern world. 5. In pursuit of all the activities mentioned in the preceding paragraphs the Institute is gradually adding to its library, is collecting learned periodicals, and is publishing a journal, Iran, which is expected to appear annually. The Institute aims at editing and translating a series of Persian texts, the first of which, the Humay-Nama, edited by Professor A. J. Arberry, has already appeared. 6. The Institute arranges occasional seminars, lectures and conferences and enlists the help of distinguished scholars for this purpose. It will also aim at arranging small exhibitions with the object of demonstrating the importance of Persian culture and its attraction for the world of scholarship. 7. The Institute endeavours to collaborate with universities and educational institutions in Iran by all the means at its disposal and, when consulted, assists Iranian scholars with technical advice for directing them towards the appropriate channels in British universities.

iii
Ia

DIRECTOR'S REPORT
June Ist 1965 to May 31st 1966 Guests Those staying at the Institute since June 1965 have included the following: Mr. David Blow (Trinity Hall, Cambridge, studying modern Persian); Mr. William Irons (Doctoral student in Anthropology at the University of Michigan); Miss Katherine Kendall (Doctoral student in Social Anthropology, University of Washington); Mr. Nigel Greenwood (New College, Oxford, studying modern Persian); Mr. V. G. Kierman (Department of History, Edinburgh University); Mr. Nazir Ahmad (Principal, Government College, Lahore); Mrs. Olga Ford (study tour in Afghanistan); Mr. Peter W. Pick (Columbia University, studying decorative motifs in Islamic architecture); Miss Elisabeth Beazley and Miss Mary E. Burkett (studying Assassincastles near Isfahan); M. Paul Bernard (Director, D616gation Arch6ologique Frangaise en Afghanistan); M. Gouin Philippe (Secretary, Ddlegation Archdologique Frangaise en Afghanistan); ProfessorE. Bickermann (Columbia University, historical studies); Miss Dorothy Marshall (British School at Jerusalem, visiting archaeological sites in Iran); Miss Sheila Morison (drawing excavated materials from Pasargadae); Dr. V. E. Crawford (Metropolitan Museum of Art, visiting archaeological sites in western Iran); Mr. and Mrs. J. Staley (University of the Punjab, Lahore, passingthrough); Professorand Mrs. M. E. L. Mallowan; Professor J. Benemann (University of Oxford, passing through); Mr. J. Hansman (visiting archaeological sites in Iran); Mr. R. H. Pinder-Wilson (lecturing at the Institute and carrying out a preliminary reconnaissanceat Siraf); Dr. S. M. Hasan (Dacca University, passing through); Mr. Pierre de Miroshchedji (D6legation Arch6ologique Fran?aise en Afghanistan, visiting museums and archaeological sites in Iran); Mr. Louis Levine (Fulbright Fellow, studying the archaeologyof western Iran); ProfessorR. C. Zaehner; Miss Beatrice de Cardi, Mrs. E. C. L. During Caspers, Mr. Peter Broxton and Mr. T. T. Strickland (members of the Bampur expedition); Dr. M. B. Nicol (College of Arts and Sciences, Pahlavi University, Shiraz); Dr. George Dales (the University Museum, Pennsylvania); MissJoanna de Groot (drawing excavated materials from Pasargadae); Mr. F. R. C. Bagley (Lecturer in Persian, University of Durham); Rear-Admiral and Mrs. Paul Furse (Botanical surveys in eastern Iran); Mr. Ian Glover (Department of Anthropology, Australian National University, Canberra); Mr. and Mrs. Cheng Te-kun (Lecturer in Far Eastern Art and Archaeology, University of Cambridge); and Dr. Kenneth A. Luther (Department of Near Eastern Languages and Literatures, University of Michigan). Visitors Visitors to the Institute during this same period have included Professor C. J. Adams; Professor Bahadir Alkim; ProfessorKemal Balkan; Dr. Klaus Ferdinand; ProfessorR. N. Frye; Professor M. Halpern; Professor A. A. Kampman; ProfessorA. K. S. Lambton; Professor Tsugio Mikami; Dr. Walter Oakeshott; ProfessorAndre Parrot; ProfessorWilliam R. Polk; ProfessorArthur Upham Pope; Dr. E. S. G. Robinson; ProfessorG. Tucci; and ProfessorL. Vanden Berghe. at Lectures theInstitute " Mespotamia and Iran-the Passageto India in about 2000 B.C." by ProfessorM. E. L. Mallowan
on October I6th. " The Evolution of the in Medieval Iran " by Professor A. K. S. Lambton, on November 6th. " by in Iran Iqt.' Mr. Ralph Pinder-Wilson, Assistant Keeper, Department of Oriental "Jade Carving Antiquities, the British Museum, on November I6th. " The Demavend and Kharraqin Tomb Towers: New Light on Biiyid and Seljuq Architecture in Iran " by Mr. Stronach, on April 13th. " Archaeology in Mainland China" by Mr. Cheng Te-kun, on May 23rd. (This last lecture represented an informal meeting, organized at short notice, on the occasion of Mr. Cheng Te-kun's recent visit.)
V

Mallowan Visitof Professor During the course of a three-week visit to Iran from October 12th to November 3rd, Professor Mallowan was able to complete a very successfullecture tour. Apart from initial lectures at both the Institute and the British Council in Tehran, he was able to speak on behalf of the Institute and the Council in Tabriz, Meshed and Isfahan. Journeys were also made to a number of archaeological sites in western Iran including Hasanlu, Tepe Nfish-i-Jin and Godin Tepe. Visitof Mr. PinderWilson his lecture at the Institute on November I6th, Mr. Pinder-Wilson was able to travel Following down to the early Islamic site of Siraf on the Persian Gulf. Accompanied by Mr. Morton, he spent several days in the area exploring the ruins and examining certain of the more practical problems that the 1966 expedition to Siraf will have to face. at Excavations Bampur From early March onwards Miss Beatrice de Cardi was able to carry out a programme of two months' excavations at the prehistoric mound of Bampur. In the course of the work Miss de Cardi succeeded in recovering a wide range of material from levels not reached by Sir Aurel Stein in 1932. In return for the Institute's assistance both in Tehran and in the field, Miss de Cardi has agreed to offer a full report on the work for publication in IranVI. A preliminary report appears in the " Survey of Excavations " in this volume. Director Apart from field studies undertaken in connection with the discovery of the Kharraqin tomb towers, Mr. Stronach has been engaged in preparing a final account of the Institute's excavations at Pasargadae. In outside lectures, Mr. Stronach spoke on " Archaeological Sites in Iran " at both the Iran America Society and the British Council. AssistantDirector During his stay in Europe, from August to November, Mr. Spooner visited departmentsof anthropology and Persian studies in both England and Germany. He was able to give a lecture on " The Tribes of Persian Baluchistan ", illustrated by slides and sound recordings,at the Institute's Annual General Meeting on November 9th as well as at the Stidasien Institut, Heidelberg University, on December 7th. During the remainder of the year, except for a short return visit to Baluchistan in March, he has been engaged in preparing for publication the ethnographical material gathered in Baluchistan during the last three years. Fellows Wolfson Mr. A. H. Morton has been working in libraries in Tehran collecting further material for his diaries. He has completed reading almost all the printed travel accounts of the study of QaOjdr nineteenth century, as well as reading Persian sources for the period. He has also gained experience in reading documents of the reign of Nisir al-Din Shah.
Mr. D. H. M. Brooks suffered from a severe attack of cholera in early September. But since that time he has been able to return to the field, joining the Bakhtiari spring migration for a second time. It is now his hope to benefit from one further year with the Bakhtiari before returning to England. Mr. R. L. Tapper has been engaged in compiling as detailed an ethnography of the Shahsavan as possible and has by no means neglected their history. He hopes to return to London University this autumn. Mr. M. B. Loraine has gained all the information he could from periodicals preserved in various libraries in Tehran and has benefited from several personal contacts, in particular with certain members of the family of the poet Bahar. He intends to present his thesis on Bahar later this year.
vi

Editorship theJournal of The Governing Council has learned with regret of the decision of Dr. Laurence Lockhart to resign from the post of Honorary Editor of the Journal. Dr. Lockhart agreed to act in this capacity from the time when the Institute was founded and was responsible for seeing the first four volumes through the press. We are greatly indebted to him for having completed this onerous task, and for his devotion to it over a period of five years during which he assured a high standard of production for the Journal. The Council wishes to record its thanks to Dr. Lockhart for the services which he has rendered and is glad to know that he is prepared to act in an advisory capacity when consulted. We extend a warm welcome to Mrs. Luke Herrmann, who has succeeded him, and is now assistedby Mr. Christopher Shackle.

vii

A LACQUER MIRROR-CASE OF 1854 By B. W. Robinson

Under the Zand and QaLjar dynasties the best Persian miniature painting is usually found not in the illustration of manuscripts, as in the earlier periods of the art, but in the decoration of objects in lacquered papier-mtch6 and enamelled gold. Both these minor arts reached a high state of development under Fath 'All Shah (1797-1834), but the best work produced under Nasr al-Din Shah in the middle years of the nineteenth century was perhaps even finer, and this was largely due to the genius of eminent and prolific lacquer painter of his time; his dated work spans the period 1815-56.' He with the punning invocation to 'All, yd Shdh i ]Najaf, whence he has sometimes been always signed wrongly called Shah Najaf, and it seems likely that this formula was inherited by one or more of his

a single family. The painter Aqa Bdba of Isfahan had two sons. The elder, Najaf, or Najaf 'Ali, became the most

pupils, some of the work on which it appears being inferior in quality to that of the master or of too late a date to have been executed in his lifetime." The American S. G. W. Benjamin, who was his country's first representative at the Persian Court in 1883-5, says that Najaf died about I845 and was a pupil of Sadiq,3 but, like other statements by the same author, these should be accepted with caution, though it appears that a pupil of Najaf did work for him, so that the informationpresumably came from what should have been a reliable source. It seemsmoreprobable, however, that the master
died about I860 or a little earlier.

Najaf had three talented sons, MuthammadKdazim,Muhammad Ja'far, and Ahmad. All three
The

were outstanding miniature painters in lacquer, and Kaizim and Ahmad also worked in enamel. former's work in this medium is among the most exquisite ever produced in Persia.4

But it is with Najaf's younger brother Muhammad Isma'il of Isfahan that we are here concerned.

Albert Museum (No. 763-1876) depicts the court of Minfichihr Khan Mu'tamid al-Dawla (" the Matamet "), the terrible Georgian eunuch governor of Isfahan under Muhammad Shah, whose names are written alongside them.

to have been a court artist almost throughout his career. A qalamddn 1264/1848 in the Victoria and of cruelties are described by Layard.5 Most of the tiny figures appear to be excellent portraits, and their the subject of this article, obviously a " court piece " and specially commissioned, probably by a
Six years later (1270/1854) he painted the mirror-case which is

On the evidence of surviving dated pieces, his working life stretched from 1840 to 1871, and he appears

member of the Royal family. Six years later still, in 1276/i86o, he painted an exquisite spectaclescase for Prince 'Abd al-'All Mirza Ihtishdm al-Dawla, son of Farhad Mirza and grandson of 'Abbas Mirza, the heir-apparent of Fatlh 'All Shah till his untimely death in 1833.6 This bears a full-length portrait of the young prince carrying his falcon and, on the other side, portrait heads of three ladies,
SSee Amir Mas'fid Sipahram, " Najaf Isfahdni QalamdanAqS saz ", in Honar va Mardom, No. 31 (Ordibihisht 1344/1965), p. 25. 2 The question of these punning signatures in the form of invocations is a constantly recurring one in the study of later Persian painting. Thus we encounter ya Sdhib al-Zaman for Muhammad Zaman, yj Sddiq al-wa'd for Sadiq, and yad Muhammadfor Muhammad, as well as Najaf's ya Shdh i Najaf. Most authorities agree that these should be regarded as signatures-the Persians having a weakness for puns-but Mr. A. Ivanov of the Hermitage, Leningrad, for whose opinion I have the greatest respect, stoutly maintains that they are invocations pure and simple, and should not be connected with the artists whose names they appear to play upon. I confess that I find it rather too much of a coincidence that these particular invocations should be constantly found on pieces made during the lifetime, and in the style, of the artists in question if they have no connection with them. In

this view I have the support of Mr. Yahya Zoka, Director of the Ethnographical Museum, Tehran, who is probably the leading authority on the arts of the eighteenth and nineteenth centuries, of Mr. Meredith-Owens of the British Museum, and of the writer of the article referred to in note I. The very plausible suggestion that these punning signatures may also have been used by pupils was kindly made to me by Dr. L. T. Gyuzalian, Keeper of the Oriental Department of the Hermitage, Leningrad, through Mrs. Ada Adamova of the same Museum. 3 S. G. W. Benjamin, Persia and the Persians, London, 1887, PP- 316, 327, 330. 4 There are several examples in the Crown Jewels Museum, Central Bank, Tehran, and one in the author's collection. 5 Sir Henry Layard, Early Adventuresin Persia, Susiana, and Babylonia,London, 1887, 2 Vol. 6 Author's collection.

JOURNAL

OF

PERSIAN

STUDIES

two of them European and the other Persian. All the above pieces are signed with the single name " Isma'il coupled, as the custom was, with some self-deprecating epithet such as kamtarin(" the least or " the most humble "). A further interval of six years brings us to 1282/1865, the date of a superb casket in the Historical Museum, Bern, Switzerland (No. 71/13) which Isma'il has covered with crowded scenes of Muhammad Shah's campaigns against Herat and Ghorian in 1837-8. These scenes, comprising literally hundreds of figures, occupy the top and all four sides of the casket ; on the interior of the lid is Mirza Aqdsi painted Muhammad Shah enthroned with his court, including his notorious vizier JHaijji also appears at his master's side in some of the campaign scenes), and even the interior surfaces (who and the bottom are finely painted with flowers, birds, and arabesques. On this piece Isma'il's signaor ture includes his first name, Muhammad, and the title of Naqqdsh-bdshi, " painter-in-chief ", a court which in the reign of Fath 'Ali Shah ranked on a par with that of Malik al-Shu'ard,or appointment " poet laureate ".' One more interval of six years divides this casket from the artist's last known dated work, a very splendid mirror-case, also in the Historical Museum, Bern (No. 73/I3). The outer surfaces are exquisitely painted with the traditional flower and bird designs (gul u bulbul) arranged in large cartouches and pendants surrounded with gilt rococo scrollwork. On the inner surface of the sliding lid is 'Ali, seated between the two young Imams, IHasanand Husayn, his famous sword Zfi'l-fiqar on his knees, surrounded by angels and some of the Companions, and, above, a miniature representation of the Mi'raj, or Heavenly Ascent of the Prophet. This is all of the finest work and in brilliant with the date 1288/1871. condition. The artist again signs himself Muhammad Isma'il Naqqdsh-bdshi, of comparatively minor importance, have appeared One or two other works of Isma'il in lacquer, from time to time,8 but the above are the ones which mainly concern us, providing as they do a regular series of milestones throughout his career. A short passage in the work of the traveller John Ussher,9 who was at Isfahan in the early I86os, may refer to him. In the course of his description of the Chehel Situn he says: " Behind this large apartment were the other rooms of the palace, in one of which lived a native artist, who, we were told, was the best painter in Ispahan. His chief employment seemed to be painting kalemduns, or cases for holding writing materials, some of which were remarkably handsome, and not exceeded in minuteness of detail or delicacy of finish by productions of the finest miniature painting. He showed us a portrait of the Shah dressed in a short blue frock coat, with a high collar, completely covered over with gold embroidery and precious stones." Muhammad Isma'il may well have been " the best painter in Ispahan " at this time, and Ussher's description fits the delicate minuteness of his work admirably. This is about all we know at present of this outstanding Persian painter of a century ago. It is confusing that at least two other painters named Isma'il seem to have been at work at much the same time. The first, Isma'il Jala'ir, son of Ha-jji Muhammad Zaman Khan Jala'ir of Khurasan, studied painting at the Dar al-Funfin Academy in Tehran, and attracted the attention of Nasr al-Din Shah, himself an amateur draughtsman of some ability. A large group of women and boys round a samovar on a terrace, now in the Victoria and Albert Museum (No. P.56-194I), is an excellent example of his style.'I Other works by him are in the Gulistan Palace collection, Tehran. Isma'il Shirazi Naqqdshbdshi is a more shadowy figure, and may perhaps be identifiable with our Muhammad Isma'il the brother of Najaf. There is a fragment of an oil-painting signed by him and dated 1277/1861 in the

8E.g. a qalamddn dated 1256/1840 (private collection, Tehran), et 7 Ange de Gardane, Journald'un Voyage dansla Turquie-d'Asie la his earliest recorded work ; another dated I280/I863 (G. Wiet, Perse, Paris and Marseilles, 18o09, p. 57. This title was in constant use throughout the nineteenth century, and was Exposition d'Art persan, Cairo, 1935, P. 91) ; and another, whereabouts unknown, undated, but signed Muhammad Isma'fl. borne at first by prominent and talented painters such as Mirzd There is also a water-colour drawing of a girl with a young Babd (c. 1798-1805) and Abii'l-Hasan Ghaffiri (1850-I86I, when he received the more exalted title of Sani' al-Mulk ; boy and a gazelle in the Gulistan Palace collection, Tehran ; it is signed Muhammad Muhammad Isma'il may possibly have taken over from him Isma'il Isfahani,but bears no date. at the dignity of Naqqdsh-bdshi this time). Its later bearers seem 9John Ussher, A Journey from Londonto Persepolis,London, 1865, to have been worthy but obscure, e.g. Nasrullah al-IjIusayni 583. p. (1870), Abfi Turab Ghafftri (1887), and Shirazi (1896). In the late nineteenth century Prince Abfi'l-.Hasan "oAnother of his paintings, " The Sacrifice of Isaac ", is reproMiniaturmalerei, Mas'fid Mirzd Zill al-Sul.tdn had his own Naqqdsh-bdshi, duced by P. W. Schulz, Die persisch-islamische Mirza Alhmad, and possibly other princes had theirs also. Leipzig, 1914, vol. I, taf. F.

Pl. I. Interiorof cover.

Pl. II. Exteriorof cover.

Pl. III. Back of mirror-case.

Pl. IV. Middle sectionof inner face of cover,shownon Pl. I.

LACQUER

MIRROR-CASE

OF

1854

Pars Museum, Shiraz; it was executed to the order of the Prince Ihtishdmal-Dawla. But we have already noted that Muhammad Isma'il painted a lacquer spectacles-case for the same prince in 1276/1860, and it may therefore seem improbable that the latter employed two different painters of the same name in consecutive years. It is also worth noting that the date of the Shiraz painting, 1277/1861, with the is use of the title of Naqqdsh-bdshi, the year in which Ghaffairi was promoted from Naqqdshwas used by bdshi to Sani' al-Mulk (see note 7). The former titleAbfi'l-.Hasan our Muhammad Isma'il, as we have seen, in 1865, so if " Isma'il Shirizi " is a different man, we must postulate a further coincidence of two within a space of four years. Isma'ils using the title of Naqqdsh-bdshi We must also consider, in this connection, a group of four very similar water-colour portraits of Nasr al-Din Shah. The first, dated 1270/1854, is signed kamtarinKhdnazdd Isma'il (British Museum and the second (British Museum Or. 4938, No. 4) is dated two years later, 1272/ No. 1947-2-IO-01) and signed bandai dargdhMuhammad Isfahani. A third, almost identical with this latter and bear1856, ing the same signature, but without date, is in the Hermitage, Leningrad (No. VP. 665). The fourth, in the Gulistan Palace collection, Tehran, is rather closer to Isma'il's than to those by Muhammad Isfahani, and is signed kamtarin Ibrahim Mirza Q ajar, with the date 1275/1859. All these four portraits show the young Shah lounging at ease on a flamboyant mid-Victorian settee, wearing a high black lamb's-wool cap at a rakish angle (in Muhammad Isfahani's portraits this is adorned with a circlet of diamonds and a jewelled aigrette), a short frock coat of Kerman shawl trimmed with fur, and (except in the Hermitage portrait) a Cossack dagger at his waist. There is also, incidentally, a head-andshoulders miniature on ivory of Nasr al-Din Shah in the Victoria and Albert Museum (No. 705-1876) which certainly belongs to the same group, though of inferior quality to the others and neither signed nor dated. This interesting group of portraits might well provide material for a full-length article, but the only aspect of the problem that concerns us here is : could the Isma'il of the earliest of them be our Muhammad Isma'il ?" Judging by the style this seems not impossible. The face, the shawl pattern of the Shah's coat, and the carpet design are exquisitely rendered, but where a broader treatment is required, as for the settee and the monarch's trousers, the artist is not quite so successful. One receives the impression that he would be more at home working on a smaller scale, and one thinks of the minute finish of Muhammad Isma'il's authenticated lacquer paintings. If we look at the small-scale, but full-length, figure of the young prince Ihtishdmal-Dawla on the spectacles-case of 1860 mentioned above, we notice a similar set of the head and slight awkwardness about the legs and feet ; the Shah's feet in the portrait are excessively small and not very well drawn, and in the spectacles-case Isma'il has almost entirely evaded the problem by covering most of the prince's feet with the gilt scrollwork that frames his figure. On the other hand it must be admitted that the signature on the British Museum portrait is unlike any of those on Muhammad Isma'il's lacquer painting, both in the style of writing (thulthin place of the usual nasta'liq) and in the wording, which uses the appellation Khdnazdd, not found elsewhere in Muhammad Isma'il's work, though a number of other nineteenth-century artists incorporate it in their signatures.'2 However there are examples of Persian painters using different scripts at different times, and Isma'il may have thought that a different type of work should be so marked. Nor is the occurrence of Khdnazddbeyond explanation. This portrait and the elaborate and historically important mirror-case (which will be fully described later) both bear the date 1270/1854. It may be that the latter was brought to the Shah's notice, as illustrating an episode of his boyhood, or even offered to him by the artist, and that as a result the latter was commissioned to paint the royal portrait and permitted the use of the title The success of this portrait and the esteem in which it was held are Khdanazdd. sufficiently demonstrated by the fact that at least three other versions of it were executed by different artists during the next five years. The spectacles-case of I86o omits the title of Khdnazdd possibly
1, Mr. Zoka identified him with Isma'il Shirdtzi. 12E.g. Husayn 'All (1841), 'Abd al-Rahim (c. 186o), Yfisuf (1863), and 'Ali b. Mahmfid Khan Malik al-Shu'ara (1883). The literal meaning of the word is " born in the house ", and it was normally used of slaves, but in the present context it seems to carry the meaning of " artist in the royal service ". The extended form Ghuldm-khanazdd was used by 'All and Bdqir, two of Fath 'Ali Shah's best painters in enamel.

JOURNAL

OF

PERSIAN

STUDIES

because it was not painted for the Shah, but for a minor princeling, and by 1865, as we have seen, Muhammad Isma'il had attained the rank of Naqqdsh-bdshi. Muhammad Isma'il's work is minute and meticulous ; he seems to have delighted in painting he was crowded scenes on a very small scale. Superficially his style is strongly westernized-indeed, nicknamed Farangi-sdz or " the Europeanizer "-but, like that of all Persian artists of the QjAjar period who adopted the western conventions of shading and perspective, it remains fundamentally Persian. A western artist of the same period, planning an historical composition, would make numerous sketches from models in costume, and would almost certainly paint his landscape background from nature, on the spot if possible. Holman Hunt painted most of " The Scape-goat " on the shores of the Dead Sea, and is said to have brought away a tray of the surrounding salty deposit, with the help of which he later finished off the foreground in the comparative comfort of Jerusalem. Persian painters would never dream of going to such lengths to achieve realism. To them the European style was a vogue which they followed by putting in shading and working out perspective more or less by rote. But they could never " think European " any more than Edmund Dulac or any of the other western illustrators of Omar Khayyam and the Arabian Nights could achieve a truly oriental style. Kipling's " East is East and West is West, and never the twain shall meet " applies to art no less than to the niceties of behaviour on the North-West Frontier. But, alas, we can only speculate on the result of a confrontation of " two strong men "-say Abfi'l-Hasan Ghaffrri and W. P. Frith-in the sphere of art. Would all the barriers have come down ? And, as the " border thief" became a " man of the Guides ', would Abfi'l-Hasan have exchanged Sani' al-Mulk for P.R.A. ? Personally, I rather hope not. We may now pass to a more detailed examination of the mirror-case of 1270/1854. It appeared amongst a quantity of other Persian lacquer at a recent London auction,'3 where it fell to a dealer at the It would not be surprising if it turned up in due course on the Tehran market, high price of ?205. where it would probably command an even higher figure. Its artistic and historical interest were obvious and, having secured photographs of it before the sale, I felt that it should be recorded before it disappeared from view in the cabinet of some wealthy private collector. It is of the usual rectangular form, measuring approximately io x 61 in. (25'5 x 16 5 cm.), with hinged cover, the hinges being of gold. Nine scenes in all are represented on it, three on each face of the cover, and three on the back, and they are divided by bands of gilt Victorian rococo scrollwork. The historical background of these scenes'4 is, briefly, as follows. In Rajab 1254 (SeptemberOctober 1838) the Tsar Nicholas I, whilst on a visit to Georgia, expressed a desire to meet Muhammad Shah. That monarch being fully occupied at the siege of Herat, a meeting was arranged between the Tsar and the Persian heir-apparent, Prince Nasr al-Din Mirza, then aged seven. Meetings took place at Etchmiadzin and Erivan in Armenia, and whilst it is doubtful whether any important political results were achieved, an atmosphere of cordiality seems to have prevailed, rich presents were exchanged, the young prince was flattered and made much of, and the Tsar Nicholas established himself as a father-figure, a position which stood him in good stead when Ndsr al-Din came to the throne. The Persian officials attending on the Prince were Muhammad Khan Zangana Amir i Ni]fdm, 'Isd Khan Qajar, Mirz~ Taqi Khan Wazir i Niadm (later Amir i Kabir), Mirzd Muhammad Hakim, 'HIajji Muhammad Tabrizi Nifzam al-'Uldma, and MirzA 'Ali Akbar Tabrizi the dragoman. In addition Muhammad Tahir Khan Qazwini (later Wakil al-Dawla) was in charge of the presents, consisting of pearl necklaces, shawls, and horses. They proceeded to Erivan and encamped on the banks of the Aras. In due course the meetings took place ; the Tsar presented the Prince with the
13 Messrs. Phillips, Son, and Neale, March 15th 1966, lot 29.

I am indebted to Mr. Christopher Hawkings of that firm for kindly providing the photographs that illustrate this article. Ethnographical Museum, Tehran, for identifying and supplying details of the main episode commemorated by this mirror-

'4 I am deeply indebted to Mr. Yahya Zoka, Director of the

case, and for valuable help with various other problems that arose during the preparation of this article, and I take this opportunity of expressing my gratitude for his scholarly cooperation. I should also like to express my appreciation of the ungrudging help I have received from my friend and colleague Mr. Meredith-Owens of the British Museum in translating Mr. Zoka's letters.

LACQUER

MIRROR-CASE

OF

I854

Order of the White Eagle set with brilliants, and a belt, Amiri Nizdmreceived a jewelled snuff-box, and a Waziri Nigadm plainer one. In more detail the nine scenes on the mirror-caseare as follows : A. Interior Cover(Plate I) of Top : The walled city of Erivan with cannon firing a salute ; Persians and Russians are strolling and lounging outside the walls. Middle (Plate IV) : Prince Nisr al-Din Mirza sitting on the Tsar's knee in a tdldroverlooking a courtyard. Before them are Amiri Nizdm sitting and Waziri Nizdm standing. An elderly Russian is in attendance on the Tsar. Below in the courtyard are members of the Persian and Russian parties, with the presentation shawls and horses. This incident is thus referredto by the traveller Dr. Moritz and Wagner (Travelsin Persia,Georgia, Koordistan, London, 1856, Vol. III, p. 161): " Abb6 Vidal had been engaged as teacher of the French language, and was to be the tutor of the Crown prince. The latter was at that time a boy nine years of age, and had lately seen the Emperor Nicholas in Armenia, when the colossal Czar, bored with the etiquette of the interview, had dandled the lad on his knee, and had suffered him to play with his Imperial mustachio. The familiarity, personal appearance and pompous entourageof the great Czar, are said to have made a great and lastingimpressionon the prince, and his suite. Accordingly, the Court of Teheran thought it expedient that the heir apparent should acquire the universal polite language of Europe, that he might utter some appropriate sentences,if the Russian ruler should ever think fit to nurse him on his knee again." (Ibid., p. 283:) " In 1837, the young Shah, then Crown Prince, and a child of seven years old, saw the Emperor Nicholas at Erivan, and was caressinglydandled upon the knees of that colossal monarch, and graciously titillated with the tips of the imperial moustachio. Even at that tender age, the successorto the Persian throne imbibed a profound impression of that august personage. The Cossack and body-guards, the thunder of artillery, and the whole military pomp, with which the Czar surrounded himself on that side of the Caucasus, created so vivid an effect upon the imagination of the young son of Iran, that it annihilated for ever any ambitious projectsin a north-westerlydirection, which he later might have formed." Bottom: The Persian camp : a meal being prepared, and Persian officers and soldiers standing about dressedin the uniformsof French type then in use. There is a solitary Russianvisitor. B. Exterior Cover of (Plate II) Top : The Cathedral of Etchmiadzin, with priests and worshippers. The three separate belltowers are exactly like those of the Armenian Cathedral of Julfa (which Isma'il, as an Isfahani, no doubt knew well by sight), and the representation of the cathedral itself, with its mosque-likedome, is wholly imaginary. Middle : Probably intended for a view of Saint Petersburg : the city and a large bridge are in the background, whilst in the foreground a Russian couple are riding in an open carriage, attended by horsemen. Bottom: Feast given to the Russian guests, twenty-six of whom, of both sexes, are represented sitting round a large table laden with kebabs, dishes of chilow, bowls of sherbet, and decanters of wine or spirits. Their horses, held by grooms, are standing round the table.
C. Back of Mirror-Case(Plate III) Top : The Tsar Nicholas, mounted, with his entourage, which includes the same elderly man who was attending him at his meeting with the young Persian prince, and accompanied by a number of dogs. Before him are two young princesses, one already ensconced in an imperial carriage and the other about to enter it. Just behind the Tsar is the artist's signature, 'amal i kamtarinIsma'il, and the date I270. In the background is a lake with a rather curious paddle-steamer approaching a town on the shore.

Middle : The Ottoman Sultan (presumably Mahmfid II, 1808-39, though it looks more like his successor 'Abd al-Majid I) inspecting a detachment of artillery, attended by guards and his veiled Sultana and her ladies, all mounted. An officer is presenting the Sultan with a pair of pistols, and each of the guards has a pair of pistols stuck in his belt.

JOURNAL

OF

PERSIAN

STUDIES

Bottom: Tsar Nicholas, seated and attended by his staff and guards, interviewing three officers who raise their hats to him. Above, a battle is taking place : an armed paddle-steamer, crowded with soldiers and with guns blazing, approaches the shore, which is lined with field-guns firing furiously. Behind these are infantry drawn up in line with shouldered arms. Though the exact subject-matter of some of these scenes is still obscure, it nevertheless seemed worth while to publish this mirror-case, partly as an historical document subject, perhaps, to further research by specialist historians with access to original Persian and Russian sources, but chiefly as an outstanding example of Persian miniature painting of the mid-nineteenth century. Artists of this period have suffered much from neglect due to the admittedly inferior quality of many surviving pieces, The western veneer of shading and which has called down upon them a general condemnation. has also earned them some sneers. But, as has already been emphasized, this is no more perspective than a veneer through which their essentially Persian characteristics are continually breaking, and painters of the ability of Muhammad Isma'il, even if they cannot be ranked with Bihzad or Sultan Muhammad, nevertheless occupy a modest but respectable niche in the long gallery of Persian painting. It may therefore be hoped, now that our own arts of the Victorian period are being viewed with an increasingly indulgent eye, that students of Persian painting may likewise relax their austere disapproval of the Persian Victorians, recognizing their naive charm and gay exuberance, and that future historians of the art will not smugly bow themselves out with the last of the Safavids.

AN INSCRIPTION OF DARIUS FROM PASARGADAE By George G. Cameron

In IranII (1964), pages 38 ff., Mr. David Stronach presented a reasonably accurate drawing and a revised reading of a few Old Persian (OP) and Elamite signs on a solid block of stone found, several years ago, by the late Djavad Zakataly in or near the ruined tower (" Zendan-i-Sulaiman ") at Pasargadae-a tower which, because of its close resemblance to the so-called Ka'bah-i-Zardusht at Naqsh-i-Rustam has already received (and merits) close attention.' Having personally examined the inscription, presently housed at Persepolis, I can testify to the accuracy of Mr. Stronach's reading of the OP signs: on what has hitherto been considered to be the first line of a two-line inscription, the signs are unquestionably ku-u-sa;just below these signs (in the supposed " second " line), I saw still visible a tiny bit (the " tail ") of a word divider, followed by ha-yaand then a second word divider. Unhappily, as Mr. Stronach relates, previous misreadings of these few signs had given rise to a belief that the inscription had been composed either by Cambyses, the son of Cyrus,' or by a later Persian sovereign who was reporting on some activity performed by Cyrus.3Unhappily also, no one seems to have considered it possible that the fragment in question was perhaps only a portion of the lasttwo lines of what was originally a much longer OP text. I believe it can be demonstratedthat the entire original text is (or was) none other than a duplicate of what has been called the " Foundation Record " of Darius the Great, of which two gold and two silver inscribed plates were uncovered in two different foundation deposits under the Apadana at Persepolis (DPh) and of which one gold and one silver plate emerged from Hamadan (DH).4 That this famous text should also be found in the ruins of Pasargadae-upon which, as Mr. Stronach has fully demonstrated, Darius laboured intensively5--should occasion no surprisewhatsoever. The OP text of DPh bore ten lines of inscription, that of DH, eight lines. If we assume that the OP Pasargadae text of the same inscription comprised five lines only, then the above reading, ku-u-sa becomes Kush (" Ethiopia ", as translated by Kent), and :ha-ya: below it becomes " who " in the full text, the complete original version of which (see P1. Ia) may be translated, in its five lines, as follows: I. Darius, the great King, King of Kings, King of countries,son of Hystaspes, 2. An Achaemenian. Saith Darius the King: This is the kingdom which I 3. hold, from the Scythianswho are beyond Sogdiana thence unto 4. Kush;from Sind, thence unto Sardis-which Ahuramazdabestowed on me5. (He) whois the greatestof the gods. Me may Ahuramazdaprotect, and my (royal) house. For the preparation of P1. Ia I have simply used a photograph of one of the ten-line silver plates found at Persepolis (DPh) and recast it into a five-line text in such a way that ku-u-saappears almost directly (as P1. Ia shows) above :ha-ya:. No real effort has been attempted to " justify " (in printer's jargon) the endings of the lines, but anyone who reads OP knows that there is always a considerable
difference in spacing individual signs in the various lines when the stonecutter or metalsmith wrote the same inscription more than once. The fact, consequently, that lines 4 and 5 of P1. Ia appear slightly longer than lines 1-3 simply means that the stonecutter, in this instance, has crowded the signs in
For the most recent comments, see Stronach, Iran III (1965), II ff. and Bernard Goldman, JNES XXIV (1965), 305 f.f ' So Djavad Zakataly, in " L'authentique Tombeau de Cyrus " (" Rapport ta sa Majest6 Imperiale Mohammad Reza Chah Pehlevi Chahinchah de l'Iran ", Teheran, I954), pp. 43 ff.; so also Ali-Sami, Pasargadae. The Oldest Imperial Capital of Iran (" Publication No. 4 of the Learned Society of Pars ", Shiraz, March 1956), p. Ioo. See also Carl Nylander, OrientaliaSuecanaXI, 1962 (Uppsala, 1963), PP. 121 ff. 3 So W. Hinz, ZDMG Io9 (n.f. 34; 1959), 125 ff. 4 For abbreviations and bibliography, see R. G. Kent, Old Persian (2nd, rev. ed. 1953), p. 109 for DPh and p. I I I for DH; see also Schmidt, Persepolis (1953), 37I 5 See esp. Iran IlI (1965), PP. 17 if.

JOURNAL

OF

PERSIAN

STUDIES

lines 4 and 5 (or, on the contrary, more widely spaced the signs in lines 1-3) than the silversmith's copy here shows.6 Below the two fragmentary lines of the assumed five-line OP text there is a line left blank (as is customary when an OP text is accompanied by its Elamite and Akkadian versions). Mr. Stronach's drawing is a definite improvement on that of all former renditions but still shows, after a vertical wedge, a sign which is, in Elamite, la. My own examination of the sign at Persepolis led me to the firm conclusion that the sign was in reality, ak (followed by part of a possible ka)-a reading endorsed independently by W. Hinz,7 and one involving .only the alteration of a single (partially damaged) vertical wedge to a horizontal. Now as a matter of fact, the Elamite la sign does not appear at all in either DPh or DH; further, there is no place, in the Elamite version of that document, where an ak could occur at or near the beginning of the text, which means that the opening phrases of the Elamite version cannot have been directly beneath the opening phrases of the OP text. The signs which my eyes saw on the rock did, however, lend themselves readily to a restoration involving [ik-ki]-[mar7l ak-F ka]-[be]-a phrase which does appear at about the middle of line four of the seven-line Elamite versions of both DPh and DH (that is, at a point approximately half-way through the Elamite version). Conceivably, this might signify that the Elamite version consisted of a single line which began far to the left of the OP text above it and ended, on the right, approximately with the ending of the lines of the OP. Since, however, at least at Persepolis, the OP texts are always given more prominence than the other two, I believe it can be proved that (except for the final sign) our presumed reading [ik-ki]-[marv ' ak-Kka]-[be] appeared at the end of a two-line Elamite text which, of course, had the beginnings of its lines to the left of the beginning lines of the OP above it.8 In view of the Old Persian predilection for balance, one may also almost automatically assume that a similar (and probably two-line) Akkadian version once appeared below and for the most part to the right of the OP, as shown in P1. Ib, in which the shaded section shows the block which may be all that is at present known to be preserved of the three original texts (see, however, below). This novel positioning of the three versions may have arisen because of some architectural or structural feature: the " gap " or opening beneath the OP text may once have sheltered a relief or, perhaps better, have been a door, a window, or a window niche. An exact computation of the width of this opening is of course unobtainable, but, by assuming the positioning of the three texts as shown in Pl. Ib, and knowing the exact measurements of the lengths of the signs ku-u-s'aand :ha-ya:9 one may make a reasonably accurate calculation that it can scarcely have been more than 2 -35 m. or less than 2 m.'0 This is "25 at least in the palace of Darius at Persepolis much too wide for windows or window niches-which, averaged no more than i95 m., and perhaps even for that structure's doorways, the largest of which is I -465 m. wide." Exact information regarding the architectural features of the Pasargadae buildings has, unfortunately, hitherto been lacking, and is only just now being supplied by the work of the British Institute of Persian Studies. Dr. Ali-Sami does tell us that the doorways of what he called the " Palace of Audience of Cyrus the Great " (Herzfeld's " Palace S ") were 2 o8 m.,"' and since lintels are normally consider6 There are no traces of signs in a small smooth area above broken left edge . . . of the OP second line to the lower tip of the division stroke, the distance is 18-7 cm ". ku-u-?a,which might seem to argue that this word appeared in thefirst line, not the fourth, of the inscription. Note, however, The variability depends on the spacing of the individual signs Io even in P1. Ia which is based on the text of the silver plate, in lines 1-3 as compared with those in lines 4-5; see above. that the beginning signs of what is here assumed to be line 3 " Schmidt, PersepolisI, Figs. 94a and 93d. The doors to what are themselves rather widely spaced. A similar spacing might m. Schmidt called the " Council Hall ", however, were well have left no traces on the preserved portion of the rock. 2"7 wide; ibid., Fig. 56a. 7 See ZDMG I09, p. 126. However, in the palace of Xerxes at Persepolis, the inscription 8 A recasting, into two lines, of the seven-line Elamite version of which is above a doorway only I -74 m. wide appears on that the same silver tablet DPh demonstrated conclusively that part of the frame which, at its widest part, gives space for a line I ended, as shown above, with via-ak-ka-be ik-ki-mar text which is 2-44 m. long; cf. Schmidt, PersepolisI, Fig. 99a -be beginning line 2, which line ended also, as vak-ka (with and P1. 77a; similarly, another doorway with opening only partially shown in P1. Ia-and as it should-with ku-ud-da -82 m. long: I *22 m. can accommodate a text which is huL.HIlg-mi) op. cit., Fig. 99b. 9 I am grateful for a part of this information to Mr. Stronach, who speedily and graciously wrote to me that " from the 12 Op. cit., p. 50.

at lower left: the signs on a block found in or " PI. Ia. Reconstructed 5-line OP text of a " Darius Foundation Document (DPh). Outlined the Pasargadae Zendan, with concluding signs of a 2-line Elamite version.

FIVE LINES OF OP TEXT (PLUS ONE BLANKLINE) TWOLINES OF ELAMITEVERSION (ASSIUMED LINES OF AKKAQLAN TBWO

DOOR,
NICHE, WINDOW, OR RELIEF

Pl. Ib. Probable arrangement of the trilingual textsfrom which the inscribed block (shaded area) derived.

AN

INSCRIPTION

OF

DARIUS

FROM

PASARGADAE

ably wider than the apertures beneath them our inscription may once have graced, say, a doorway in that structure. It is far more likely, however, that our inscription appeared above a door or other architectural feature in what Herzfeld denominated the " Residential Palace " or " Palace P ". In his Altpersische Inschriften (1938), Herzfeld published photographs and drawings of a number of fragmentary but inscribed stone blocks which, he says, were " im Feld iiber den K6nigsfiguren in den Thuiren des Saales; nur wenige Bruchstticke im Schutt gefunden "3--all of which came from this structure. Perhaps much too hastily he presumed, first, that they all represented the fragments of a single trilingual inscription, and second, that the text began with the name of Cyrus. For the latter conclusion there is not a shred of evidence. His block labelled " Kyr. Pas. Pal. P, Ap. frgmt. I " which he read as [ku]-u-r[u-u-sa] (and which would thus bear witness to Cyrus) could be rendered just as accurately [da-ha-ya]-u-v[a] which occurs in innumerable texts and which appears in line I of our P1. Ia.'4 His " frgmt. 5 ", bearing parts of two lines (below which there is a blank line) is with almost equal certainty to be read as parts of the concluding signs of lines 4 and 5, also as shown on P1. Ia.'5 In my opinion, then, some of the inscribed fragments published by Herzfeld may actually have belonged to the same text from which the block which began our investigation derived, and there may have been many copies of it above a number of apertures, reliefs, or what not. It would be folly, however, to insist that all of his fragments belonged to one and the same trilingual inscription; as at Persepolis, so at Pasargadae various trilingual inscriptions probably appeared on most structures. It will of course be recalled that all of the windows and niches of the portico and main hall of Darius' palace at Persepolis were adorned with trilingual inscriptions; there, however, the OP text always appears horizontally on the lintels, with the Elamite version forming a vertical band on the left-hand frame and the Akkadian version, balancing it, in the same position but on the right-hand frame.'6 At Pasargadae, then, the architect and stonecutters were merely experimenting with exclusively horizontal text lines; a further calculation of the overall length, from the left edge of the Elamite version to the right end of the assumed Akkadian version would yield about 9 5 m. Thus, although our fragment may have been found in or near the ruined tower, the very dimensions of the original text preclude it from ever having been written on that structure, the walls of which measure in length only 7-25 m. (on the north-west and south-east) and 7-22 m. (on the north-east and south-west).'7 Further, Mr. Stronach has himself determined that the block is much too thick ever to have been a part of the original door of that building.'8 There is also a real question regarding the exact findspot of the block. In a letter to me dated at Persepolis on June 8th 1953 Mr. Zakataly wrote that he had found " dans le tombeau de Cyrus " (italics mine) an inscription hitherto unknown, and sent me a handcopy which showed only the OP signs u-sa on one line and, below it, ha-ya. Subsequently (letter dated November 25th 1953), in response to my expressed astonishment that the inscription had been found " in the tomb of Cyrus " he clarified his meaning: what he considered to be the real tomb of Cyrus was " la tour 't demi-ruinde de Pasargade. Les inscriptions que je vous avais envoydes, furent decouvertes par moi, cette annie, dans ce tombeau(italics mine) et non pas dans celui qu'on attribuait "t Cyrus ." Still later, on p. 43 of his 1954 monograph (see above, n. 2), Mr. Zakataly described his discovery in the following terms: "Je me suis contente de l'examen des pierres qui gisaient en desordre au pied de la tour ruinie (italics
13

Op. cit., pp. 2 f., with Abb. 3 and Tafel 11i; cf. also his AMI I (1929-30), 14f. 14The tiny top of the wedges of Herzfeld's " second " line could easily be, e.g., a part of the logogram for " king " in our line 2. 15 In his Tafel II, this is the block in the centre of the top photo. Directly behind it is a block (pictured again in both the centre and the bottom photos!) with Elamite signs; it is virtually impossible to distinguish the two blocks from each other, but for Herzfeld's OP ka-ra we should restore either to 4 [fa]-ra-[a-ba-ra] 5 [vi]-Vi 7-Oa-ma

or to

4 [fa-ra-a-b] a (!?)-ra [:]

with a blank line below, just as is visible at the lower right corner of P1. Ia
16

5 [vi]-F-i--Oa-[ma]

Schmidt, op. cit., I, p. 223. Much the same is true of the doors, windows, and niches in the palace of Xerxes: ibid., Pls. 176 ff.

'7 See Stronach, Iran III, 11.

18Personal communication dated June 3rd, 1966.

10

JOURNAL

OF

PERSIAN

STUDIES

mine) sur l'une de ces pierres dont les bords 6taient casses j'ai d6couvert les traces bien conservees d'inscriptions cundiformes." Dr. Ali-Sami, in his own report on Pasargadae (see above, n. 2), in one place (pp. 99 f.) gives a slightly differing version of the find: " A skilful constructor of models, Mr. Zakataly, who was for four years making a model of Persepolis under the writer's direction, and had visited Pasargadae a few times, saw a stone among other fallen stones on the southernside of the Tower there (italics mine), which was discovered in 1952 [sic!], and has three lines on it with a few cuneiform letters of an Old Persian inscription." Later in the same monograph (p. 14o) the discovery is reported as follows: " A piece of broken stone 70 centimeters long and 40 centimeters thick was found near (italics mine) the ruined tower known as the Tomb of Cambyses, on which parts of three lines of cuneiform, the first two in Persian, and the third in Elamite, can be seen." Thus the evidence respecting the exact location of the block remains somewhat obscure: it may have been found within the ruin, or simply near it. In either case, however, it could have been (and most certainly was) carried there from some other ruined structure of the Pasargadae building complex. In any event, if my assumptions are correct, the apparently identical texts of DPh, DH and this DM (" Darius, Murghab ") inscription bear out a comment of Ernst Herzfeld to the effect that " all examples (scil. of these documents) were prepared from one and the same original text of a central chancellery ".'9

'9 Altpersische p. Inschriften, 19.

11

THE IRANIAN MIGRATION INTO THE ZAGROS, By T. Cuyler Young, Jr.

Introduction The Iranians entered written history when they collided with the Assyrian Empire as it and they expanded into the Zagros mountains, an encounter first recorded in the Royal Annals of Shalmaneser III." The following two centuries witnessed an increasing struggle between the Iranians, the indigenous Zagros peoples, and the Assyrians for control of the western Iranian plateau. In time, with the founding of the Median Kingdom and its successor state, the Achaemenid Empire, the Iranians triumphed. In the process, cultural elements brought by the newcomers blended with those of both the older inhabitants of the land and the Assyrians, and the resulting fusion set a pattern for western Iran which has persisted in part through many vicissitudes to the present day. Here, therefore, is an area and a time deserving the closest attention by anyone interested in Iranian history. Unfortunately, relevant written sources are few, usually secondary, often prejudiced and frustratingly fragmentary. Recent rapid developments in our understanding of the archaeology of western Iran in the millennium from c. 1500 to 500 B.C. suggest, however, that now might prove a profitable moment to initiate a discussion of at least the mechanics of the Iranian migration into western Iran by attempting to combine archaeological and written data into a coherent historical picture. The discussion which follows falls into three parts: (I) a review of the geography and chronology of the Iranians in western Iran prior to Achaemenid times as we know them from the written sources, (2) an outline of the cultural dynamics of the period as seen in the archaeological record, and (3) a test of the extent to which the patterns drawn from the written sources and the archaeological data agree or disagree by superimposing the one upon the other. The Written Sources What little evidence can be gleaned from the Iranian and Classical sources on the earliest Iranians in the west relevant to our problem has been summarized recently by R. Frye.3 In essence what these data indicate are a certain connection between Iranians on the east and the west of the plateau prior to Achaemenid times. Most important are the close linguistic affinities between Old Persian, a southwest Iranian dialect, and Sogdian, an eastern Iranian language. The connection is sufficiently strong to suggest that the two groups must have once been closely associated, assumedly in the east.4 Such a conclusion is perhaps partially supported by the statement of al-Biruni that the Khwarezmians of the east were once closely linked with the Persians. Suggestive but not conclusive evidence for some link between the Parthians in the east and Parsua/Parsa or Persians in the west, and the identification of Persian elements amongst the Massagetai, a tribal confederation of eastern Iran, also point to an eastwest connection early in Iranian history. In general, of course, there is no question about the broader bonds which tie together eastern and western Iranians, both of whom have their linguistic roots in a period before the differentiation of the Indian and Iranian branches of the Indo-Iranian language group.5
' Given the close and fruitful informal co-operation amongst those who work in Iranian archaeology, I cannot thank all who have contributed to my thinking about the problems discussed here. Profs. Mark J. Dresden and Robert H. Dyson, Jr., who directed the dissertation from which this article is largely drawn and who have nurtured my interest in Iran, have my special thanks. David Stronach and Louis Levine travelled with me over much of the area involved and will no doubt find in what follows a number of plagiarized thoughts which had their birth in animated discussions on tea house terraces. Prof. A. K. Grayson has kindly read the article in its initial draft and has discussed at length with me several of the Assyriological issues involved. z The Persians have the honour of being first as they are mentioned in the report on the campaign of year sixteen of Shalmaneser, 844 B.c. The Medes were encountered in year twenty-four, 836 B.c. See D. D. Luckenbill, AncientRecords of Assyriaand Babylonia,Chicago, 1926, Vol. I, paras. 581 and 637. Hereafter, LAR. 3 R. N. Frye, The Heritageof Persia, Cleveland, 1963, pp. 36-49. 4 W. B. Henning, " Mitteliranisch ", in B. Spuler, et al., Handbuchder Orientalistik,Bd. 4, Ab. I, Leiden, 1958, pp. o05-8. 5 For what is still an excellent and clear statement on basic Indo-Iranian connections, see V. G. Childe, The Aryans: A Studyof Indo-European Origins,New York, 1926, pp. 30-41.-

12

JOURNAL

OF

PERSIAN

STUDIES

One could do little with these very broad inferences were the Classical and Iranian sources all the written evidence available. Fortunately, the Medes, Persians and Scythians, the principal Iranian groups in the west, are comparatively much better documented in the Assyriansources which concern affairs in the Zagros following or during the migration of these groups into western Iran. But before turning to these western Iranians in detail, brief consideration should be given to the fundamental geographic characteristicsof western Iran relevant to any discussionof the movement of a new people into that area. The dominant natural feature of the mountainous Zagros landscape is the great range of Mount Alvand, anchored in the south-east on the high Bakhtiari massif, stretching northwestwards behind Hamadan, and eventually becoming entangled in the knot of mountains where the modern borders of Iran, Iraq and Turkey meet west of Lake Urmia. South-west or west of the Alvand barrier the land falls gradually away to the Mesopotamian alluvium in a series of parallel ridges.6 Here the human settlement pattern follows a series of long, narrow, cramped intermontane valleys, all trending northwest-southeast, the most important of which is the Mahi Dasht or Kermanshah Plain. Northeast or east of the Alvand alignment the geography and settlement pattern are rather different. In the north is the Lake Urmia Basin, the greatest concentration of settled population in the Zagros, which is only partially separated by a comparatively low range of mountains or hills from the long, broad Hamadan Plain which sweeps from Bijar in the north to Arik in the south.7Numerous routes permit easy movement to and from the Hamadan area and the open country of the Tehran-Qazvin-Zanjdin regions to the east: moving from south to north there are the Arak-Qom, the Arik-Siveh, the SavehHamadan, the Thkestan-Hamadan, the Abhar-Kharragan-Hamadan, and the Zanjdn-Bij-r roads. The best of these, that from Sdveh to Hamadan, involves no mountain crossing at all. On the other hand, it is not easy to penetrate west of the Alvand barrier. The best and only direct route acrossthe mountains and down to the Mesopotamian Plain is the ancient Khorisdn Road running Hamadan-Kermanshah-Shdhaibd-Sar-iPul-i Zuhib-Khinaqin. The next best route west, and a poor second choice, is the road which crossesthe Alvand barrierjust east of Sanandaj, winds its way through the mountains west of that town, and crossesinto Iraqi Kurdistan to the north of Lake Marivan. South of Hamadan a route through Malayer gives relatively easy access to the Borijerd Valley, but west of Borfijerd a mountain range the equal of Alvand blocks the road to Khorramabad. In the far northwest, access to the west and the north Mesopotamian Plain out of the Lake Urmia Basin can be gained either through the Rowandfiz Gorge (not an easy pass) or by following the more circuitous valley of the Little Zab River south and then turning west at Sar Dasht. The accessibility of the Hamadan Plain from the east, the ruggedness of the Alvand barrier, the
ease with which one can move northwest-southeast along the inner face of the barrier from Lake Urmia to Arak, the difficulty in passing beyond it to the west, and the brokenness of the country west of the barrier are all important geographical factors which have shaped the history and culture of western Iran.8 The Medes. The Medes are one of the best documented of the Zagros mountain peoples mentioned in the Neo-Assyrian sources.9 They are often described as "the distant Medes", the "mighty (greater)
6 Parts of some of these ridges are, of course, even higher than Mount Alvand itself: e.g. those west of Boruijerd, Bisitun Mountain, the Kfih-i Shdhfi southwest of Sanandaj and the crest of the range west of the upper valley of the Little Zab River. Topographically these peaks are actually part of the Alvand range, but for our purposes the dividing line between the eastern and western Zagros is the eastern (or more correctly the northeastern) face of the Alvand range. 7 In Medieval Islamic times the principal caravan route north from Hamadan passed through Bij-r and Takab and then followed the Zarineh Rfid into the Lake Urmia Basin. What is now the main road via Sanandaj and Saqqiz carried only local traffic prior to the introduction of motor vehicles, the need to approach Iran's western borders for defensive purposes and the discovery of oil in the southwest (the road is most heavily used today by oil tank trucks feeding the population of Azerbaijan from sources in Kermanshah and further south). 8 It is significant, for example, that practically nowhere west of the Alvand barrier is there a village today whose mother tongue is either Persian or Turkish, the dominant languages east of that mountain alignment. The reverse, of course, is not true. Pockets of Kurdish and Kurdish-related dialect speakers are found east of Alvand. The important point is that Persian, and particularly Turkish, languages which entered the Zagros from the east, have failed to penetrate significantly west of the Alvand barrier. 9 The Medes, unlike many of the other peoples of western Iran familiar to the Assyrians in the ninth century, are mentioned at least once by every Assyrian king from Shalmaneser III to Ashurbanipal.

THE

IRANIAN

MIGRATION

INTO

THE

ZAGROS

13

the Medes "whose country lies far off", the "distant Medes on the borders of Medes of the and Mount Bikni",,east", the Medes who live near the salt desert.'o Now although efforts to see in these terms different groups or tribes of a Median confederation are certainly misplaced," nevertheless, such expressions in reference to the Medes had definite meaning for the Assyrians. When the Assyrians pushed east beyond a certain point in the Zagros they seem to have encountered only Medes for as far east as they could go, and this depth of Median occupation was so impressive as to warrant special mention."2 A geographic reconstruction of these campaigns suggests that in the central Zagros all of the non-Iranian groups encountered in the sources can be located west of the Alvand barrier, while the bulk of the Medes were in the region stretching east of Alvand all the way to the Qazvin-Tehran area. Medes, however, did appear in the central Zagros west of the Alvand barrier in limited numbers, where they came into contact with indigenous Zagros peoples, the most important of which were the Kingdom of Ellipi and the province or city of Harhar. Ellipi was the principal non-Iranian kingdom in the central western Zagros. We hear of it first from Assurnasirpal II, who mentions "captives from the city of Elipi",'3 and it remained important down to the reign of Esarhaddon, at which time it disappears from the Assyrian records. ' Talta of Ellipi was a faithful vassal of Sargon II, and his power was often shored up with military assistance from his overlord.'5 On his death, Ispabara, one of his sons, was placed on the throne and, at first, continued his father's Assyrian alliance.'6 He revolted during the reign of Sennacherib, however, and Assyrian influence in Ellipi had to be established once again by force of arms.7 That the maintenance of such influence was vital if the Assyrians were to check Elamite ambitions in the central Zagros is clear: we are specifically told that Ellipi, and the kingdom of Rashu, were on the Elamite frontier;'8 Elam had interfered in Ellipi at the time of Talta's death on behalf of Ispabara's brother, Nibe, the anti-Assyrian candidate for the throne;'9 and the last reference to Ellipi in the Assyrian Royal Annals is the report of Sennacherib on the crushing of an alliance between Elam, Ellipi and Babylon at the battle of Hadule on his eighth campaign.2o Elamite power appears to have extended as far north in the western Zagros as at least the upper end of the Saimarreh River Valley, where the city and passes of Bit-Bunakki were located, and Rashu was probably the name for the lower mountain regions between Bit-Bunakki and Der.2' All of these considerations point to a
LAR I, 795 and 812; LAR II 54, 82, 519, 540 and 566. 672 B.c., and in the Vassal Treaties. See J. A. Knudtzon, AssyrischeGebetean den Sonnengott, Leipzig, 1893, No. 75 and D. J. Wiseman, " The Vassal-Treaties of Esarhaddon ", Iraq, 20, 1958, 82. LAR II, 23 and 58. 5s 16LAR II, 65. Ispabara is a good Iranian name, *aspabara-. We may assume that Ispabara's mother, only one of the wives of Talta, a non-Iranian, was an Iranian. The mother of Nibe, Ispabara's brother and a non-Iranian, was probably also a non-Iranian. Here is a glimpse of one way in which Iranian and indigenous Zagros cultures became blended in this period. '7 LAR II, 237. The account indicates that Ellipi was a sizable kingdom, containing several cities important enough to name and a subdistrict, Bit-Barru. 1aLAR II, 54, 82 and 96-98. I9 LAR II, 65. 20LAR II, 252 and 352. See also, A. K. Grayson, " The Walters Art Gallery Sennacherib Inscription ", AFO, 20, 1963, 85-96. 21 On Bit-Bunnaki see LAR II, 248, 351 and 355. The passes and the city were usually occupied by the Elamites and appear to be the first truly Elamite territory encountered by the Assyrians on their campaigns against Elam in the mountains: See HEI, p. I65. Rashu was probably located in the region south of modern Ilim where several mounds testify to an important ancient occupation. See LAR II, 805 and L. Waterman, Royal Correspondence the Assyrian Empire, of Ann Arbor, 1930, No. Ioo7. Hereafter, RCAE.

0oE.g.

"1 stressed by G. Cameron, History of Early Iran, Chicago, As

1936, p. 149, n. i9. Hereafter, HEI. "zHow often, how far, and in what numbers the Assyrians penetrated western Iran beyond Alvand is an open question. Only Sargon II and Esarhaddon actually claim to have campaigned as far east as Mount Bikni or Mount Demdvand (LAR II, 79 and 519). Other references to Demdivand speak in vague terms of bringing Medes that far east under Assyrian rule or of receiving tribute (trade ?) from such distant Medes. See LAR I, 784, 787, 795, and 812; II, 54, 82, 96-99, 102, 540 and 567. Tiglathpileser III mentions the salt desert which is on the edge of the land of Rua (LAR I, 795), but could be referring to that arm of the Dasht-i Kavir in the neighbourhood of Saveh. Esarhaddon's reference to the salt desert could mean the Kavir proper, since he mentions DemBvand in the same passage (LAR II, 519). We may be fairly certain that campaigns as far east as Demrivand were little more than raids to gather cavalry mounts: cf. the worried questions put to the god Shamash in the Omen Texts (E. G. Texte aus de Sargonidenzeit, Klauber, Politisch-religi6se Leipzig, 1913, Nos. 21 and 22) with the more boastful account of that raid in the Annals (LAR II, 519). See also, HEI, pp. I72-73. '3 LAR I, 542'4 References to Ellipi from the reign of Esarhaddon are found in

the Omen Texts, one of which has been attributed to Ashurbanipal as crown prince and hence is dated shortly after

14

JOURNAL

OF

PERSIAN

STUDIES

location for Ellipi in the Mahi Dasht, with a possible extension into the Shahabid Valley. Numerous large mounds with evidence of occupation from earliest times into the first millennium testify to the importance of this area in the past.2 Though Ellipi maintained a separate political existence under Assyrian control and some of its preIranian ethnic identity into the early seventh century, it was closely linked with the Medes and under considerable Median pressure from at least as early as the mid-eighth century and probably even earlier as well. The two peoples are first mentioned together by Tiglathpileser III;23 Sargon II speaks of Ba'it-ili, a province of Media, on the borders of Ellipi;24 and Sennacherib received the tribute of the Medes on his return from smashing the revolt of Ispabara.25 Even more revealing is the Royal Correspondence. Medes are either directly or indirectly involved in the affairs of Ellipi as messengers working for the Assyrians, as collectors of horses in the same neighbourhood, or possibly even as allies of Ispabara at the time of his revolt.26 When we last hear of Ellipi, Esarhaddon is concerned about the growing threat of an alliance between that kingdom and the Medes."7 The city of Harhar was closely associated with both Ellipi and the Medes. It is first mentioned by Shalmaneser III in the context of a campaign against the Medes.28 In the sixth year of Sargon II the city revolted and sought vassalage under Ellipi, but Sargon quickly strangled the rebellion, took and rebuilt the town, changed its name to Kar-Sharrukin and placed Ellipi under the control of the Assyrian governor of the citadel.29 In the following year the king tells us that he strengthened the fortifications of the city "against the Medes".3o Sennacherib, at the time of Ispabara's revolt, placed a district and several cities of Ellipi under the governor of Harhar.3' Finally, in letters from the governors, we hear that the Medes who live in the immediate neighbourhood of the citadel are sometimes troublesome, and sometimes (can one detect here a tone of thankfulness?) are quiet and tending to their business.3" Expeditions are sent out from the town against the Medes to collect horses; and Talta's suspicious reception of tribesmen (Medes?) is reported in one letter, while in another we find the old man under house arrest for not having paid his taxes on time.33 The location of Kar-Sharrukin depends, to a considerable extent, on the location of the main routes used by the Assyrians to move troops into central western Iran. Two natural roads suggest themselves. The first of these cuts southeast from the Assyrian fortress of Dur Assur in the land of Zamua (southern Iraqi Kurdistan),34 crosses the Diyala near the Iran-Iraq border, and eventually enters the Mahi Dasht at its northwestern end near Ravansar. The second enters Iran from the relatively open country east of Sulaimaniya just north of Lake Marivan and continues east to Sanandaj (there are at least two relatively easy roads from the lake basin to the modern town). From Sanandaj one can either continue east over the Alvand barrier into the northern Hamadan Plain, or turn south and follow an easy river road into the Mahi Dasht. The first route, from Dur Assur to Ravinsar, is topographically much the more difficult and any stiff local resistance would make it untenable. The road crosses the Diyala at least once, and in 1965 the river was unfordable along this trail even in early August. The necessity of permanently bridging the river would, of course, give still further advantage to local resistance. Nevertheless, this route does give fairly direct access to the Mahi Dasht. The second route, from Marivan to Sanandaj, though more circuitous, offers the invader a number of choices, is less rugged, involves no major river crossings and hence puts the heavier strain on the defenders. Other factors being equal, it seems more likely that the Assyrians usually used the Marivan-Sanandaj road
The same is true for the extension of this area toward Karind. The western border of Ellipi might be the edge of the plateau above and east of Sar-i Pul-i Zuhib. The latter area has few mounds with first millennium occupations and seems to have not been densely occupied until Sasanian times. 23LAR I, 795.
22

27HEI, p. 180 and Knudtzon, op. cit., No. 72. 28LAR I, 581. 29LAR II, II.
3o LAR

II, I5.

31LAR II, 237.

32RCAE, Nos. 128, 556, 713 and 145433RCAE, Nos. 126 and 129, and ND 2655, Saggs, op. cit., loc. cit. 34E. A. Speiser, " Southern Kurdistan in the Annals of Ashurnasirpal and Today ", AASOR, 8, 1926-27, 16 and map.

24LAR II, 23. 25LAR II, 238.

26 RCAE, Nos. 174 and 1454, and ND 2655 in H. W. F. Saggs, " The Nimrud Letters, 1952-Part IV", Iraq, 20, 1958, 191.

THE

IRANIAN

MIGRATION

INTO

THE

ZAGROS

15

into central western Iran.35 Kar-Sharrukin should be located at or near the Iranian end of one of these routes with good lines of communication back to Assyria but close enough to Ellipi and the Medes to function as a seat of local government. It should also be located on a river, since we have a relief of the citadel showing a quay or retaining wall protecting it from a stream and a letter giving some details of the construction of that structure.36 Considerable search has failed to reveal any large citadel site in the area between Ravansar and the Iran-Iraq border on the route to ancient Dur Assur. On the other hand, there are several large citadel mounds in the northwest end of the Mahi Dasht which were occupied in roughly the eighth to seventh centuries B.C. and which are strategically located in terms of controlling both Ellipi and the Dur Assur-Ravinsar route back to Assyria. None of these sites, however, is on a river,37 and there is some evidence that Kar-Sharrukin was close to, but not actually in, Ellipi and therefore should not be located in the Mahi Dasht.38 A mound closer to Sanandaj controlling the northernmost reaches of Ellipi as well as the traffic north and south through the Sanandaj area would be as likely a location for ancient Kar-Sharrukin.39 Unfortunately, no site has yet been found in that region which yields pottery in the correct time range. Further survey may yet find Kar-Sharrukin; for the moment, the best theoretical reconstruction points more in the direction of Sanandaj than toward the Mahi Dasht.4o Thus the Medes in the central western Zagros. A somewhat similar situation obtained to the northwest in the southern Lake Urmia Basin, where the Iranians, given the absence of any real geographic barrier between the Hamadan Plain and the Lake Urmia area, were putting even more pressure on the indigenous peoples, the most important of whom were the Manneans. The exact limits of Mannea remain somewhat uncertain. The centre of the kingdom appears to have been in the rich lands along the southeastern shore of Lake Urmia, probably extending north up the east side of the lake into the area of Mardgheh and south to Bfikhan and perhaps to Saqqiz.4i There is good evidence for consider35The Assyrians could, of course, have entered the central western Zagros along the Khorasin road from Khdnaqin to Shhabaid via Sar-i Pul-i Zuhib. Their usual route to Der probably did take them into the Khanaqin region by a line of march northeast of the Jebel Hamrin (ShamshiAdad V seems to have followed such a route on his fourth campaign against Babylonia in order to take the enemy in the rear and to separate the Elamites and Babylonians. LAR I, 723.) Assyrian campaigns against Elam could also have moved southeastwards still further inside the mountains down the long open valley north of Sar-i Pul-i Zuhr b and south to modern Ilfm. A sharp left turn at either Khdinaqin or would have put the Assyrians on the high Sar-i Pul-i Zuheb road to Ellipi. Such routes might be used if campaigns in Ellipi were part of a larger effort against the Elamites, but the more direct routes into central western Iran suggested here would be the sensible lines of march for a campaign directed specifically against the mountain peoples in the Kermanshah-Sanandaj regions and it is here, not along the Khorftsn road, that we should look for Harhar. 36 See E. Herzfeld, Archaeological History of Iran, London, 1935, p. 15 and RCAE, No. 126. 37 Rivers, of course, move. Air photographs of the Mahi Dasht, were they obtainable, would permit the tracing of older river beds, one of which might have passed near one of the large citadel sites in the area. 38 Knudtzon, op. cit., No. 72, " The state of Sissirti, a fortress of Harhar, which is situated on the border of the land of Ellipi... 39 There is some evidence for an extension of Ellipian territory considerably north of the Mahi Dasht proper. Shalmaneser III received tribute from Paru of Ellipi in the pass of the land of Tukliash after completing a campaign against Namri (LAR I, 637). Namri is almost certainly the Marivdn Lake Basin and surrounding mountains. It is tempting to see in the passes of Tukliash one of the routes out of the Marivdn region in the direction of Sanandaj. That Kar-Sharrukin could be as far north as the neighbourhood of Sanandaj is perhaps supported by the fact that it was close enough to Mannea (south of Lake Urmia) for the Assyrian governor to expect reinforcements from troops stationed in Mannea: see RCAE, No. 556.
40

Cameron, following Streck, locates Kar-Sharrukin southeast of Lake Marivin: see HEI, p. 15o and " Armenien, Kurdistan und Westpersien nach den Keilinschriften ", Zeit.f. Assyriologie 15, 1900, 348-50. A few miles south of the lake is a large citadel site which, unfortunately, neither yields pottery of certain first millennium date nor is on a river. In any case, a fortress in the Marivtn Lake Basin itself would not give the Assyrians the control of the central western Zagros that they sought by fortifying Harhar. The site must be further east. Herzfeld has pointed out the similarity in name of Harhar and the modern village of KhSlkhiel located at the east end of the Kermanshah Plain: see op. cit., p. 14. No large site with first millennium pottery and no river is located near this village. Salnmatibod Tepe, strategically located south of Bijsir at the junction of the Rfidkhineh-i Uzfin Darreh and the Rlidkhineh-i Shfir, and on the river, yields the right kind of pottery in quantity. It is, however, probably too far north and east of Sanandaj, and thus too far distant from Ellipi and Assyria, to be more than an outside candidate for Harhar. G. A. Melikishvili, " Some Problems of the History of the Mannean Kingdom ", VestnikDrevnei Istorii, I, 1949, 57-72 (in Russian). M. van Loon kindly loaned the author his MS translation of this article. The valley of the Tdtdfi Chai is thick with mounds south of the Midndaib Plain to Bfikhin and it is this area which probably formed the Mannean heartland. No Iron Age pottery has yet been found on any of the three mounds in the immediate neighbourhood of Saqqiz, but on geographical grounds it is reasonable to include areas perhaps even south of Saqqiz in Mannea proper.

4I For a general discussion of the kingdom of Mannea, see

16

JOURNAL

OF

PERSIAN

STUDIES

able contact between Manneans and Iranians, many of whom were certainly Medes. In the reign of Shalmaneser III an Iranian, Artasari, ruled a petty kingdom on the borders of Mannea; Sargon II knew an Iranian, Baghdatti, as the ruler of Uishdish also near Mannea; and in the reign of Esarhaddon Manneans and Medes were allied against Assyria.42 We learn in greater detail about at least the geography of Mannean-Median relations, however, from the account of the eighth campaign of Sargon II, when the Assyrians defeated the rebel king Mettati of Zakirtu, a sometime Mannean province. Mettati was an Iranian and Zakirtu was probably a small Median kingdom on the eastern border of Mannea.43 A reconstruction of the relevant section of the eighth campaign allows us to locate Zakirtu with reasonable accuracy.44 After Sargon received the gifts and good wishes of the people of Gizilbunda, described as "a distant place among the far off mountains, shut in alongside the land of the Manneans and the land of the Medes as a bolt", the Assyrians departed the fortress of Zirdiakka and marched thirty beru between the land of the Manneans and the land of the powerful Medes to the fortress of Panzish on the border of Zakirtu, which was then provisioned with supplies.45 Wright, probably correctly, has suggested that this march took Sargon up the Tdtafi Chai as a flanking operation against Zakirtu,46 which was thus located, at least in part, along the upper and central reaches of the Zarineh
in the neighbourhood of Mardgheh: so E. M. Wright, " The Eighth Campaign of Sargon II of Assyria (714 B.C.) ", JNES, 2, 1943, 178. On Baghdatti, see F. W. Konig, "Alteste Geschichte der Meder und Perser ", Der Alte Orient,33 (3-4); 1934, 58. Onomastic evidence on Iranians in the Mannean area must be treated with caution. A local chieftain, Bagbarana, has a name that at first looks Iranian, but which is certainly a compound from Urartian Bagbartu, the wife of the god Haldi: see LAR II, 173. On the coalition at the time of Esarhaddon, see HEI, pp. 177-179. 43Mettati is first mentioned by Sargon II on his sixth campaign: LAR II, io. Zakirtu has been equated with O. P. Asagarta-with some hesitation: see Frye, op. cit., p. 67. A similar equation has been suggested for the city or country Zakruti: see LAR I, 784 and 8i1 ; II, 147; and HEI, p. 149. The possibility that Zakirtu and Zakruti are the same place, which is tempting, seems ruled out by Sargon's reference to both places on the eighth campaign : see Saggs, op. cit., p. 192. 44For geographic reconstructions of Sargon's eighth campaign this discussion relies primarily on Wright, op. cit., supplemented by the author's own experience with the area. Wright, in contrast to F. Thureau-Dangin, Une Relationde la Huitilme de Campagne Sargon,Paris, 1912 and H. A. Rigg, Jr., " Sargon's Eighth Military Campaign ", JAOS, 62, 1942, 130-38, possesses unrivalled knowledge of the terrain and hence of the physical possibilities and necessities facing Sargon. 45LAR II, 149-50. Estimates of the length of the beru in the Neo-Assyrian period range from 6,ooo metres (ThureauDangin, op. cit., p. iv, n. 2) to 10,692 metres (E. Weidner, " Assyrische Itinerare ", AFO, 21, 1966, 43, n. 23). The CAD has it a measure of distance greater then Io,ooo metres. All such estimates, particularly those in the 0,000ooo metre range, which are ultimately based on the known length of a cubit (F. Thureau-Dangin, Textes Mathematiques Babyloniens, Leiden, I938, p. xiii), are much too precise. As implied in its original meaning of" double hour " the beruwas the distance a man could walk in two hours. Thus it was a measure of time, not of linear distance, and its actual length will vary according to circumstances. When the Assyrians marched thirty beru from the fortress of Zirdiakka to Panzish no one went along with them to measure the distance covered with a metre tape, and so Sargon can only mean that it took him sixty hours to make the march. Today, in travelling the same route, a question about the distance from X to Y gets an answer in time not distance, for the important point is how long it will take you to go from X to Y given local conditions.

42For Artasari, see LAR I, 587. Uishdish was probably located

It is possible that on regularly travelled roads in Mesopotamia actual linear distances between points were known, but this could never have been the case in western Iran. Common sense, a knowledge of the terrain and some appreciation for the known possibilities in terms of ancient sites make possible rough estimates for the length of the beruin each case, but one cannot use a certain length for the beruas a means of establishing locations. If we accept 10,692 metres as the length of the beru, Sargon's march from Zirdiakka would take him a total of 320 kilometres. If he went south up the Tatfiu Chai this would land him somewhere deep into the Hamadan Plain. If he went east, as Thureau-Dangin has suggested, he probably could have bathed in the Caspian Sea. Tactically, of course, this makes no sense at all. A march of thirty " double hours ", even if a purely cavalry operation was undertaken, fits the tactical and strategic requirements. 46Wright, op. cit., pp. 18o-8i. Panzish is probably correctly located by Wright in the Takab area. A more correct translation of the phrase " over against the lands of Zikirtu and Andia " in LAR II, 150, used to describe Panzish would be " up stream from the lands . . . " (oral communication from A. K. Grayson), thus supporting the suggestion that Sargon marched up the Tatdfi Chai to get in behind and above the centre of Zakirtu on the Zarineh Rfid. This reconstruction of the flank attack on Zakirtu opens the question of the location of Gizilbunda as " a district which is located in a distant place among the far-off mountains, and is shut in alongside the land of the Manneans and the land of the Medes, as (with a) bolt... " How could Gizilbunda be too far distant and still touch Mannean and Median territory? Sargon's historical memory, of course, failed him when he had written that " none of the kings who went before me had ever seen their dwelling-place, heard their name or received their tribute ". Shamshi-Adad V passed through Gizilbunda en routefrom Mesai (Meshta, the Midndfidb region) to the lands of the Medes (LAR I, 719), and Adad-Nirari III mentions Gizilbunda in a list, bracketing it between the Medes and the Manneans (LAR I, 739). It is, therefore, tempting to suggest that Gizilbunda was not so far away after all, and may have been the mountain district between the Ttdfia Chai and the Zarineh Rfid or the mountains separating the Qizil Uzun and Zarineh Rfid drainage basins. If the latter, more likely, suggestion is correct, then Panzish was under considerable threat from these hill people, and Sargon, by his friendly reception of the ambassadors from Gizilbunda, was seeking a safe conduct for his army on the first leg of the flanking operation against Zakirtu.

THE

IRANIAN

MIGRATION

INTO

THE

ZAGROS

17

Rild, the natural route into the Lake Urmia area from the Hamadan Plain.47 The fortification of Panzish prevented Median reinforcements from reaching Zakirtu from the south, "the land of the distant Medes in the east", while Sargon, countermarching toward Lake Urmia down the Zarineh Riid, stormed the cities and the capital of Zakirtu from the rear, thus clearing his own rear prior to a push up the east side of the lake for a final and decisive reckoning in the Maragheh region with Ursa of Urartu and his ally Mettati. Of particular significance is the fact that Zakirtu along the Zarineh Rfid and Uishdish near Maragheh are the most northern points in western Iran where the Assyrians encountered Iranians, even though Sargon's account of this campaign gives numerous details of the peoples and kingdoms north of Lake Urmia. The Medes are never mentioned in the Urartian records, and do not seem to have been associated with the Urartians in the Assyrian sources except in this one instance. The Persians. The land Parsua occurs numerous times in the Assyrian sources.48 References from the

reign of Shalmaneser III indicate that it should be located somewhere in the northwesternZagros, for it is associated with the Manneans, Allabria and Missi (Urartian Meshta). Later, on the third campaign of Shamshi-Adad V, it is again found in the neighbourhood of the Manneans, and in a list of
Adad-Nirari III it is bracketed by Mannea and Allabria.49 The evidence from the reign of Tiglathpileser III is somewhat more enigmatic. On the one hand, Mannea is never mentioned in the

inscriptions of this king, and it is possible to interpret all of his campaigns as having been directed against the central Zagros. On the other hand, Parsua is twice mentioned in conjunction with BitKapsi, certainly located in the northwest on the basis of the eighth campaign of Sargon II.5o There can be little doubt that Sargon II found Parsua in the northwest. On the eighth campaign, after passing through Allabria and Karalla, the left and right banks of the upper reaches of the Little Zab respectively, the Assyrian army marched on to the city of Laruete, still within Allabria, and from there descended to the land of Parsua.5' From Parsua Sargon moved to Missi or Meshta near
Miandfiib. Here the king received the people of Gizilbunda and placed them "under the hand of the Assyrian officials, the prefects of Parsua".52 In doing so Sargon was following a precedent he himself

established on his sixth campaign when he placed several captured cities of the Manneans under All the administrationof "the province of Parsua".53 of this evidence points to a location for Parsua in the basin of the Gadar River in what is today the Soldfiz and Ushnfi Valleys.54 Such a location not
47It is impossible to establish from the written sources how far east into the mountains from the Zarineh Rfid Zakirtu may have extended. Geographically it is reasonable to assume that Medes were in the region of Midineh, which they could have reached either by moving down the Qizil Uzun from the Hamadan Plain or by spreading up the long, narrow Zanjain Valley. Thus the Iranians in the Zarineh Rfid-Mardgheh area (Zakirtu and Uishdish) would have links with fellow Iranians both south to the Hamadan Plain and east to Midneh, and the territory of Zakirtu might actually have extended east to Midineh. Sargon was concerned only with those parts of Zakirtu which touched on Mannea and which affected his planned campaign up the east side of Lake Urmia. 48 Several forms of the name Parsua occur in the Assyrian records: Par-su-a, Par-su-as, Par-su(sa)-mal. The distinction is orthographic only, and the simpler Parsua is used throughout this discussion.
49 LAR I, 581, 637, 718 and 739. On the location of Allabria,

though a location at the uppermost end of the Little Zab Valley would do equally well. 52LAR II, 148-49.

53 LAR II, io and 56. There is a hint in the record that Parsua
may have been under direct Assyrian control as early as the reign of Tiglathpileser III. See LAR I, 784. The implication of Sargon's action on his sixth campaign is that there was in being an Assyrian administration in Parsua. A letter from the Assyrian governor in Parsua concerning a delay in forwarding some horses to Assyria caused by the neighbouring Manneans testifies to the implementation of Sargon's directive. See RCAE, No. 16554The suggestion is by no means original. See Wright, op. cit., pp. 178-79; S. Smith, " Parsuash and Sulduz ", Professor Poure Davoud Memorial, II, Bombay, 1951, pp. 60-67; and J. V. Kinnier-Wilson, " The Kurba'il Statue of Shalmaneser III ", Iraq, 24, 1962,

see below and n. 51. Urartian Meshta has long been identified as Tash Tepe near Miindfi~b: see F. W. Konig, Handbuch der chaldischen Graz, 1955-57, pp. 6 and 59-60. Inschriften, 50LAR I, 784 and 81 I.

5sLAR II,

145-47. See also, Wright, op. cit., pp. I77-78. The location of Karalla is perhaps confirmed by the letter ND 2677 which described the route from Karalla to Calah as passing through Kakzu: see Saggs, op. cit., pp. 195 and 211. Wright puts Laruete near modern M0ahtbid. This is reasonable,

that the site of Qal'eh Pasveh, a mound in the northeastern corner of the Little Zab Valley, preserves in its modern name ancient Parsua. See V. Minorsky, " Mongol Place-names in Mukri Kurdistan ", BSOAS, 19, 1957, 78-80. Qal'eh Pasveh unfortunately is today topped by the house of the local Khan and sits in the centre of a village. Thus it yields no pottery that is not recent. The size and shape of the site, however, suggests a firstmillennium B.C.date. See A. Stein, OldRoutes Western in Iran, London, 1940, pp. 408-412. Wright's identification of the site with the fortress of Sinihinu of the eighth campaign is perhaps 23preferable: op. cit., p. 178, 11n.

I I-12.

Minorsky

argues

L KSDE URAA

PA RS UA

vv*~

4d4f

c Ir
ASSYRIAU

s0

10 so 801m,0

40

N1 Ilk
A

WM'LAND OVER 2000 METRES ALIGNMENT ALVAN4D

;:-:

BITBUNAKKI

DER0
z;F

Iran. of Fig. I. The AssyrianGeography Western

THE

IRANIAN

MIGRATION

INTO

THE

ZAGROS

19

only fits the progress of the campaign, but also would be a reasonable base from which Assyrian officials could administrate Mannean territories and Gizilbunda and at the same time maintain good lines of communication west to Assyria.55 After the beginning of the seventh century Parsua was definitely located in the southwestern Zagros. On his eighth campaign Sennacherib found Parsua allied with Anzan, Ellipi and Elam, all areas in or south of the Mahi Dasht.56 Records of Ashurbanipal associate Parsua with Huban-nugash of Elam and with the city of Hidalu, probably located on the middle Kirfin River.57 Though these references to a southwestern Parsua are too vague to permit its location with any precision, they leave no doubt that from the reign of Sennacherib to Ashurbanipal there was a land called Parsua somewhere in the neighbourhood of Elam. Eventually, of course, we hear of yet a third Parsua, Parsa or the homeland of the Achaemenid Persians in the province of Fars. From these facts only one of two conclusions is possible: either Parsua moved from the northwest to the southwest late in the reign of Sargon II or early in the reign of Sennacherib, or there were three kingdoms or peoples in the Zagros with the name Parsua/Parsa. Much has been made of the first of these possibilities as evidence for the migration of the Persians from the Lake Urmia area to Fars.ss This is an attractive theory, but seems to strain the geographical and sociological probabilities, if not possibilities. To move a tribe of people from the Soldfiz-UshnCit area southeastwards through the twisted and tangled intermontane valleys of the western Zagros, each occupied either by firmly entrenched indigenous non-Iranian or recently arrived Iranian farmers or pastoralists, to a new land somewhere on or within the borders of Elamite territory which would have to be wrested from yet another established power, all in a matter of two decades at most, would be a task beyond even the prowess of one of the mythical Persian tribal leaders celebrated by Ferdowsi. On the evidence available a good case can be made for the theory that there were three Parsuas or groups of Persians in western Iran at roughly the same time; one in the northwest, one on the borders of Elam, and one in Fars.59 The evidence for a Persian migration from the northwest to the southeast may be only an artifact of the sources. Assyrian power in the northwest declined considerably during the reigns of Sennacherib and Esarhaddon, and was but partially restored by Ashurbanipal.6o Thus the disappearance of Parsua in the northwest might reflect only poor Assyrian intelligence on that area during the seventh century. The appearance of a Parsua in the southwest in the records of Sennacherib and his successors might be explained in part on similar grounds. Since the growing threat in the seventh century appeared to be Elam, the southwestern Zagros assumed strategic priority in the thinking of the Assyrians. More and more attention was given to Elam, and as a result of the accompanying increase of Assyrian intelligence on the southwest a kingdom of Parsua on the borders of Elam and in Fars made its first appearance in the sources. A reconstruction which allows for three groups of Persians in the Zagros is perhaps the simpler explanation of the limited evidence at hand.
Either through the Rowanduz Gorge or, more probably, down the Little Zab and into Zamua by the routes over Mount Kullar west of Sar Dasht. One of the many mysteries of the Assyrian campaigns in western Iran is that no clear evidence is forthcoming for the Assyrian use of the Rowanduz Gorge, one of the natural routes linking northern Mesopotamia with northwestern Iran. Perhaps Urartian power, based on Musasir near the western end of the Kel-i Shin pass immediately north of Rowanduz, made the Rowanduz Gorge untenable. Partial confirmation of this suggested location for Parsua is found in the Urartian sources, where Parsua is mentioned twice, once in conjunction with Meshta, the Miindfidb region: see Konig, op. cit., p. 40 (Inscription no. 7, II and IV) and p. 90 (Inscription no. 8o, 5, V). 56LAR II, 252, 352, 356. 57RCAE, Nos. 1309 and 1311 On the location of Hidalu see i. HEI, p. 165. 58 E.g. HEI, pp. 179-81; R. Ghirshman, Iran, Harmondsworth, 1954, pp. 118-19 ; and G. Husing, " Vorgeschichte und
55

Wandering der Parsawa ", Mlitt. d. Anth. Gesell in Wien, 60, 1930, 246-68. 59Originally suggested by Dyakonov, who would, on etymological grounds, include Parthia as yet another variant on Parsa. The idea has most recently been discussed and amplified by Frye, op. cit., pp. 45-46, where a summary of Dyakonov's suggestion is given. 6o The campaigns of Sennacherib and Esarhaddon against the Manneans were none too successful: see LAR II, 364 and 383 for rather general references to Mannea under Sennacherib, and ibid., 517, 533 and 786 for reports on the activities of Esarhaddon in the northwest. The true situation during the reign of Esarhaddon is revealed in the Omen Texts: see Klauber, op. cit., No. 8 and Knudtzon, op. cit., No. 6 for evidence of the growing co-operation between Kashtariti the Mede and the Manneans. Ashurbanipal admits to this striking decline in Assyrian power in the northwest when he refers to the cities which the Manneans had seized " in the time of the kings, my fathers, and appropriated for themselves ". LAR II, 852.

20

JOURNAL

OF

PERSIAN

STUDIES

The Scythians.Though perhaps somewhat tangential to a discussion of the main migration of the Iranians into the Zagros, the Scythians are an important group which plays a role, albeit rather poorly defined, in first millennium western Iran. The earliest evidence for the appearance of the Scythians in the area would seem to be a relief of Assurnasirpal II from the second quarter of the ninth century.6' Indeed some Scythians may have been present in the Zagros that early, particularly if they were in fact responsible for driving the Cimmerians onto the Urartians. But, since they were not found by Sargon II on his march to the north of Lake Urmia in 714 B.C., it is perhaps more cautious to consider their role in events of even the late eighth century as somewhat limited. Certainly they could not have been present in any numbers prior to their first appearance in the texts from the reign of Esarhaddon.6 2 Herodotus tells us in some detail of the Scythian movement into the Zagros from across the Caucasus, probably by the passes toward the Caspian end of that range, and describes their activities in terms typical of marauding warrior bands.63 This description is reinforced by the evidence of the Omen Texts from which we learn of potential coalitions of raiders consisting of Scythians, Medians and even Manneans.64 Unlike the Medes and the Persians, these warrior nomads cannot be associated with any specific areas in the Zagros, but, as is to be expected if they entered the region from the north, most of their activities seem to be concentrated in the central and northwestern mountains. The impression gained from the written sources is that they played a comparatively secondary role in the "Iranization" of the western plateau on a pattern rather different from that of other Iranian groups. Certainly they were relatively late in arriving on the scene.65 From this cumulative evidence there emerges a general picture of the mechanics of the Iranian migration into western Iran. The first phase of that migration brought the Iranians into the Hamadan Plain in numbers sufficient to overrun that area completely. No significant natural barrier to the east hindered this movement. Such was not the case west of Hamadan, for, at least temporarily, the Alvand range acted as a brake on further mass movement toward Mesopotamia. This geographic barrier was further strengthened by strong indigenous peoples to the west backed by the power of Assyria. Chronologically, we may assume that this initial phase of migration was completed by the ninth century, for Assyrian references of that date to lands east of Alvand already mention only Medes. Indeed, it is probably necessary to conclude that these events began some little time prior to 900 B.C.66 The second phase of the migration, chronologically overlapping in part with the first phase, involved a three-pronged spread outward from the Hamadan Plain. In the central western Zagros limited numbers of Medes filtered through and over the Alvand barrier along the two principal routes to the west. Moving down the Khorasdn Road, they came in contact with Ellipi and settled in considerable numbers in and along the borders of that kingdom, most probably in the Asadhbdd, Kangdvar and Nehavand Valleys, and quite possibly in the upland valley of the Gaveh Rfid as well (one of these valleys might be the Median province of Ba'it-ili). Medes also crossed over the pass east of Sanandaj and filtered west into the region along the headwaters of the Diyala in the neighbourhood of KarSharrukin. The initial penetration west of Alvand probably occurred in the centuries prior to 900 B.C. at the same time that the Iranians were flooding into the Hamadan Plain. Median penetration west of Alvand gained momentum between the reigns of Tiglathpileser III and Sargon II, and grew towards and other indigenous Zagros groups a climax under Sennacherib and Esarhaddon. Since Ellipi, HIJarhar begin to disappear from the Assyrian Annals in the reign of Sennacherib and from all Assyrian sources
6, T. Sulimirski, " Scythian Antiquities in Western Asia ",

ArtibusAsiae, 17, 1954, 290-93.

62LAR II, 517. That the Scythians were a military force to be

reckoned with is evidenced by the apparent willingness of Esarhaddon to marry one of his daughters to a Scythian: see A. T. Olmstead, Historyof Assyria,New York, 1923, P. 36o and Wiseman, op. cit., p. 10 o. 63Herodotus, I, lo4-6. 64HEI, pp. 170o ff. and H. W. F. Saggs, The Greatness that was Babylon,New York, 1962, pp. 126 ff.

65The Scythian interregnum in Median history and the problem of the Scythian association with the fall of Nineveh are important issues deserving a detailed discussion elsewhere.
66 It is perhaps significant that the original movement of Iranians

into western Iran was an event of such antiquity that none of the Median informants of Herodotus were aware of the fact that the Medes had not always lived in Media. Similarly, no tradition of the Persian migration into western Iran is preserved in any of the Old Persian or Classical sources. One can only regret that the Iranian National Epic deals so exclusively with eastern Iran.

THE

IRANIAN

MIGRATION

INTO

THE

ZAGROS

21

in the reign of Esarhaddon, we may assume that the absorption of these groups by Iranians was fairly complete by the end of the second quarter of the seventh century. In the northwest, Iranians spread along the northeast face of the Alvand barrier until they spilled over into the valley of the Zarineh Riid and the Lake Urmia Basin. Here they may have met other Iranians coming into the same region from the direction of Mianeh. These events must have begun before the ninth century, as indicated by the probable location of Parsua as an established kingdom in the Soldfiz-Ushnfi area by the time of Shalmaneser III and by the number of Iranian rulers and petty principalities associated with Mannea from fairly early on. The written evidence is quite clear on the point that the Medes and the Persians were not found north of Lake Urmia and that Urartian power was sufficient to keep them below a line drawn roughly from Marigheh to Rezd'iyeh. This kept the burden of dealing with the newcomers on the Manneans, who, in spite of heavy pressure, managed to maintain something of their identity down to the fall of the Assyrian Empire.67 Yet the Omen Texts document how, in the seventh century, they were drawn increasingly under Iranian influence. The Manneans of the late seventh century were certainly neither culturally nor politically the Manneans of even the days of Sargon II. During the reigns of Esarhaddon and Ashurbanipal the situation was further complicated by the appearance of the Scythians in some numbers in the central and northwestern Zagros. This nomadic warrior band of Iranians from the north may have first appeared in the area as early as the ninth century, but by the seventh century had considerably reinforced the Iranian pressure on the Manneans to say nothing of the Assyrians themselves. Though there is less evidence for a spread of Iranians into the southwestern Zagros, such a movement, as a parallel to that which took the Iranians into the northwest, is hinted at in the references to a Parsua on the borders of Elam and in the known presence of Persians in Fars, and may be reasonably assumed. It is probable that the migration of Iranians into Elamite territory occurred somewhat later than the initial phases of the Iranian penetration into the central and northwestern Zagros. For one thing, the country is more rugged: we observed that access into the Borfijerd area west of the Alvand barrier is fairly simple but that between Borfijerd and Khorramdbhd rises one of the more formidable natural mountain barriers in the western Zagros. More important, Elam was and remained, down until the time of Ashurbanipal, the strongest of all the indigenous Zagros kingdoms and so reinforced geographic difficulties with political strength. But, by the beginning of the seventh century Iranians had moved into the central western Zagros in sufficient numbers to gain access to the natural routes which lead southeast from the Mahi Dasht and the Shdh5bid Valley into what might be called Elamite Luristan (Pusht-i Kfih). Once Ellipi, supported by Assyria, no longer blocked these routes, the door to the southwestern Zagros was open. When Ashurbanipal destroyed forever the power of Elam between 642 and 639 B.c., he removed the remaining political power barrier to further penetration of the southwest, and Iranian groups which had probably moved into the highland country north and northeast of Elam proper by the first decades of the seventh century overran Khuzistan itself. Thus the Assyrian sources reveal the principal Iranian migration into western Iran as an east to west movement probably beginning as early as the late second millennium, and yield no evidence for any Median or Persian (as opposed to Scythian) movement into the Zagros from the north. This main migration struck the great natural barrier of the Alvand alignment, penetrated fairly soon due west of that barrier along the best routes, and then followed up that initial penetration at a more gradual pace while spreading northwest and probably southwest as well along the barrier in a continuing effort to push west. While these events were still in progress, Iranian Scythians from the north reinforced the main Iranian migration by warrior raids into the Zagros. By the early 7th century the thrust of the migration had carried the Iranians in dominant numbers into the central and northwestern Zagros, and the push into the southwest, which was to be completed later in that century, had begun. The Assyrian sources, therefore, tend to agree with and bring into sharper focus most of the broad inferences of early east-west Iranian connections drawn from the Iranian and Classical written evidence.

67Some Manneans were allied with the Assyrians at the time of the fall of Nineveh: see LAR II,

158.

22

JOURNAL

OF

PERSIAN

STUDIES

The Cultural Record68 Dynamics theArchaeological of grey pottery. Five type shapes define the period, all of which are found at the three principal excavated sites (Hasanlu V, Sialk V and Giyan 14-I3) with Iron I occupations, and numerous other ceramic parallels link these and several secondarysites. Equally typical of the culture of the period is the practice of burying the dead in cemeteries, usually in simple inhumations (Hasanlu, Giyan, Sialk, KhorvinChandar, Yanik Tepe).69 No evidence for any widespread constructionof fortificationsin the period has yet been established, and all sites typically yield little evidence for the use of iron. Perhaps the most remarkable trait of the period is its broad cultural unity: parallels between the several sites, particularly in pottery, are numerous and striking, and the impressiongained is one of a minimum of local variation on a widespread theme. Iron I pottery has been found over much but not all of western Iran (Fig. 2).70 Sialk representsits furthest extension east, and Marlik documents its spread at least in modified form part of the way down Valley, as does Dinkha Tepe in the Ushnu area.7' Geoy Tepe and Yanik Tepe indicate its extension part way up both sides of Lake Urmia, and Bfikhan on the Tdtf-a Chai may have an Iron I occupation, though grey ware is absent on most of the sites between Midndfiib and Bfikhdn.The data from eastern Azerbaijan is more enigmatic, but some evidence, principally in the form of pattern burnishing, indicates an Iron I influence if not occupation in the Ardabil Plain and down the corridorlinking that Asaddbid, Kangivar, Nehdvand and Borfijerd Valleys in central western Iran, as well as at several sites in the Hamadan Plain and along the northeasternface of the Alvand Alignment. Of equal significance are those areas which have not yielded evidence of an Iron I occupation: western Luristan, central western Iran west of the Kangivar Valley (extensive survey in the Mahi Dasht has failed to
area with the central reaches of the Qizil Uzun or Safid Rfid. Grey wares of the period are found in the to the Caspian shore. In the northwest, Hasanlu and other sites yield Iron I occupations in the Soldfiz The Iron Age I Period (c. 1300/1250 to ooo1000 B.c.): This period is characterized by the use of plain

produce a single sherd of Iron I pottery), and Kurdistan west of the Alvand barrier. Though survey of comparable intensity has yet to be conducted north of Lake Urmia, no Iron I grey ware has yet been found there, and some evidence suggests a quite different ceramic horizon in that area in Iron I times.72 The widespread distribution of Iron I pottery and the marked evidence for considerable

cultural uniformity in the period points to a fairly recent common origin for the Iron I culture, and
68The archaeological discussion which follows is based principally on two articles which review the comparative data in detail: see T. C. Young, Jr., "A Comparative Ceramic chronology for Western Iran, 1500-500 B.C. ", Iran 3, 1965, 53-85, and R. H. Dyson, Jr., " Problems of Protohistoric Iran as Seen from Hasanlu ", JNES, 24, 1965, 193-217. The rather cumbersome terminology proposed in the first of these articles should be amended following Dyson, ibid., p. 211: the Early Western Grey Ware Period is now Iron Age I, the Late Western Grey Ware Period is Iron Age II, and in part the Late Western Buff Ware Period is Iron Age III. Note that Late Western Buff Ware originally included Dyson's Iron Age III and the early historical material. Since Hasanlu IIIB and IIIA are now much better defined, Dyson's distinction should be followed. The documentation and discussion of points of fact or controversy mentioned here and not specifically footnoted will be found in the two above articles. '69 Marlik is certainly in part related to and contemporary with other sites of Iron I date and its stone-lined tombs are, therefore, an exception to the pattern. Occupation at Marlik seems to have continued down into Iron II times as well: see E. Negahban, " A Brief Report on the Excavations of Marlik Tepe and Pileh Qal'eh ", Iran 2, 1964, 18-19, and E. Porada, The Art of AncientIran, New York, 1965, pp. 90-o104. 70Often it is not possible to distinguish Iron I and II occupations from survey surface collections. Grey pottery of the correct fabric establishes with certainty the occupation of a site in one or the other period, but since a number of the more

common shapes are found in both periods it takes a particular sherd to fix a more precise dating. One should also be warned of the possible confusion between certain grey ware bowl shapes of Iron I date and similar pottery comparable to the Early Bronze Age grey wares at Yanik Tepe, a ware which is fairly common over parts of western Iran, particularly in the Hamadan Plain. Such a confusion may have occurred at two or three of the sites in the AsadaMbid-KangSvar-NehavandBorfijerd region originally surveyed by the author in 1961. For details of Iron I-II occupations in this area, see T. C. Young, Jr., " Survey in Western Iran, 1961 ", JNES, 25, 1966, 228-35. 7' Personal communication from R. H. Dyson, Jr. 7 A buff and light or dark brown plain ware sometimes decorated with monochrome or bichrome painted designs and occurring in shapes which are only suggestive of some of the vessel forms found further south in the late second and early first millennium has been found at JUtim Tepe, a large citadel site in the Shdpfir Valley at the northwestern corner of Lake Urmia. Traces of this pottery occur at other sites north of Lake Urmia and over toward the Ardabil region near modern Sardb. The painted variety is probably a northern variant of Iron III wares, while other shapes suggest dates in the Iron I and II periods. The pottery is sufficiently distinct that any precise chronological sorting on the basis of comparisons to excavated sites further south is impossible without excavation. Some of the painted ware has been appearing in antique shops in Tehran in the past two or three years, one suspects from illegal excavations between Tabriz and Ardabil.

AA
SCAS
0

P8 EA

AN

%S
,

*TABRIIZ

0 0

00

<EE
I

00

E
-

rree
I R A Q
/

I"r"

TErl S

AN

A?

0 00?0 0 O0

SEE

E
~060 G

g*

I~I

E:

EXC4vATrED

SITES .I]RON
o

SITEA eoio AE.I-iL

so

40

eokm

Iran Fig. 2. Western : SitesDating to IronI and H Periods.

24

JOURNAL

OF

PERSIAN

STUDIES

suggests a rather rapid diffusion of that culture over much of western Iran. The appearance of Iron I grey ware in western Iran marks an almost complete truncation of earlier painted pottery traditions at the principal sites where it has been excavated. One expects such a break at Sialk, for a long hiatus separates periods IV and V. Yet at Hasanlu the cultural break between periods VI (in the painted pottery tradition) and V is equally sharp without any significant chronological gap. The same is true for the shift from the painted pottery of Giyan II to the Iron I pottery of Giyan 14-13, and at Geoy Tepe in the shift from periods C to B. Thus, in a short period of time, old ceramic traditions in the Zagros are quite literally cut off, and new ones introduced. Indeed, in the entire prehistoric ceramic sequence of western Iran there is no sharper or more rapid cultural shift. The important ceramic links between Iron I sites and earlier sites in northeastern Iran, particularly Tepe Hissar, indicate a possible origin for these new ceramic traits on the western plateau. Of the five type shapes of the Iron I period, three are found at Hissar in periods IIIB or IIIC, and only one of these, the simple cup with a handle, appears at Hissar in a slightly modified version. The appearance of grey ware alone in the west is enough to make one look to Hissar, but more significant is the presence of pattern burnishing at almost all Iron I sites in the west (Giyan excepted), for this technique of surface decoration is a hallmark of Hissar III. Further evidence of the northeastern origin of the Iron I ceramic tradition can perhaps be seen in the fact that the sites closest to Hissar (Sialk and Khorvin-Chandar) show the greatest number of ceramic links to Hissar III.73 These same data can also be read as an indication of a chronological slope for the spread of the Iron I tradition into western Iran from east to west. Sialk V may have begun some hundred years before Hasanlu V and as much as a hundred and fifty years earlier than Giyan 14-13, and Khorvin-Chandar may have been occupied even earlier than Sialk V.74 A considerable hiatus exists between the final occupation of Hissar III and the beginning of the Iron I period around 1300/1250 B.C., which makes all the more striking the distinct links between late Hissar and the Iron I sites in the west.75 Sites which bridge this chronological gap will almost certainly be found as more exploration is conducted west of Hissar and east of Khorvin-Chandar.76 to The Iron Age II Period (c. I oo1000 800 B.c.). Typical of all sites with occupations in Iron II times is the continuation of the plain grey pottery tradition that began in the Iron I period. This uninterrupted development is best documented at Hasanlu and Giyan, but is also characteristic of Sialk in spite of the probable hiatus between periods V and VI.7z Extramural cemeteries also suggest a continuation of earlier traditions. In contrast to the Iron I period, however, iron comes into widespread use; several close parallels between Hasanlu IV and Sialk VI suggest something of a common metallurgical tradi73 Marlik also shows certain traits paralleled at Hissar but not

is the appearance of a four-spoke wheel on what may be a horse-drawn chariot on a seal in Hissar IIIB: see E. F. found in other Iron I sites. See esp. the tanged spear-heads and R. H. Dyson, Jr., " Notes on Weapons and Chronology Schmidt, Excavations at Tepe Hissar Iran, 1931-1933, Philain Northwestern Iran Around 1000 B.C. in M. Mellink, ed., delphia, 1937, fig. I i8, H. 892. Spoke wheels are otherwise ", unknown in the Near East prior to c. 2009 B.C., and are not Dark Ages andNomads,Istanbul, 1964, pp. 40-43. 74 Khorvin-Chandar is difficult to date with any precision since commonly documented until a century or two later: see a great deal of the material attributed to the site comes from C. Singer, et al., ed., A History of Technology, Oxford, I958, pp. 212-13 and 721-22. illegal excavations. Discussion should be confined to the 76 A few sherds scattered on the surface of Hissar today in a grey evidence from actual tombs excavated by L. Vanden Berghe, ware quite unlike that of the Bronze Age levels suggest that Le ne'cropole Khurvfn, de Istanbul, 1964 and from sherds collected there was a badly ersoded stratum at that site perhaps dating at the site. 75 Suggested dates for the end of the Hissar sequence have ranged to the early Iron Age. The site of Murteza Gert on the southB.c. (D. McCown, The Comparative western edge of Tehran has also yielded an early Iron Age all the way from c. 2300oo see Schmidt, of grey ware which might fit into this gap Stratigraphy Early Iran, Chicago, 1942, p. 42) to as late as " Dating the Hissar Sequence-the 1500 B.C. (S. Piggott, op. cit., p. 319, n. 2 and p. 323. 77 Sialk VI is certainly a very long-lived period which ends as Indian Evidence ", Antiquity, 17, 1943, I69-82, and D. H. late as the beginning of the seventh century. Evidence confirmGordon, " The Chronology of the Third Cultural Period at ing this late end for the period does not, however, tell us Tepe Hissar ", Iraq, 13, 1951, 40-61). A recent restudy of the anything about when the period began. On the evidence of the problem, taking advantage of the new comparative evidence from Yarim Tepe suggests a compromise between the two strong tradition of grey ware (24 per cent of the assemblage) with a few parallels to Hasanlu IV and Giyan extremes of a date for the end of Hissar IIIC around 1900and the II-Iz " Problems in the Relative definite links to Hasanlu IV for a good many of the nonI8OO B.c.: see R. H. Dyson, Jr., ceramic finds (see below, n. 78) the beginning of Sialk VI and Chronology of Iran, 6000-2000 B.C. ", in R. Erich, ed., in perhaps a large portion of the total period must be put in Chronologies Old World Archaeology, Chicago, 1966, p. 241. Iron II times. One small bit of evidence apparently thus far unnoticed

THE

IRANIAN

MIGRATION

INTO

THE

ZAGROS

25

tion in Iron II times.7" Fortified citadels are also now documented at one site at least and perhaps at others.79 Certain fundamental traits thus bind the several assemblages together sufficiently to warrant treating the period as a meaningful unit. Nevertheless, the principal characteristic of the Iron II period is the development of regionalism and the disappearance of the cultural uniformity which marked the Iron I period. At Hasanlu a wide range of new, complicated and unique pottery shapes makes its appearance in period IV. Several non-ceramic artifacts cannot be paralleled at other western Iranian sites in the same time range, and a strong current of Assyrian influence which is not found elsewhere can be traced in the assemblage. In architecture, evidence points to the introduction of features which are so far unique to Hasanlu.'o At Sialk a type of gabled tomb appears which is distinctive, often containing skeletons of an unusually brachycephalic people; a good many non-ceramic small finds are unique; and, most important of all, a distinctive painted pottery appears side by side with the ongoing grey ware tradition which began in Sialk V. Giyan does not yield sufficient data to reveal its full character in the Iron II period, but on the whole Iron I ceramic traits seem to continue in detail longer here than elsewhere, and Giyan's regional variation is perhaps a kind of archaism. What little evidence is available for Iron II material from the Ardabil region suggests that here too is found a considerable amount of ceramic variation distinctive to sites in the immediate neighbourhood. Thus the period is one in which each site represents something of an entity in itself. Because regional variation is the main characteristic of Iron II times, surface survey is severely limited as a means of identifying occupations for this period, and little can be said positively about a distinct distribution of the Iron II culture (Fig. 2). On the basis of excavated data alone a distribution not unlike that of the Iron I period is suggested, since Iron II materials are known from Geoy Tepe, Hasanlu (and other sites in Soldfiz), Dinkha, possibly Yanik Tepe, Giyan, Sialk and probably Marlik. Some indications point to a definite Iron II occupation in the Ardabil region. On the absence of grey wares alone we can be as certain as survey permits that Iron II occupations are missing in those same areas which yielded no Iron I pottery. Fundamentally the Iron II culture represents a development from that of Iron I times. Thus there is no break in the archaeological sequence c. iooo B.C. to match that which occurred c. 1300/1250 B.C. The regional variations which develop in Iron II times can probably be traced to three forces of perhaps roughly equal weight. There is little question that outside influences played an important role. The Assyrian impact on Hasanlu IV is obvious. Less certain in terms of their origin are those traits at Hasanlu, ceramic and nonceramic, which point in general terms to influences from the north and the Caucasus (some metal forms

78 Parallels for the following objects from Sialk VI can be found

at Hasanlu IV. All references are to R. Ghirshman, Fouilles de Sialk II, Paris, 1939; hereafter Sialk II. (i) Swords with bronze handles and iron blades, P1. L, S. 546d; (2) bronze mirrors, P1. LII, S. 58o; (3) bronze armour buttons, P1. LV, S. 796c, 592a, 591; (4) dangles, P1. LVI, S. 589; (5) bronze buttons, P1. LVI, S. 595, 794; (6) simple two-piece horse bits in both bronze and iron, P1. LVI, S. 588, 841 and P1. LXVIII, S. 715; (7) iron sickles, P1. LVII, S. 832a; (8) lance heads, P1. LVII; (9) plain copper or bronze buckets, P1. LVII, S. 828 and P1. LXVIII, S. 717; (io) bronze bells, P1. LVI, S. 833; (I1) decorative furniture knobs, P1. LXII, S. 761; (12) two-pronged iron forks, P1. LXVIII, S. 711; iron chisels, P1. LXXI, S. 893a; (14) bronze sword with (IS3) inlaid handle, P1. LXXV, S. 910; (15) bronze tacks, P1. LXXV, S. 917; (16) leaf-shaped arrow heads, P1. XCII; in bronze, P1. LVI, S. 835; (I7) triple-holed bit attachment (18) leaf-shaped spear heads, P1. XCII; and (i9) plain and knobbed bronze toggle pins-see esp. P1.XCII, S. 1785a and d. 79Evidence for any large fortification at either Sialk or Giyan dating to Iron II times remains enigmatic. For Giyan, see Young, Iran, 3, 1965, 64. The "grand massif" at Sialk

(Sialk II, p. 23 and P1. XXXIV) unfortunately had no structures or occupational debris left on its upper surfaces when excavated, and no occupational debris or slump containing pottery of Sialk VI date is reported from the areas excavated at the sides of this great brick platform. The date of the structure is apparently based on what has been described as a handful of sherds found loose on the surface of the platform; too few sherds, in fact, to suggest any distinction between Period V and VI until after Necropoles A and B were excavated in a later season: see R. Ghirshman, " Rapport preliminaire sur les Fouilles de Tepe Sialk (Kashan) ", Syria, 16, 1935, 229-46. Such sherds might easily have washed out of the brick work of the platform. Masses of Sasanian (?) and Islamic pottery litter the platform today, and on the basis of possible parallels with somewhat similar structures of known Sasanian date, of which several are in the Tehran Plain south and southwest of the city and near Verdmin, a much later date for the "grand massif" than the Iron II period should not be ruled out.
8o T.

C. Young, Jr., "Thoughts on the Architecture of Hasanlu IV ", IranicaAntiqua.(In press.)

26

JOURNAL

OF PERSIAN

STUDIES

few artifacts from that site have parallels in the north and the Caucasus.82Several features indicate some link between Sialk VI and Hissar III.83 Since there can have been no direct contact between Hissar III and Sialk VI on chronological grounds, we may assume either that these traits in Sialk VI representhold-oversfrom period V not documented in the fifteen tombs excavated for the latter period, or that Sialk VI experienced a limited renewed contact with the area of the old Hissar tradition in the east and a cultural descendent of Hissar III which has yet to be documented through excavation. It is significant that most of these "eastern" traits appear only at Sialk and not at Iron II sites furtherto the west. One suspectsthat it is also to the north that one should look for the origin of some of the materials
in the Ardabil region dating to Iron II times.84 Giyan, however, appears to be relatively free of features

and the rare object with Scythian affinities) and from Anatolia (perhaps certain architectural innovations).8, A foreign point of origin might perhaps be sought for the painted pottery of Sialk VI, and a

that cannot be traced back into the Iron I period, though admittedly the data are quantitatively limited. Altogether the foreign influences at work in the Iron II period seem to be as diverse and as varied as the several assemblages defining the period, and cannot be organized into any consistent pattern. Other influences were also equally important. The cultural unity so marked in Iron I times could not last. Indeed, it is remarkablethat it lasted as long as it did, for as the carriersof that culture, now widely distributed over western Iran, increasingly lost contact in time and space with their common cultural origin and with each other, it was only reasonablethat regional variationson that common base
8i On possible Anatolian connections for the architecture see

n. 8o and Dyson, JNES, 24, 1965, 198. Most of the northern parallels to Hasanlu IV are found in the Talysh: see C. F. A. London, 1948, pp. 404-43. Schaeffer, Stratigraphie Comparde, The dating for much of the Talysh material remains uncertain since the bulk of it comes from cemeteries used over long periods of time. In dealing with this area and the Caucasus it is important to keep in mind that here is one of the most striking " refuge " areas in Iran, a region which has throughout history been a cultural backwater: see Frye, op. cit., pp. 8-9. Recent Iranian excavations in cemeteries along the Iran-U.S.S.R. border in the Dasht-i M6ghdn under the direction of S. Kambaxsh have demonstrated anew difficulties in dating on typological grounds alone much of the material from this area. Several of the tombs opened contained bovine figurines in pottery definitely in the Marlik tradition. One of these on typological grounds might have been dated early in the first millennium B.c., but the skeleton in the tomb, clearly associated with the figurine, had its eyes closed with two Parthian coins! (S. Kambaxsh kindly showed me these finds and discussed them with me in the autumn of 1965.) In any case, the northern influences on Hasanlu IV do not seem to have been particularly strong as compared with either the greater mass of material of local origin with cultural roots in part in Iron I times, or even with the more numerous parallels to Assyria. " Foreign " influence from the north on Hasanlu IV might mean principally renewed contact with related cultures which had experienced a different regional development on a common Iron I cultural base, and which in turn had been partially influenced by outside influences from beyond the Caucasus. For a discussion of the Iranian influences on the Talysh and the Caucasus in an earlier period which probably came from the direction of Hissar, the source here suggested for the Iron I tradition in general, see J. Deshayes, Les Outils de Bronze, de L'Indus au Danube, Paris, 1960, pp. 414-17. 82 Schaeffer, op. cit., pp. 475-76 and Sialk II, pp. 81-92. Again, as in the case of Hasanlu IV, the northern parallels for some aspects of Sialk VI certainly are there (some metal types, stone tombs, etc.), but they do not constitute those objects which define the principal character of the assemblage. Connections between Sialk VI and Marlik, in turn related to the north,

have been discussed by Porada, op. cit., pp. 10 o5-6. Ghirshman has noted in some detail the limited links between the painted pottery of Sialk VI and that of Phrygian Gordion: see Sialk II, PP. 92-94. A most mysterious fact is that SialkVI remains, after the survey of literally hundreds of sites, the only occurrence of this painted pottery in its fully developed form. To the best of my knowledge no Sialk VI painted ware has even appeared on the Tehran antique market that could not be traced directly to Sialk itself. Dyson, while pointing to the possible Anatolian connections of this pottery, stresses the possibility that much of its inspiration derived from a resurgence of native Iranian painted pottery traditions: op. cit., p. 2ol. The origin of some of the motives, as Porada has shown, can certainly be related to Marlik: op. cit., loc. cit. and p. 94. 83References are given first to Hissar from Schmidt, op. cit. and second to Sialk from Sialk IL. In pottery, cf. shoulder spouts -P1. XLI, H. 4296 (IIIC) and P1. LXVII, S. 731; pedestal fruit stands-Pi. XXVI, H. 4161 (IIB) and Pl. LII, S. 568; jars-Pl. XXXVII, H. 5126 (IIIB) and P1. LIX, S. 649a; and jugs-Pl. XLI, H. 2871 (IIIC) and Pl. LXIII, S. 853b. For various non-ceramic small finds, cf. bronze vessels with trough rim spouts-Pl. LVII, H. 4883 (IIIC) and P1. LXXI, S. 855; bronze mirrors-Pl. LIV, H. 4872 (III?) and Pl. LXVIII, S. 727; adzes-Pl. LII, H. 3247 (IIIC) or P1. LII, H. 2710 (IIIB) and P1. XCIII, S. 1467a or S. 1467b; leaf-shaped arrow heads-Pl. LII, H. 1856 (IIIC) and P1. V, no. i; pins with curled eyes-Pl. LII, H. 3141 (IIIB) and Pl. XCIII, S. 1440; needles-Pl. LIII, H. 5265 (III?) and XCIII, S. I785c; elongated bracelets of thick coiled wireP1. LIII, H. 4263 (IIIB) and P1. XCIV, S. 1384; shallow concave buttons-Pl. LV, H. 2631 (IIIB) and Pl. LVI, S. 819a; bronze tacks-Pl. III, H. 2244 (III?) and P1. LXXV, S. 917; snailshell beads, worked-Pl. LXX, H. 1755 (IIIB-C) and LCV, S. I353a and no. 21; and bronze human figurines -P1. XLVII, H. 3272 (IIIC) or H. 3200 (III) and Pl. LXXV, S. 918 or LII, S. 582. 84 Schaeffer, op. cit., pp. 500-4 and Childe, op. cit., pp. I 17-24. The question of foreign influences on the Talysh and the Caucasus in this period in relation to ongoing developments in Iran and Anatolia is discussed by Deshayes, op. cit., pp. 418-i19. It is not always clear in which direction the influences were flowing.

THE

IRANIAN

MIGRATION

INTO

THE

ZAGROS

27

should have developed in a natural evolutionary process even without the spur of foreign stimulus. Finally, and perhaps most important of all, much of the cultural breakup which characterized Iron II times can probably be traced to indigenous influences which were at first masked by the new Iron I culture but which were undoubtedly not totally lost. In time, these native traditions would reassert themselves and have a considerable modifying effect on the new culture introduced in the Iron I period. It is perhaps as much in forces such as this as in foreign inspiration or contact that we should seek the origins of the Sialk VI painted pottery traditions. Regional variation in Iron II times can be attributed to new and renewed but less sweeping foreign contacts, to natural cultural evolution and to an indigenous cultural revival. The Iron Age III Period (c. 750 to 550 B.C.). This period is characterized by the use of plain buff coloured pottery which is sometimes either painted or incised, usually with a motif of hanging triangles left blank, filled in solid, or cross-hatched. Little evidence beyond pottery is available for the period, though extramural burial remains the custom at Sialk, where period VI definitely continues well into Iron III times, and fortified citadels are found at Hasanlu, Ziweyeh, and the Zendan. All of the sites with materials in this time range are closely linked by their ceramic assemblages. Some evidence suggests a possible northern (Hasanlu IIIB, Ziweyeh, the Zendan, Geoy Tepe A in part and Yanik Tepe) and a southern (late Sialk VI, Giyan I and Achaemenid Village I) clustering of sites, with late Sialk VI perhaps acting as a link between the two groupings. In general the period may be described as one of considerable cultural unity throughout western Iran-something of a return to the conditions which were typical of Iron I times. In Iron III times there are more sites more widely spread over western Iran than in either the Iron I or II periods (Fig. 3). Of particular interest is the appearance of Iron III wares in quantity in the Mahi Dasht, at one site (and probably more) high in the mountains west of Saqqiz, and at Ziweyeh. In the north, Iron III pottery in a variant form is found at the northwest corner of Lake Urmia as well as at Yanik Tepe. It is widespread over the whole of the Hamadan Plain. In the southwest it appears at Susa and, on the basis of survey work by Miss Clare Goff, is common in eastern Luristan between the Mahi Dasht and Khuzistan.8" This broad distribution pattern, combined with considerable ceramic unity, is the dominant cultural feature of the period. The relationship between the Iron II and III periods is a question on which considerable discussion will focus until excavations are conducted which better document the archaeology of the eighth century. In the northwest a large chronological gap separates Hasanlu IV and IIIB: the latter period may have begun as late as the last quarter of the seventh century, whereas the former ended c. 800oo This B.C. gap is partially filled by the Zendan. This site, though firmly anchored in the Iron III period, shows the continuation of several ceramic traits characteristic of the Iron II period, which suggests that perhaps the Iron III pottery tradition grew out of that of Iron II, even though grey ware went out of fashion in the half century or more between the end of Hasanlu IV and the beginning of occupation at the Zendan. At Hasanlu itself there is considerable evidence for ceramic continuity from period IV to IIIB which gains in significance when one considers the hiatus separating the two periods. 86 Continuity is also suggested in the evidence from sites further south. It has been suggested on reasonable grounds that the painted ware of Sialk VI is perhaps ancestral to the painted and incised triangle ware tradition of Iron III times, and the tradition of painting pottery, in some cases with motifs paralleled by those used in the Iron III period, began in Iron II times. Certain features connect Sialk VI and Giyan I' with Achaemenid Village I, perhaps in terms of a progressive development, and Giyan I', of course, has its roots in earlier sub-divisions of Giyan I which date to Iron II times. In pottery, the only foreign connections which can be traced in the Iron III period are those with late Assyrian pottery at Nimrud, though a single incised vessel of the type documented at the Zendan and Ziweyeh has been found at Toprak Kaleh in eastern Anatolia. The absence of any significant foreign influences, taken together with the available evidence of basic continuity from the Iron II period, suggests that the development of the Iron III ceramic tradition represents only a gradual change
85 Personal communication. 86Further evidence for continuity is perhaps found at Geoy Tepe

A, but the unstratified nature of the data confuses the issue considerably.

A9

SC

S
S

P
E

I
A

STABRIZ

t EE

I
4

EE

S
e

--1
?O0

?TEHRAN 94 S

2E
s

"

". ,-,
Ee

'_".
I 0

S0

\,

)
%

'%

E: EXCAVATED SITES

'4

AGE II _IRON
'4

SITES
40 6o0kr

2o 10 o

20

Iran: Sites Dating to the Iron III Period. Fig. 3. Western

THE

IRANIAN

MIGRATION

INTO

THE

ZAGROS

29

of fashion in western Iranian pottery and not a sharp break in the sequence such as might be attributed to the appearance of a new people. In this regard, and with respect to the Sialk VI painted ware as the possible inspiration for the painted and incised ware of Iron III times, one wonders to what extent older indigenous traditions in western Iran may have reasserted themselves, as they already had in Iron II times, thus further modifying the grey ware ceramic tradition which traced its origins back to the beginning of the Iron I period.8" In dealing with the ceramic tradition of the following historical period we are on firm ground. Hasanlu IIIB, Ziwiyeh, the Zendan and Achaemenid Village I all have strong links to assemblages such as Hasanlu IIIA, Agrab Tepe, Achaemenid Village II-III, Persepolis and Pasargadae which can be dated to Achaemenid times. The origins of Achaemenid pottery in Iran lie in the Iron III period, and no break in the record marks the transition from pre- or proto-historic times to the historic age of Cyrus and his successors. A sharp truncation of older ceramic traditions and the introduction of something entirely new instituting a period of considerable cultural unity, followed by a period of regional cultural variation resulting from natural evolution and a reawakening of older indigenous patterns and fresh but less ordered and drastic foreign influences, followed by a period of further evolution and eventual cultural reunification over an even wider area than ever before summarizes at least the ceramic history if not the broader cultural history of western Iran in the centuries from roughly 1300 to 550 B.C. It remains to test how this archaeological pattern fits with the written evidence on the migration of the Iranians into the Zagros. and WrittenEvidenceCompared The Archaeological If we accept as a working hypothesis the proposition that the migration of the Iranians into western Iran was the kind of event which would be reflected in the material and particularly in the ceramic record, then the Iranians were responsible for initiating the Iron I or the Iron II or the Iron III traditions. Each of these possibilities must be examined in turn and tested against the evidence from the written sources. Three major arguments can be marshalled against a theory that the beginning of the Iron III tradition represents the initial appearance of the Iranians in western Iran. First, and perhaps in itself decisive, there is the problem of timing. The earliest that the Iron III tradition can be firmly docuB.C. and, in the southwest, at mented is, in the northwest, at the Zendan perhaps somewhat before 70oo0 Achaemenid Village I possibly slightly earlier. Yet it is clear from the Assyrian sources that the Iranians were present in western Iran en masse long before this. Second, there is the evidence of geography and site distribution. The Assyrians seem clear on the point that the early Iranians were not found in certain major regions of the western Zagros: i.e. in the area under Elamite control, in Ellipi, along the present Iran-Iraq border, and in the Mannean heartland. All of these areas, however, reveal archaeological evidence for occupation in the Iron III period. Third, there is the evidence of essential cultural continuity from Iron II to Iron III times. Thus, if the appearance of the Iron III tradition represents the first appearance of the Iranians in the Zagros, then their impact on the culture of western Iran is not observable as something new, at least in the ceramic record. This is not to say that the culture of Iron III times was not associated with Iranians (a point discussed further below), rather, simply, that it was not brought to western Iran by the earliest Iranian migration into the area. As might be expected, we find in turning to the complex cultural situation which obtained in Iron II times that the issue is not so clear. Some scholars, working principally with the evidence from Sialk, have put forward a case for associating the first migration of the Iranians into western Iran with the start of the Iron II period.88 In favour of this suggestion is the matter of timing: at least chronology is not an argument against an Iranian-Iron II equation. It has been traditional to date the appearance of the
87For example, differences between late Sialk VI and Achaemenid Village I and between the latter site and more northern Iron III sites can, in part, be traced to local Elamite pottery traditions, particularly with regard to the common ware of Achaemenid Village I.
3

88 R. Ghirshman, Iran, Baltimore, I954 and The Arts of Ancient Iran, New York, I964.

30

JOURNAL

OF

PERSIAN

STUDIES

Iranians in the Zagros to roughly the turn of the millennium and the Assyrian evidence cannot be construed as denying such a suggestion. More important are the undoubted foreign and therefore new elements in the Iron II assemblages, particularly Hasanlu IV and Sialk VI, which point to cultural influences if not an infusion of new peoples from beyond the immediate limits of western Iran. Since stress has been put on those parallels which point to the north and to Anatolia, it has been argued that the Iranian migration came into the plateau from the north and from the direction of the Caucasus. Finally, the certain links between Sialk VI and the Iron III period have been noted and are taken as evidence for the gradual spread of the Iranians deeper into western Iran during the seventh and sixth centuries. A number of perhaps more persuasive arguments, however, can be brought against this reconstruction. One crucial issue is an assessment of the basic continuity in the ceramic sequence from the Iron I to the Iron II period and of how important the new elements in Iron II times are in defining the culture of that period in contrast to those elements which can be traced back into Iron I times. On the one hand, how "foreign " are many of these new features (involving a fairly limited number of items, the bulk of which are metal artifacts) which come from the north? Do they represent a major influx of traits from beyond the Iranian plateau, or do they point principally toward contact between Sialk VI and Hasanlu IV and cultures in the Talysh and the Iranian Caucasus which are themselves related to Iron I or earlier traditions on the plateau? On the other hand, are the parallels sufficient to establish an actual migration of people from the Caucasus area into western Iran, particularly since metalworking traditions could easily have spread through trade (less likely for pottery) or through the influence of itinerant metal smiths ?89 One must also consider the foreign elements in the Iron II period which perhaps point to Anatolia: the possible common ancestor of Phrygian and Sialk VI painted pottery, the hyperbrachycephalic population of the Sialk VI cemetery, and the suggested architectural connections of Hasanlu IV. Should these hints of Anatolian influence prove correct with more excavation and data analysis, is one prepared to suggest that the Iranians came into the Zagros from Anatolia? More important perhaps, if the people who brought painted pottery to Sialk VI were Iranians, where were the rest of them? Except to the extent that this pottery, in combination with a general revival of painted pottery traditions, may perhaps be ancestral to some of the decorative traditions of Iron III times, Sialk VI painted ware is a unique phenomenon found only at one site-certainly a description the Assyrians never gave to the Medes. To some extent the same criticism could be made of any attempt to associate the new features at Hasanlu with the Iranians, for Hasanlu IV has yet to be defined, except in very variant forms, much outside of the Soldfiz-Ushnfi Valley. Finally, some consideration must be given to the elements in Sialk VI which point somehow to the east. These foreign influences on Iron II culture are as varied as the regional manifestations of that culture in its totality, and it is difficult to co-ordinate them into a coherent pattern which could fit the evidence on the Iranian migration drawn from the written sources. Our reconstruction of that migration from Assyrian sources suggested a predominantly east to west movement bringing large numbers of Iranians into at least the areas east of Alvand sometime well before 900 B.c., and yielded little or no evidence for Iranians further north in the western Zagros than Uishdish (Maragheh) and Zakirtu (the central Zarineh Rfud) until the later appearance of warrior bands of Scythian raiders. This picture corresponded with the limited Iranian and Classical evidence for early links between east and west Iranians. Certainly it is too much to ask the few eastern or Hissar connections of Sialk VI to bear the weight of a major migration of Iranians from east to west. If the foreign influences in the Iron II period which point north represent the main Iranian migration, why did the Assyrians find no Medes or Persians

89 Trade

influences along routes of ore distribution should be stressed. Much of the iron ore used in Iron II times was probably coming into Iran from the general direction of eastern Anatolia. Itinerant metal smiths, called " Gypsies " by the locals, operate today in a symbiotic relationship with the settled population in much of Kurdistan and Luristan.

In the autumn of 1965 a large group of these Gypsy smiths camped for some three weeks on the flat in front of the town of Kangavar and advertised their wares by holding nightly dances in their own tradition which the townsmen enjoyed greatly and considered highly colourful.

THE

IRANIAN

MIGRATION

INTO

THE

ZAGROS

31

north of Maragheh in the Lake Urmia region, the richest lands in the northwest and an area which would have immediately attracted a foreign invader coming from the north in large numbers? One might answer, because the Urartians firmly controlled these regions; which only raises the question, why are Iranians not mentioned in any Urartian sources except for Parsua southwest of Lake Urmia? Furthermore, are those northern elements in Iron II culture sufficiently numerous or of a type to mark a majormigration at all? Finally, what evidence is there from fully historic times for any important invasion of Iran bringing cultural changes of a magnitude comparable to those introduced by the Iranians having crossed the Caucasus ?90 Perhaps the crucial question, however, involves the basic characteristic of the Iron II period, its regional variation. Is the arrival of large numbers of Iranians in the Zagros the kind of event which would create such a period of strong cultural regionalism? Perhaps we should rather expect to find that the impact of such a new group would result in considerable cultural uniformity rather than the opposite. To phrase the issue another way, is not regionalism under the conditions which obtain in the Zagros, where the mountain landscape tends to break people up into separate cultural (and political) units, what might be expected of an older culture, not a new one? But the argument approaches an equilibrium. The data remain elusive and vague precisely because the period is one of such diversity in terms of both the individual assemblages and the foreign influences operative on those assemblages. The decisive factor in assessing the evidence, therefore, becomes the proposition that perhaps a more comprehensive case can be made for a theory which equates the initial Iranian migration with the appearance of Iron I culture. Timing is again no problem. Indeed, one might claim that to push the main Iranian migration back yet another two centuries would fit the sources better, since more time would be allowed for the undoubted Median consolidation east of Alvand prior to the ninth century. We have argued that the introduction of the Iron I ceramic tradition, if not that culture as a whole, represents the break in the archaeological sequence in western Iran, that after this break there is essential ceramic continuity through to Iron III times, and that such a break could only be explained as the appearance of a new people. Perhaps one should not demand too much of the written sources so far back into prehistory, but if it is not the Iranians who brought the pottery of Iron I times to the Zagros, then Who did? becomes a crucial unanswered question in Iranian archaeology. The written evidence points to the east as the direction from which the Medes and the Persians came, and the Iron I ceramic tradition has roots in Hissar III and related sites in the northeast which were occupied long before Iron I grey ware appeared in western Iran.9, That tradition and those who bore it spread fairly rapidly over much of western Iran in numbers which perhaps match what the written sources tell us of the earliest Iranians in the Zagros. Furthermore, the areas where this pottery is found match in considerable detail those areas where we have postulated that early Iranians were found by the Assyrians. As the bearers of Iron I grey wares spread out of the northeast (Hissar and beyond) they first established themselves in the Tehran-Qazvin area (Khorvin-Chandar) and perhaps moved southward to the fertile country along the edge of the Dasht-i Kavir (Sialk). They then moved into the Hamadan Plain (survey sites), while others of their numbers went down to the Caspian by

90 The

Mongols, perhaps comparable to the Scythians since they represented a warrior band of raiders, crossed the Caucasus from south to north. Russian advances over the Caucasus represent quite a different kind of event, the slow, steady, planned advance of a centralized eighteenth-nineteenth century empire. By contrast the east has often been the direction from which invaders entered on to the plateau in historic times. 9' It is beyond the scope of the present discussion to take up the issue of an association between the Indo-Europeans and the grey ware culture of Hissar III. One may assume that at this date one would be dealing with Indo-Iranians rather than Iranians. On chronological grounds it would not be un-

reasonable to associate the appearance of Indo-Iranians in northeastern Iran with the Indo-European migrations which brought the Hittites and the Luwians to Anatolia, the Mitanni ruling classes to Mesopotamia, and perhaps even the earliest Greeks to the Aegean. For a recent discussion of these possibilities, see H. L. Thomas, New Evidence Dating the Indofor European Dispersal in Europe(mimeographed paper read to the 1966 Conference on Indo-European Studies, Philadelphia), pp. 1-7. Unpublished evidence from Yarim Tepe will be vital in this regard, as indicated by the suggestion of the excavator that there the grey wares may have developed directly out of earlier Chalcolithic painted wares: see V. Crawford, " Beside the Kara Su ", Bull. Met. Mus. of Art, 21(8), 1963, 271-72.

32

JOURNAL

OF

PERSIAN

STUDIES

the Safid Riid Gap (Marlik).92 These are the regions where the Assyrians found only Medes in considerable depth and numbers. Iron I pottery also appears west of the Alvand barrier part way down the Khorisan Road, where the Assyrian sources indicate Medes did penetrate into the central western Zagros (Giyan and survey sites). The carriers of this grey ware also appear in the northwest in Soldfiz and other areas on the borders of Mannea proper, again regions where the Assyrians found Iranians (Hasanlu and survey sites). Those areas where Iron I pottery is not found are equally important for these are regions where the Assyrians seem to have found no early Iranians. There is no Iron I grey ware south of Bfikhdn and very little if any between Bfikhdn and Miandfiab (the Mannean heartland), none in the high mountains along the Iran-Iraq border (Namri, etc.), none in the north except for the isolated material at Yanik Tepe (areas under Urartian control and influence), none in the Mahi Dasht or ShThdbdd Valley (Ellipi) and none in the southwest (areas under Elamite control). Two locations pose problems. First, no Iron I pottery has yet been found in the area along the headwaters of the Diyala where Kar-Sharrukin and a considerable number of Medes were probably located. Second, at least some Iron I grey ware and related cultural elements are found in the Ardabil and Talysh regions, but there is no solid evidence from the written sources for Iranians in that area.93 It is possible that the kingdom of Zakirtu extended some little distance east from the Zarineh Rfid in the direction of Ardabil, but we shall probably never be better informed on this point since it is unlikely that the Assyrians had much first hand experience with northeastern Azerbaijan. Yet, with these two exceptions, the geography of the early Iranians as it can be reconstructed from the written sources and the geography of Iron I ceramics seem to match with very little overlap. Finally, and in the light of comments above on regionalization vis-a-vis the Iron II-Iranian equation, the relative uniformity of Iron I ceramic culture is perhaps what we should expect from the initial impact of the main Iranian migration. The whole period shows a broad character which fits a reconstruction of the Iranians as taking over fairly rapidly and en masse those areas where they were first found by the Assyrians. Given this suggestion that the Iron I culture represents the initial and major migration of the Iranians into the Zagros, problems associated with the Iron II period become perhaps more explicable. We can now consider that period as one in which the cultural impact of the original Iranian migration begins to be absorbed and modified by natural evolution, by geographic necessity, by two centuries of contact with the indigenous culture and by secondary and diverse foreign influences. In the more western Zagros, the most important of those outside influences naturally come from Assyria, and at Hasanlu IV we have ample evidence for the penetration of western Iran by the Assyrians under Shalmaneser III, Shamshi-Adad V and Adad-Nirari III. Thus the Iranians of the northwest experienced fully both the political and cultural force of Neo-Assyrian expansion. Their cousins living" at Sialk early in period VI were little affected by these influences, as witnessed by the written evidence for probably no more than Assyrian cavalry raids that deep into the plateau and by the absence of any
92 Though

there do not appear to be many Iron I sites in the Hamadan Plain in the light of the known concentration of Medes in that region, proportionate to the total number of sites in that area dating to any period earlier than Islamic times the Iron I period is well represented. Large stretches of this plain are unsuitable for agriculture and the total settled population in any given period in the ancient past was never very high compared to many other areas in western Iran. Modern farming methods, particularly well digging techniques and the ability to plough very large areas mechanically on the offchance of reaping a crop watered exclusively by rainfall in the occasional year, have greatly expanded the population of the Hamadan area in fairly recent times. One suspects that in earlier periods most of the plain was suitable onlyfor pasturage, and if the sheep and goats were kept off the land, pasturage for horses. May this not have been the reason why the Seljuks in their earlier years chose Hamadan as their capital and as the gathering point for their large mounted military units?

Iron I elements in the Marlik culture could, of course, have also penetrated up the Safid Rfid Valley from the Caspian shore from an ultimate origin in earlier sites related to Hissar III in Gurgdn. 93 The geographical mechanics of a penetration of Iranians into the Ardabil region from the south and east are relatively simple. Some could easily have gone up the Zanjdn Valley to the middle reaches of the Qizil Uzun and thence into the Ardabil Plain by the easy routes connecting those areas along the west face of the Elburz. Others might have followed either up the Safid Rtid Valley from the Marlik area to the same routes into the Ardabil region, or up the Caspian Coast to the passes into the valleys west of the Elburz behind Resht (since the actual coast was a thick, almost impenetrable jungle perhaps as late as the time of Shah Abbas, the main medieval Islamic caravan routes from Resht to Ardabil and the eastern Caucasus crossed the mountains in the area of Mdsfileh and did not follow up the coast to Astara).

THE

IRANIAN

MIGRATION

INTO

THE

ZAGROS

33

marked Assyrian cultural elements in the archaeological record. The same was true for Marlik, where the few traces of Assyrian contact perhaps go back to earlier and more peaceful times in the Iron I period.94 Those foreign influences pointing toward Anatolia tell another story. The invasion of the " Sea " at the end of the thirteenth and early in the twelfth People century brought troubled times to the eastern Mediterranean and to western Anatolia which lasted as late as the migration of the Dorians into Greece in the eleventh century. It is reasonable to expect that these events might have had delayed repercussions further east, and that Anatolian elements in the Zagros in Iron II times may reflect pressures on groups in Asia Minor which had their origin far from the Iranian scene. Turning to the north and those influences on Iron II culture which may come from beyond the Caucasus raises the issue of the Cimmerians and Scythians. Recent reconstructions of the archaeology of the northern Caucasus suggest that here there developed throughout the second millennium a local culture called North Pontic (closely linked with the Kuban culture) which can probably be associated with the Cimmerians. The Timber Grave culture which had developed over the same period in the area between the Volga and Yenesei Rivers is usually associated with the Scythians. This latter group moved into the north Pontic steppe and the northern Caucasus around or shortly after the start of the first millennium, displaced the north Pontic and Late Kuban culture and drove elements of that group (the Cimmerians) across the Caucasus into the Urartian and Phrygian regions of Anatolia. Some Scythians followed these displaced persons across the Caucasus in limited numbers to become the Scythians who appear in the Assyrian sources perhaps as early as the reign of Assurnasirpal II and certainly by the time of Esarhaddon.95 Here, in these events, is perhaps an explanation of those elements in Iron II culture which point to the north, and of the one or two specifically Scythian elements in Hasanlu IV. Such disturbances would both displace peoples originally located south of the Caucasus barrier (e.g. in the Dasht-i Mogan and the Ardabil region) whose own roots would go back to Iron I cultural origins, and would bring south a few features originally at home north of the Caucasus. Though there were, therefore, Iranian elements mixed with some of these foreign influences, the contribution of the newcomers from the east (the Hissar III-SialkVI parallels), from the north, and from Anatolia was to augment and to deflect slightly cultural trends already established in western Iran and to add to the diversity which was so marked a feature of Iron II culture. The gradual evolution of the Iron III tradition out of that of Iron II times, the growth of comparative ceramic unity in this period, and the spread of Iron III pottery over most of western Iran is an archaeological development which perhaps matches the political unification and spread of the Medes. The Assyrian sources suggest that the Manneans maintained something of their ethnic identity despite increasing Median pressure during the seventh century. Thus it is fairly late in that century that Iron III pottery appears at Ziwiyeh.96 The presence of wares related to those of Iron III northwest of Lake Urmia perhaps testifies to a still wider expansion of the Medes in the later seventh century, as would the appearance of Iron III pottery in the high mountains along the Iran-Iraq border in Kurdistan. To the south, Ellipi came under increasing Median pressure under Sennacherib and disappeared
94Porada, op. cit., p. 91. For the best summary of this reconstruction see G. Clark and S. Piggott, PrehistoricSocieties, London, 1965, pp. 274-78. Further details are found in M. Gimbutas, Prehistory Eastern of Europe, Vol. I, Cambridge, 1956: Gimbutas, " IndoEuropeans; Archaeological Problems ", Amer.Anthropologist, 65 (4), 1965, 833; and Sulimirski, op. cit. Note especially the latter's conclusions that many of the dates for materials north and south of the Caucasus should, in general, be lowered on the basis of recent Soviet excavations. If the Timber GraveScythian equation is correct, then here is evidence for an indigenous group of Iranians located north and east of the Caspian in the second millennium B.c. (the Timber Grave culture appears to have been a local development). If the cultures north of the Caucasus are to be associated with the Cimmerians (a non-Iranian group), then there would seem to be no Iranians in that area other than the Scythians, who arrive sometime after the start of the first millennium (when other Iranians were already in the Zagros). Thus, for the major Iranian migration into the Zagros to have crossed the Caucasus and come down from the north, the Cimmeriannorth Pontic equation would have to be in part denied, and some specific cultural group north of the Caucasus labelled Median-Persian. 96 The Zendan was in territory controlled by the Medes from early on and thus it is not surprising to find Iron III wares at that site earlier than they appear in the central Mannean area. It is possible that the appearance of Iron III peoples in Hasanlu IIIB represents a return of Iranians to that site after the Mannean expansion under Sennacherib and Esarhaddon which would cover the later part of the hiatus between Hasanlu IV and IIIB.

95

34

JOURNAL

OF

PERSIAN

STUDIES

as an independent state sometime in the reign of Esarhaddon, leaving the Medes free to move further west down the Khorasan Road. So it is that we find Iron III pottery in quantity in the Mahi Dasht and west into the ShahTbhd Valley where Iron I and II wares never penetrated. As the bulwark of Elam declined under repeated Assyrian blows, eventually to vanish between 642 and 639 B.C., even the relatively inaccessible southwestern Zagros was open to the Iranians, and Iron III pottery spread into western Luristan and to Achaemenid Village I. It is in these same decades before the death of Elam that the written sources tell us for the first time of a Parsua in the southwestern Zagros. As for foreign influences, it was during Iron III times that the Scythians were most active in western Iran, though on the whole the archaeological record has yet to yield much evidence of their presence. The Scythian elements in the Ziwiyeh Treasure, the horse burial found in the Iranian excavations at Hasanlu (if it does not belong to Hasanlu IV) and perhaps certain of the Luristan Bronzes are almost the sum total of evidence for the Scythians (and Cimmerians perhaps) in the Iron III period.97 If the picture of these people as warrior bands of nomads is correct, perhaps their apparent limited impact on the culture of the Zagros is to be expected, but one suspects that more evidence of their presence may be forthcoming as excavation gradually defines the Iron III period in terms broader than pottery. Assyrian cultural contact with the expanding Iranians was maintained during the seventh century, as illustrated by Assyrian elements in the Ziwiyeh Treasure, by individual pieces which have been ascribed to the Medes and which show Assyrian influence, and perhaps by the few ceramic links between late Assyrian pottery from Nimrud and Iron III wares. Thus the people associated with Iron III pottery were indeed Iranians, though not the original Iranians who migrated into the Zagros. Culturally they were linked back through Iron II and I times to those earliest Iranians, and they provided the ceramic basis for Achaemenid pottery just as the Medes laid the groundwork for the empire of Cyrus and his successors. Conclusion The historical reconstruction proposed above is undoubtedly too simple. One point is clear: the centuries from 1300oo 550 B.C. witnessed events in western Iran of such complexity that the available to data must sometimes be strained beyond what they perhaps can bear in any attempt to create a coherent explanation of those events. The rather assured tone in which the case has been stated at times masks a number of problems which cannot be disregarded, and of which the author hopes he is aware. To give but one example: partly because the detailed study of the non-ceramic evidence has not yet been attempted in a systematic fashion (particularly for unpublished and recently excavated materials) and partly because pottery is our best indicator of cultural change and hence perhaps of historical events, the argument on the archaeological side has been built mainly on the ceramic evidence. Other types of data should be ordered and brought to bear on the issue with due consideration given to the nature and import of the evidence they supply. But perhaps the broad outline reconstruction suggested here will provide something of a framework for the strategic planning of further excavation and for such data analysis. As stated in the introduction, the operative phrase remains, " to initiate a discussion ".

97 On the Scythians in the Iron III period, see Porada, op. cit.,

pp. I23-141 and Ghirshman, op. cit., pp. 98-125. The latter should be used only in conjunction with Dyson, op. cit., pp. 208-212. Ghirshman has also suggested that the bulk of the Luristan Bronzes were the work of the Cimmerians: op. cit., pp. 41-82. Porada has clearly demonstrated that indeed many of the bronzes may date to Iron III times, but many styles and motifs have their origin in Luristan or neighbouring areas

earlier in the first millennium. She rightly stresses the point that though nomadic invaders such as the Scythians or Cimmerians, or less transient newcomers such as the Medes, may have stimulated a renewed interest in bronze working in Luristan in the seventh century, the tradition so revived was essentially indigenous. Here is yet another example of the debt which the Iranians owe to the local cultures they overran. See Porada, op. cit., pp. 75-89.

35

THE SUFI MASTER AS EXEMPLIFIED IN PERSIAN SUFI LITERATURE

By Seyyed Hossein Nasr
The Sufi master is the representative of the esoteric function of the Prophet of Islam and by the same token the theophany of Divine Mercy which lends itself to those willing to turn to it. The shart'a, the Divine Law, is meant for all Muslims, and in fact from the Islamic view, for all men if its universal meaning is considered. But the tariqa,or spiritual Path, is only for those who seek God here and now and who search after that immutable Truth which although present here and now is also the eternal is source of all revelation. The tariqa thus a means whereby man can return to the origin of the Islamic and become in a spiritual sense both a companion and a successorof the Prophet and revelation itself the saints. or The role of the spiritual master, the shaikh,murshid, per, is to make this spiritual rebirth and himself connected through the chain of initiation (silsila)to the Prophet transformationpossible. Being inherent in the prophetic mission itself, the Sufi master is able to and the function of initiation (waldya)' deliver man from the narrow confines of the material world to the illimitable luminous space of the spiritual life. Through him, acting as the representativeof the Prophet, spiritual death and rebirth take present in him. place by virtue of the baraka Fallen man grows old, decays and dies, whereas the regenerated spiritual man is always inwardly in the prime of youth. Having drunk from the fountain of eternal life and having gained access to the elixir of immortality he lives in the perennial spring of the soul even if his body passes through the winter of life. And that is why he is able to endow the disciple with youth, whatever age he may formally be. To behold the perfect master is to regain the ecstasy and joy of the spring of life and to be separated from the masteris to experience the sorrowof old age. "I aged with his affliction, but when Tabriz You name, all my youth comes back to me.'' Man may seek the fountain of life by himself. He may seek to discover the principles of spiritual regenerationthrough his own efforts.But this endeavour is in vain unless the master is present. Without the philosopher's stone no alchemical transformationis possible. Only the power of the shaikh can deliver man from himself-from his carnal soul-and enable him to behold the Universe as it really is and to join the sea of Universal Existence. " Without the power imperial of Shamsu'l-Haqq of Tabriz One could neither behold the moon nor become the sea."3 Of course there are those who are initiated into the way by Khidr or the " men of the invisible such as the Uwaisis, or, in the case of Shi'ism, by the Hidden Imam who is the hierarchy " (rijdlal-ghaib) of the Universe. But all these ways, which are not for men to choose and to seek spiritual pole (Qutb) but for which they must be chosen, nevertheless concern the universal initiatic function of which the Sufi master is the embodiment on earth. Inwardly one with the invisible hierarchy and the Truth
of (haqq) itself, he appears outwardly among men as the sign of the supreme mercy (ralhmat) God, as the
The term wall in the context of Sufism means " saint ". The Arabic root besides meaning " friend " also possesses the meaning of dominion or power. From this root the terms wildya and waldya are formed, the first meaning sanctity and the second the initiatic power or function as such. In Shi'ism especially, the distinction between " the cycle of " " prophecy (dd'iratal-nubuwwa)and the cycle of initiation" (dd'iratal-waldya) is emphasized. R. A. Nicholson, Selected Poemsfrom the Divdni Shamsi Tabrfz, Cambridge, I952, p. 25. Shams-i Tabrizi was the spiritual master of the greatest Sufi poet of the Persian language, Jala^l al-Din Rifmi, and also for him the perfect theophany of the Divine Names and Qualities, the total exemplar of the Universal Man. The Divdn especially, contains some of the most beautiful and profound verses in Persian on the function of the spiritual master and the relation between master and disciple. The name of Shams-i Tabrizi (Shams al-Din meaning the " sun of religion ") is itself highly symbolic and Rumi often uses the symbolism of the name in verses which seem to refer to both the master and the Divine Truth Itself alluding again and again to the inner union of the master with God. 3Nicholson, op. cit, p. 79.

36

JOURNAL

OF

PERSIAN

STUDIES

means whereby man can gain access to the spiritual world and be admitted to the company of prophets and saints. He is the door through which one must pass in order to enter " the garden of the Beloved" while at the same time he is the guide to the inner court of this garden. To become initiated into a Sufi order and to accept the discipleship of a master is to enter into a bond that is permanent, surviving even death. For the disciple the shaikhis always mysteriously present especially during the rituals. The shaikh never dies for the disciple even if he has left the world. His spiritual guidance (irshdd) and assistance continue after death. The spiritual master, whom Rflmi calls the heavenly rider, comes and goes, but the dust of his gallop remains. His effect upon his disciples is permanent and the seed he has sown in their hearts continues to be nurtured and cared for, even after the temple of the body has fallen into dust. Under his care even from this ruin of earth, that seed can grow into a tree which stretches to heaven and extends from the Eastern to the Western horizons.4 " Slave, be aware The lord of all the East5 is here; The glittering storm-cloud of eternity Reveals his lightning-flash to thee. Whate'er thou sayst Is but as inference has guessed; He speaks upon the eye's experience, And therein lies the difference. The heavenly rider passed; The dust rose in the air; He sped; but the dust he cast Yet hangeth there. Straight forward thy vision be, And gaze not left or right, His dust is here, and he In the Infinite."6 Likewise the assembly (majlis) of the Sufi master is a terrestrial image of the heavenly assembly of the saints. The disciple (faqfr) who gains the right of entry into this assembly, by virtue of having been initiated by the master, also gains for himself a place in the assembly of paradise provided he remains faithful to the master and his instructions. Once he fulfils the conditions of discipleship and reaches perfection in the assembly, his station becomes of permanent importance and gains a significance beyond the life of this world and the grave. The master leaves a permanent mark upon the disciple by virtue of which the disciple who has reached perfection again joins the assembly of his master in the other world. By the perfection gained in the majlis of the Sufis, thefaqir gains access to the royal assembly of heaven and constructs for himself an exalted abode in the after life. "And, O friend, if you reach perfection in our assembly (majlis) Your seat will be the throne, you will gain your desire in all things. But if you stay many years more in this earth, You will pass from place to place, you will be as the dice in backgammon. If Shamsi Tabriz draws you to his side, When you escape from captivity you will return to that orb."7
4 It must, of course, be remembered that initiation into Sufism 5 Since the name Shams al-Din means the " sun of religion "

by a master does not in itself guarantee realization. The disciple must be firm in his devotion to the master and in performing his religious and initiatic duties. He must love God more than the world and be attached to Him not only through a theoretical comprehension of Sufi metaphysics but also by a total " ontological " attachment. There must also be present Divine succour and aid (taufiq) without which nothing is possible. The gardener sows many seeds in the ground: not all of them grow to be plants that bear fruit.

and the sun rises in the east, Rfimi often employs the universal symbolism of East and West as domains of light and darkness and refers to his master as the " lord of the East " again implying his union with the source of all light. 6 A. J. Arberry, The Ruba'iydt of Jaldl al-Din Rumi, London,
1949, P- 19.

7 Nicholson, op. cit., p. I87.

THE

SUFI

MASTER

AS

EXEMPLIFIED

IN

PERSIAN

SUFI

LITERATURE

37

Not only is the influence of the Shaikh permanent but his light is everywhere. Though distinct as a spiritual personality, he is inwardly identified with the light that shines upon the land and sea and illuminates all things by the disciple who is closely attached to him. "From Tabriz-ward shone the Sun of Truth, and I said to him: 'Thy light is at once joined with all things and apart from all.' "8 It is this light that shines upon the heart of the disciple and converts it from something corruptible and perishable to an incorruptible and eternal substance. It is the spiritual influence of the Shaikh that transforms the transient, passing into the abiding and everlasting. " The sun of the face of Shamsi Din, glory of the horizons Never shone upon aught perishable but he made it eternal."9 Not everyone who claims to be a Sufi master possesses all the qualifications that make for the perfect master. Not only must the disciple seek a master, but he must also be sure that the shaikh or pir to whom he is surrendering himself is a seasoned guide who can lead him through the dangerous precipices of the way to the final goal of realization. Otherwise there is a danger of deviation and the corruption of what is most precious in man. With an incompetent guide, it is best not to climb mountains but to remain on flat ground where the danger of a fall would be much less fatal. The shaikh or Sufi master must have a regular connection with the chain of initiation and a realization of the truths of the Path. Moreover, he must be chosen from on high to fulfil the function of guiding others. Even among those who have advanced on the Way, not everyone is qualified to become a master. The master is chosen by the hand of Divine Mercy to guide men. He cannot claim to fulfil this function simply through his own will. If he is to be an effective master he must know the details and intricacies of the Path and also the soul and psychic substance of the disciple whom he is to guide. Not only is the presence of a perfect shaikh necessary but he must also be a master who has the qualifications for guiding a particular disciple. Not every shaikhis a master for every disciple. The disciple must seek and find the master who conquers his soul and dominates him as an eagle or falcon pounces upon a sparrow in the air. " O Splendour of the Truth, IHusamu'ddin,Io take one or two sheets of paper and add (them to the poem) in description of the Pir. Although thy slender body hath no strength, yet without the sun (of thy spirit) we have no light. Although thou hast become the lighted wick and the glass (lamp), yet thou art the heart's leader (the Spiritual Guide): thou art the end of the thread (which serves as a clue). Inasmuch as the end of the thread is in thy hand and will, the beads (of spiritual knowledge) on the heart's necklace are (derived) from thy bounty. Write down what appertains to the Pir (Guide) who knows the Way: choose the Pir and regard him as the essence of the Way. The Pir is (like) summer, and (other) people are (like) the autumn month; (other) people are like night, and the Pir is the moon. I have bestowed on (my) young Fortune (IHusamu'ddin) the name of Pir (old), because he is (made) old by the Truth, not (made) old by Time. So old is he that he hath no beginning: there is no rival to such a unique Pearl. Verily, old wine grows more potent; verily, old gold is more highly prized. Choose a Pir, for without a Pir this journey is exceeding full of woe and affright and danger. Without an escort you are bewildered (even) on a road you have travelled many times (before): Do not, then, travel alone on a Way that you have not seen at all, do not turn your head away from the Guide.""
8 Nicholson, op. cit., p. 27. 9 Ibid.,p. III. 1o When writing the Mathnawi, Rfimi's spiritual pole of attraction was Husaim al-Din, the figure who is so often cited in the Mathnawt. SRfimi, Mathnawi, trans. R. A. Nicholson, vol. II, London, I926, pp. 16o-6i.

38

JOURNAL

OF

PERSIAN

STUDIES

To accept initiatic guidance from one who is not a perfect master is most dangerous for it may completely spoil the possibility of spiritual realization and even open the soul to demonic influences. The potentiality of growth is there within men, but if it is not trained correctly, it will become spoiled and like a spoilt seed never be able to grow into a tree. " As for the master (murdd), in the sense of one who is initiated and followed, he is one whose initiatic power (waldya) in influencing others has reached the degree of perfecting those who are imperfect and who has seen [initiatically] the different kinds of capabilities and ways of guiding and training disciples. Such a person is either a traveller attracted by Divine Grace (sdlik-i majdhdb)who has first traversed all the deserts and precipices of the qualities of the carnal soul through travelling upon the Path (suldtk),and then with the help of Divine attraction has returned from the stations of the heart and ascensions of the spirit, and has reached the world of vision and certainty and joined the state of contemplation and examination. Or he is one attracted by Divine Grace who travels on the Path (majdhdlb-i sdlik), who first through the help of Divine attraction has traversed the extent of the stations and has reached the world of vision and unveiling of the Divine realities and then has crossed again the stations and stages of the Path through travelling (suldk) and has rediscovered the truth of contemplation in the form of knowledge. " The degree of being a spiritual guide, worthy of imitation, is certain in the case of these two types of men. As for the destitute traveller (sdlik) who has not as yet left the narrow strait of spiritual struggle and endeavour to reach the space of spiritual vision or the destitute person attracted by Divine Grace who (majdhdTb) has not as yet become aware of the intricacies of travelling upon the Path or of the realities of the stations and stages, and of the pitfalls and dangerous passages of the Way, neither has as yet the right to the station and rank of spiritual master (shaikhdkhat).The initiatic power of influencing the spiritual capabilities of the adept and training the disciples according to the laws of the Path (tarfqa) has not been entrusted to them. Whatever conquest they make in the souls of men in this sense is more harmful than beneficial. " The existence of the disciple and the potentiality of spiritual perfection in him is like an egg in which the potentiality of becoming a bird is present. If it becomes capable of receiving the spiritual power, the influence of the spiritual will (himmat) and the protection of a mature bird in whom the power of procreation and causing the egg to hatch has become predominant with ebullition, and if for a period the influences of the spiritual life and the characteristics of the perfection of being a bird affect him, then finally the form of egg is taken away from him. He will be dressed in the form of a bird and made to reach the perfection of his capabilities. And if the egg is placed under a chicken who does not possess the power of flight or has not as yet reached the degree of maturity and power to make the egg hatch and this goes on for some time, the potentiality of becoming a bird is destroyed in it and then there will be no way of restoring it. " Likewise, if the sincere disciple places himself in complete obedience and submission under the control of a perfect master who has attained the degree of perfection and in whom travelling upon the Path, flight, spiritual march and attraction to the Divine Grace are combined, from the egg of his existence the bird of the truth ' Verily God created Adam in His image ' will come forth, a bird which will then fly in the space of spiritual identity and reach the degree of procreation and generation. But if he comes under the influence of a destitute traveller (salik) or one attracted to Divine Grace (majdhd/b) then the possibility of the perfection of the human state will become spoiled in him. He will not reach the excellence of spiritual men (rijidl)or the station of perfection."'2 The disciple must surrender himself to the perfect shaikh, without any reserve. In the hands of the master he must be like a corpse in the hands of the washer of the dead without any movement of its own. The master is the representative of the Prophet and through him of God. To shake his hand is to accept the " Hand of the Divine ". zz'Izz
al-Din Kishani, Misbdh al-hiddya, ed. J. Homd'i, Tehran, I323 (A. H. Solar), pp. Io8-Io9 (translated from the Persian

by S. H. Nasr).

THE

SUFI

MASTER

AS

EXEMPLIFIED

IN

PERSIAN

SUFI

LITERATURE

39

"God has declared that his (the Pir's) hand is as His own, since He gave out (the words) the Hand of Godis abovetheirhands. The Hand of God causes him (the child) to die and (then) brings him to life. What of life ? He makes him a spirit everlasting."'3 The role of the Sufi master to whom one must make perfect surrender and his significance in delivering the disciple from bewilderment in the world of multiplicity to contemplation in the world of Unity are well exemplified in the spiritual testament of Shams al-'urafh' (" the sun of the gnostics "), one of the leading Sufi masters of the present century in Persia. Shams al-'urafA' described his meeting with his master and subsequent transformations that overcame him as follows: "This humble faqi'r, Sayyid Husain ibn al-Ridla al-Husaini al-Tihrani al-Ni'matullThi, was blessed with divine favour in the year 1303 (A.H.) when I met His gracious Highness, the model of gnostics and the pole of orientation of the travellers upon the Path, the honourable direction of prayers Shaikh 'Abd al-Quddflis Kirmanshahi. At this time all my attention was directed to the study of the formal (traditional) sciences and I possessed some knowledge of medicine, philosophy, mathematics, geometry, astronomy and astrology, jurisprudence and its principles, grammar, geography and prosody and was occupied with studying and teaching. But I had no knowledge of the problems of Sufism and the laws of spiritual poverty and gnosis and was unaware of the science of the truth and the intricacies of Divine knowledge. My attention was turned only to the problems of the formal sciences and the debates and discussion of text-books but not to inner purification, embellishment and contemplation. I had made no endeavour on the path of purifying the soul and cleansing the inner being, thinking that the way to know the truth is none other than pursuing the formal sciences. " Thanks to Divine Grace and the aid of the Pure Imams-upon whom be peace-I met that great man on the above-mentioned date near Imam-Zadih Zaid.'4 He did with me what he did. Again within the distance of a week I was blessed with his presence near Imam-Zadih Zaid. " After some conversation I expressed the wish to become initiated. On Thursday night I went to the bath at his side and received the ritual ablutions that he had ordered. After the bath he took my hand in the customary fashion and after performing the formula of repentance he instructed and initiated me to the invocation (dhikr) of the heart with the litanies (awrdd), the particular initiating acts and invocations. I obeyed. After fifteen nights [the minor retreat (khatwat-i saghir)] near the hour of dawn, while in contemplation, I saw all the doors and walls of the dark room in which I was placed participating in the invocation with me. " I fainted and fell. After sunrise my corporeal father because of the great love he had for me did not stop at any measure in bringing a physician and calling those who attract the jinn (psychic forces) or write prayers to cure illness. My corporeal mother also did all possible in the way of administering different medicines, fumigations and nourishments. " For twenty days I was in such a condition. I could not perform the duties laid down by the nor was I aware of formal customs. I spoke to no one concerning this matter. After this period my shari'a condition returned somewhat to normal and I became free from the state of' attraction '( jadhba). I went to the bath and purified myself. I felt the desire to meet that great master and for a few days I wandered as a mad man in the streets and bazaars seeking him. Finally I succeeded in meeting him. I kissed his hand and he expressed his benevolence towards me. " To summarize: for two years I travelled upon the spiritual Path under his care and following his instructions. I turned away completely from the formal sciences and endeavoured to understand questions of gnosis and march upon the Path of certainty. Whatever he ordered I obeyed without saying yes or no. If some of the things I heard or saw appeared on the surface to be opposed to the shari'a, I considered it a defect of my own ears and eyes and did not fail in any way to serve and obey him. In service, conversation, solitude and retreat I obeyed as completely as I could. I also obeyed all that he had ordered as necessary in the six kinds of invocation: the manifested (jali), the hidden (khafi),

'3 Mathnawt,vol. II, p. 162.

14A tomb of a saint near Tehran.

40

JOURNAL

OF

PERSIAN

STUDIES

the informal (hamd'ilf), the obscure (khumdli),that connected with the circle (halqa) and with the gathering (ijtimd'). I was also made to realize the four houses of death.... " Thanks be to God, through his spiritual will and the assistance of the saints I realized all the seven states of the heart and fulfilled in the way of actions and litanies whatever was required for each station. I performed the minor, middle and major 'forty days' (arba'fndt) [of spiritual retreat]. In the year 1309 (A.H.) I accompanied him to the city of faith, Qum, and there performed two consecutive major' forty days'. His Highness joined the Divine Mercy (died) there and I became very ill without my intimate friend and comforter. I passed days and nights in hardship at the corner of the mosque of Imam until my poor mother discovered my condition and sent someone to Qum who for a .Hasan while treated me. "After some improvement I returned to Tehran with that friend. Thank God through the spiritual will of that great gnostic I came to know of the details of spiritual poverty, gnosis, the subtleties of realized knowledge and certainty, and reached the station of annihilation in God (fand') and subsistence in Him (baqd'). I travelled the seven stations of the heart each with its special characteristics. With his esoteric aid and assistance from the intermediate world (barzakh) whatever order I received of commands or prohibitions, cleanliness, worship, asceticism, spiritual retreat, or self-purification I performed fully and did not fall short of serving God's creatures as far as I could.""' In speaking of the Sufi master in the Persian context one must also remember the role of the Twelfth Imam, who is the Hidden Imam, in both Shi'ism and Sufism as it exists in the Shi'ite world. In as much the Imam, although in concealment, is alive and is the spiritual axis of the world, he is the pole is to with whom all Sufi masters are inwardly connected. He is to Shi'ism what the supreme pole(Qu.tb) Sufism in its Sunni context. In Shi'ism the Imams, especially 'Ali, the first, and the Mahdi, the last, are the spiritual guides par excellence.The Hidden Imam, representing the whole chain of Imams, is the pole that attracts the hearts of the believers and it is to him that men turn for guidance. Moreover, the Imam also exists within the hearts of men. He is the inner guide who can lead man on his journey beyond the cosmos and also into the inner dimensions of his own being. If only one could reach this inner pole. That is why certain Shi'ite gnostics and Sufis have instructed the disciple to seek the "imdm of his being ". The possessor of the power of waldyaor initiation, by virtue of which the Imam in fact becomes the Imam, is the esoteric interpreter of things, of religion and of nature. And it is, in the Shi'ite view, his inward connection with the Sufi masters that enables them to gain the power of initiating and guiding men so that they can in fact reach the inner pole of their being. It can therefore be said that despite a difference of view in Shi'ite and Sunni Islam as to the identity of the Qu.tb, which Sunnism does not identify distinctly but Shi'ism considers to be the Hidden Imam or the function of the ImAim as such, both agree to the presence of this universal initiatic function (waldya) which exists within Islam as within every integral tradition. The Sufi masters are those unique individuals who through their connection with this golden chain of initiation are called upon by God to keep the presence of the spiritual Way alive on this earth and to guide upon this royal Path those who possess the necessary qualifications. They are thus the princes of the spiritual world. In their hands the desert blooms into a garden, base metal is turned into gold and the chaotic state of the soul is crystallized into a pattern of beauty reflecting the perfume of Unity (tauhid). " Make a journey out of self into self, O master, For by such a journey earth becomes a quarry of gold. From sourness and bitterness advance to sweetness, Even as from briny soil a thousand sorts of fruits spring up. From the sun, the pride of Tabriz, behold these miracles, For every tree gains beauty by the light of the sun."'6
'5

al-.hunafd' al-'urafa', Tehran, 1327 (A. H. Solar), pp. 232-34 from the Persian by S. H. Nasr).

'Abd al-Hujjat Balighi, Maqdldt

ft maqdmdtShams
(translated

i6 Nicholson, op. cit., p. 1II.

41

THE EVOLUTION

OF THE IQTA' IN MEDIEVAL IRAN, By Ann K. S. Lambton

ceased to be comThe title of my lecture may, perhaps, have puzzled some of you. The term iq.td' who came to power in the second half of the thirteenth the time of the monly used in Persian from Ilkh5ns, century A.D., and was superseded in its various aspects by the terms tuyal and soyfirghdl. The former of these, still part of the vocabulary of modern Persian, is sometimes translated as fief, as also is the term and the systems of government and societies to which they severally belonged are sometimes iqdtd', the parent or somewhat loosely described as feudal. My purpose this evening is to examine the iq.td', in order to show to what extent, if at all, medieval Persian society can be called forerunner of the tuyal, feudal in the narrow sense and Persian institutions, in a legal sense, feudal. I am not concerned with the all too frequent current misuse of the word feudalism, which has become a term of abuse, just as it did during the French revolution when it was adopted as a general description to cover the many abuses of the ancienregime. in its various aspects, military, administrative, economic, political, and social, was the The iq.td' dominant institution of the Great Saljfiq Empire (A.D. IO87-I 157). Iqtd's had been granted by rulers prior to the Great Saljfiqs but it was under them that the institution became, as it were, regularized. In to the beginning of the the terminology changed-down its essentials the system continued-though present century. Among the superficial resemblances between society and tenures in medieval Persia and the neighbouring lands of the 'Abbasid caliphate on the one hand and feudal society and tenures in Western Europe on the other are the following: the fragmentation of authority in the tenth century A.D. and again on the break-up of the empires which were subsequently established in Persia ; the grant of land in return for the performance of military service ; the grant of land to certain officials of the bureaucracy in lieu of salary ; the grant of immunities ; the supremacy of the military class ; the existence of a subject peasantry ; private armies ; and private courts. None of these in themselves constitute or confirm the existence of feudalism. As against these similarities there are certain obvious differences between conditions in the 'Abbasid empire and Western Europe in feudal times. Among them is the fact that the dominant military class from the ninth century A.D. onwards was composed to a great extent of slaves and freedmen ;2 and they were concentrated in towns and not dispersed in castles as was the case in Western Europe. In practice, however, the slave or freedman who rose to high administrative or military office differed very little from the freeman so far as his power and authority were concerned ; and the appointment of slaves and freedmen under the Great Saljfiqs to the office of atdbag meant that their status was second only to that of the sultan. Another difference is that commerce was relatively highly developed in the 'Abbdsid empire. These facts do not preclude the existence of feudalism ; but they may, conceivably, make it less likely. Feudalism is generally believed to have existed in pre-Islamic Persia, but the available evidence makes it, I think, clear that the iqtd' was not a continuation of Sasanian feudalism in Islamic dress. Under the Sasanians there was a strong antagonism between two mutually opposed tendencies, feudalism and bureaucratic centralization, the struggle between which dominated the social and political development of the Sasanian Empire until the reign of Khusraw Anfishiravan.3 By that time the great nobles had become a court nobility and had lost the character of a true feudal nobility. The lowest in rank among the inferior nobility were the dihqdns,village chiefs, whose power came from their hereditary possession of certain local administrative functions, the chief of
SThe substance of this lecture, given at the British Institute of Persian Studies, Tehran, on Saturday, November 6th 1965, has " already been published under the title of Reflectionson the " in ArabicandIslamicStudiesin honor HamiltonA. R. Gibb of Iq.td' (ed. George Makdisi), Leiden, 1965. The Editors thank the President and Fellows of Harvard College, the holders of the copyright, for their permission to publish Professor Lambton's article here. 2 This was true of post-'Abbasid Persia only at certain relatively limited periods. 3 See Christensen, L'Iran sous les Sassanides,Copenhagen, I966, p. 45.

42

JOURNAL

OF

PERSIAN

STUDIES

which was the collection of taxes. They were in fact not so much members of the feudal nobility as the link between the government and the population living on the land. When the Sasanian empire was conquered by the Arabs the nobility as such disappeared and their lands either became state lands or were left in the hands of their occupiers. The dihqdnson the other hand retained their importance and continued to act as a link between the government and the people. Their functions, however, were bureaucratic and not feudal. Similarly, when the 'Abbasid caliphs later began to imitate Sasanian practice in some respects, it was not the feudal tendencies of the Sasanian empire which they imitated but its bureaucratic centralization. Land after the early Arab conquests fell into two broad categories : (i) land which had not been abandoned by its holders on the conquest continued to be held by them as non-Muslims (though later they were often converted), and paid khardj(land tax), which in the early period carried the stigma of inferiority; and (ii) land which had been abandoned by its owners, former crown lands of the Byzantine and Sasanian empires, land which had belonged to the religious institution, and land which had no owner (i.e. waste lands, etc.). From the second category grants, known as qati'a, were made to Muslims. Such grants were subject to 'ushr (tithe), which was regarded as an alms tax payable only by Muslims. The Arab junds (citizen armies) on the morrow of the conquests were largely remunerated by tax-grants ; but since at this period virtually all Arabs were soldiers and the notion of possession was barely separated from property, such tax-grants tended to be regarded as qati'a, i.e. land grants also. In theory Muslims could not acquire khardj land ; but in practice it proved impossible to prevent the conquerors from obtaining large areas of land as private property. Land thus obtained became indistinguishable from land granted as qati'a. Both were alienable, hereditary, and subject to 'ushr. Such lands were seldom in the early period cultivated by the Arabs personally but rented in various forms to the peasants, the owner being responsible for the payment of 'ushr, but exercising no rights other than those of an owner ; he enjoyed no immunities, fiscal or otherwise, and had no was linked to the tribal structure of early judicial or administrative functions. In some cases the the early period was not for money, but for Muslim society. The primary need of the Arabs in qa.ti'a pasturage ; and a qati'a was accordingly sometimes given to the leader of a tribal group and exploited by the group collectively. In general, apart from the grants to tribal leaders and a few leading the area of a qati'a was usually relatively small, sometimes only enough to support one personages, family. So far as this was the case the qati'a grants differed from both the large emphyteutic grants of the Byzantine empire and the large estates which had formerly existed in the Sasanian empire. In the course of time grants also came to be known as iqtd'. Quddma, writing in the tenthqa.ti'a uses the term iqtd' to mean a hereditary grant subject to 'ushrwhereas he calls a non-hereditary century, grant tu'ma. Both were made from land other than kharajland, and were under the diwdn al-diya. Qudima also mentions grants called ighdr, i.e. land which paid a fixed sum to the treasury and was immune from the entry of the tax collector and taswij, i.e. annual, but renewable, tax-free grants. Ighdr and taswij were both made on khardj land.5 began to emerge in its medieval form. A brief examination of the By the tenth century A.D. the iq.d' contemporary social and economic institutions of the 'Abbasid empire will show, I think, that the new in the iqtd' did not arise, as had the feudal system in Western Europe, from a need for developments " but that the needs which gave rise to these developments were of another kind and the " protection institutions seized upon different from those seized upon in Western Europe by the need for protection. In the later Roman and early medieval society in Europe the prevailing need was for protection. The institutions which this need seized upon when it first began to turn away from the state were twofold ; one related to the person, the other to the land, the patrociniumand the precarium. In early Islamic society there was an institution not wholly dissimilar to the Roman patrocinium,namely the relationship of patron and client. This had two aspects, one Arabian in origin and social in its nature, and the other legal in origin and arising from slavery and influenced by Roman institutions. At the beginning of the first century of Islam a Muslim was, practically speaking, an Arab. On the expansion of Islam converted non-Arabs were affiliated to an Arab tribe and known as mawdli.6 The term carried
4 See further Cahen, L'evolution de l'iqta du IXe au XIIe sicle, in Annales,E.S.C., I953, pp. 27 ft.
6 Sing. mawld.

5 Ibid., p. 28.

THE

EVOLUTION

OF

THE

IQT'

IN

MEDIEVAL

IRAN

43

with it, in the eyes of the Arabs, a social inferiority, and for some two centuries the distinction between Arabs and mawdli was a live issue. By the tenth century A.D. the social differentiation between them had virtually disappeared and mawld was used only as a legal term to mean a freed slave. Clientship was, thus, in Islam a relationship between two parties arising from a need for social status. It did not carry primarily the implication of service and support ; and it never became associated with land as did the patrocinium. The precariumas used for protection in Western Europe gave rise to the commendation of lands (patrocinium fundorum). In early Islamic times an institution known as talji'a7 which resembled the also existed. By it the weak landowner committed his estate to the protection of patrociniumfundorum the strong to indemnify it. The man who ceded his estate by such a transaction lost his ownership in it, but in some cases he retained the right to sell it and to transmit it by inheritance. The motive for this transaction was usually to secure protection from either attack or extortionate demands by the government. It was found, inter alia, in frontier districts and in tribal areas, where security was, presumably, precarious. It was often members of the military classes holding adjacent estates who undertook this form of protection, and responsibility for the payment of the government tax. Conditions for a talji'a contract varied : in some cases the former owner agreed to pay a special due to the " patron " for protection.8 There are also instances of landowners transferring their property on some sort of life rent. The 'Abbasids when they came to power confiscated many Umayyad taldji; but various new were created during the 'Abbasid period. Under the later 'Abbasids references to taldji are raretaladji perhaps because the numbers of the small landowners had decreased. From time to time in the later periods, notably under the Qajars, occasional instances of a similar practice are recorded ; but on the whole this type of tenure was not widespread. It was a sporadic rather than a permanent feature of the land tenure system. This suggests that protection was not the dominant need in Islamic society ; and that the need or needs which gave rise to the system must be sought elsewhere. iq.td' The accession of the 'Abbaisids had been followed by a great economic revival " based on the exploitation of the resources of the empire through industry and trade, and the development of a vast network of trade relations both within the empire and with the world outside. . . . The Islamic town was transformed from a garrison city to a market and exchange, and in time to the centre of a flourishing and diversified urban culture."9 With this development of trade and its concentration in Baghdad and the large Persian cities, a change took place in the financial administration of the state. The old silver standard which had prevailed in the central and eastern part of the Sasanian empire was replaced The decisive step was taken in Baghdad between A.D. 874 and 915; though the by the gold standard.'o silver standard continued in existence beside the gold standard. The situation, because of the diversity of coins in circulation and their fluctuating relative values, was, thus, a complicated one. Taxes in various areas, notably Fars, Isfahan, and Ahwiz, which had formerly been paid in kind, were paid in cash ; in Syria and Mesopotamia they were reckoned partly in cash and partly in kind." The process of conversion, was performed by agents known asjahdbidha;,2 and a diwdnal-jahbadhawas set up, the first mention of which occurs in A.D. 928.13 The functions of the jahdbidha concerned the administration, remittance, and supply of funds. Prominent officials all had their own jahbadh, or agent, with whom they deposited funds for safe-keeping and administration. By the tenth century A.D. it was customary to pay debts by letters of credit (suftaja) as well as in cash. The use of suftaja in private commercial transactions and the financial administration of the state-particularly with regard to the remission of taxes from the provinces-greatly facilitated the mercantile relations and government business of the 'Abbasid empire. The money requirements of the caliph and the state had become by this time very considerable ; and a rapid supply of funds, specially for military purposes, was necessary. Various devices were adopted to obtain money ; tax-farming, the putting up of offices to auction, the sale of
7 Plural taldji. 8 Lokkegaard, Islamic Taxation in the classicperiod, Copenhagen, 1950, pp. 67 ff. 9 Encyclopedia Islam (new ed.), Art. on 'Abbasids. of and political life of Medieval 'o Fischel, The Jews in the economic Islam, p. 3 ; B. Spuler, Iran infriih-islamischer Zeit, Wiesbaden, 1952, Pp. 408 ff. " des Mez, Die Rennaissance Islams, p. 104. 12 Sing.,jahbadh. '3 Fischel, pp. 4-5.

44

JOURNAL

OF

PERSIAN

STUDIES

crown lands, and the confiscation of private fortunes. Nevertheless the budget was often unbalanced ; and in spite of the economic revival there was a progressive deterioration in the financial stability of the state from the ninth century A.D. onwards. This was due to a number of causes, inter alia the civil war after the death in A.D. 809 of Harfin al-Rashid, internal revolts, foreign wars, the growing luxury of the court, over-taxation, the extortion of officials, and the consequent decline in agricultural production, notably in 'Iriq. Considerable changes were also meanwhile taking place in the social structure of the caliphate. The most striking was in the composition of the military forces of the caliphate and the increase in the numbers and power of the Turkish military slave troops, noticeable especially from the reign of al-Mutawakkil (A.D. 847-61) onwards and which culminated in the emergence of the amir al-umardin A.D. 935. Against this background of change in the economic and social environment (though not in the nature and purpose of government) the iqtd' began to emerge as one of the main institutions of the state. The dominant need of the state at this time was for money to finance its operations ; and the crucial problem the payment of its civil and military officers. Hence, it was the financial and administrative institutions of the state that this need seized upon and transformed ; from this stems a fundamental difference between the feudal institutions of Western Europe and the iqtd' system, namely that whereas a contractual relationship was an essential characteristic of the former, the element of contract never became a feature of the system. In Islam all authority is delegated authority and excludes, iq.d' the question of contract. Al-Mdwardi, it is true, claims that the imamate was a contract, therefore, The authority of but this claim had ceased to have any validity in practice by the tenth century A.D. the governor, the tax-collector, and the judge, was in all cases delegated. If, therefore, it was the financial and administrative institutions of the state which were the real parents of the iqtd' it is not surprising that the element of contract never became a feature of the new system. The institutions which were seized upon by the need to finance the state and transformed by the were the damdnand qabdlaon the one hand and the provincial governorate new development of the The damdnwas a tax-farm, which often involved the putting up to auction or amirate on the other. iq.d' of the right to collect the taxes of a given area, usually annually. Under the early caliphs the taxes of certain districts were handed over to certain Arab towns or individuals. Thus, Nihavand and later also Hamadan were handed over to Basra by Mu'awiya ; and Dinavar to Kfifa. Out of this practice grew the tax-farm proper. The whole of Egypt under Manstir (A.D. 754-75) was offered to a taxfarmer; and in the time of Harfin al-Rashid (A.D. 786-809) tax-farming was common ; and by Sometimes whole provinces were involved and the tax-farmers, as witness A.D. 900 it was widespread. the case of the Baridis in Khfizistan, often exceedingly powerful. The qabdlawas not a tax-farm proper, but the undertaking by a local notable to pay the tax quota of the local community, which would be i.e. assessed by the method known as muqd.ta'a, assessment at a lump sum, payable according to the lunar contradistinction to assessment by measurement (misdha) and assessment by a share of the year,'4 in and payable according to the solar year. The muqdta'amethod was adopted in many crop (muqdsama) areas for the sake of the convenience of the government. The Sunni jurist, Abfi Yfisuf, recognized it, but disapproved of it and recommended control of the activity of the man who accepted the qabdla by a military leader (amir) or agent acting on behalf of the imam, whose expenses were to be defrayed continued to be common in Persia until by the state treasury.,' The assessment of taxes by muqd.ta'a the twentieth century. The second institution seized upon was the provincial governorate. This under the 'Abbasids had been through various vicissitudes. The keystone of the administrative system had been the separation of military and financial affairs, with the amir as the head of the military administration and the 'dmil of the financial. In the provinces there was a network of officials, the provincial government being virtually a replica of the central administration. The existence of a large body of officials placed a heavy burden on the finances of the province, upon which they were a charge. This burden was further aggravated by the fact that the officials of the bureaucracy were not, even in theory, the servants of the public ; they were the servants of the caliph (and later of the sultan), partaking according to
14

Cahen, p. 29.

5~ Lokkegaard, pp. 96 ff.

THE

EVOLUTION

OF

THE

IQTA'

IN

MEDIEVAL

IRAN

45

their degree, of his absolute authority. Appointment to an official position was broadly regarded as a since all, or virtually all, officials passport to wealth-hence the establishment of the diwdn al-musddira, were expected to get rich by illegal means. The burden imposed by the official classes was still further added to by the existence of factions : each important official had his own supporters who rose (and fell) with him. The strain imposed by the bureaucratic administration eventually became too great and the system broke down. In the first place, the difficulty of paying the officials of the bureaucracy and the army led to a wholesale extension of the farming of taxes, thus setting up a vicious circle. The situation was, moreover, complicated by another tendency : already by the death of Ma'minin n A.D. 833 the balance between the civil and military arms of the administration, whether at the level of the central government or the provincial, had been upset, and the money received from the farming of the taxes soon ceased to be sufficient to pay the army leaders and their troops. As the revenue came in more and more irregularly, so the practice grew of assigning the taxes, not to the tax-farmers, but to the military themselves. Once a military leader was assigned the right to collect the taxes of a large area, it became relatively easy for him to establish his semi-independence. Morethe matter did not stop with the assignment of the taxes. When the taxes still failed to come in over, regularly, rights over the land itself were then assigned: in other words the tax-farm swallowed up the land revenue system and itself, in turn, became assimilated to the provincial governorate, which had become militarized. The result of these various developments was the ruin of the land on the one hand and the failure of the military to support or defend the central government on the other. This militarization of the state, which is marked not only in 'Iraq and the neighbourhood but also in the east under the Samanids and more especially under the Ghaznavids, and the growing tendency of the military to be occupied, not only with the arts of war, but also with administration, have obscured the true nature of the system, which in its origins was bureaucratic and not feudal. iq.td' to the The assignments made which military were called iqtd's but they differed from the old of land were made probably mainly from khardjland (but the distinction between these two types iq.td's, had by this time become blurred). They were, therefore, an extension of the ighdrrather than of the original calling the latter iqtd' iqtd'. The jurists, in fact, distinguished between the new and the old iq.td', al-tamlikand the former al-tamlik al-istighldl. The difference in theory between them was that the iq.td' iq.td' was a grant of ownership, its purpose the extension of cultivation, and the al-istighldl a grant of the iq.td' of usufruct, its purpose remuneration for services. Under the Bfiyids the military, to whom the the new type were given, did not normally live on their grants but merely sent their agents to iq.td's collect their revenues. If the historians of the period are to be believed, they exploited their holdings to the maximum and ruined the countryside. In theory the iqtd' al-istighldlwas not hereditary, or even granted died his children did not receive his as a life-rent ; there were periodic redistributions and if a muq.ta' but pensions. The soldier in return for his iqtd' had to perform military service and was, in theory, iqtd' subject to detailed regulations and inspection. The amirs, who also received iqtd's, had no responsibility for the payment of the soldiery, who themselves received pay (rizq) or an iqtd' from the state. There was, thus, no sub-infeudation as in Western Europe. A provincial governor could distribute but he did this as an official of the state and not because the area the area under his jurisdiction as iq.td's, did not give the beneficiary any formed part of his private domain. Legally, possession of an iq.td' rights over the inhabitants. But, in practice, it clearly contributed to the spread of patronage juridical and was, in the Bfiyid period, accompanied by widespread acts of usurpation by the military. Moreover, the tendency for the function of the provincial military commander, the tax-collector, the taxto farmer, and the muqpta' be combined in one person led to the emergence of large properties, virtually of the central government. The tendency for the defacto powers of the muqita' increase to independent was also strengthened by the fact that governors frequently received in their provinces some areas by way with their of ; in such cases they combined in these districts their economic power as muq.ta' iq.td' functions as governor ; and tended to exercise the powers of both throughout the province.'6 Under who had in addition to their obligations as the Bfiyids and in Fatimid Syria there are cases of or the administrative duties and obligations ofmuq.ta's governors also. In such cases the wdlis muq.ta's muq.ta,
I6 Cahen, pp. 35-6. 4

46

JOURNAL

OF

PERSIAN

STUDIES

carried out all governmental functions ; all the towns and fortresses in the province would be under his jurisdiction ; all the costs of the provincial government would be defrayed locally ; no taxes would be paid to the central government, but in time of war a military contingent would be furnished ; and the ruler's name would be mentioned in the Friday prayers and on the coinage. Further, the muq.ta' was free to constitute the area under him into iqtd's. Such a muqta'or wdli was, in the view of Cahen, in many respects the brother of a European feudatory.'7 But the great difference between the two, or government and what would seem to be a crucial difference, was that the muqta' held his iq.td' at the will of the sovereign and as a matter of grace. This type of iqtd', which was the exception solely under the B-iyids, became widespread in the Saljfiq period. under the Bfiyids was the military iqtd'. These iqtd's were controlled The dominant type of the military diwan, theiq.d' diwdn al-jaysh, at the head of which was the 'drid or muster master. The by military diwan was thus concerned not only, or even primarily, with military administration, but rather with the fiscal value ('ibra) and characteristics of each iqtd' and the reallocation of iqtd's as they fell vacant.,' This intimate connection between the assessment of taxes and the levy of troops continued in Persia down to the twentieth century, but did not of itself constitute feudalism. When the central government was strong control over the allocation of iqtd'swas close. The military iqtd' did not in itself involve decentralization or a relaxation of the authority of the central government. to the governorate Under the Great Saljfiqs, there took place an assimilation of the military iq.td' tended or " administrative " iqtd'. With this the careful estimate of the exact fiscal value of the iq.td' to be replaced by merely an approximate value ; and the iqtd' to be defined not by fiscal to become, by usurpation, a hereditary domain over value but by service ; and for the had governmental iq.d' which the prerogatives.'9 The situation, however, varied in different muq.ta' of the empire. Broadly speaking, five main types of iqtd' can be distinguished ; but the parts in the same category conform distinction between them was not absolute and nor did all iq.td's to the same pattern. Moreover, the situation was further confused by the fact that the same exactly First there was the grant of an iqtd' by the sultan persons sometimes held more than one type of of his family. These grants wereiq.d'. to a member probably not intended originally to be hereditary, but for certain branches of the family to regard certain districts as its own and the tendency developed iq.d', of Kermdn and the Salj faqsof Syria, to set up virtually independent sometimes, as in the case of the Salj fiqs dynasties. Originally the grants to members of the Saljfiq family were perhaps not unlike the qati'a grants to Arab tribal groups in the early centuries of Islam. The dominant need of the Saljfiqs in the early period of their expansion was also for pasture for their flocks. Moreover, the authority of the Saljiaq sultan in the early period especially was essentially personal and the kingdom was probably not regarded in territorial terms but rather as being co-extensive with the area over which the flocks of the Saljfiqs and their followers roamed. This phase was rapidly succeeded by the " settled " phase ; in which the sultan became the ruler of a " territorial " empire. Alp Arslan, the second Saljfiq ruler (A.D. IO63-72) assigned parts of the kingdom to various members of his family. By the end of the " " period the iqtd's granted to Saljfiq princes were barely distinguishable from administrative iq.td's and did not or provincial governorates. These grants were in all cases simply delegations of authority held the area at the will contain any implication of vassalage or permanent rights. The Saljaiq prince of the sultan, who could, and frequently did, revoke the assignment. The usual method by which a prince was dispossessed was to assign the area to a third person who would then take possession of it by force. The terms of these grants varied. In most cases complete financial control was handed over to the prince, in the sense that all revenue was collected by him ; he was usually instructed, however, not to increase taxation and was, in any case, to some extent limited in this by local custom and Islamic theory. The degree to which he could sub-assign the area under him varied. In some cases the right to was accorded ; in others not. The Saljfiq sultans also assigned iq~td's their wives and other Salj fq were in the nature of pensions or personal estates and fall into a different women ; but such iq.t's category or form a sub-division of this category. The practice under which the Qjijir rulers in the
17 Ibid., p. 36. i8 Cf. Ibid., pp. 36-7.
'9 Cf. Ibid., p. 43.

THE

EVOLUTION

OF

THE

IQTA'

IN

MEDIEVAL

IRAN

47

nineteenth century appointed their sons and relatives to provincial governments was in some respects not unlike the grant of by the Saljfiq sultan to members of the Saljfiq family. Neither were in a feudal or technical senseiq.d's grants. legal was the " administrative " which was in effect a provincial governThe second type of iq.d' of the iq.t', this type of grant was normally called an iqtd', but in ment. In the western part Saljfiq empire the eastern provinces it was known by the traditional term for a provincial government, namely wildya, 'amal, riydsa,or niydba,the term iqtd' being used for the military iqtd' which continued to exist as a separate category in spite of the tendency for it to be assimilated to the " administrative " iqtd'. In the " administrative " the holder, known as the or was granted complete control ; and iq.td' muq.ta' wdli, was able to make sub-assignments. This is especially noticeable in the case of who attained to muq.ta's No effective system was devised for maintaining control over the virtual independence. In muq.ta'. some cases there appears to have been an attempt to exert this by " reserving " certain taxes, retaining or appointing a deputy (nd'ib) to reside in the certain " crown " lands within the area of the on iq.td', iq.td' behalf of the sultan. Under the later sultans the power of the amirs relative to that of the sultans grew ; and with this the grant of an iqtd' tended to become merely official recognition of the possession of a given district by an amir. Simultaneous assignment of the same iqtd' to two amirs, who were then forced to fight it out, became increasingly common. With the increase of the power of the amirs a " " also ; but this was by usurpation hereditary tendency began to appear in the administrative iq.d' and not by right. who held an " administrative " iqtd' were not The relations between the sultan and wdli or muq.ta' those of lord and vassal. The source of all grants of authority or land was the absolute sovereignity of the sultan ; and the grant a matter of grace. Moreover, the fact that the amirs, who were the dominant group among the assignees, were for the most part slaves or freedmen, militated against the emergence of a feudal society in which there was a union of vassalage and benefice ; and in theory, a freely negotiated contract between equals. By Islamic law on the death of a slave his possessions escheated to his master. In practice it often happened that on the death of a royal slave the sultan would grant his possessions to one of his descendants. Many amirs and atdbagssucceeded in establishing hereditary dynasties, but they did this by usurpation. The holder of a provincial government or an " " or military, owed the obligation of obedience and service to the sultan, but iqtd', administrative there was no obligation of protection or maintenance upon the sultan. There was, in fact, no contract involving mutual fealty. The grant was arbitrary, subject to regrant at irregular intervals, and to revocation without cause. There is in some of the Saljfiq documents an injunction to the assignee to " consult " ; but this, I think, derived from the practice of the steppe and was not part of the iq.td' system. Under the Safavids there appears to have been a council of amirs (umard-yijanki). Details of its procedure and activities are scanty ; but in the absence of evidence there seems no reason to " regard it as a feudal court ". Under the Qpjdrs the jdnki was not a regular part of the organization of the central government, but a tribal council dealing with affairs concerning the Qjarx tribe, which, appears to have sat, not under the shah, but under the head of the tribe, the ilkhdni. It seems likely that in Persia, at least, the fact of the amirs in medieval times often being slaves or freedmen had very little influence on their status. There are two trends running through society, mutually opposed, but in fact seldom coming into conflict namely the " equalitarian " theory of Islam reinforced by the general levelling tendency of Turkish military government on the one hand, and on the other the " social ", but not functional pre-eminence of "old" families (the buyitdt). There was, however, nothing in the nature of a hereditary aristocracy. Repeated conquest and the inheritance laws of Islam effectively prevented the emergence of such a class. The sayyids occupied a special position, deriving from hereditary descent, but this was not combined, except by coincidence, with landownership. The only other class, which had, perhaps, some claim to be regarded as a hereditary aristocracy was that of the tribal leaders. So far as they constituted such a class, their authority was essentially personal, and derived from their possession of flocks and tribal followers, not from the possession of land (though tribal leaders often acquired land). Further, the vicissitudes of conquest, invasion, and economic pressure resulted in a frequent dispersal of tribal groups, and the disappearance of chiefly families.

48

JOURNAL

OF

PERSIAN

STUDIES

The third type of iqtd' was the military which was a continuation of Bfiyid practice. This iq.d', " " type, as I mentioned earlier, became assimilated to the administrative iqtd' ; and it is not always was a grant on the revenue and the easy to distinguish between the two. In theory the military iq.td' beneficiary did not live on the land on which the draft was made ; but in practice he very often did. A hereditary tendency appeared in the military iqtd' also. In addition to the military assignments to individuals there were a number of areas throughout the empire in the time of Malikshih (A.D. 1072-92) constituted into iqtd's for the standing army so that when passing through such an area on campaigns, fodder and whatever was necessary for its immediate expenses was ready. This too was an " administrative " device to ensure that commissariat difficulties would not impede the progress of the army when it was marching through the countryside, and not a feudal arrangement. " " Fourthly, there were iqtd's granted to officials of the civil government. Such grants resemble, perhaps, the serjeanty of Western feudalism. These iqtd's were either grants on the revenue or grants We have no of land. The wazir apparently received one-tenth of the produce of the soil as an was alienated from details of what was implied by this. In the first place much of the empire iq.td'. precise the control of the central government in the form of " administrative " iqtd's and therefore the sum involved may not have been very large. Secondly as the alter ego of the sultan, the wazir was probably expected, from the proceeds of such a levy, to carry out some of the traditional duties of the sultan, such as the provision of pensions for the religious classes, the upkeep of hospices, bridges, roads, and so on. The leading officials of the Safavid empire (A.D. 1502-1736) also received a percentage levied on various grants, taxes, and funds.2o In addition to grants of revenue, the Saljfiq wazirs also held " administrative " and military assignments of land. These differed from the iqtd's in that there was no explicit or implicit obligation to furnish the sultan with troops ; but the distinction was obscured " by the fact that the maintenance of" private armies was the rule for all prominent persons with the of the religious classes. Various members of the religious classes also held iqtd's ; those held exception by qddis or others of the religious hierarchy may be regarded as being in lieu of salary ; others were rather pensions, and belonged to the fifth category of iqtd', namely the " personal " granted to there was iq.td' another type of private persons and others as a gift on a life, or hereditary, tenure. Lastly which was a personal estate of a rather special nature, namely the iqtd' granted to the caliph to iq.d' the meet expenses of his position. The Great Saljfiq period, broadly speaking, marks not the dissolution of the fiscal administration the details of the local arrangements for assessment and collection of of the central government-in taxes there is a striking continuity-but its militarization and a retrenchment in the area of its direct operation. This retrenchment was, however, accompanied by the emergence of what tended to become a " hereditary " domain, over which the muqta'had governmental prerogatives, which included the collection of taxes, the holding of the court, and general supervision of security and religious a technical sense. If a comparison is to be sought, it is to be affairs. This system was not feudalism in mazdlim found with the Byzantine pronoia rather than the Western European fief. In the early stages the system did not coincide with weakness in the control of the central government; adoption of the a change in the military forces of the state : the basis of the Great Saljfiq state when but rather with iq.td' it was first established was the Turkoman tribes ; they proved an unreliable source of power and eventually the military forces of the state came to be largely composed of slaves and freedmen. With this change in the basis of the state the problem of the payment of the military forces and the financing was a device designed to solve this of the administration in general received a new urgency. The iq.[' problem. It was only partially and temporarily successful. In the long run it contributed to a decline in the relative power of the sultan vis-e-vis the amirs and eventually to the disintegration of the Tribal forces also formed the A not dissimilar course of events occurred in Safavid times.' basis of Safavid power ; and were eventually replaced by an army composed largely of slaves original and freedmen, paid by assignments of the revenue and of land. A new terminology to describe the various tenures had meanwhile developed, but there is unfortunately a certain lack of precision in its
20

Saljfiqempire.

See Tadhkiratal-Mulik, Persian text in Facsimile, translated and explained by V. Minorsky (G.M.S., I943), f. 85a ff.

21

The question of space forbids a discussion of Ilkhan practice and the reforms of Ghdzdn Khan.

THE

EVOLUTION

OF

THE

IQTA'

IN

MEDIEVAL

IRAN

49

use. Broadly the term tuyal covered grants of lands, which were not in theory hereditary, and made, for the most part, in outlying areas, amounting in effect to provincial governments. The holder was entrusted with the administration of the area and provided military contingents when called upon and presents (pishkash) to the shah at the New Year. This was similar to the Saljfiq " administrative" iqtd'. Secondly the term tuyal was used to designate land assigned in lieu of salary ; some areas, presumably originally crown lands, were permanently attached to certain offices. Thirdly it covered the grant of immunities in private property; and fourthly the grant ofcrown lands to the standing army. The latter were hereditary and passed to one of the male heirs of the holder unless these refused to bear arms. Originally these were merely grants on the revenue, made, in some cases, jointly to a group ; the proceeds were collected, not by the soldiers personally, but by a tahsi4lddr. The term soyirghdlwas used mainly to designate hereditary or life grants of crown land, waqf land and usufructuary property ; they often involved some sort of grant of immunity." As the central government weakened in the later Safavid period the general tendency was for both to tuyal and soytrghadl be usurped and converted into private property over which the officials of the state exercised little or no control. But there was not, as there had been at the end of the Great Saljiiq period, the same tendency for the empire to break up into semi-independent kingdoms but rather for frontier areas to be taken over by neighbouring powers. In the nineteenth century under the Qfjars, with the tendency towards a centralization of the administration in this sense that the revenue demand was drawn up and allocated to different provinces and districts by the central government, the term tuyal ceased to be applied to provincial governments. The provincial governor, however, was frequently assigned a tuyal in lieu of salary. Moreover in the second half of the century with the increasing difficulty which the central government experienced in financing the administration the circle came full round: many government offices and provincial governments were in effect farmed and the position of the provincial governor, whether he farmed the or tuyilddr of provincial revenue or not, was in practice not very different from that of the muq.ta'/wdli earlier times. He had virtually a free hand in the collection of the quota due from him and of what was required for the expenses of the local administration. Any surplus remaining after the payment of the central government's quota was retained by him. He usually exercised rights of jurisdiction within the province ; and if called upon had to provide military contingents. In the frontier districts and tribal areas tuyals continued to be granted in return for the provision of military contingents. From the middle of the century onwards the tendency was to assess villages in terms of taxes and military contingents (though some areas were exempted from the provision of the latter), and to assign them as tuyal. Cause and effect were thus reversed : originally the military iq.td' had been given for service and not vice versa. Under the later Qajars service was thus, in some cases at least, given for the tuyal. Such grants usually carried with them certain immunities. In the course of time it not infrequently happened that the liability to provide contingents was neglected and the grant was converted by usurpation into private property in which the holder continued to enjoy immunities and exemptions, and in which he tended to exercise virtually arbitrary power. There are in Qajir times cases of tuyil or grants of land being assigned to certain offices. These also tended to become assimilated to private property. Lastly, grants known sometimes as tuyal and continued to be made on crown land, waqf land, and usufructuary property. sometimes as soyfirghdl in its various forms on the one hand There is, I would suggest, a striking continuity between the iq.td' and the tuyil and soyirghdl on the other. Both were in the main bureaucratic devices for financing the administration-or rather for running the administration in the absence of a proper provision for financing it-and for providing for the upkeep of the army. Under the Great Salj fiqs there was a general tendency for the military leader to become an administrator and to exercise jurisdiction. This was also true of the subsequent periods of Persian history ; though by the second half of the nineteenth century it would perhaps be more accurate to say that the tendency was rather for the administrator also to become a military leader. There was throughout a connection between military functions and the possession of land. But this was not based on a feudal contract involving mutual fealty between
22 There were various other grants made on the revenue and grants of immunities, known as muqarrari, mu'aff,and hamasdla.

50

JOURNAL

OF

PERSIAN

STUDIES

sovereign and vassal. There was nothing in the relationship between the muqtd' and the sultan or tuyalddrand the shah like the " foi " which bound together lord and vassal in Western Empire and to which the high importance attached in Western Europe to fidelity in part goes back. Nor did the system give rise, as feudalism did in Western Europe, to a belief in the binding force of engagements iq.td' entered into or to the idea that one is not bound to obey an order incompatible with one's dignity freely as a free man. The source of all grants of iqtd' or tuyil and soyarghdl was the absolute sovereignty of the sultan or the shah. They were entirely matters of grace and lacked any element of vassalage or contract ; and the fact that, when the central government relaxed control over the system in iq.d' Great Saljfiq times and over tuyals and soyarghdlsin Safavid and Qajdr times, those who usurped power exercised this in an arbitrary fashion was due, not to the fact that these systems were feudal systems, but to the nature of the conception of power which prevailed in society.

51

NOTES ON THE BALUCHI SPOKEN IN PERSIAN BALUCHISTAN By Brian Spooner

I i. The material which follows has been gathered during some twelve months anthropological field work (spreadover the last three years) in the Persian Province of Sistan and Baluchistan, excluding the sahristdn Zabol or Sistan (hereafterreferredto simply as the Province). The language of the Adminisof tration is of course Persian, but outside the major settlementsfew people are bilingual, and probably as much Urdu as Persian is known-which is very little. Baluchi is the lingua franca of all natives of the Province (except a few Persian and Sikh merchants in the main towns) both in the settlements and the nomad encampments. Another language, Jadgili-said to be derived from Sindhi, is spoken by a few thousand people-the Jadgil tribes-in Dashtiari in the extreme southeast of the Province, but these also use Baluchi as a lingua franca, and all the men are bilingual in these two languages. Brahui is not spoken in the Province (as defined above, though I am told it is spoken by nomadic communitieswithin the area of the .ahristdn Zabol) but its influence is evident in the vocabulary of all the Baluchi dialects. of This influence has been noted on the other side of the border in Pakistanin particularby Morgenstierne and b), and in a recent published comment on the relationship between the two languages (esp. 1I932a it has been suggested that future research " will bring to the attention of linguistic scholars something unprecedented (I believe) in linguistic reporting, an instance of symbiosis over a large area of two genetically unrelated languages, between which there has taken place not only structural borrowing two languages) but even a structural borrowing which shows (demonstrated previously for these " parallel divergenceson the two sides (Emeneau, 1964: 75-76). Morgenstierne (I 932a: 9 ff.) noted the scale of interchange of vocabulary between the two languages, and considered it likely that some communities (in Pakistani Baluchistan) had changed from one language to the other more than once. Furthermore,he did not consider that either of the two languages possessedany cultural predominance as compared with the other. 2. This is not the place to attempt to account for the modern line of division between Persian and Pakistani Baluchistan. It is a complicated story and we do not possess sufficient historical evidence to account for it satisfactorily.However, two things are relatively certain: i. the main factors were political-in the form of alliances and enmities between Baluch leaders. (I am preparing a monograph on the subject of Baluch political activities and concepts.) ii. the process may be said to have started with the " legitimization " of the Khanate of Kalat by Nadir Shah in 1739, which was later confirmed by the British in India. The British may be said to have fossilized the western extent of the influence of the Khan of Kalat and treated it as their far northwesternborder, although the border was not actually defined until much later (187I, cf. Aitchison, C. U. (Comp.) A Collection Treaties,Engagements Sanads. . . Calcutta, and of Vol. XIII: 14) when renewed Persian interest in the area necessitated a settlement. The 1933, Khanate of Kalit was the only paramount native Baluch political power which can be said
ever to have attained any degree of legitimacy, and certainly the only one to have had that legitimacy recognized formally by a foreign power.

3. The relevance of this to a discussion of language is that the institution of the Khanate of Kalat gave rise in time to the beginning of a consciousness of nationhood and unity even among those Baluch who did not own allegiance to it, which in turn was the basis for the first attempts to write the Baluchi language and therefore a gradual tendency towards standardization of the language. But all this was confined to British (now Pakistani) Baluchistan. There is now an official monthly magazine published in Quetta in Baluchi, and other occasional publications appear from time to time in Quetta and Karachi. The standard language which is slowly being formed approximates to what is generally thought of as the Makrani dialect and contains a large number of new loan words from Urdu. Nothing

52

JOURNAL

OF

PERSIAN

STUDIES

comparable has happened in the Persian Province, and although there is a growing awareness there now of what is happening over the border very few Persian Baluch, though literate in Persian, when confronted with the Pakistani publications are able to read their own language-even though it is in the same alphabet (with a few minor modifications). 4. The Baluchi spoken in Persian Baluchistan then is an unwritten, unstandardized and unrationalized language. There is no conception among its speakers of what is correct and what is incorrect speech. Rather they think, obviously, in terms of what is intelligible, or unintelligible, or funny because it is an unusual mode of expression, or marks the speaker out as coming from a particular area or class. There is tremendous pride in poetry, and particularly in the traditional epic poetry (cf. Dames, 1907). This may be considered something of a standardizing influence, but as is often the case-even though poetry is very much a living art-it tends to be more conservative than everyday speech and even to cultivate archaisms. The meaning of some of these archaisms has been forgotten, with the result that a transcription and study of the old epic poetry still recited in the Province today would require expert " textual " criticism. Therefore, while it is definitely a linguistically unifying factor poetry by no means controls the development and diversification of ordinary speech. 5. A corollary of this is that although the Baluch take great pride in their language and in the art of expression in prose and poetry, they tend not to be conscious of lexical or even structural borrowings -which in the present context means mainly Persianization. For these reasons it is often not feasible in a must remain basically a description of recorded speech-either to description of the language-which from or account for all elements. We must be content rather in some cases simply to say " I generalize heard this said ", and indicate similar modes of expression and contexts in the same and (sociologically) related languages. 6. The bibliography given at the end of this paper is designed to include all the significant or useful publications relevant to the study and analysis of the Baluchi spoken in the Persian Province. It will be noticed that none of the items listed refers specifically to this form of Baluchi (with the minor exceptions of items 15 and I16).In fact, there is (so far as I have been able to trace) no significant published information based on first-hand information about it. I have taken this in itself as a sufficient justification for publishing the material which follows. 7. This article therefore is designed to be complementary to already existing publications on what are generally known as the Western or Southern dialects of Baluchi-in particular, the Makrani, Panjguri, Marwi dialects. Items 14, 19, 26, 40 of the bibliography are useful as general introductory descriptive grammars. Short comments are given to indicate the relevance and usefulness of each individual item in the bibliography. In general it must be remembered that each of these grammars attempts to reduce the language to general grammatical principles or rules. Although it is of course the job of the philologist to attempt this type of reduction, as we have already pointed out it may be misleading so far as the present usage of particular forms of the language are concerned. In such a language, speech may be divided into two categories: (a) Sentences, phrases and idioms which have become stereotyped-the natural way to express certain ideas, and (b) new modes of expressing new ideas. The former may contain modes of expression which compared to the latter are archaisms. The word list in item 26 is particularly useful. After these works I found item 2 the most useful publication in my own efforts to acquire facility in the use of the language. 8. The main section of this article is in the form of an annotated word list, but before proceeding to this a few remarks concerning general grammar and dialect differences may be useful. My aim is to present those items and aspects of the linguistic idiom of the Province which differ from already published information on the language in general. Purely philological comments are reduced to a minimum, since I write as a social anthropologist who learnt to use the language as a tool for research I have no formal training in philology or linguistics. Nevertheless I hope the article will also show that the social anthropologist may make some contribution to the study of languages-if only because of the opportunities (and the need) he has for acquiring familiarity with them, and because basically his job

NOTES

ON

THE

BALUCHI

SPOKEN

IN

PERSIAN

BALUCHISTAN

53

is comparable to that of an interpreter anyway. For this reason and since it is natural that dialect variation should first affect the colloquial core of a language, I should like to think that the data given in this article, without constituting a general purpose phrase book, also represents a cross-section or " characterization " of everyday linguistic expression through the medium of Baluchi in the Province. the article, references are made to other works only where the material seems to be compleThroughout mentary, corroborative or contradictory to them. Basically, then, I list and comment briefly on grammatical and vocabulary usage which I have found which is not published elsewhere, or which differs from or may serve to illuminate obscurities or insufficiencies in published sources.

II
I. The vocabularies used to express time, distance, and kinship and marriage relations and the analysis of the basic conceptions which these vocabularies represent are reserved for future publication. The repertoire of personal names in use and the local toponymy will also be discussed at greater length elsewhere, but a few general remarks may be relevant here. The Baluch are predominantly Hanafi muslims, and naturally a large proportion of their personal nomenclature is derived from those Arabic names which figure in the history of the first century of Islam. However, they also include a few traditional Baluch and a number of" heroic " Persian names, and in general the nomenclature in use shows great variety. They are particularly fond of compound names. Thus, from Muhammad which they pronounce Mahmad they form All Muhammad, Din M., Dorr M., Dust M., Feiz M., Golam M., Gol M., Hatd (A. 'Ata) M., Jan M., Heir M., LMl (A. La'l) M., Nik M., Nizar M., Nur M., Pakir (A. Faqir) M., Seih M., Sir M., Ta-j M., Wall M., M. Akbar, M. Ali, M. Amin, M. Hasan, M. Karim, M. Morad, M. Omar. Several of these compounds are (to my knowledge) either non-existent or rare in the rest of Persia, but are found among other tribal communities in Afghanistan and West Pakistan. Cf. also Imim-bahv, Karim-bahv, Lil-bahs, Nabi-bah', Pir-bahk, Rahim-bahs, Taj-bahv; Allah-did, Kdder-did, Halek-did, Mauli-did, Sah-dad; Dad-allah, Did-vh, Ddd-rahman; Jom'a, Samba, Dosamba, Cdr-'amba, Panj-vamba, but never yak-vamba, or sei-lamba, though it is said that these also are used in Bashikerd.
2.

Place names may be divided into three types: (a) Old names of presumably pre-Islamic origin, most of which feature in one or other of the early geographies (e.g. Iudtid al-'Alam, trans. etc. V. Minorsky, E. J. Gibb Memorial New Series XI, 1937: 123, 373). These were there before the Baluch entered the Province and have remained unchanged to the present day. They are almost exclusively names of agricultural centres. (One in fact called in Baluchi " Wash(t) ".) Cf. also Kant, Naskant, Kaserof these-Khwdysh-is kant (P. Qasr Qand) and Misc.: 289. (b) Names of Baluchi origin-or at least Baluchi in form. These are predominantly names of rivers, streams, rocks, mountains (but not the two highest peaks in the Province-Kuh-i Taftin or Ceheltan and Kuh-i Zenda or Bazmdn-which must have provided unforgettable landmarks and perhaps watering places, from very early times on important routes which passed to the north and south (?) of them) and small areas. That is, the names of Baluchi origin or form suggest the toponymy of a pastoral, nomadic people. (c) New Persian names, mostly ending in "-abad ", which belong almost entirely to small, new agricultural settlements started in the present century, or at least since the renewed Persian interest in the Province in the middle of the last century.

3. A few general points regarding vocabulary are interesting: The usage of loan words of Arabic and Persian origin, though they are generally thought to have come to Baluchi through Persian, often does not coincide with standard Persian usage. For instance, P. taqsimkardanis in Baluchi bdharkanag, and for P. 'avaz kardanBaluchi uses badal kanag. In several cases Persian Baluchi shares similar usage of A./P. loan words or phrases with Afghan Persian, Urdu, and/or Persian dialects spoken in Eastern Persia, which are not found in standard or other forms of Persian. Some of these similarities are indicated in the word list.

54

JOURNAL

OF

PERSIAN

STUDIES

4. It is practically certain (though difficult to prove to any great extent-a problem which will be discussed in a future publication) that the majority of speakers of Persian Baluchi are not of Baluch ethnic origin (whatever that may be) and did not speak Baluchi when they entered the Province; also that immigrants into the Province have represented each of the languages which now neighbour Baluchi, even though now all call themselves Baluch, are excepted as Baluch, and (except for the small Jadgli-speaking minority) all speak Baluchi as their mother tongue. 5. Finally, names of plants and varieties of pedigree date-palms are excluded from the list, mainly because of difficulties of identification. III I. Baluchi is a very conservative Iranian language. Phonetically and grammatically (both syntax and accidence) it is also very simple. Its historical development, its diversity and adaptability are seen in its vocabulary and wealth of idiomatic expression. Allusion has already been made to the amount of lexical borrowing, and it is mainly for this reason that the main section of this article is in the form of a word list. It is the vocabulary and idiom which presents the major barrier to acquiring a useful degree of proficiency in the language-together with the diversity of usage from one dialect to another. Nevertheless, certain details of grammar seem to me (as a non-philologist) still to require explanation. The following points which I have observed in the usage of the language in the Persian Province are offered as a contribution. 2.1 Past stems. It is generally thought that the past stems of all transitive verbs are used as though they were passive only. Morgenstierne, who of all writers has probably the broadest first-hand knowledge of Baluchi, writes (1947): ' " The so-called ' passive construction of transitive verbs in past tenses was, no doubt, the original and still prevails in some Western dialects. Thus, e marddd mardjat, this man struck that man; one, j marddhamd mard kushtagant, this man killed these [sic!] very men. In some dialects, however, the kushta, the man killed the horse; or the construction object is put in the accusative dative: marddapsdrd has been completely changed into the ordinary Persian one: zdgdzadgdjat, this boy struck that boy; " jan u mart gJ wat kaul kurtant,the woman and the man made an agreement between them (for more detail cf. also Misc. 260). In the Persian Province also the usage varies. Analysis is complicated by the fact that the pronominal suffix may be added to the verb to indicate the (logical) subject, and for the first person the pronominal suffix and the personal ending of the verb are identical. Thus, the first person and singular of the preterite Zodrt-on, the perfect zorte-onor zortdg-onmay be interpreted as either active In the second person singular there is a complication: the second person or passive constructions. singular pronominal suffix is identical with the third person singular ending of the verb for the pluperfect: e.g. zort6-at or zortdg-at. This form tends to be reserved for the latter purpose, which means that it is difficult to say " you have taken " in the passive construction without emphasizing the personal pronoun, thus: tau zorta. This point must surely be taken into consideration when accounting for the active construction zortj or zortagi. In the third person singular zo'rt-ior zortdg-imay again be interpreted as passive-i.e. he took it or he has taken it-or may be compared with Persian use of the enclitic ke personal pronoun to indicate the (at least, logical) subject, e.g. godft-as . . . -he said that. . . Note also that two pronominal suffixes may be added to the verb to indicate the (logical) subject and object respectively: e.g. z(orteAi they took it. cf. G(onabadi): vd-stundd-i-i. 2.2 The use of the plural is mentioned again below in connection with the noun. In general speakers tend to avoid the plural in the first and second persons and the third person used for people, but tend towards it in the third person for inanimates and animals. With regard to the analysis of the construction similar ambiguities arise. 2.3 Pronominal suffixes apart, it makes sense to treat the active past construction as a Persianism. It is for the most part confined to Saravan and bilingual speakers in Persian administrative centres. 3.1 The usage of the tenses of the verb is almost identical to that of Persian, except that there is no specific future tense or periphrasis and there is only one subjunctive tense-that formed from the present or imperfective stem. Other differences are peculiar to the Sarivin dialect and will be discussed below (IV).

NOTES

ON

THE

BALUCHI

SPOKEN

IN

PERSIAN

BALUCHISTAN

55

3.2 The subjunctive may or may not be distinguished from the present/future,imperfective tense the addition of the prefix " be-". The principle involved seems to be similar to modern Persian by usage. For example: dast dpbeydr o dimashuddn bring some water for me to wash my face and hands; berawi?- lit. are you homesick to go somewhere? keja-j zahir-et-int These are the normal ways of saying two very common sentences. It would be equally " correct " but unusual to use be-in the first example. It would change the meaning to omit it in the second example. The imperative usually only takes be-in verbs beginning with a vowel, e.g. beyd-come, beydr-bring, but pdd-d-get up, bar-take away. When used with the imperative be-is not stressed. " " 3.3 The variety of forms used to expressthe various tensesof" to be and to become" may make clearer the usage of these three-tenses-in-one. Basically, in normal usage, these ideas are expressed by the defective verb beag,and a series of enclitic forms, thus:

I am becoming, will become, will be, may be bin we etc. (a) bdn bi bit you etc. you etc. bant bu(t),bi(t) he, she, it, etc., also, it is possible cf. P. misavad they etc. be-may be used in the subjunctive but is omitted when the distinction it affordsis not explicitly required. (b) bo = imperative: be, become. am -in are (c) -on -4 are -it are -int is (Saravani-o) -ant are With past participle of other verbs, this form is used to form the perfect tense. Negative: nd-(h)on,etc. (d) ha(st) it exists, there is, i.e. emphatic form of the above enclitic. The other persons are said to exist, but seem to be rarely used. Negative: nist. was -atin were (e) -aton -atj were -atit were -at was -atant were With past participle of other verbs, used to form pluperfect. etc. Negative: nd-(h)aton, (f) hdsta(t)= it existed, there was-as hast. Negative: nestat. etc. = I became, etc. (g) but-on, (h) buta(g) = perfect participle: having become, having been. or etc. (j) butdg-on butg-on, = (i) I have become, etc.: (ii) I was, etc., in the sense in which P. budd-am used to expressa fact is
in the distant past, as opposed to a step in a narrative. (Cf. Lazard, G. Grammairedu Persan Contemporain, Paris, 1957: I44-45.) I had been or had become, etc., i.e. previous to remainder of

(k) butdg-aton, or but6-atonetc. context.

Where not shown the negative of each of these tenses is formed simply by prefixing " nd- ", which takes the stress. The negative of the enclitic forms thus becomes non-enclitic. I believe the distribution of the meanings of " be " and " become " through these forms (with the additional help in some cases of context) in fact precludes ambiguity, despite the apparent confusion which confronts the Persian speaker attempting to learn the language.

56

JOURNAL

OF

PERSIAN

STUDIES

4.1 Little needs to be added to what is already written about the noun. The following inflections are used in the Province: (i) -j indefinite article (enclitic). (ii) -dn plural. to which are added: (iii) -i genitive (unstressed). (iv) -ig, -ig absolute genitive, usually only used with personal pronouns and proper nouns-manigmine, Bahrdmig Bahram's. This inflection seems only to be used in the oblique case (i.e. with -d added) or in front of the copula. (v) -d, -a oblique case. (Cf. Morgenstierne, 1947, paragraph commencing "The grammar is simple . . ." and Misc. 257.) An interesting point of Baluchi idiom is that, whereas for instance Persian speakers always 4.2 tend to use the singular of nouns and third person singular verbs and pronouns for plural inanimate objects, Baluchi tends firmly towards the plural. For instance, bring the tea (for more than one person) is always cah-an-a beydr.On the other hand Baluchi has no use for honorifics, and never uses the plural for a single person in either the second or the third person of the verb. rd seems not to be used except to form the acc. and dat. cases of the 4.3 The enclitic particle personal pronouns tdrd, mdrd,iomdrd. " he " and "she "-" it " tends to 5.1 The third person pronoun is di (cf. below, VI (i) 1.2), be i~i (cf. below VI (i) 13.2) rather than di (cf. P. dn/un, isdn, G. unhdfor " he " or " she" ; P. dnhdfor " "they "; more normally in for it "). The second person is tau, gen. tei, and ti which are also used after prepositions-but ti is often 5.2 heard for the nominative also, particularly in Saravan. Absolute genitive-te-ig. 6.1 Adverbs are formed with the suffix -ak -+ the obl. case ending -d, thus: hurakd together, = zutakd quickly, tahnakd- alone. 6.2 The suffix -ok may be attached to any noun or adjective to form a diminutive e.g. hordokende a small one. 6.3. All adjectives take the suffix -in when used attributively, even when the noun is understood, tuhnj int - this is a large one. Numerals also take the suffix before a definite noun (which may e.g. ifSi be understood), e.g. pakat md sei-in (napar) rawin - just the three of us are going; d dau-inpalds - those two tents. IV
I. I That usage and vocabulary varies from area to area and village to village is certain, but pinning down the variations with any certainty is difficult (without spending much longer periods in individual villages and communities than I have been able to do) because of the perpetual and complicated patterns of movement which affect not only the whole population of the Province but also involve linguistic influence from dialects on the other side of the border in Pakistan. In this situation there is a continual give and take between dialects. Nevertheless two major differences stand out: 1.2 In any area there is always some dialect variation between the pastoral (nomadic) and the agricultural (settled) sections of the population. It is natural that the range of vocabulary should vary from one to the other. However, in some areas there are more general dialect variations between the two which must be due to patterns of migration and historical origin. For instance, in the Saravan area the settled agriculturalists use mdt, pet, brdt, while the nomadic pastoralists use the forms mds,pes, brds (cf. Misc. 258). 1.3 There is a definite dividing line to be drawn between the dialects spoken in the Saravan area and those spoken in the rest of the Province. That several of the stronger tribes in the Saravan area most likely have Persian or Afghan origins may be the main cause of these differences. The dialects of the Sarhladd similarly distinguish themselves from the rest of the Province but I have not spent sufficient time in that region to be able to give a useful account of them.

NOTES

ON

THE

BALUCHI

SPOKEN

IN

PERSIAN

BALUCHISTAN

57

2.0 The following are the main featureswhich distinguishthe Saravani dialects from the speech of the Makran: In general these dialects may be said to be closer to Persian. 2.1 2.2 The infinitive is formed from the perfective (or past) stem - Jn-e.g. kortgn to do, make rather than the imperfective (present) stem + ag-e.g. kanag. 2.3 The Persian iddfat (P. ezdfa) construction tends to be used instead of the normal Baluchi genitive construction.All the dialects of Baluchi throughout the Province are mutually comprehensible, but this particular feature of Saravani can lead to confusion. For instance, the following was an actual exchange between a Saravani and a Makrani with whom I was recently discussing the usage of the word nei~dr: A son's wife is called neidr Saravani: int jan-i cokne!dr Makrdni: A wife's son... ! ? cok... !? jdni A wife which a son takes (lit. took-notice jan-i-kecdka ge(pt) Saravini: tendency always to drop final consonants.) 2.4 In perfective stems of verbs -tk- (which becomes -xt- in the Eastern dialect) becomes -ht- in Saravani (cf. Misc. 255), e.g. Saravani Makrani Eastern axta dyag,dtka ydyag,yahta dutka doxta ducag,dohta ? gejag, gehta gitka ritka rexta ricag,rehta sutka soxta sucag, sohta

cf. verb list in Grierson, 192 1: 350-52. Other verbs which might be expected to change in the same way-e.g. bujag-are not used in Saravani. It is interesting also to compare the Marwi formsof the past participles of these and similar verbs (cf. Marw,passim): they lack the -h- of the Saravani forms, but have compensated by lengthening the preceding vowel. 2.5 The continuous tenses formed from the infinitive of the imperfective stem + the obl. d + the enclitic form of the verb " to be "-waragd-on, waragd-aton-are not used in Saravani. Instead a periwdrt-on. phrasis is formed with the adjective goldaYi P. geldviz),thus: mangoldis-on wardn, (cf. goldei-aton Thus the simple past tense is used in Saravan as an aorist and an imperfect. 2.6 Saravani drops the final g of the past participle before a suffix,thus: yaht6-on- I have come; mdddatd-at we had stopped. = 2.7 Saravani makes far more use of compound verbs, for instance: pac kanag for bujag;prdmu. for (P.fardmui)kanag s'amusvag. 2.8 Some Saravani dialects do not use the absolute genitive, but instead employ ceisimilarly to the Persian mdl,thus: i~i cei mdnaint- this is mine. V The representationof the language in writing is problematical because:
(a) the pronunciation varies from place to place, (b) the precedents already set are not always consistent or satisfactory, (c) it is desirable to spell in such a way as not to disguise philological connections. If a standard orthography had been evolved in the Urdu script by the Baluchi publications in West Pakistan, the task would be easier, but this is not the case. The system used here has been worked out with these points in mind. It is felt that for the reader with a knowledge of Iranian languages the values of the letters used should be self-explanatory, except for the following points: I. Initial h is often dropped, or added where it does not belong, i.e. it is not phonemic. 2. Any final consonant (but not clusters) may apparently be doubled at will, except when followed by a word beginning with a consonant.

58

JOURNAL

OF

PERSIAN

STUDIES

3. Final consonants (inc. clusters) at a pause tend to be dropped, the preceding vowel being nasalized when the consonant is an n (or the first consonant of the cluster is an n). 4. Any word ending in final short a may take a final g (e.g. past participles) and in most dialects (except Saravani) normally does so before an inflexion or a word beginning with a vowel.

of variation in pronunciation: i and e; e and ei. (My ei is normally represented elsewhere as ai or ay.) It is similarly often difficult to distinguish between long and short a, and in the obl. case ending this difference does not appear to be significant. 6. In general Sardvdni keeps long u while most other dialects have changed it to i. There does not seem to be a phonemic difference between this long u (where it is retained) and a sound which is something between a long o and long u. For this reason I have represented all long back rounded vowels as u, and short ones as o, since there seems to be no phonemic differentiation within each. 7. It is generally thought that f and x not only do not exist in Western Baluchi but cannot be pronounced by Western Baluch. However, in the Persian Province bothf and x are occasionally heard. The actual situation would seem to be that there is no phonemic distinction in the province between p andf, and h and x, but that the tendency is always to pronounce p and h respectively. It is interesting to note that there is (at least) one instance of x becoming k-in kandag- to laugh (cf. Misc. 254).

5. The following vowels are particularly difficult to distinguish between in many cases because

8. The combinations 'a and a' in A. or P./A. loan words (exceptwhere 'is intervocalic) becomed, e.g. drusi.Intervocalic ' is simply dropped, e.g. Ra-is. mdlum, " " " 9. The traditional distinction of" long and short vowels in both Persian and Baluchi does not of the ordinary spoken language (cf. Sokolova, 1953: 9 ff. and seem to be adequate for the analysis 15 ff.). In Persian speech in general the distinction is one of quality, not of quantity. We have already in noted that in Baluchi inflexions the Province there does not seem to be any distinction between long and short a. In some positions long a is reduced to ah, e.g.yaht fordtk(cf. above, IV, 2.4), also -ahn for annin Pahnuc for Fannuc, and -oh- for u in Banfohl for Bampur-though the form Banfohl is not now known in the Province and I have only met it in Persian texts, and below (VI (vii), 49) pohn and pun. 10. Baluchi uses a large number of loan words which are predominantly of Arabic, Persian, Brahui, Urdu, or Sindhi origins. It is interesting that although the Arabic, Persian and Brahui loanwords contain phonemes which are strange to Baluchi it has not adopted them but has modified them to fit its existing phonemic system, whereas in the Urdu and Sindhi loanwordsit has retained the retroflex dentals (cf. Geiger, 1898-1901: 234, Notes: 38 and Misc. 256). In Baluchi words in the following word list d, and r representthe retroflexconsonants. In non-Baluchiwords the diacritical dots are used .t, according to the normal canons of Arabic/Persiantransliteration. I I. The first person singular personal ending of the verb is pronounced -dn when stressed (in the present/future/subjunctive),but on when not stressed (in all other tenses). (It will be remembered that the traditional transcriptionvaries between a and u.) The n is usually only pronounced before a suffix; otherwise the vowel is nasalized. I2. The stressis always on the final syllable unless otherwise shown. VI
In the word list, which is given here, words which are obviously (at least in part) Persian and not Baluchi in form whether of Iranian or other origin, and are likely to be old loanwords from Persian Of verbs the infinitive only is given, except where the third person singular present and/or are marked t". the past participle are formed irregularly from it, in which case the irregular parts are also given at the first occurrence. The arrangement is in twelve sections: (i) Adverbs, prepositions, particles, demonstratives. (ii) Verbs. (iii) Adjectives. (iv) Abstract nouns.

NOTES

ON

THE

BALUCHI

SPOKEN

IN

PERSIAN

BALUCHISTAN

59

(v) Domestic. (vi) Domestic animal terms. (vii) Pastoraland agricultural. (ix) Institutions. (x) Parts of the body. (xi) Food. (xii) Time. Within the sections the order is alphabetical except that synonyms and certain other closely related groups of words are kept together. It is felt that in a list of this length (only 455 items) the inconvenience of the lack of strict alphabetical order is outweighed by the simplification of cross referencing.
(i) I.I d that-demonstrative adjective. 1.2 adi that, he, she-demonstrative pronoun; inflects: dia,dig,didn, didnd, aidnig. (viii) Social relationships and statuses.

d-dim= on that side, over there. d-go(r)- in that direction, there (d + gwar). 2.3 d-ke'-ter= a little further that way. 3. aca= well, then; e.g. aca bwaroda - well, eat it then. 4. I cinka/inka how much, how many/some. Note: dauinka 4.1
2.1 2.2
4.2

kaY,kel, cf. above 2.3 and 9.I. Alsoyak-kdsa - continuously, without break. 17. olda/ode well, then. 18. peim manner, way. Note: har peimin = whatever sort of; -peimd = (postpos.) like; peim kanag synonym of a~d kanag (ii), 21.1 and P. dorostkardan. 19. rand(a) after(wards)-postpos. and adv. Cf. Marw: 67. = till now. 20.1 tanigak. 20.2 taniwahd(i) (P./A. waqt) 2 I. trika = in the middle, adv. and postpos. = 22. wa = enclitic particle inserted after word requiring emphasis, used as ke in P., e.g. di wa nadyt he certainly will not come.
i6.

4.3 con.t how much, how many. -5. cos(i) = thus, like this (co/cun+ (i)s(i)). Cf. Marw: 29. 6.1 dar - out, e.g. dardyag- to come out. 6.2 ddrd outside, as also, ddnnd lit. in the plain. = ku 6.3 jdn-dar= naked (jdn = body). Also, jdn dard = he undressed. Cf. Gershevitch, 1962: 82-84. he won't be offended. dardin pa ddrda ndbi(t)= 7. 8.I daul-way, manner, e.g. d da laa in that way, cf. peim(below 18); also Marw: 34. 8.2 daul-ddr " posh "-looking, impressive. = 9.1I --this, - demonstrativeadjective; forms g-dim, -go,e-ke-teras d above 2.1-3. Cf. this, it-demonstrative pronoun. Inflects: isia, isdn,iisdna. above 1.2. 9.2 iS, is, is Io. go/gun with. Note: d dagdr mdna(int) - that plot of land is mine (sc. to till, though I may go not actually own it); d dagdrmanig int - that plot of land belongs to me. Also, mdnazar zar ha(st) = I have (own) money; mdina gun (int) -I have money on me. Also, e.g. Gol Mahmad at gun = (sc. we went somewhere and) Gol Mohammad came with us. gun godda = afterwards, then. Cf. Marw: 35. Ii. 12. hanga(t)= still (P. hanuz).Cf. Marw: I4, dngdt. hannu= now. 13. 14. jam (P./A. jam') in pajam = (all) together; synonym of hurakd. 15. ji =--particle showing respectful attention, with which speech addressed to superiors tends to be started. Standing on its own it signifies a positive answer. The negative is expressed na, or by.ji fi inna.

cinco/inco cinka/inkac}

twice as much.

}-=

60

JOURNAL

OF

PERSIAN

STUDIES

(ii) = I. I. drag/ydrag,(y)drta/aorta bring. Cf. P. dvardan/dvordan. dvardan. 1.2 paja(h) drag- recognize, remember (sc. call to mind). Cf. P. be-jd
2.1

2.2 3.2

3.1 barag,bart, borta

4.1
4.2

4.3 5.
6.1

6.2
7.1 7.2

8.1 8.2 8.3 8.4 8.5 9.

S10.1

10.2
I i.
I I11.2

padddyag- to get up (from sleep or from lying position). Cf. Marw: 62. bring. dam barag be tired-temporary tiredness, out of breath (e.g. after climbing hill, as distinct from the adjective lzand).Note inflexion: manddm-a-bardn- I shall get tired, it will tire me; borta?- are you tired? mandamndborta- I am not tired; ddmet Marw: 27). cand(in)ag(cf. sor(in)ag = shake (trans. and intrans.). larz(in)ag cdrag, cdrta- look (at). (as P. cidan)pick-senses include chicken picking up grain with its beak, and cinag,cint, cita- counters in a game or plates and cutlery on a table, i.e. picking up one thing, down setting or setting down a number of things in place. cinm roll up trouserlegs. s'alwara bdld(d) ddrdn I will hold the water-pipe (for you). (b) stop, e.g. ddrag, ddita- (a) hold, e.g. mancelima mdddst- we stopped (i.e. counts as transitive). show. peJdarag-ddta- give. deag,dant, bondeag- light (a cigarette). durdeag= throw (away). rahdeag= send. suj deag lit. give guidance, advise, help. to meet with unexpectedly. dik beag/waraggallag - flee. gallinag- chase out/away. find, seek. guide, or find path for (an animal), down a mountainside. Cf. Elfenbein, 1961: 95 n.- 7sargerag,gepta- to start making preparationsfor setting off on a journey. to make fun of (kaldg- P. masxara). kaldg geraggesag,gesenag- finish (trans. and intrans.). hallet it is finished, there is no more. Past: halleta. " join (trans. and intrans.), e.g. a group; hurlit. means mixed together ". hur/hawdr beag/kanaghelds finished, over (P./A. xaldswhich is used similarlyin Afghan Persian). manheirdn-on I do not understand, I do not know what to do or say. heirdn: jakkag cough. Cf. Marw: 42..
to abort (of a sheep or goat) gjag, gehta/gitka- on the cir gejag ground. Cf. Marw: 28. spread

= is Saravani only. dyag/yayag, ket/yjt, dtka/yahta come. The second form

11.3 dar gejag

I 1.4 Fr gejag
12.1

12.2

13. 14. 15.

16. 18.
19.

S17.

jallag = stay over. kneel (of a camel). 20. jokkag 21.1 a d kanag, kant, korta/kota kanag (cf. above, (i), 18) = P. dorostkardan = make, put right, --peim repair, prepare, etc. = 21.2 bad kanag P. kul kardan take on one's back. Cf. Elfenbein, 1961: 94 n. 6. -= sell. 21.3 bahd kanag bdharkanag divide, share out. 21.4 turn back, return. 21.5 biru kanag (Dashtiari) 21.6 cist kanag = remove, lift, take away. Cf. Elfenbein, 1961: 94 cikut-context seems to demand cist kut. = back. 21.7 danikanag (cf. above 3.2 dam barag) rest a while to let temporary tiredness pass, get breath

NOTES

ON

THE

BALUCHI

SPOKEN

IN

PERSIAN

BALUCHISTAN

61

long will it take you to get to Karachi. 21.9 dapkanag- put in the mouth. 21.Io jr kanag= put down. 21.11 gap talk, hold conversation. Cf. P. gap kardan/zadan-more common in Afghan Persian. kanagidle talk, P. sJuxi. Cf. also gappo rap on --joke, 21.12 gwar (clothes). put kanag2 I. 13 habar/hawal kanag- talk, speak. 21.14 Jadgdli speakJadgili. kanag21.15 jost kanag-ask. and bejeh)= run. 21.16 meiddn (also pa kanag, meiddnddyag/rawag in Saravan, as imperatives, betac urinate. mes 21.17 kanag/mizag21.18 kapkanag- crack or split open (e.g. pistachio nut). Cf. Marw: 48. one's nose. In Sardvdn sar ric is used instead of pacdnk(pacdnk=nasal 21.19 pacdnkkanag -blow mucus). 21.20 pddkanag= (a) wake, get up (trans.), (b) put on (shoes). an extra blanket; (b) to prepare a water pipe; (c) to rise (of a river in 21.21 pir kanag- (a) put on, e.g. flood). 21.22 se(h)i kanag/beag = inform, let know/know (in the sense of " to have information "). Cf. Marw: 69, si. in 21.23 seil kanag- watch. Cf. seil kardan Afghan Persian, and Marw: 72. call, summon. Cf. Marw: 78 (tawdr- noise). 21.24 tawar kanag= ka?idan. 21.25 teckanag= spread (trans.) . Cf. also teckasag- P. derdz and wadiin Marw: 79. 21.26 wadikanag(Saravan) - find. Cf. G. vadikardan 21.27 warid kanag- pound (in mortar). 21.28 was kanag= try, make effort.
22.1

= td 21.8 dar kanag- take off (clothes, shoes, etc.). Also in cinkarucddr-a-kant birasJ Karddca how pa

22.2 22.3 22.4

22.5

23. 24. 25. 26. 27. 28.

29.

30.

31. 54. 32.1 lutag 32.2 lutdanag = summon. which is impersonal in the present as well as taking the 32.3 Saravani prefers the defective verb pdestin, usual passive construction in the past, e.g. pdsJ mdna pei = I want a sheep/goat; pelt-et koja what do you want ? blrawJ? - where do you want to go ? Also ce taupei? min beag sink (in mud). 33. and 20. 34. ultag, ultdta stand (i.e. the act of standing, to get up), imp. bust. Cf. Marw: 18, 19, I am sowing seed. scatter, e.g. t6ham pardkandn 35. parakanag = Cf. pat.t o luft = research. 36. = search (P. taftif). paft.tag enter (in a manner that normally requires stooping). Cf. Marw: 65. The phrase 37. potrag (lug) = also used as an equivalent to bdhutbelow, (ix), 3. is lug potrag
5

jr kapag- climb/come/jump down. sar climb/walk up. kapagpac kapag= fall behind. pe-kapag get ahead, go on ahead. = pa kdrbeag be necessary. = kandag= laugh. Cf. Marw: 14. kelterbo. kenzag= move (a little), bekenz= kammdkid Cf. Marw: 50; (b) be finished, as heldsbeag,hallet. kottag= (a) (Saravan) pound. ladlag - load, decamp. Cf. Marw: 52. = (a) meet with, hit, touch; (b) start (suddenly), as in laggitpa kanddgd he started to lag(g)ag- Cf. Marw: 52 (?), and Geiger 1898-1901: laugh. 234. lakolag-- slip, slide. Siterag (Dashtiari) - synonym of 29. climb. lek beag= want. Cf. Marw:

kapag, kapta = fall.

62 38. 39.

JOURNAL

OF

PERSIAN

STUDIES

sekkag= learn. tahag to get on with, P. jur dmadan. 40.-1 tarr(jn)ag, tar(ri)ta/tarrinta- P. gaitan,garddnidan. 40.2 pac tarrag- P. bargaitan,return. 40.3 tarruk in training, P. varzida. ydd in tdraydd/hus ha(st) do you remember? 441.2 hus' 1I.2 kardan. 42.1 yala kanag- give up, P. tark 42.2 yala deag- let go, P. velkardan. 43.1 zurag,zorta pick up, take. 43.2 pa bahdzurag = buy. ndzurj= you do not listen/take any notice of what I say. 43.3 mdnihabardna (iii) dgdh= awake. bdz = (too) much, equivalent of P. zidd.Cf. Marw: 25. cedd- marshy, waterlogged. dadd - well, strong, fit, skilful. del-gui (a) thinking to oneself alone, lost in one's own thoughts. (b) Sympathetically attentive to what someone else is saying. del-mdnag expectant, in the sense of the P./A. motavaqqe'-butperhapsnot so strong. drost all. cf. Marw: 32. koll- all. ganuk- mad, crazy-not so rude as puk, for which cf. Marw: 62. gdr lost, disappeared. gwand.Cf. Notes,46: and Marw:39. = short. loka

i.
2.

3. 4. 5. 6.
7.1 7.2

8.1
8.2

9.
10.

Ii.
I12.

pa.tak 13. jur -- well (especially in greetings). Cf. Marw: 42. 14. kaland old, worn out. 15. plenty. Cf. Marw: 54, and Afshar, 1333: 461. moc= 16. nizur = weak. = visible. 17.I peiddg 17.2 zdher 18.1 prdh= wide. 18.2 prdhi= width. in pottarik-on= I have indigestion. 19. pottaruk man 20. sell- bad, dirty. 21. sobak light (P. sabok). = dark. 22. 23. 24. 25.2 25.3 26.1 26.2 26.3 27.1 27.2
25.I

tahdr taklipi (cf. A./P. taklif) = uncomfortable, troublesome. Also in Afghan Persian. tayar ready. b Cf. Marw: 78 = big, large.

mazan

.tuh. zand zabr

Jarr
wal jahl juhl
= =

= good (wal also includes the sense of P. xwol). low ;jahld = low down ;jahld-ter - further down. deep, of well or gwalm, (below (vii) 24) etc.

NOTES

ON

THE

BALUCHI

SPOKEN

IN

PERSIAN

BALUCHISTAN

63

2.

I.

3. 4. 5. 6. 7. 8. 9. Io. I I.
12.

13. 14. 15.

(iv) bongej foundation, P. bonydd. = and dozman insult; dolmdn deag = swear at. Cf. P. dosndm, Marw: 33. dalwat = riches-usually in the form of flocks,from A. daulat.Cf. P. mdl. humdn-desire-appears sometimes to include the meanings of hobb,haydl(P./A. khidl)and zahir, below, 14. Cf. Marw, 19, 6man. hobb(P./A. inclination, desire, P. meil. hobb)+ pronominal suffix self. Cf. Marw: 43, jind(i). jend kalah = message. leib = (a) game, P. bdzi (gwdziis also used). (b) A scene, uproar, great fun, as P. ma'raka. mesr =a beautiful, luxurious place. molk = one's own country, area-e.g. rawdn(wdti) mdlka I am going home. sdh = soul, spirit, P./A. ruh. sir - consciousness; bi-sdr unconscious. Sohrat shame, P. xejdlat. = zahir = longing, homesickness-e.g. zahir-et-int are you homesick? in zahrbeag/gerag be not on speaking terms, sulking/sulk. = zahr (=P. qahr)

The firsttwo terms, which are Persianborrowingsthough I have not come acrossthem in use in any form of Persian, appear to be restricted to certain tribes in the Sardavinarea; the third term is more general. The conception is in fact superfluous in most of the Province. 2. baca(g)(P. bacca)= any male servant, slave (goldm) otherwise, particularlyas a term of address. or Cf. similar usage in Afghan Persian. 3. bdkes-match(es). 4.1 bdn= room built of mud or mud brick. (In this word the n is not dropped or reduced to nasalization.) 4.2 dawdr generic word for dwelling, also extended to mean wife. 4.3 geddn- tent, which may be either of black goat-hair or pis (23.1 below). Cf. Marw: 35. 4.4 ges - general word for dwelling used by the pastoralistsin the Sarhadd. in (I 4.5 kdpar have also heard kafar).Cf. P. kapar,and Bashk. kavdr Gershevitch, 1959: 219 = shelter of branches and pis'. 4.6 - a generic word for dwelling. Cf. Gershevitch, 1959: 219. ka.t kuti = bdn(above 4.1) but bdnis usually bigger. 4.7 4.8 koll= type of squarehut built of branchesand mud with pis'roof,mainly in south-eastof Province. Also name of tree. 4.9 palas = tent. 4.10 tup =- as koll, but circular, larger and better built. Cf. Gershevitch, 1959: 218-19. 4. 11 lug = the most general word for dwelling throughout Baluchistan. sediment which formsin water from inside of water-skin(maik). 5. bolgdr6. bdpdri= merchant, tradesman. 7.1 bond 7.2 gandal(used mainly in Ldshdr)I = the traveller'sbaggage and, esp., his blanket(s). 7.3 lihp (cf. Marw:53, lep) (Saravan) 7.4 napdd Marw: 60, nipad)(Makrin) (cf. 8. celim water-pipe, P. qalydn. Cf. Marw: 27, and Majalla-i Ddnilkada-i Adabfydt,Tehran, Vol. XIII, no. 3, p. 81. eikar - live embers or charcoal. 9.

I.I tdb-rjz 1.2 tadab-hdna 1.3 dim-s'ud

(v)
lavatory, latrine.

64
10.1
10.2 I I.I

JOURNAL OF PERSIAN STUDIES

garib(A./P. yarib) = stranger,with associationsof" poor ", "helpless ". dar-d&tk-instranger (without the above associations),visitor. Not used without the attributive =
suffix.

halk(P./A. xalq) (Sarhadd) = e encampment. 11.2 mitag 12.1 hdkum(P./A. hdkim)-used in the Province to mean ruler of a not always well-defined area, traditionally always centred on a fort in an agricultural settlement; implies the right to of collect I/ 10 (dah-yak) the agriculturalproduce of the area in tax. Cf. Marw: 12. kamda (a) one of the socially and politically important men in their community but only slang 12.2 for important chiefs; (b) the senior man in any particular situation. 12.3.1 wdja (P. xwdja) = (a) of a third person, very respectful term of reference; (b) respectful term of address; (c) neologism for Mr. as P. dqd. but more respectful. 12.3.2 waja-kdr as kamdi, = and has hand-mill, quern. Cf. Marw: 15, ashksh parallels in Bashk, Kurd., and P. 13.1 of has,in use in Sarbiz area. jenter-synonym 13.2 heik = egg. 14. the long, full shirt, which together with salwdrmake up Baluch male attire. Cf. 15. jdma Marw: 43. 16. joll - woven rug, P. gelim-the only form of floor covering, besides piA mats, made in the Province. Cf. Marw: 43kalla-ddr rupee. Cf. Marw: 46. = 17. 18. kapal pottery made without wheel in Kalpurkdn (Saravan). -= 19.1 makesk fly/flies, mosquitoes. Cf. Marw: 55, magis. fly (Saravan only). 19.2 malAMk mosquito-mainly Saravdn. = 19.3 mauled female yoldm.Cf. Marw: 55, molid. 20. 2 I. pdg a type of hat with turban. pdg bandag= to appoint as headman of village or tribe. Cf. Geiger 1893-1901: 234. pandul- intricately embroideredpocket on front of woman's shift. 22.I 22.2 jig (in Ldshar-zih) = the embroidered bodice of the woman's shift. 23.1 pis - the leaves of two species of wild palmetto Chamaerops (ddz and pork)which are put to innumerable uses, and of which the following items are woven. 23.2 duzmdn ppi string for sewing the seams of basketswoven of piS. cilik (mainly Saravdn) = rope ofpI. Cf. sddbelow, 23.11 and Afshar, 1333: 46323.3 dried and pounded ready for making sawds(below 23.12). jatk = piA 23.4 kac = donkey panier ofpiS. Goat-hair panier is gwdl-cf. P.jovdl. 23.5 23.6 kamag= pzi sieve for flour. bag for storing dates. kapdt= pz-s 23.7 23.8 lac camel panier of piz. pdt = pBlcontainer for storing wheat. 23.9 23.10 pat =- strip ofplaitedp-lapprox. 4 in. across. sad = rope ofpi. Cf. cilikabove. Also traditional measurement of length equal to 50 bagal (P. 23.11 bayal). sandals. sawas 23.12 shallow pil tray used for scales. iahim =-piH 23.13.1
amud = wooden bar from which are hung the two slahim,thus forming a pair of scales. = son~d pis basket. tagerd = mat woven ofp i&. 23.15.I or spread out for diwdn assemblyof important man. 23.15.2 pattar = tagerd = container for grain. 23.16 ticag pi: 23.17 tijig = rope on which sawds are woven. 23.18.1 sabd = tray of pif for feeding donkeys.
23.I3.2

23.14

NOTES

ON

THE

BALUCHI

SPOKEN

IN

PERSIAN

BALUCHISTAN

65

23.18.2 tug = same as sabd.

24. 25. 26.

27.

28.

poloh = nose ring for women. Cf. Marw: 63. por = ash. ragu = guinea worm. sakk = comb. sirken= stone beside tarun(= P. tanur)on which dough is spread in bread baking.

(vi)
I. 1.2 1.3 1.4 1.5
1.6
2.1
2.2.2

pas = sheep/goat, generic. gattor= young male goat. gwarag(ok)= baby lamb, unweaned. pala.t= female goat not yet pregnant. senik= baby goat.
trot
=-

1-2

guk = ox (male or female), generic.

year old male goat-that

most commonly eaten.

= calf up to three months old. Cf. Marw: 36. gwalk) = 2.3 porrdp young female calf. 2.4 rong= heifer. = young male calf.. 2.5 2.6 .tdb = ox used for ploughing. kdrgar 3.1 gwavz w (Saravmn) 3.2 gudarwasp.

2.2.21 gwak

2.2.1 gulu

1.1 agalinc i
1.2

(vii) mushroom edible edible generic poisonous poisonous

katingbu (Dashtiari)

ko.tomb zur 1.4 zamznz

2.1 2.2

1.3

Cf. 1i.5 kirdn-haruk. Morgenstierne1957: 453 aces(Saravan) fire.

as
I

L.

Teger(uk)

3.I bdj 3.2 kor~t

3-4

S
KORT'

RIJ

UI

3.3 sah-ju 3.4 sar-ju sar-ju 3.5 tig-band 3.6 teger(uk)

= irrigationterms, see Figure.

BAJ

TTgband

4.
5.1
5.2

Balucused among the Baluch themselves means pastoralist/nomadas opposed to ?ahri= settled agriculturalist.
(dp)band sendat
=

dam.

66 6. 7.1 7.2 8. 9.
S10.

JOURNAL

OF

PERSIAN

STUDIES

i i.
12. 13.2
13.1I

13.2

14.1

14.2 seng o 1i5.I gadd grdnddn

16.

summer rain when it continues long enough (i.e. on and off for several weeks) to spoil or bas'endanger the date crop. cdk small piles made at reaping. }-.= gwdp (Sarv~n). Cf. Notes:37 and 45 cammag (cam= eye) = spring dagdr= cultivated land. Cf. Marw:34. dakk = pass dann= plain (cf. danna= outside) daryd river, sea, as in Afghan Persianfor the Helmand. = ddz = two species of wild palmetto Chamaerops,see above (v) 23.-1. pork duk Marw:34, donk?) ? -- stone (cf. stone

15.2 seid fast high stepping walk (of horse). o 15.3 gadd waddI1 17. 18.
I19.

game

20.
21.1

21.2

21.3
21.4

21.5

21.6
21.7

21.8
22.

23. 24. 25. 26. 27. 28.

29.

30. 31. 32. 33. 34. 35.1

35.2

35.3 35.4 36. 37. 38. 39.

garr Also in e.g. edamiwa/njwag gatt int it is impossible to get gatt impassable, of a mountain path. fruit here. Cf. Elfenbein, 1961: 93 n. 6. (cf. gram= piles of gwdp/cdk above 7) made at reaping. Cf. Misc.: 288. guj = lizard. guric north, north wind. nambi= south. = ru-dar-dt east. ru-i-rod= west. magreb west. jahl-gwdt= south wind. jr-gwdt = east wind. sar-gwdt = west wind. we were cold. gwahr gwahr= cold (noun). mdra buta= gwdl(ag) = donkey pannier (P.juwdl). Cf. Notes:4 , Misc. 288, and Marw:39, and above (v) 23.5. gwalm = water left standing in river bed. Cf. Misc.: 288. gwan wild pistachio tree (P. ban). Cf. Notes: 5o. gwdng= date palm sapling. Cf. Misc.: 288. gwap mat of reeds for stranding fish in a flood. hdr= flood-of river in spate. haur= (a) rain cloud; (b) rain. Cf. Notes:46, and Marw: 19. Also includes land irrigatedfrom flood water. holap = dry farm land, equivalent of P. deim. kahn= P. qandt. = kandeg steep path up to mountain pass. = sorghum (as a plant), especially when used as fodder. karab kaur= river, river bed. = dates from the time they turn colour (about one month before ripening) until they kolon.t ripen. papok = dates before they become kolon[t. (Sarav~n) I hormd ripe dates. na kontak= thorn. kruc= non-pedigree date palm and its fruit. Cf. Misc.: 289. kucek= dog. Cf. Marw: 45. ku the heart or cabbage (P. panir) of date palm or palmetto. --

crocodile. gdndu= rock or mountain which serves as landmark.

NOTES ON THE BALUCHI

SPOKEN IN PERSIAN

BALUCHISTAN

67

40. 41. 42. 43. 44.-I

44.2

45. 46. 47. 48.1

48.2

49. 50. 51.

52.

lat/lot = stick. Cf. Marw: 54. lewdr= the hot NNW wind which blows during the summer in the Jaz Mfirian depression. Cf. Marw: 54. lor = muddy (of water). lur dust devil. mac n mok(Saravan). Cf. Redard, 1962: 2 i5 n.2. palm. -datealm mahdr reins of camel. nud - low grey cloud, overcastsky, Scotch mist. oldk= general term for any riding animal. pal- furrow, also used to refer to sowing in furrows,as opposed to sarcand = broadcastsowing. pogol/pogal frog. = pohn/pun blight; windfall dates; dates turned sour on tree because of damp during bas (above 6). = Cf. Misc.: 290, generally equated with P. pusida. Iahr = patch of cultivation and/or palm grove, not including dwellings. Cf. Gershevitch, 1959:
217.

= sadnz drizzle, rain not heavy enough to cause rivers to flood (and thereforeno use for agriculture throughoutmost of the Makrin). sink - small white (non-rain) clouds. 53. minor river/ravine in mountains. 54. fp - green, algae, e.g. d gwalmasunz gepta = that pool (of stagnant water) has turned green 55. sunz with algae. 56. i ur mountainside from which the rock splinters and breaks away. 56.2 taldr= smooth, firm rock. Cf. Marw: 76. siah gwdt - dust storm. 57. or 58. talk/tak - pond for storing water at end of meagre qandt spring. = 59. trampan rain drops. 6o. hail. trongal= 61. uteg -(artificially dug) hole for storing rain and flood water (Dashtiari) for drinking and domestic purposes, P. berka. (viii) - close friend, comrade. Cf. etc. parallel P. barddar-khwdnda, 1.2 tdaz brddar dazgohdr- synonyms (male and female) of brdhondag. and
i. Itbrdhondag I

2.2 mardin = son, boy. cok 2.3 janin cok- daughter, girl, spinster. 3.1 detdr - fiancee. Cf. Notes, 37, 44 and 52, and Marw: 33 and 34. In the Persian Province it is
3.2 3.3 3.4 3.5 4.I 4.2 5. 6. 7.1 7.2 8.I applied to the girl only. Also used now to mean " girl friend " by Baluch in Tehran. = Cf. parallel P. khwdstdr suitor (zdmdd also has the same senses as wastdr (SarivSn) fianc&. j P. ddmdd). zdmdd bdnur - bride i.e. just before and just after the wedding night only. sdliik =groom bibi = respectful title for (usually elderly) lady. bdnok= as bHbi. doksic = sister-in-law (i.e. WZ and HZ, not BW). Cf. Notes: 43-44, and Marw: 33. hapuk = co-wife. Cf. Notes: 47. jan - wife. one of the species " woman ", a woman. janen widow, divorcee. Also found in Brahui (Marw: 44). januzdn -

2.I

cok = child.

68

JOURNAL

OF

PERSIAN

STUDIES

8.2 saruzdn- widower, divorced man. 9. jar = twin.

Io. I I.

12.1 mardom= a person (i.e. without understanding whether man or woman). 12.2 mard= husband.
12.3

jowdn young, in prime of life (as warnd below); good, as becomes a warna. a girl who has " gone sour " (tropsvbuta), past marryingage. i.e. kalmdnt-

wdres (A. wdrith) = synonym of siydd. 16.1 tru = aunt (FZ and MZ). Cf. Notes: 51, and Marw: 77. 16.2 ndku = uncle (FB and MB). 17. warg = (Sardvan) sister. Elsewhere gohdris used.
I5.2

mardin one of the species " man " (i.e. sex differentiated), a man. == 13. mirwandi extra share of inheritance given to eldest son before generally equal division of remainder between all sons and daughters. nea-r= sister-in-law, daughter-in-law (i.e. BW, SW; contrast with doksicabove (5), and cf. 14. P. A. 'arfis).Cf. Notes:44, and Marw: 61. 15.1 siydd= relative, any person thought to be related by blood or marriage even though degree may not be known.

18.

warnd= youth, man in the prime of life. Cf. Marw: 80.

I.
2.I

2.2 sfir, sir

(ix) dirdt killing animal and cooking and distributing the meat, as at eid Qorbdn. drasi (mainly Sardvdn) wedding. = asylum. Cf. Ulus,July, 1964, p. 20. bdhu.t = mutual bejjdri help, especially contributions of relatives towards cost of wedding, to be repaid when the contributorstands in similar need. Cf. Ulus,ibid. hdl = exchange of news on meeting. Also simply news, information. = in nadr(A. nadhr) phrasenddr-on A./P. qorbdn a term of address. as in phrasepeil kan = forgive, let bygones be bygones, especially as greeting on eid Qorbdn. peil sding= marriage, A./P. ezdevadj.
saun = divorce.

3. 4. 5. 6. 7. 8.
9.

i.
2.

(x) dap - a mouth. Fig. ddpaas a postposition of place-at the mouth of, and in place names, e.g. Dap-kaur (Sarbaz) = lit. mouth-river (should we expect Kaurdap?). as dgm= face. Fig. dnmd postposition of place-in front of, and in place names-e.g. Tip-dem = throat. Cf. Marw: 39. go.t in gwar mdnigwdra = beside me. Cf. also gwar kanag, (ii) 21.12, igo, (i) 2.1 and 13.1.
lit. facing the sun. Cf. also ddim, etc. Cf.
(i) 2. i

and 13.1.

3.
4. 5. 6. 7. 8. 9.

hdruk= gum. kuc = eyelid. kolop = cheek. ldp = stomach. Also as postposition: ldpd = in the middle of. = mocac/mecdc eyelash. = (a) foot; (b) spoor, also used for track of motor vehicle. Cf. also pddd (with short first a) as IO. pdd postposition = on the tracks of, after, and adv. again, back. Cf. also sarpadbeag = to understand. I I. pMidnig/pitok= forehead. 12. sar, as postposition and adverb: sard = on, above; sarater = higher up. 13. I sont = beak. Cf. Marw: 71. 13.2 son.tJg= a particular form of mosquito with a fierce sting in Dashtiari. 14. srin = the small of the back, the waist. Cf. Notes: 50, and Marw: 71.

NOTES

ON

THE

BALUCHI

SPOKEN

IN

PERSIAN

BALUCHISTAN

69

(xi) = I.I ndn/nagan bread. The general word used for " a meal ". Owing to the fact that breakfast is seldom eaten as an independent meal and the midday meal may be eaten at any time between o0 a.m. and 4 p.m., there is some confusion in the usage of the words for these two meals. However, 1.2 nahdri generally and a used to mean " breakfast ". are na~t.(in Saravan) Saravan) '1.3 ndi'ta(in .3 the midday meal, which is the main meal of the day. (Cf. Marw: 29, 1.4 cdit(Saravan) and where cdat given as meaning " noon, till noon is 1.5 sobdreg (Makran) is 1.6. idm used to mean the evening meal. 2.1 = bat.tg eggplant. 2.2 gombak onion seeds. 2.3. gorja= tomato. Cf. P. gauja (-ifarangi). 2.4 kusic= marrow. Cf. 2.5 kuteg= water melon, P. hendavdna. Marw: 50, kutigayt. 2.6 limbu= lime. 2.7 panir = a berry mixed with fresh warm milk to cause it to set, in which condition it also is called panir. 2.8 pimdz = onion. 2.9 sob = local species of apple; P. sib is used for other (imported) varieties of apple. 2.10 tejag= melon, P. xarboza. 2.1 I ddr-tejag paw paw, P. deraxti. = xarboza-i 2.12 titok= pomegranate. cooked rice, P. celau. Cf. Marw: 25, which gives meaning " gruel ", and compares 3. ba.t bhatt. Si. 4.1 boc= dug (P. duy) when boiled and thickened, but before being dried-when it becomes dildnj (P. kalk, qorut). 4.2 sar-du -= fresh (goat's) milk milked directly on to the previous day's dug-a very refreshing, frothy drink. 4.3 sott = water with a little dug added. This is thought to be more refreshing, and it improves the taste of bad water. Cf. the A. shanin Thesiger, Arabian in Sands,1959: 122. However, the Baluch often drink sottwith sweet water. 5. cah = tea, P. cdi,cd-i. Cf. Marw: 26, cd, compensatingfor dropped " h ". 6. cangdl= preparationof dates and ghee, worked almost to paste consistency. 7.1 sohru= red sorghum. Cf. P. sorx = red. 7.2 mdk - types of bean. 7.23 md 8.1 wdd= salt.
8.2 wdd-dp = soup. Cf. P. db-gust.

(xii)
I.

= bdndad tomorrow.

Cf. Marw: 23.

2.1 bJ-gdh= late afternoon, towards sunset. 2.2 bd-wahd(i.e. P./A. bi-vaqt) = late. 3.I dir, ddr-wahd= a long time ago. 3.2 zzt = a short time ago, not long ago. 4. hdmin = the period of the date harvest, which is generally July and August. Cf. Notes: 39 and 52. 5. imbiri, imbirigin sal = this year. Cf. Marw : 17. 6. mahalla = (very) early.

70
7.

JOURNAL

OF

PERSIAN

STUDIES

8.

sabdhi = in the morning, tomorrow morning, as in Afghan Persian. Adverbs of time generally end the day after tomorrow; in -z: e.g. i-ap = tonight; marzici, mdrci, mdru today; pusiduli - last night; parrid,peiri = the day before yesterday; pdri = last year; pirdri = the year before last. Cf. also zi = yesterday and nahdri - lunch. sarwa-in = a period of (relatively) cool weather with better pasture after rain-normally in the spring, but may be at any time of the year.

BIBLIOGRAPHY

(Works referred to in the text are marked with asterisk) *1. Afshar, I., 1333 A.H.S. " Diddri az Sistdn-i Veirdna in ", raymd, Vol. VII, no. Io. *2. Baluct Zahg Balad, Baluchi Academy, Karachi, 1963. Intended to serve as a first-year school text-book for Pakistani Baluch children in Pakistan. Standardized Makrani dialect. 3. Bray, Sir Denys de Saumarez, I909, Brahui Language, Calcutta. The standard work on Brdhui. 4. Bruce, R. I., 1874, Manual and Vocabulary the Biluchi of Dialect, Lahore. Eastern dialect. Balochi language, 5. Dames, M. L., 188i, A Sketchof theNorthern containing a grammar, vocabulary and specimens of the language, Calcutta. Eastern dialect. 6. Dames, M. L. (Compiled) 1904, Text-book... of theBaluchi language,translated into English by R. S. Diwan Jamiat Rai, with the assistance of Munshi Dur Muhammad, Lahore. Eastern dialect. *7. Dames, M. L., 1907, Popular Poetry of the Baloches, Royal Asiatic Society. 2 vols. Standard published work on Baluch poetry but transcribed from Eastern dialect. 8. Dames, M. L., I913, " Balochistan " in Encyclopidiede l'Islam. Includes account of Baluch history and language (eastern dialect), and bibliography complete up to 19139. Dames, M. L., 1922, A Text-bookof the Balochi language, consisting of miscellaneous stories, legends, poems, and Balochi-English vocabulary, Lahore. Eastern dialect. 10. Elfenbein, J., 1960, " Baluchistan. B. Language ", in of Encyclopaedia Islam. General, philological account. *I . Elfenbein, J., 1961, " A Baluci Text, with translation and notes " in BSOAS, XXIV, I, pp. 86-103. Popular story in Makrani dialect. * 2. Elfenbein, J., 1963, A Vocabulary Marw Baluchi, Naples. of Etymological glossary of published texts of the Marw dialect. *13. Emeneau, M. B., 1964, " Linguistic Desiderata in Baluchistan" in Indo-Iranica, Milanges presentesa' Georg Morgenstierne,Harrassowitz, pp. 73-7. On the relationship between Baluchi and Brahui. yazyk, Moscow. Latest *14. Frolova, V. A., 1960, Beludisky" and most lengthy (approximately 70 pp.) descriptive grammar; bibliography complete to 1960. *I5. Frye, R., 1331 A.H.S. (a) " Safar-i Biyabdnak va Sistan va Balucistan ", Ddnesh,Vol. II, nos. io/II, Tehran. *16. Frye, R., 1331 A.H.S. (b) "Tatbiq-i lahja-ha-i Khwuri va Baluci ", Mehr,vol. VIII, nos. 3 and 4, Tehran. des I7. Geiger, W., 1890, Etymologie Baluci, Munchen. des 18. Geiger, W., 1891, Lautlehre Baluci, mit einem Anhange fiber Lehnw6rter im Baluci, Miinchen. " *i9. Geiger, W., 1895-1903, Die Sprache der Balutschen " in Grundriss Iranischen der Philologie,Band I, Abteilung 2. These three works of Geiger's may still be considered standard general works on the Baluchi language.
*20. Gershevitch, I., 1959,

21. 22.

23. 24.

25. *26.

27. 28.

29.

30. 3I. 32. *33. 34. 35. *36. 37. *38.

" Travels in Bashkardia " in RCAS, and 4, pp. 213-25. Includes some philological and XLVI, 3 ethnographical material from Bashkardia relevant to Baluchi and Baluchistan. Gershevitch, I., 1962, " Outdoor Terms in Iranian " in A Locust'sLeg, Studiesin honour ofS. H. Taqizadeh,pp. 76-84. Gershevitch, I., 1964, " Iranian Chronological Adverbs " in Indo-Iranica, pp. 78-88. These two works include discussion of some Baluchi and Bashkardi terms. Gilbertson, G. W., 1923, The Balochi Language:a grammar and manual, assisted by Ghano Khan, Haddiani. Gilbertson, G. W., 1925, English-Balochicolloquialdictionary, assisted by Ghano Khan. Both these works are confined to the Eastern dialect. Assisted by Hetu Gladstone, C. E., 1874, Biluchi Handbook. Ram and Mean Jiwan, munshis, Lahore. Eastern dialect. " Grierson, G. A., 1921, Balachi " in The LinguisticSurvey of India, Calcutta, vol. X. Standard work, includes general grammar, word list, irregular verb stems, texts, and indicates main differences between eastern and western dialects. Hetu Ram, Rai Bahadur (Compiler), 1885, Translation of Bilochi-Nama, Calcutta. Includes grammar, verb-list, conversation, stories, vocabulary-in eastern dialect only. of Hughes, A. W., 1877, The country Balochistan,its geography, topography,ethnologyand history; with a map, photographic illustrations,and appendices containinga short vocabulary the of principaldialectsin useamongtheBalochis... London. Leech, R., 1838, Epitomeof the grammarof the Brahuiky,the Balochky and the Panjabi languages . . . Calcutta, Asiatic society of Bengal. Eastern dialect. Lewis, A., 1885, Bilochi storiesas spokenby the nomadtribesof theSulaiman hills, Allahabad. Eastern dialect. Marston, E. W., 1888, Lessons in the Makrani-Baluchi Dialect, Karachi. and Marston, E. W., 1877, Grammar vocabulary the Mekranee of Beloochee dialect, Bombay. Marw-see Elfenbein, 1963. Mayer, T. J. L., 1900-190 Balich Classics, Fort Munro I, and Agra. Eastern dialect. Mayer, T. J. L., 191o, English-Biluchidictionary,Calcutta. Eastern dialect. Misc., see Morgenstierne, 1948. Mockler, E., 1877, A grammar the Baloochee of languageas it is in spoken Makran. . ., London. Morgenstierne, G., I932a, Report on a linguistic mission to N. W. India, Oslo. Starts with observations on Baluchi and Brahui contacts.

NOTES

ON

THE

BALUCHI

SPOKEN

IN

PERSIAN

BALUCHISTAN

71

" " *39. Morgenstierne, G., 1932b, Notes on Balochi Etymology in Norsk tidsskriftforsprogvidenskap, pp. 37-53. Discussion V, of words from his material " which appear to be of genuine Bal. origin, and which are not found in Geiger...". *40. Morgenstierne, G., 1947, "Balochi Language" in the Britannica (also in I957 ed.). Best short general II Encylopaedia account of the language. *41. Morgenstierne, G., 1948, " Balochi Miscellanea " in Acta Orientalia, XX, pp. 253-92. *42. Morgenstierne, G., 1957, " ' Mushroom ' and ' Toadstool' in Indo-Iranian " in BSOAS, XX, pp. 451-5743. Morgenstierne, G., 1958, " Neu-Iranische Sprachen " in Handbuch Orientalistik der Abteilung I, Band 4, Abschnitt I, pp. 155-78. "44. Notes-see Morgenstierne, 1932b. 45. Pierce, E., 1875, " A Description of the Mekranee-Belochee Dialect " in JRAS (Bombay), XI.

*46. Redard, G., 1962, " Le Palmier a Khur. Notes de dialectologie iranienne I", in A Locust'sLeg. Comparative Iranian vocabulary relative to date palm culture. 47. Sokolov, S. N., 1956, " Grammati6eskij ocierk jazyka Beludiej Sovetskogo Sojuza ", in Trudy Instituta Jazykoznanija, VI, pp. 57-91. Marw dialect. 48. Sokolova, V. S., 1953, " Belud'skij jazyk " in OQerki po fonetikeiranskixjazykov,I, pp. 7-77. Concerns phonetics of Marw dialect only. " 49- Weryho, J. W., 1962, Sistini-Persian Folklore" in IndoIranianJournal, V, No. 4. 50. Zarubin, I., 1930, "K izucieniju jazyka i Fol'klora ", in gapiski kollegii vostokovedov, pp. 653-79. belud.skogo V, skazki I (1932), II 51. Zarubin, I., 1932, Beludbskdje (1949). Both restricted to the Marw dialect.

73

THE SHERLEY MYTH By R. M. Savory

Introduction In the spring of 1514 the Ottoman sultan Selim I mustered an army of 200,000 men for the invasion of Persia. The fundamental reason for the outbreak of war was the establishment of the Safavid state in Persia in 1501 by Shdh Isma'il I. One of Shah Isma'il's first acts had been to declare the Shi'i form of Islam the official religion of the new state. His main purpose in taking this step had been to differentiate his dominions from those of the Ottoman Empire, and to create a sense of unity among his subjects. The creation of a militant Shi'ite state in Persia had caused alarm to the Ottomans, the more so because large numbers of Turcoman Ottoman subjects in Anatolia sympathized with Isma'il's pretensions to semi-divinity and supported his claim to the throne of Persia. The active subversion of these elements by the politico-religious propaganda of the Safavids, and the fear that the Safavids, if not checked, might succeed in detaching large areas of Asia Minor from Ottoman control, led the Ottomans to try and crush the Safavid state. By their signal defeat of the Persian army at Chaldiran in north-west Persia on August 22nd 1514, the Ottomans came near to achieving this end. After his victory, Selim occupied Tabriz, at that time the capital of the Safavid state. He himself wished to winter at Tabriz, and complete the conquest of Persia the following year, but his janissary officers mutinied at the prospect, and, eight days after he had occupied Tabriz, Selim was forced to evacuate the city, and to withdraw to Amisiya. Although Walsh' sees the Ottoman victory at Chaldiran as primarily a triumph of logistics, the fact remains that the most important single factor in the Safavid defeat was the Ottoman artillery, supported by hand-guns. Isma'il possessed two commanders, Muhammad Khan Ustijlfi and Nfir 'Ali Khalifa, who had first-hand experience of Ottoman methods of warfare, and these officers urged Isma'il to attack at once, before the Ottomans had had time to complete their dispositions. Isma'il disregarded their advice, and the Ottomans were able to station their musketeers behind a barrier of gun-carriages linked together by chains. Mortars of various sizes were placed on the gun-carriages. This barrier presented an insuperable obstacle to the Safavid army, which was composed mainly or entirely of cavalry. The sources give conflicting accounts of the course of the actual battle. A majority seem to agree that the Safavids gained an initial success on the right, where IHasan Pasha, the commander of the Ottoman left, was killed. The Ottoman centre, however, held firm, and Sinan Pasha, in command of the Ottoman right, brought into action the Ottoman artillery, consisting of 200 cannon and oo mortars. According to Caterino Zeno, Selim, " seeing the slaughter, began to retreat, and to turn about, and was about to fly, when Sinan, coming to the rescue at the time of need, caused the artillery to be brought up and fired on both the janissaries (sic!) and Persians ".2 The effect was devastating. Muhammad Khan Ustajli, in command of the Safavid left, was killed, and his men, after suffering heavy casualties, fell back in disorder. Isma'il rallied his troops, and led them with desperate valour in repeated charges against the Ottoman guns, hacking in vain with his sword at the chains which linked them. From behind the protection of this barrier the Ottoman musketeers directed on the Persian cavalry a deadly fire to which the Persians had no answer. After further heavy casualties had been inflicted on his men, Isma'il was compelled to retreat. The Safavid dead included many leading amirs and high-ranking officers of state, but the Ottomans had not escaped unscathed. Knolles, in his GeneralHistory of the Turks, states that " besides his common footmen, of whom he made least reckoning, he (Selim) lost most part of his Illyrian, Macedonian, Servian, Epirot, Thessalian and Thracian horsemen, the undoubted flower and strength of his army, which were in that mortall battel almost all
of ' J. R. Walsh, article " Cdldirin ", in Encyclopaedia Islam, new edition, vol. II, pp. 7-8.
2

A Narrativeof Italian Travelsin Persia in the15th and i6th Centuries, Hakluyt Society, London, 1873, p. 61.

74

JOURNAL

OF

PERSIAN

STUDIES

slaine or grievously wounded".3 Ayalon is of the opinion that " had the Ottomans not employed firearmson such a large scale at the Battle of Chdldirdn,it is reasonablycertain that their victory-even if they had been able to win-would have been far less decisive. In other words, the Ottomans would have acquired far less Safavid territory in that event, and a much stronger Safavid army would have been left intact to prepare for a war of revenge ".4 TheSherleys The question arises,why did the Safavidsnot use artilleryand hand-guns at the Battle of Childirdn ? The answer traditionally given is that the Persianswere at that time not familiarwith the use of firearms. Firearms, it is alleged, were only introduced into Persia nearly a century later, in the reign of Shah 'Abbas the Great, by two English gentlemen-adventurers,Sir Anthony and Sir Robert Sherley. It is my contention, however, that the Safavids did not use firearms at Chaldiran because they did not choose to use them, not because firearmshad not been introduced into Persia by that date, or because they were unfamiliar with these weapons. There is abundant evidence, both in the European and the Persian sources, that the Persianswere familiar with the use of artillery, and that Persian troops were equipped with hand-guns and skilled in their use, long before the time of 'Abbas I. The Persians had an innate dislike of firearms, the use of which they considered unmanly and cowardly. In this opinion they were supportedby the Maml-iksof Egypt and Syria, who were similarly defeated by the Ottomans a few years later and for the same reasons. In particular, the Persians disliked artillery, which hampered the swift manceuvresof their cavalry. As Minadoi put it, in his Historyof the Warsof the Turksand thePersians,published in Rome in 1587, " the people of Persia are afrayde of Artillery beyond measure, and yet sometimes they have not been afraide with sudden assaults to assaile their enimies trenches, and lodgings in their Campes. And although they be so timorous and fearefull of that Engine, and yet know of what moment it is in a battel; yet have they not hitherto received the use thereof, being rather obstinate in their blind ambitious conceite, that it is a sinne and shame to exercise so cruell a weapon against mankind, than ignorant how to make it, or destitute of matter to cast it ".5 In other words the Persians, like the Mamlilks, preferrednot to take cognizance of military developments in Europe and the Ottoman Empire during the fourteenth and fifteenth centuries. The first cannon is said to have been cast at Florence in 1326, although Rathgen, claims that the Germans had produced one a few years earlier." The im in his Das Geschiitz Mittelalter, used cannon at Crecy in 1346,7 Spain had artillery by 1359 (if not earlier), and Portugal by English 1385. The Ottomans probably had artillery by about 1424, and the important part played by the Ottoman heavy guns at the Siege of Constantinopleis well known. Hand-guns, mainly matchlocks or arquebuses, in general came in somewhat later than cannon. The earliest instances of the use of genuine matchlocks, as opposed to devices for the discharge of naphtha bottles and other incendiary missiles, seem to date from the early fifteenth century. The flintlock did not come into use until a century later, about 1510. Leonardo da Vinci was responsiblefor the original design. Hand-guns seem to have been in use in the Ottoman army from about 1440, or perhaps a little later. Yet they were not introduced by the Mamliks for another fifty years. Arquebuses are first heard of in the Mamliik Empire about 1490, during the reign of Sultan al-Ashraf Qaytbey. Even then, they were not used effectively or on an adequate scale, and in 1516 an Ottoman army of 12,ooo arquebusiersrouted a
much larger force of Mamlfks. It is inconceivable that the Safavids should have been unaware of these developments in Asia Minor, Syria and Egypt during the latter half of the fifteenth century. Certainly their predecessors, the Aq Qoyinlf or White Sheep Turcoman dynasty, rulers of Mesopotamia and the whole of Persia except Khurisin, were well aware of them. In 1471, for instance the Seigneur of Venice sent the White Sheep
3 Ibid., p. 62, note I. 4 D. Ayalon, Gunpowder and Firearms in the Mamlak Kingdom, pp. Ig9-I Io. 5 Quoted by L. Lockhart, " The Persian Army in the Safavi Period ", in Der Islam, Band 34, p. 90. 6 Berlin 1928. Quoted in J. R. Partington, A Historyof Greek Fire and Gunpowder, Cambridge, 1960, pp. 96 ff. 7 This is denied by Rathgen, op. cit.; quoted in Partington, op. cit., p. 10o7.

THE

SHERLEY

MYTH

75

ruler Uzun IHasan six bombards, 600 culverins (spingarde),matchlocks (schiopetti),etc. This consignment was intercepted at Cyprus, and never reached Tabriz. Two years later, in 1473, when Aq Qoyiinlf aspirations for westward expansion were finally frustrated by their defeat at the hands of the Ottomans on the Upper Euphrates, the principal cause of their defeat was the Ottoman artillery. We know that the White Sheep Turcomans drew the obvious conclusion from this defeat, for a few years later, in 1478, the Venetians sent to Uzun Hasan " Ioo artillerymen of experience and capacity ".s Unfortunately, we do not know what became of these men, but we do know that their efforts bore some fruit. The Tdrikh-i AmZnz, Persian chronicle completed in 1490, contains several references to the use a of firearms by the Aq Qoyinl- in the last quarter of the fifteenth century. Here, for instance, is a description of a battle fought on July 15th 1478 between two rival Aq Qoyfnli princes: " First Ya'qtb's skirmishers attacked and reached Khalil's artillery (tzp), but, as Khalil had strengthened his centre with cannon and hand-guns, the gunners scattered the skirmishers with their hand-guns ".9 In 1485, the Aq Qoyfnlf prince Ya'qfib was laying siege to the fortress of Akhisqt (Akhaltsikhe) in Georgia: " Such a fortress could not be taken by means of (ordinary) warfare. The heavy artillery (tap-i girdn) was battering the walls hourly, but when the structure was about to collapse, the infidels repaired the damage with iron implements and solid material ,".' In 1487 the Aq Qoyinli commander reported to the ruler on " the preparations made for the recapture of Tiflis, such as the building of artillery ". The same year, Aq Qoyinl- forces laid siege to a fort near Tiflis, but "the destruction caused by the artillery (tap) did not bring the garrison to their senses ".11 Obviously at this stage siege guns were not very formidable. In 1489, Aq Qoyiinlfi forces laid siege to Tiflis itself. The Georgians abandoned the town and took refuge in the castle. " The fire of war was raging, and the cauldrons of the guns (qazghdn-i tap) were at the boil." But the besiegers were held up by the fact that " all the artillery equipment (by which most probably is meant the wooden gun-carriages) was destroyed by fire ".12 But most significant of all, in view of later events, is the statement that the Safavid leader IHaydar himself used cannon (tap) at the Siege of Gulistan in 1488.13 Minadoi is quite right, the absence of Persian artillery at Chtldiran was not due to ignorance, but to the " blind ambitious conceite " of the Persians in their own superiority. During the sixteenth century, the backbone of the Persian army was the qizilbdsh troops. The term qizilbdsh, " redheads " derived from their distinctive red hats with twelve folds commemorating the twelve Shi'i im5ms. Originally a pejorative term used by the Ottomans, who delighted in such jingles as qizilbash-i awbdsh, " scoundrelly qizilbdsh ", etc., it was adopted by the qizilbish as a mark of pride. The use of the term " Roundhead " affords an obvious parallel. These qizilbdsh troops were not Persians at all, but men of Turkish, or more accurately Turcoman, stock. Although Isma'il, the founder of the Safavid dynasty, derived his support in part from the local populations of north-west Persia, the dlite of his fighting men was drawn from the Turcoman tribes of Asia Minor, Syria and the Armenian highlands. These men, who were essentially cavalry troops, looked, as Minorsky rather inappropriately puts it, " like walking arsenals ". They wore light armour, and were armed with bows and arrows, swords, daggers and sometimes battle-axes as well. They were expert horsemen, skilled in firing from the saddle, and they excelled in cavalry warfare. The qizilbdsh troops were levied on a feudal basis, for there was no standing army until the time of Shdh I. Before his reign, it was the responsibility 'Abbts of the military governors of the provinces to furnish the Shih with sufficient troops in time of need. The qizilbdsh had a strong esprit de corps, and constituted a formidable fighting force which was respected even by the Ottomans, and which was appreciated by those Europeans who had first-hand experience of it. Busbecq, Ferdinand's ambassador at the Ottoman court, reported: " 'Tis only the Persian stands between us and ruin. The Turk would fain be upon us, but he keeps him back." In addition to the qizilbash troops, which constituted the greater part of the Safavid armies during the sixteenth century, there were various contingents of local militia recruited on an ad hoc basis. Many of these local units were infantry.
8 See my article " Bdrfid. V. The Safawids ", in Encyclopaedia of Islam, new edition, vol. I, p. IO67. 9 V. Minorsky, Persia in A.D. 1478-9o, London, 1957, p. 36. 1oIbid., p. 51.
I, 12 Ibid., p. 13 Ibid., p.

Ibid., p. 89. 90. 74-

76

JOURNAL

OF

PERSIAN

STUDIES

In broad terms, this picture of Safavid military organization obtained until the reign of Shah 'Abbas the Great (1588-I629), a monarch worthy to be ranked with other great rulers of the sixteenth century, such as Elizabeth I of England, the Mughal Emperor Akbar, and the Ottoman sultan Stileyman the Magnificent. Shah 'Abbas created a standing army consisting of: first, a cavalry corps of ghuldms,that is, Georgian, Armenian and Circassian prisoners, converts to Islam. The strength of this corps was eventually increased to Io,ooo men; second, a personal bodyguard for the Shah, numbering 3,000 men, also ghuldms; third, a regiment of artillery, with 12,000 men and 500 guns; and fourth, a regiment of musketeers, also I2,000 strong, recruited mainly from the Persian peasantry. 'Abbas thus had a standing army of some 37,000 men, paid directly from the Royal Treasury, and owing allegiance only to him. It is at this point that the Sherley brothers come upon the scene. In the first place, on the strength of a letter written by the traveller Pietro della Valle on April 22nd 1619, Sir Anthony Sherley is credited with giving Shah 'Abbas the idea of creating a regiment of musketeers. This letter states that the regiment was created by Shah 'Abbas " a few years ago " on the advice of Sir Anthony Sherley. Secondly, apparently on the basis of the statement of Sir Anthony's steward, Abel Pingon, that at the time of his arrival the Persians had no artillery at all, and in view of the fact that Sir Anthony's party of twenty-six persons contained " at least one cannon-founder ", the Sherleys are credited with the introduction of firearms into Persia. Before these claims are refuted, it is necessary to place the Sherley brothers in their context. In December 1598, Shah 'Abbas returned to his former capital, Qazvin, after a victorious campaign in Khurasan against the Ozbegs (the powerful Ozbeg confederacy in Transoxania constituted one of the two principal threats to the security of the Safavid state). On his arrival in Qazvin, Shah 'Abbas received a party of twenty-six Europeans led by the English soldiers-of-fortune Sir Anthony and Sir Robert Sherley. The Shah took this party with him to Isfahan, the new capital of his kingdom. Five months later, in May 1599, Sir Anthony was despatched by Shah 'Abbas to Europe, bearing letters of friendship from the Shah to the Pope and to various European princes, including the Holy Roman Henry IV of France (1589-i6IO), Philip III of emperor Rudolf II, King of Bohemia (1552-1612), the King of Scotland, the King of Poland, the Queen of England, the Seigneur of Spain (1598-162 ), Venice, and the Grand Duke of Tuscany. Sir Anthony was charged with the task of enlisting the support of these princes against their common enemy, the Ottomans. His brother Robert remained behind at the Persian court, ostensibly to console the Shah for the absence of Anthony, but in reality as a hostage for Anthony's good behaviour. Anthony was accompanied on his mission by a Safavid officer, IHusayn 'Ali Beg Bayat. We are not primarily concerned here with Anthony Sherley's adventures in Europe, fascinating though these are, but the main points must be mentioned: after six months in Moscow, Anthony embarked at Archangel early in the year 16oo, sailed to Emden, and in October 16oo, reached Prague where he was received by the emperor Rudolf II. This circuitous route was of course necessitated by the fact that the Ottoman Empire lay astride the regular routes between Europe and Persia. From Prague he went to Rome (April 1601), where his fellow ambassador, .Husayn 'Ali Beg Bayat, who had constantly quarrelled with Anthony as to which of them should take precedence at the various courts they visited, dissociated himself from Sherley and went on alone to Spain (May 29th I6oI). Sir Anthony's fortunes declined from this point. In March 1602 he went to Venice, and from there he carried on a correspondence with the King of Spain. Some of his letters were intercepted by English agents, and were deemed treasonable. His request to be allowed to return to England was peremptorily refused, and English ambassadors abroad were instructed to repudiate him. Financial troubles began to worry him. In April 1603 he was arrested in Venice and imprisoned, either as an insolvent debtor or as a seditious person-it is not clear which. He was released after the accession of James I, and was granted a licence by the British Government to " remain beyond the sea some time longer ". This document further recommended him to the consideration of the " princes and strangers by whom he might pass ". In the spring of 16o5 he was sent by the emperor Rudolf to Morocco to report on the state of that country. In Morocco he ransomed some Portuguese prisoners and sailed with them to Lisbon, hoping to reap the handsome reward. Unhappily, he did not even recover the large sum of money which he had laid out on this enterprise. From Lisbon he went to Madrid, where he was placed in command of an

THE

SHERLEY

MYTH

77

armada bound for the Levant to attack the Turks. Reaching Naples in July 1607, he made a flying visit to Prague, and was made a Count of the Empire by Rudolf, and dubbed himself " Admiral of the Levant Seas ". But the expedition ended in disaster. In 1609 he led his forces in a futile descent on the island of Mitylene, and was dismissed from his command. Anthony's failure either to return to Persia or to report on the progress of his mission naturally imperilled Robert's position at Isfahan, and it says much for Shah 'Abbts's sense of justice that he did not vent his displeasureon Robert to a greater extent than he did. As it was, after sufferinga temporary loss of royal favour, Robert was reinstated, and in 1608, nearly ten years after Anthony had set out for Europe, he was despatched by 'Abbas on a mission similar to that of his brother, with the additional task of finding Anthony and reporting to the Shah on the success or otherwise of his mission. Robert caught up with his brother in 161I in Madrid, to which Anthony had returned discredited and in abject poverty. After fruitlessrecriminationswith his brother, Robert went on to England, where he was received by James I, but all his efforts to establish trade relations between England and Persia were thwarted by the powerful lobby of Levant merchants, and Robert returned to Persia in 1615. Later the same year he was sent back to Europe on a second mission which lasted until 1628. This second absence proved to be too long, for powerful enemies at the Persian court, who resented his influence with the Shah, secured his downfall, and he died shortly afterwards in penurious circumstances. The Shah, lamenting his death, said that Robert had done more for him than any of his native subjects. Anthony still eked out a meagre existence on a pension of 300o ducats a year from the King of Spain -most of which went to defray his heavy debts. Calling himself the Conde de Leste, he remained at Madrid until his death about 1635. This, in the barest outline, is the history of the missionsof Anthony and Robert Sherley in Europe. In this article, however, we are more particularly concerned with the claim made on their behalf that they were responsible for the introduction of firearmsinto Persia, and that they played a leading part in the reorganization of the Persian army during the reign of Shah 'Abbas, and in the training of the new regiments then created. Who is responsible for making this claim? There is little doubt that the man most to blame is his Samuel Purchas, the enthusiastic editor of Purchas Pilgrimes Hakluytus or Posthumus, published in ten volumes in 1625. The praise lavished on the Sherleysby Purchasis fulsome in the extreme: " Amongst our English Travellers, I know not whether have merited more respect than the Honorable, I had almost said Heroike Gentlemen, Sir Anthony and Sir Robert Sherleys. And if the Argonauts of old, and Grecian Worthies, were worthily reputed Heroicall for European exploits into Asia: what may we think of the Sherley-Brethren, which not from the nearer Grecian shoares, but from beyond the toto have not coasted a little way (as did those) but European World-Et penitus divisisOrbe Britannis;14 the very bowells of the Asian Seas and Lands, unto the Persian centre: and that not by a pierced combination of Princes, as those former; but (God directing their private Genius for publike benefit) to kindle a fire betwixt the two most puissant of both Asian and Mahumetan Princes, that by their division and diversion of Turkish invasions, Christian Princes, Countries and States might be indebted to their private undertaking: nor hath ten yeeres detained them at one Citie, or one Voyage finished their worthy indevours, as theirs at Troy and to Colchos; but about twentie yeeres altogether, all Turkie hath
groaned, in which shee hath lost two millions of her unhallowed children; remoter India, Moscovia, Africa hath felt the Sherleian working, Poland, Spaine, the Emperour and Pope have admired and adorned the English name of Sherley .... Who ever since the beginning of things and men, hath beene so often by Royal Employment sent Embassadour to so many princes, so distant in place, so different in rites ? Two Emperours Rudolf and Ferdinand, two Popes Clement and Paul, twice the King of Spaine, twice the Polonian, the Muscovite also have given him audience. .. . in terrisSherlii nonplena .uae regio laboris? ". And, finally, the substance of the claim: " The mightie Ottoman, terror of the Christian world, quaketh of a Sherly-Fever, and gives hopes of approaching fates. The prevailing Persian hath learned Sherleian Arts of War, and he which before knew not the use of Ordnance hath now 500oo Peeces of Brasse, and 60,000ooo Musketiers; so that they which at hand with the sword were before dreadfull to
,4 See over.
6A

78

JOURNAL

OF

PERSIAN

STUDIES

the Turkes, now also in remoter blows and sulfurian arts are growne terrible."'4 This, then, is the grandiose claim advanced by Purchas on behalf of the Sherleys. On what evidence did he base it ? In support of Purchas, we have the statement of Sir Anthony's steward, Abel Pingon, already referred to, that at the time of the arrival of the Sherleys the Persians had no artillery at all.15 On the other hand, Anthony's interpreter, Angelo, stated in Rome in November 1599 that Shah 'Abbas did have some cannon, " having captured many pieces from the Tartars. Moreover, there is no lack of masters to manufacture new ones, these masters have turned against the Turks and have come to serve the King of Persia." Angelo further stated that Shah 'Abbas had 50,000 arquebusiers, and added: " at one time the King did not use arquebusiers, but now he delights in them."'" It seems unlikely that a corps of 50,000 men could have been organized during the (at most) four months Sir Anthony spent in the Persian capital. Our case does not, however, rest on mere surmise. George Manwaring, one of Anthony Sherley's companions, the sobriety of whose account encourages one to take his True Discourseat its face value, states explicitly that the Persians were already " very expert in their pieces or muskets; for although there are some which have written now of late that they had not the use of pieces until our coming into the country, this much must I write to their praise, that I did never see better barrels of muskets than I did see there; and the King hath, hard by his court at Aspahane, above two hundred men at work, only making of pieces, bows and arrows, swords and targets ". The similarity in the wording tempts one to think that this may be a direct refutation of Purchas, whom Manwaring perhaps hesitated to mention by name because of the prestige of the Pilgrimes. Whether this is so or not, it is extraordinary that Purchas should have ignored what Anthony Sherley himself said on the subject. Referring to the Persian army which defeated the Ozbegs in 1598, Anthony states: " thirty thousand men the King tooke with him for that warre, twelve thousand Harquebusiers which bare long pieces, halfe a foot longer than our muskets, sleightly made . . . which they use well and certainly."" So the claim that the Sherleys introduced firearms into Persia-a claim, significantly, never made by the Sherleys themselves, is refuted by the testimony of Anthony Sherley himself and by that of a majority of those of his companions who have left us accounts of their journey. But even if we did not have this evidence, there is no lack of other evidence, in both the Persian and European sources, to prove that firearms were known and used in Persia nearly a hundred years before the advent of the Sherleys. Twenty-seven years before the Sherleys reached Qazvin, the Venetian ambassador d'Alessandri wrote: " they (the Persians) use for arms swords, lances, arquebuses, which all the soldiers can use; their arms are superior and better tempered than those of any other nation. The barrels of the arquebuses are generally six spans long, and carry a ball a little less than three ounces in weight. They use them with such facility, that it does not hinder them drawing their bows nor handling their swords, keeping the latter hung at their saddle-bows till occasion requires them. The arquebus is then put away behind the back, so that one weapon does not impede the use of another."'s The earliest European reference to the use of hand-guns in Persia can be dated about 1507, that is, ninety-one years before the arrival of the Sherleys, and there are frequent references in the Persian sources to the use of hand-guns from 1520 onwards. For instance, in 1520 hand-guns were used by the Safavid garrison at Harat; infantry units equipped with hand-guns constituted part of the Safavid garrison at Harat in 1523; the royal army led by Shah Tahmasp against the Ozbegs in 1528 included detachments of musketeers, apparently organized on a tribal basis;19 and so on. Artillery was accepted by the Persians even more unwillingly than were hand-guns. Apart from their aversion to firearms in general, the Persians, as mentioned earlier, particularly disliked cannon, because, as Lockhart has pointed out, the Persian army traditionally relied on " extreme mobility and the ability to carry out swift tactical manoeuvres. The cannon of those days were extremely heavy and cumbrous weapons ".~20 In a pitched battle with the Ozbegs near Mashhad in 1528, Shah Tahmasp
'4 An echo of Virgil, Eclogues, i, 67: Et Penitus toto divisos Orbe '7 Ibid., p. 29. Britannos. I8 Ibid., p. 222. 19 See my article " Birad ...", I5 Purchas, op. cit., X, pp. 374-6.
20

referred to above.

z6Sir E. Denison Ross, Sir Anthony and Sherley His PersianAdventure, London, 1933, P- 163-

Op. cit., p.

90.

THE

SHERLEY

MYTH

79

stationed wagons in front of his army. On these wagons were mounted a type of light cannon called darbzan. He also employed in this battle guns described as " Frankish cannon ". In this battle, the Persian gunners were unable to bring their guns to bear, because the Ozbegs did not attack from the direction anticipated.21 This seems to have discouraged the Persians from persevering with field artillery, and we hear no more of it in the Persian sources until the time of Shah 'Abbts I. In general, little attempt was made to emulate the Ottomans in the use of artillery in the field. There is plenty of evidence, however, in the Persian chronicles, that artillery was regularly used in siege warfare from 1528 onwards. We read, for instance, of the presence of a master-gunner in the Safavid army which laid siege to Damghan in 1528, and we are told that in 1538 Safavid artillery fire destroyed the towers of a fort in Shirvan. In I539 we hear for the first time of the existence of a tfipchi-bdshi, commander-inor chief of artillery. " Frankish cannon " were used again in Shirvan in together with a type of 155I, cannon called bddlj",and mortars, which are mentioned for the first time.22 Since none of Sir Anthony's companions claims that the Sherleys either instructed the Persians in the casting of cannon, or were responsible for the creation of the regiment of musketeers, how could Purchas have built up this fabulous picture of the brothers ? The answer is to be found, I believe, partly in the character and upbringing of Anthony himself, and partly in those of the worthy Purchas. It is clear that Anthony and Robert were completely dissimilar in character and temperament. Robert impressed everyone he met as a man of integrity. The British ambassador at Madrid, who met him there in 1611, describes him as of " wise and discreet carriage, and both modest and moreover grave in his speech, diet and expenses ".23 During the long years at the Persian court waiting for news of Anthony, there is little doubt that Robert employed himself usefully in directing the drilling and training of the new regiments formed by the Shah. In fact, Shah 'Abbas appointed Robert " Master General against the Turks ". In this capacity he commanded Persian troops in battle against the Ottomans in the campaigns of 1604 and I605, and distinguished himself by his valour. Anthony, however, was cast in the mould of his elder brother Thomas, who, after a chequered career as a gentleman-adventurer and privateer, gave up the latter profession to become a Member of Parliament. Anthony was awarded his B.A. at Oxford in 158 1, when he was sixteen years old, and was made a Probationer-Fellow of All Souls. As he was a kinsman, through his mother, of the founder of All Souls, Archbishop Chichele, he readily obtained leave of absence from his academic duties to engage in military service. He fought in the wars in the Low Countries under Robert Dudley, Earl of Leicester, and in 1591 went to Normandy with the Earl of Essex in support of Henry of Navarre. The Earl of Essex made a great impression on Anthony, then about twenty-six years of age. Anthony desired to make the Earl " the pattern of his civil life, and from him to draw a worthy model of all his actions ".24 While in France, Anthony was made a Knight of St. Michael by Henry IV, but on his return to England he was promptly thrown into the Fleet prison for his temerity in accepting such an honour without the Queen's permission. Elizabeth admonished him in her usual mordant style: " As a virtuous woman ought to look on none but her husband, so a subject ought not to cast his eyes on any other sovereign than him God has set over him. I will not have my sheep marked with a strange brand, nor suffer them to follow the pipe of a strange shepherd."25 Anthony was released, on condition he surrendered his knighthood. Nevertheless, he continued to allow himself to be commonly known as " Sir Anthony ". An unhappy marriage to a cousin of the Earl of Essex led Anthony " to undertake any course that might occupy his mind from thinking on her vainest words ",26 and he returned to the life of a gentleman-adventurer. In 1598, at the invitation of the Earl of Essex, he led a small company of twenty-five men to assist the bastard son of the Duke of Ferrara against the Pope. On his arrival in Italy, he found that affair settled, and, casting around for something to keep his men occupied, he decided, after a chance meeting with a Persian merchant, to go to the court of Shah 'Abbas. He had two objects: to persuade the Shaih to ally himself with the Christian princes of Europe against the 21 cit., p. 90o. Op.
22 23

24 25

See my article " Brfid ...". Ibid., p. 80.

Dictionaryof National BiographyXVIII, London, 1909, p. 136. Ibid., p. 121. 26Gentleman's Magazine, New Series XXII, 1844, PP. 474-5.

80

JOURNAL

OF

PERSIAN

STUDIES

Ottomans; and to promote trade between England and the East. Anthony attempted to father his decision to proceed to Persia upon the Earl of Essex, but the English Government was not consulted, and viewed his enterprise with suspicion. The fact that Anthony had no official sanction doomed his mission to failure. As we have seen, he was later refused permission to return to England, and was repudiated by the English Government. At first, however, all went well. After drawing on the Earl of Essex's credit with English merchants to the tune of ?400 at Constantinople and ?500 at Aleppo, he arrived in Persia, where he was well received by the Shdh-though not so well as he would have us believe, otherwise why in all the contemporary Persian chronicles is there only one reference which may possibly allude to the Sherleys? It is true that the Shah issued to Anthony afarmdn, granting in perpetuity to all Christian merchants freedom from customs, religious liberty, and the right to trade to all parts of the Shah's dominions; but no immediate advantage was taken of this concession, owing, as we have seen, to the hostility of the Levant Company. There is no doubt that Anthony was a plausible rogue. He was also a born intriguer. Wadsworth, in his English and Spanish Pilgrim (1625), summed him up thus: " This Sir Anthony Sherley is a great plotter and projector in matters of state, and undertakes by sea stratagems, to invade and ruinate his own country, a just treatise of whose actions would take up a whole volume.""27 Samuel Purchas, and to some extent Pietro della Valle, were taken in by his pretensions. Not so R. B. Wernham, the editor of State Papers of the Elizabethan period, who says, with reference to an earlier exploit of his: " On the whole, I have the impression that the story of Sherley's adventures is a fictitious account fabricated in Then 1622 rather than a genuine, though muddled, account of actual events in the years 1586-7." the testimony of the eccentric Englishman Thomas Coryat, born at Odcombe in Somerset, who we have walked 3,300 miles from Aleppo to India, and hence gave himself the bizarre title of the " Odcombian leg-stretcher ", and who published Coryat's Cruditiesin 1611. Coryat met Robert Sherley in 1615, somewhere between Isfahan and the Indian frontier. He comments: " It is said, on slender evidence, that Robert Sherley played an important part in reforming the Persian army and equipping it for the first time with artillery."28 Sir John Malcolm, author of the classic History of Persia which appeared in 1807, when the Sherley myth was well established, will go no further than to say: " the new corps of infantry raised by Abbas to make himself independent of his turbulent chiefs, and to oppose the Turkish janissaries, probablyowed their discipline to the counsel and aid of the two Sherleys and their military followers. We are told, indeed, that they not only formed this force, but taught the Persians the use of artillery." And in a footnote Malcolm quotes Purchas, with the remark: " the following passage from A Memorial of the Travels of Sir RobertSherley,written by a contemporary, appearsto prove this fact.""29 In other words, remove the corner-stone of Purchas's account, and the whole edifice collapses. As for Anthony's behaviour after he left Persia on his mission to Europe, even the Gentleman's Magazine of 1844, which has a glowing account of the Sherleys beginning with a series of noble apostrophes: " Those three brave Sherleys! Each separate history a romance! How proud must the old knight their father have been, living at Wiston with his noble sons! ", etc., is unable to find any excuses for Anthony's desertion of his brother and disloyalty to the Shah. After saying that " what appears to be most unaccountable and mysterious is, that not a word is ever said of Sir Anthony's return to Persia; Persia, the Shah, and his younger brother, seem to have passed away from his mind ", it concludes charitably: " one must suppose that many links are wanting in this part of his story ".30 Most telling of all, perhaps, in regard to the claim that the Sherleys instructed the Persians in the use of artillery, is Anthony's own naively frank account of his early meetings with the Shah: (the Shah), he says, " had divers discourses with me, not of our apparell, building, beautie of our women, or such vanities; but of our proceeding in our Warres, of our usuall Armes, of the commoditie and discommoditie of our Fortresses, of the use of Artillerie, and of the orders of our government: in which, weregreaterthan myjudgment,yet I had that felicitie weresuch, that I knew my errours thoughmy unskilfulnesse of a good time, that I gave him good satisfaction, as it seemed ".31
27Dictionaryof National Biography,loc. cit., p. 12 I. 28Ibid., p. 123. 29 Malcolm, History of Persia, London, 1815, vol. I, p. 533, note I.
30 Vol. I, p. 351. 3' Op. cit., p. 474; P. 479.

THE

SHERLEY

MYTH

8I

It is clear, then, that Anthony was an impulsive opportunist who was not averse to embroidering the facts to gain an advantage. Is there any evidence to show why Purchas should have magnified the Sherleys' role to such an absurd degree ? There is no direct evidence on this point, but there is evidence that Purchas was not a reliable editor. Unlike his predecessor Hakluyt, a grave, sensible and restrained man, whose labours were " animated by a desire to benefit his fellow-countrymen by the promotion of commerce and manufacture and the spread of colonisation ", Purchas was " very far indeed from a faithful editor or a judicious compiler ", and " he took little pains to arrive at a accurate knowledge of the facts " (thus Sir John Laughton in the Dictionary of National Biography).32 A lot of Purchas's work was based on the copious materials collected by Hakluyt and unpublished at the time of the latter's death in 1616, but volume X, in which the Sherley eulogy occurs, contained, according to Purchas, material which came to hand later, and " is therefore a supply to all than any well-ordered part of the worke ".33 In other words, the Sherley eulogy formed no part of the material carefully sifted by Hakluyt. Moreover, Purchas admits that he is attracted by the strange and the marvellous: " My genius delights ", he says, " rather in by-wayes than high-wayes."34 We know that Purchas had met Robert Sherley. It needed little to fire Purchas's imagination, and, once fired, it expressed itself in the extravagantly exaggerated style of which we have had adequate examples.35 The question remains, if the Sherleys were not responsible for the introduction of firearms into Persia, who was? The answer is, I think, first, the Venetians, and second, the Portuguese. The important trading centre of Hormuz in the Persian Gulf was in Portuguese hands from 1515, when it was seized by Albuquerque, until 1622, when a combined force of Persians and English drove the Portuguese out. We are told that in 1548 the Portuguese furnished Shah Tahmdsp with Io,ooo men and twenty cannon at the time of Siileyman the Magnificent's second invasion of Persia, but the number of men at least seems greatly exaggerated. Herbert, in his A Relation of Some Teares Travaille (1634), states that the Persians " got the use of cannon from the vanquised Portugal ", and the Spanish ambassador Don Garcia de Silva Figueroa, in his Commentarios la Embajada al Rey Xa Abas de Persia, de first published in the first half of the seventeenth century, states that the Persian artillery was manipulated by Europeans " and particularly by the Portuguese ".36 The Rev. John Cartwright, in The Preachers Travels, published in 1611, describing the siege of Tabriz by Shah 'Abbas in 1603, states: " in which siege he for battery used the helpe of the Canon, an engine of long time by the Persian scorned, as not beseeming valiant men, untill that by their own harmes taught; they are content to use it, being with the same, as also with skilfull canoniers, furnished by the Portugals from Ormuz."37 On the other hand, Professor Boxer, in a personal communication, gives it as his opinion that " during the 16th-I7th centuries the Portuguese were very poor gunners, notoriously so, in fact, and the overwhelming majority of their skilled men as gunners, whether ashore or afloat, were Germans, Flemings, and Italians ". As a postscript, in making an assessment of Anthony Sherley, the following point may be thought not to be irrelevant: in 1888 the Rev. Scott Surtee, of Dinsdale-on-Tees, published a pamphlet entitled William Shakespeare Stratford-on-Avon, which he assigned to Sir Anthony Sherley the honour of in of writing Shakespeare's plays.~8

32Purchas, op. cit., IX, p. 415. 33Quoted in HakluytSociety,series II, no. XCIII, p. 49. 34Ibid., p. 57. 35Purchas, op. cit., I, p. xliii.

36 See my article " Bdrfid ..." 37Quoted in Purchas, op. cit., IX, p. 503. 38Dictionaryof National Biography,loc. cit., p. I24.

83

SOME ANCIENT METAL BELTS: THEIR ANTECEDENTS AND RELATIVES By P. R. S. Moorey

In a country like Iran, where the output of ancient metalsmiths never ceases to surprise by its inevitably involves a whole vigour and variety, consideration of any individual items in their repertoire complex of problems.' The three objects to be discussed in this paper are united by function and to some extent by form, but in style they reflect three distinct cultural milieu. It is hoped that by examining each piece in detail some light may be thrown first on the interrelation of metal workshops in Western Iran in the period immediately preceding the genesis of the Achaemenian style and secondly, in considering the silver gilt belt, on the Achaemenians' legacy to the Asiatic periphery of their Empire. As no general survey of evidence for the history of the metal belt in the ancient Near East is at present available, the main discussion is prefaced by a brief review of evidence for its development in this area before the eighth century B.c., to set the objects under consideration in their widest context.

I Evidence for the history of the metal belt takes three main forms : belts found in situ on skeletons, pictorial or sculptural representations of belts and detached fragments of sheet metal. Except when found in situ the evidence is usually far from satisfactory. Representations of belts in art, however skilfully drawn, rarely make clear beyond all doubt whether the belt is made of metal or fabric and usually leave details of fastening and decoration obscure. Detached metal plaques may well have served a multitude of functions and their use as belt fittings is never more than one among a number of possibilities. The same is to some extent true of continuous strips of sheet metal and it is here that considerations of size play an important part in their identification. The great length and height of certain " belts " reported from Urartian sites indicates that they are better identified as decorative fittings for furniture or chariots.2 It is less easy to define the dimensional limit between a " diadem " and a " belt ", when there is clear evidence from Crete in the second millennium B.c.3 that extremely constricting metal belts were quite commonly worn. Wasp-waisted statuettes from Western Asia suggest that metal belts were quite often short and that a length between sixteen and twenty inches does not necessarily indicate a " diadem ", particularly when the band of metal is three or more inches high. The ends of metal belts may often not have met round the waist. As with so many metal artefacts later to become characteristic of the Western Asiatic tradition, the earliest available evidence for metal belts was found in two late Early Dynastic III graves at Ur. In grave PG/58o " lay what had been a belt or baldric of leather plated with silver "4 attached to it was a mounted cylinder seal (U.934I) and a gold dagger with its sheath (U.936 I). In the grave of Mes-kalamdug (PG/755) were the traces of an identical belt which had once carried a cylinder seal, a dagger and perhaps a whetstone and toilet reticule.5 These two graves establish at the outset a close link between the metal belt and military equipment. In Sumer, from at least the late prehistoric period, the woven girdle appears to have had a distinct ritual significance6 and even in the early historic period the multiple woven girdle appears only on nude statues, in metal and stone, designed to carry or support objects as part of the temple furniture.' Although they carry military equipment, the nude male
i My thanks are due to R. W. Hamilton, Keeper of the Ashmolean Museum, for permission to publish these belts and for his comments on a draft of section III ; and to my colleague Dr. H. W. Catling for discussing the silver gilt belt with me. The drawings were done by Mrs. P. Pogson. 2 R. W. Hamilton, A.S., XV, 1965, PP- 50-1. 3 A. Evans, The Palace of Minos, III, pp. 446 ff. 4 C. L. Woolley, The Royal Cemetery, Excavations,II (henceUr forth U.E.II), p. 51, pl. I3b.
5 Ibid., p. 156, fig. 35-

6 B. L. Goff and B. Buchanan, J.N.E.S., 15, 1956, p. 231. 7 H. Frankfort, Sculptureof the Third Millennium B.C.,..., p. i iff., pl. 26-7, 98-o103.

84

JOURNAL

OF

PERSIAN

STUDIES

figures of copper in the Judaidah hoard" wear belts which were probably made of leather or fabric (rather than of sheet metal) like the identical ones on the virtually contemporary stone and metal " caryatids " from Tell Asmar and Khafajah. The next stage in the history of the metal belt is confused by the fact that certain objects found out of context have been identified as metal belts when they are more probably " diadems ". At Byblos, in found a very late third millennium B.c. contexts-the Egyptian First Intermediate Period-Montet fine complete rectangular strip of silver with both ends rolled back and fragments of others in " foundation deposits ".9 It is clear from their size and form that these pieces are probably not belts. The discovery of an identical object of copper still in place on the skull of a skeleton in a ninth dynasty burial at Qau in Egypt endorses their identity as " diadems ".'~ Such objects were not confined to Egypt and the Levant, for the ruler portrayed in the well-known bronze head from Nineveh" wears such a diadem, with the ends rolled back, to support the large curls into which his hair is dressed across the brow. A number of sheet metal fragments from Alaga Huyuk in Anatolia may have served a similar function." True metal belts may be represented by thin gold sheets on two silver statuettes of the early second millennium B.c. found by Schaeffer at Ras Shamra.'3 The first evidence for the round-ended metal belt, which was to become the most common type throughout the Near East, was found in Tomb J3 at Jericho.'4 This fragmentary bronze belt is decorated with a repousse' pattern of concentric circles and bosses. It was fastened by passing a tie through metal set back from either end of the belt. The decoration is almost certainly a reproduction entirely loops in metal of a leather belt fortified with circular metal discs.'5 A better preserved, but plain belt, of the same general type was found in tomb A at Tell Far'ah in Jordan.'6 Though that at Jericho is slightly earlier than the one at Far'ah, both graves belong to the local Middle Bronze Age IIB, c. 1750-1650 B.c. Both belts were associated with the burial of a single individual whose equipmenta battle-axe and dagger-and skeleton proclaimed him to be a young warrior. The form and fastening of these belts is clearly represented on the approximately contemporary " Stele of Baal " from Ras Shamra.'7 During the Late Bronze Age in the Levant and Anatolia representations of what are certainly metal belts increase, but the archaeological evidence is as scanty as ever. A number of statuettes of Hittite manufacture show the metal belt with curved ends worn by warriors. On the figure on the King's Gate at Boghaz-k6y,'8 a bronze statuette from Latakia'9 and an ivory from Nuzi'0 the belt appears to be kept in place by its own elasticity, but on a bronze statuette from Boghaz-k6y2" in the same style as that from Latakia, the belt is tied with thongs. In contrast an actual belt of the period, excavated at Boghaz-kdy, appears to have been fastened with a hook and eye.2" This very interesting object is made of a thin silver sheet encased in two sheets of bronze with a sunk design of interlocking spirals picked out in gold wire on the outer surface. Belts with rounded ends were used in North Syria at this time"3 and fragments of such a belt were found at Ras Shamra.'4 Further to the east, at Susa, a round ended belt, presumably of metal, is worn by a small figure in ivory, of uncertain sex, found in the Elamite levels of the "Ville Royale ". The context suggests a date at the end of the second
8 R. J. Braidwood, Excavationsin the Plain of Antioch, I, 1960, fig. 240-2-Phase " G ". The representation of the belts is very schematic ; it is possible that they have metal clasps. 9 H. Hubert, Syria, VI, 1925, fig. I ; Montet, Bybloset l'Egypt, pl. LXVIII, 579, 580-6. I am most grateful to Miss O. Tufnell for allowing me to read a typescript of her forthcoming article on " Montet's Jar ". G. 0o Brunton, Qau and Badari, I, p. 66, pl. xxix, 13. 11H. Frankfort, The Art and Architecture theAncientOrient,pl. 43 of -it is possible that this head is Old Assyrian rather than Akkadian, see H. Potratz, Die Kunst des Alten Orient,p. 171. 2H. Kosay, Alaca Hoyiik Kazisi, I951, pl. CXLIX; pl. CXLI (gold). 13 C. Schaeffer, Ugaritica, II, 1949, fig. 31, pl. XVIII-XIX, XXI. 14K. Kenyon, Jericho, I, p. 313, fig. 1I17. 15Dunand, Fouillesde Byblos, II, pl. LVII : 8354-8 ; pl. LXIX: 0oo93-5, pp. I89 ff., 338, 397. A terracotta figurine wearing such a belt, pl. LIV, pp. 154 ff16 de Vaux, Rivue Biblique, 54, I947, P. 432, pl. XX. '7 C. Schaeffer, Ugaritica,II, 1949, pl. 24. i8 Vieyra, Hittite Art, fig. 17. '9 Vieyra, op. cit., fig. i0o. Berlin, 1964, pp. 155 ff-., 2o M. Mellink, Vorderasiatische Archiiologie, pl. 2o. see also Anatolia, 6, 1961-2, 21Vieyra, op. cit., fig. i09; pl. XXIII : bronze figure. 22 K. Bittel, MDOG, 74, 1936, pp. 25 ff-, fig. I9. 23 H. Th. Bossert, Alt-Syrien,nos. 579-80. 24 C. Schaeffer, Syria, XIX, 1938, p. 240, fig. 32 W ; Caveau LVI-seventeenth century B.C.

Pl. Ia. Bronze belt: Ashmolean Museum, Oxford.

Pl. Ib. Bronze belt: Ashmolean Museum, Oxford.

Pl. Ic. Bronze belt terminal: London: PrivateCollection.

Brussels(Cinquantenaire of Cinquantenaire, 0.2665). P1. Id. Fragment a bronzebeltfromLuristan,Musde

Pl. Ha.

Pt. IIb.

Pl. IIc.

P1. IId.

Loan). Museum,Oxford(de Walden PI. II. SilverGilt Belt in theAshmolean

Pl. MIla.

Pl. IIIb.

P1. IIIc. Museum,Oxford. P1. III. Details of theSilverGilt Belt in theAshmolean

Pl. IVe.

Pl. IVb.

PI. IVd.

Pl. IVa. Pl. IVc.

Pl. IV. Achaemenian WingedBulls.

SOME

ANCIENT

METAL

BELTS:

THEIR

ANTECEDENTS

AND

RELATIVES

85

millennium B.C.'5 Apart from this there is, as yet, no evidence from controlled excavations in Iran for the form of metal belts there before the very late second millennium B.C. In the course of his clearance of room AB in the N.W. Palace at Nimrud in 185 I Layard found " two long ornamented bands of copper, rounded at both ends, apparently belts, such as were worn by warriors in armour ".26 These belts probably belong to the reign of Sargon (721-705 B.C.) or his successors. To judge from Neo-Assyrian palace reliefs they were common pieces of military equipment from at least the reign of Assurnasirpal II (883-859 B.C.) for heavily armoured troops."7The other belts on these reliefs, as on those at Carchemish28 and Sinjirli'9 are apparently of leather or cloth with rounded end and tapes or clasps to fasten them. Certain contemporary North Syrian representations may show metal belts.3s It is also reasonable to suppose that an as yet unknown Levantine prototype inspired the makers of girdles like that found in tomb P at Fortetsa in Crete.3' In Urartu a number of metal strips have been excavated in graves which were evidently those of soldiers.32 Some indeed were probably belts, like those of Transcaucasia, but others are certainly too large and may have been decorative fittings for chariots.33 During the late eighth century and early seventh century a form of metal belt with elaborate hook clasp was used in Phrygia and Ionia, but its relation to metal belts further east is still obscure, as comparable fittings are not yet known on Urartian belts.34 A simple hook seems to have prevailed there.35 Apart from Iran, Caucasia has produced the widest range of metal belts yet found in the Near East. According to de Morgan36 sheet bronze belt fittings are confined to the Iron Age, not only in Talish, but in the whole of Transcaucasia. In Talish the belts are made of separate plaques of metal set side by side on a leather backing or of long sheets of metal with similar backing. There seems to be no evidence in Talish for the finely chased decoration found on belts in Transcaucasia. In his excavations in Russian Armenia de Morgan found a similar range of belts in situ, but only on male skeletons.37 As swords and daggers were always carried between body and belt they are clearly military equipment, as was often the case in the Levant. De Morgan classified the belts he found at Akthala, Sadakhlo and Mouci-Yeri into three main groups. The most recent were the most elaborately decorated with chased geometric and figure-designs. They were large, reaching 19 cm. in height and varying in length from 88 to 92 cm. The earlier belts were smaller, thinner and less ornate. They were all fastened by a leather thong passed through a pair of holes pierced in either end. One cast belt with a hook was found at Akthala, similar to Koban forms of the eighth to seventh century B.C.38 In addition to the decorated belts found by de Morgan in western Russian Armenia, another group, from the eastern part of the same area, was fully published by Virchow in I895.39 These belts are 70-90 cm. long and 8-15 cm. broad. The ends are rounded, rarely straight, and pierced with one or three holes. Though the technique of decoration is the same in both eastern and western Russian Armenia, the subject matter varies and the style is slightly different. In the west animals and human figures, as on the Akthala belts, are highly conventionalized with very geometric or angular profiles. The body areas are filled overall with chased zigzag lines and dots. By contrast the Virchow group have animals, though still slightly conventionalized, rendered in lively, contorted forms. Their bodies are drawn with a narrow perimeter margin filled with incised lines. Within the margin the body surface is divided into carefully delineated areas filled with blocks of semi-circular lines or bands of dashes. Though the forms are not so stylized and the body patterning less intense on West Iranian metal belts there is a distinct
XXV, 1934, pl. X 4, 5 ; p. 208-9. A. H. Layard, Ninevehand Babylon, 1853, P. 180. 27 A. H. Layard, Monuments Nineveh,I, pl. 18-19, 28. of 28C. L. Woolley, Excavationsat Carchemish, III, pp. 196 ff.-full explanation with modern analogies; Vieyra, Hittite Art, des fig. 39, 40 ; B. Hrouda, Die Kulturgeschichte Assyrischen Flachbildes,1965, pl. 7, PP- 47-8. 29Vieyra, op. cit., fig. 76. 30 Vieyra, op. cit., fig. 41, 55. 3' Brock, Fortetsa, 1957, PP. 197 ff., pl. Ii5R. D. Barnett, AS, XIII, 1963, fig. 41, 46, 47 ; G. R. Meyer, 32 Altorientalische Museumzu Berlin, Denkmalerim Vorderasiatischen 1965, pl. I33-5.
25 M.D.P.,
26

33R. W. Hamilton, AS, XV, 1965, PP- 41 ff. 34J. Boardman, Anatolia, VI, 1961-2, pp. 179 ff. 35R. D. Barnett, AS, XIII, 1963, fig. 30 ; B. B. Piotrovskii, IskusstvoUrartu VIII-VI, Leningrad, 1962, fig. 4236 La PrehistoireOrientale,III, 1927, Pp. 276-8. But see A. A. Martirosian, Armeniain the Bronze and Early Iron Age, 1964 (Russian), fig. 57, p. 136; fig. 65, p. 16o. 37de Morgan, Mission au Caucase,I, 1889, fig. 17-19, 23, 27; p. 1I6. 38A. M. Tallgren, E.S.A., V, 1930, pp. 165-6. 39Uber die Culturgeschichtliche Stellungdes Kaukasus,Berlin, 1895.

86

JOURNAL

OF

PERSIAN

STUDIES

affinity. In many ways even closer to the Iranian belts with zoomorphic decoration is a belt from a tomb at Maralyn Deresi in the Trialeti district4odecorated in two registerswith a hunting scene. It will be easier to assessthe relation between these Caucasian belts and those of north-westIran when the belts reported from Marlik4'and Hasanlu4'are fully published. A metal belt from the Japanese excavations in Dailaman indicates that the relationship may well prove to be very close.43

II

Belt No. I (Ashmolean Museum 1965.833)44--PlateIb As in the following example the belt is made of a single strip of sheet metal, about 49 5 cm. long, 6 -75 cm. high. Each end is cut to form a double-spiral. A single, small hole is pierced through either end of the belt at the base of the spiral to take a cord or wire for fastening the belt. On the upper and lower edge there is a margin of repoussi dots in a double row, broken only above and below the central floral motif. At either end of the belt the dots are arrangedin triple lines. Within this frame is a chased design of two gazelles grazing, one behind the other, on either side of a central " tree". The hide of each animal is decorated overall with tiny sets of double lines. Only the shoulder, marked out by a double line, and the rib cage are specially emphasized. The ringed horns and the short tail ending in a tuft, characteristicof the gazelle, are clearly shown. The central lotus flower rises from a sub-circular core set on a low trunk. In the flower five leaves are set on either side, curving away from three pointed petals in the centre.

Fig. I.

Belt No. II (Ashmolean Museum 1965.834a, b)-Plate Ia This belt is made from a single strip of sheet metal six centimetres high and at least 57 centimetres long. One end is broken off and does not join the main piece exactly, though it is certainly part of the same belt. Both ends are cut as double spirals. At either end of the belt, set at right angles to the terminals, is a male figure with arms raised on either side of his head, very simply rendered in repousse'. He stands on a row of four repousse' blobs and the field behind him is decorated with tiny punched holes. An exactly similar figure appears in the central panel of the belt, but this time set parallel to the blobs. The area between these figuresis terminals, framed on either side by a double row of four repousse' filled by three horizontal repoussi out by rows of small punched holes. lines, marked
4oB. A. Kuftin, Archaeological Excavationsin Trialeti, I, I941, pl. XXV. 4' E. O. Negahban, A PreliminaryReport on Marlik Excavation, Teheran, 1964, fig. 55. 42R. Dyson, Expedition,I960, p. Io. 43N. Egami, Dailaman, I, Tokyo, 1965, pl. LXXV, 36; XXXII, 2. 44Both these belts form part of the Bomford gift of Persian bronzes to the Ashmolean Museum, Oxford. Also in the gift was a third belt from north-west Persia with double spiral terminals, this time undecorated: i965.832: -044 m. wide, '450 m. long. There is a similar belt in the British Museum : B.M. 134366.

SOME

ANCIENT

METAL

BELTS:

THEIR

ANTECEDENTS

AND

RELATIVES

87

Fig. 2. The distinctive terminals of these two metal belts may be paralleled exactly in the cemetery at Khurvin.45 A complete belt and a fragment of another (Plate Ic) reported without exact provenance

from Luristan, have double-spiralterminals and a very simple design of four horizontal rows of repouss6 blobs along the belt.46Analysis of the fragment has shown that it is definitely bronze not copper.47 There seems no reason to doubt the reported source of these two belts. A number of sheet metal bracethis particular type of sheet metal bracelet to the cast bronze work most characteristicof Luristan in the eighth and seventh centuries B.c.50 This pair of terminals are decorated in repoussd a pair of with " lions " rendered in the exotic style reserved for this beast in the whole range of opposed rampant Luristan cast bronze work at this time. Characteristicof it are the raised coils which frame the eyes and form the jaw. In cast bronze they reflect the original wax model, but they passed into the chased
and repousse designs on sheet metal as a purely decorative device.5' Flat-headed pins, their heads cut in a double-spiral, have also been reported from Luristan.52 Such pins have distinct Caucasian affinities.53 Double spiral bracelets are typical of the Koban culture in the early first millennium B.C., but apart from an example reported from the Lechkhum hoard54 are very lets with double-spiral terminals have been reported from Luristan, some of them from the earliest period of clandestine excavation in the area.4' One set of these bracelet terminals49 indisputably links

West Persia,which normally have elaborate overall designsin which men and animals, vividly rendered,
play an equal part."6 The posture of the naked man, however, indicates that the belt is Iranian, comparable in date to the one reported from Khurvin. This distinctive pose, with arms raised on either side of the head, is assumed by figures dressed in long robes represented on unprovenanced seals of
45Vanden Berghe, La Ne'cropole Khurvin, XXXIX, 277. de pl. 46Brussels, Speleers, Bull, MusdesRoyauxd'Art et d'Histoire, 1932, p. 68, fig. 19-terminals not shown; possession of Mrs. R. Maxwell-Hyslop, London, with whose kind permission I include it and the analysis, done by the Research Laboratory for Archaeology and the History of Art, Oxford. 47CU : 92 - 2 Sn : 7-I ; Pb : nd ; As : o - I28 ; Sb : nd ; Bi: nd; Fe: o-II3; Zn: nd; Ag: Ni: o-025; 48Godard, Les Bronzes du Luristan, 1931, pl. XXVII, 0o203. 81; Genava12, 1934, p. 81, fig. I (13) ; Survey,I, p. 226, fig. 50 ; Buhl, Acta Archaeologica, XXI, 1950, no. 63. 49Arne, E.S.A., IX, I934, P- 278, fig. 5. E. 50o
5' E. Porada, Dark Ages and Nomads, 1964, p.

Luristan, to be discussed later, or that on the many sheet metal vessels now known from sites in North-

rare in Transcaucasia. It is possible that the cast bronze belt buckles from Koban,55with spiral terminals, mark the final stage in the history of this fashion in Caucasia. The geometric simplicity of the design on belt II is unlike the decoration found on belts from

; E. Herzfeld, 2I Iran in theAncient East, 1941, pp. 163 ff. figs. 279, 282. 52P. Jacobsthal, Greek Pins, 1956, p. 127, 147, pl. 452-3. 53F. Hangar, E.S.A., VII, 1932, pp. 162 ff., fig. 25-29. 54R. D. Barnett, A.S., XIII, 1963, P. 172. 55Chantre, Mission en Caucase, III, pl. XII, 3 ; also Kazbek Treasure, Tallgren, E.S.A., V, 1930, fig. 6o. 56E. 0. Negahban, A PreliminaryReport on Marlik Excavation, Teheran, I964, pl. II, IV, V, VIII ; fig. 103-5, Io7-II5, D. K. Shepherd, Bulletin of the Cleveland Museumof 136-I42 ; Art, 1966, pp. 38 ff.

Porada, AncientIran, pp. 75 ff.

88

JOURNAL

OF

PERSIAN

STUDIES

Sialk " B " type;"5 in one case the scene is one of dancing, the other a hunt. A similar gesture is sometimes adopted by the " master or mistress-of-animals " on bronzes produced in Luristan about the same time."8 A bronze statuette of a warrior reported from Khurvin is shown in this position.59 It is probably a gesture of worship, or perhaps of mourning.6o Some of the female figurines from Piravend6' are cast in the same position, but in this case there is reason to believe it may be a posture associated with childbirth, as elsewhere in the ancient Near East.6' It is unfortunate that the date, often even the authenticity, of a wide variety of anthropomorphic metal figurines reported from Caucasia is in doubt, for crudely cast male figures with their arms raised appear frequently among them.63 The other belt (I) may be more readily attributed than its companion. An identical " tree " appears on a silver vessel reported to be from Ziwiyeh.64 Allowing for the difference in material, virtually identical renderings of the gazelle appear on a number of the ivory plaques reported from the same site.65 It seems likely that these ivories were not made in the Assyrian workshops which produced the ivories in a " Neo-Assyrian " style found in Assyrian palace excavations,66 for though motifs are often very similar certain points of style and technique suggest provincial workmanship.67 An even more definitely Iranian workshop is represented at Ziwiyeh by another group of ivories. The iconography still owes something to Neo-Assyrian patterns, but the designs are executed in a manner which has little regard for form or proportion. They follow an independent Iranian tradition of ivory and wood carving represented by finds in level IV at Hasanlu,68 c. 900-800 B.C., where the same distinctive human type also appears. Such parallels suggest a date in the eighth century for the belt. There is no reason to suppose its companion (belt II) as other than approximately contemporary. Belt number I is important as direct evidence for Neo-Assyrian influence on local Iranian smiths working in base metals. Although the motif is clearly Neo-Assyrian in origin, both the execution of the design and the object to which it is applied are part of the Iranian tradition of metal working. This is not an isolated phenomenon. A small bronze ritual bucket, now in the Metropolitan Museum, New York, and a number of other vessels without exact provenance from this area69 clearly illustrate the currency of Neo-Assyrian iconography in local workshops. As with the ivories from Ziwiyeh small departures from the Assyrian artistic canon reveal the hand of provincial craftsmen: the scale of the constituent parts of a motif are altered for dramatic effect, details of dress and hairstyle are modified and the floral motifs are less formalized than their Assyrian prototypes. As so little decorated metal work, which may certainly be ascribed to local workshops, has been reported from sites in Assyria,7"we remain ignorant of its main styles and forms; but there seems little doubt from what we have, and from the evidence of the " Neo-Assyrian " ivories from Nimrud, that in style and iconography the minor arts closely followed the canon found in the monumental sculptured reliefs. In Iran the finds in level IV at Hasanlu and at Ziwiyeh, as well as a number of other sites still to be identified, indicate the extent to which vigorous indigenous traditions of craftsmanship in ivory, wood, metal and baked clay, with a long history only partially revealed at Marlik, were subject to the influence of craftsmen and imported objects from Assyria and Syria. A parallel phenomenon on Assyria's western frontier has long been clear.7' As in Syria, the process began in Persia in the second
57M. and L-H. Erlenmeyer, IranicaAntiqua,V, 1965, pp. I ff ; pl. IV, 20, pl. V, 21. 58Ibid., pl. X, 53, 59-60 ; Godard, Les Bronzesdu Luristan, 1931, pl. XXXVII, 157. 59Vanden Berghe, Archdologie l'Iran Ancien, 1959, pl. i56b, c. de 60 Ghirshman, Persiafrom the Originsto Alexander,1964, fig. 495. 6l Survey,IV, pl. 73, D, F. 62Egypt : Spiegelberg, AnnalesduService, XXIX, 1929, p. 162 ; Mesopotamia-M. Lambert, R.A., 42, 1948, pp. 198 ff. 63A. Zakharov, Swiatowit,XV, 1932/3, fig. 56, 63, 71, 72-5. 64A. Godard, Le Tresorde Ziwiyi, I950, fig. 53, 54. 65Godard, op. cit. fig. 69, 70 ; D. von Bothmer, AncientArtfrom New YorkPrivateCollections, No. 32. 66M. E. L. Mallowan, Nimrud and its Remains,II, pp. 588 ff ; R. D. Barnett, A Catalogueof the Nimrud Ivories in the British Museum,pp. I86 ff., pl. CXIII-CXXI. 67E. Porada, AncientIran, 1965, p. 128. 68R. Dyson, Archaeology, 1964, fig. I, 2, 8, 12 ; J.N.E.S., 17, XXIV, 1965, p. I99. Recent C'4 dates may indicate a lower date for the end of Hasanlu IV: Radiocarbon,8, 1966, PP. 349-50. 69C. K. Wilkinson, Bulletin of the MetropolitanMuseum of Art, BronzenderSammlung 1963, pp. 282 ff. ; Calmeyer, Altiranische 1964, Nos. 106, 107 ; 104, 105. Brickelschen, 70H. Frankfort, Art and Architecture the Ancient Orient, 1958, of pp. 102-3 ; the bowl published by Frankfort in J.N.E.S., V, 1946, p. 155, pl. II is unlikely to be Assyrian; it might be a product of the workshop producing the decorated situlae of Zalu Ab and other sites in the central Zagros-P. Calmeyer, BerlinerJahrbuch, 5, 1965, PP. I ff. M. E. L. Mallowan, Nimrudand its remains,II, pp. 430 ff. ; fig. 356-7 ; E. Porada, Bulletin of the Museum of Fine Art, Boston, XLVIII, 1950, pp. 2 ff. 7' H. Frankfort, A.A.A.O., p. 166.

SOME

ANCIENT

METAL

BELTS:

THEIR

ANTECEDENTS

AND

RELATIVES

89

half of the ninth century B.c. and was still vigorous in the late eighth and early seventh century B.C. when many of the objects found at Ziwiyeh were made.72 This is a period for which there is reasonably good documentary evidence of Assyria's relations with her eastern neighbours. Zamua, the vassal state closest to Ziwiyeh, had been incorporated into the Assyrian system c. 856 B.C.73 by Shalmaneser III and formed a vital source of'horses for the Assyrian armies until the third quarter of the seventh century B.C. Throughout the intervening years political and diplomatic activity no doubt stimulated a vigorous trade between the Assyrian homeland and the tributary states on the eastern frontier. The gazelles on this belt, and those on the Ziwiyeh ivories, are closely paralleled on a sheet metal belt reported from Luristan to the south.74 A number of other belts and sheet metal fittings which may come from belts have been reported from this very important centre of bronze production. These metal belts are usually decorated in repoussewith chased details. Small punched holes of varying disposition round the perimeter of the surviving fragments show that they were often stitched to a backing of leather or fabric. If original, the plain wire loop on a belt in the Voss collection75 suggests a very primitive form of fastening with hook and eye, though no trace of the former survives. A very similar form of fastening exists on an Early Iron Age sheet metal belt from Cyprus.76 The plain curved or flat metal ends surviving on a number of the Luristan belts and fragments, without any evidence for cast buckles" or hooks and eyes, may mean they were retained in position by their own elasticity. The belts from Luristan may provisionally be divided into three main groups. In the first group78 the use of gold and silver, rare among the sheet metal work reported from Luristan, marks their affinity to artefacts from north-west Iran where the use of precious metals is much more common. The iconography of these belts lends additional weight to this suggestion. They are decorated with scenes which might loosely be described as ritual presentation processions. They consist of self-contained groups, either variants of the " master-of-animals " theme or groups of offeringbearers, arranged in a continuous frieze. Characteristic of the figures on these belts, and the related flat-headed pins with circular heads, are their angular movements, the hair swept up at the back, long robes with tasselled girdle and hem, and a facial type distinguished by prominent straight noses, large eyes, thin lips and straggly beards. Identical figures appear on ivories in a local style from level IV at Hasanlu79 and the Ziwiyeh hoard.8s Such parallels are very important for the chronology of this sheet metal work in Luristan."8 Carbon 14 determinations indicate that level IV at Hasanlu82 ended c. 800 B.C.; the ivories were probably produced sometime in the preceding half century. In the same level were found a group of " Syrian style ivories of the Loftus group ".83 It has been suggested by a number of scholars that the ivories from Ziwiyeh in a Neo-Assyrian style were made in the second half of the eighth century B.C.84 The local ivories found with them may then be later than those from Hasanlu, but not significantly so. As the evidence stands at the moment it is probable that the sheet metal work associated with the shrine at Dum Surkh and from elsewhere in Luristan may be dated to the eighth or early seventh centuries B.C.
72Although much attention has been paid to when and why the Ziwiyeh hoard was buried-see L. Byvanck-Quarles van Ufford, B. A. Besch. (Hague), XXXVII, 1962, pp. 25 ff.: XXXVIII, 1963, pp. Ioo ff. for full bibliography up to I963-less attention has been paid to dating the various objects in it-see E. Porada, AncientIran, 1965, pp. 123 ff. and articles by C. K. Wilkinson listed there, p. 264. 73D. J. Wiseman, Iraq, 2o, 1958, pp. 9 ff. for history of eastern frontier. 74W. Nagel, Altorientalisches Kunsthandwerk, Berlin, 1963, No. 57, pl. XXIX-XXXI. 75 7,000ooo Jahre Kunst in Iran, Essen, 1962, No. 16 (plate). In discussing the Luristan evidence I include only the rectangular metal strips whose length and height suggest belts ; among a great variety of smaller sheet metal plaques reported from the region there may well be some which are belt fittings, but it is now impossible to identify them with certainty. 76Myres, Cesnola Collection, 1914, No. 4755, fig. on p. 489. Persiafrom the Origins to Alexander,fig. 510 published a fitting of this kind said to be from Luristan. It is of bronze inlaid with a snake in iron. If in fact it does come from the central Zagros it is almost certainly an import from North Caucasia where such fittings are not uncommon: A. M. Tallgren, E.S.A., V, 1930, pp. 165-6. 78R. Ghirshman, ArtibusAsiae, XIII, 1950, p. 194, figs. 18-i9gold strip ; R. Dussaud, Syria, XXVI, 1949, P. 210, fig. Iobronze ; also perhaps some plaques-P. Ackerman, Archaeology,8, 1955, fig. I--silver; R. Ghirshman, Bichdpour, 1956, II, pl. XXIII, 4. 79R. Dyson, Archaeology, 1964, fig. 2, 12. 17, de 8oA. Godard, Le Trdsor Ziwiyi, 1950, pp. 105-6, fig. 91-2. 8s There is no comprehensive study of this material as yet; Dussaud reviews some of it in Syria, XXVI, 1949, PP- 196 ff. 82R. Dyson, J.N.E.S., XXIV, 1965, p. 202-3 ; see note 68. 83Ibid., p. 199. 84E. Porada, AncientIran, p. 128.
77 Ghirshman,

90

JOURNAL

OF

PERSIAN

STUDIES

The second group of belts from Luristan, all in bronze, provide a striking contrast to the preceding group both in style and iconography and have two additional points of interest: they seem to be a distinctly local type and they have clear links with the cast bronze industry of Luristan.85 The individual figures and animals are often very crudely rendered and the designs loosely organized. The primary motifs are scenes of war and hunting, with the individual animals or heraldic groups of " hero and beasts " found so often on the cast " standard-finials " and openwork pinheads from Luristan. There is no trace of the ritual scenes of the previous group. This may well signify the distinction between objects of daily use and decorative fittings for the ceremonial garments of priests ; a difference in part reflected in the change from a base to a precious metal. The belts in this group, like those from the Caucasus, may be regarded as the equipment of soldiers and huntsmen. A belt of this type is reported from clandestine excavations in what may have been a mountain shrine at Tang-i-Hamamlan in Hulailan, Luristan.86 Though isolated indications appear-the characteristic lion mask found both on the flat-headed disk-pins"7 and the belt in Berlin88-there are too few motifs in common and no prominent stylistic affinity between these belts and the decorated disk-pins to allow for certainty over a common centre of production. If it may be assumed that the " Dum Surkh style ", characterized by elaborately decorated bronze disk-pins, developed in south Luristan89 not too far distant from the shrine itself, these belts may be more closely associated with workshops in Hulailan to the north. The finest of them,90 now in the Louvre, has clear affinities of style and iconography with a group of decorated situlae and sheet metal objects decorated with hunting scenes reported from Zalu Ab and other sites in South Kurdistan and North Luristan.9' The cuneiform inscriptions on some of the situlae indicate a date in the ninth or eighth centuries B.C. for their manufacture. On the belt in the Voss collection,9' said to be from Tepe Giyan, appears the semi-human figure with horned crown and curling sidelocks who is one of the most common motifs on the cast bronzework from Luristan. It appears on harness rings,93 on open cast pinheads (" Wands "),94 on the tubular sockets for " standard-finials "95 and as the face of the sphinx on certain cheekpieces for horse-bits.96 This is an isolated, but by no means insignificant indication that, as accumulating evidence tends to demonstrate, the period when Luristan supported a major centre of bronze producof tion with a distinctive repertoire its own falls largely in the eighth and early seventh centuries B.C. It was an industry, parallel in date, as in range of artefacts, to that excavated from Cemetery " B " at Tepe Sialk.97 Even when these two major groups of decorated metal belts from Luristan have been considered there remains a third group of sheet metal belts of a much cruder kind with very simple geometric or floral decoration.@"There is no independent evidence for their date so far, but they are unlikely to differ significantly from those already discussed and may be attributed to the first three centuries of the first millennium B.C. The decorative styles of the belts and related plaques from Luristan represent the final phases of a local tradition. In the century after c. 650 B.c. the Luristan industry, working mainly in bronze, was eclipsed by another great metalworking industry, more often using gold and silver, established in and around Ecbatana as a consequence of the growing political coherence of the Medes and Persians. It
85For belts-Dussaud, Syria, XV, 1934, p. 187, fig. I, pl. XXV ; P. 192, fig. 3. 7,000 Jahre Kunst in Iran, Essen, 1962, No. 16 Berlin, (plate) ; W. Nagel, AltorientalischesKunsthandwerk, I963, No. 57, pl. XXIX-XXXI ; Brussels, Cinquantenaire o0 2665, 35 cm. long, 5 -6 cm. wide, pl. I D. I am grateful to Mme D. Homhs-Fredericq for this photograph and to Dr. P. Gilbert for permission to reproduce it here. Also perhaps the fragments-H. Potratz, Praehistorische Zeitschrift, 30/31, 1939-40, p. 179, fig. 4 ; Survey,I, p. 269, fig. 61 ; Godard, Les Bronzesdu Luristan,1931, pl. XXXIX, 162, also Survey,IV, pl. 56 E. 86H. Thrane, Acta Archaeologica, XXXV, 1964, p. 157. 87Dussaud, Syria, XXVI, 1949, fig. 3-4. 88Nagel, op. cit., drawing on end plate. 89Such a view is perhaps strengthened by the elements of Elamite influence in their decoration, E. Porada, AncientIran, 1965, pp. 88-9. 90R. Dussaud, Syria, XV, I934, pl. XXV. 9' P. Calmeyer, BerlinerJahrbuch, 5, 1965, PP. I ff 92 7,000 Jahre Kunst in Iran, 1962, No. 16, plate. 93Godard, Les Bronzes du Luristan, 1931, pl. XXXII, 115-17. 94Ibid., pl. XXXV, 150, 15I ; pl. XXXVI, 15495 Ibid., pl. LI, 192 ; LVIII, 211. 96Ibid., pl. XLI, 167. 97R. Dyson, J.N.E.S., XXIV, 1965, pp. 207-8 for the dating of Sialk " B " accepted here ; also R. Boehmer, Arch. Anz., 8o, I965, pp. 802 ff. 98e.g. Herzfeld, Iran in the AncientEast, 1941, fig. 263-now in the British Museum.

SOME

ANCIENT

METAL

BELTS:

THEIR

ANTECEDENTS

AND

RELATIVES

91

was the artisans of the Medes and their associates who were to be the heirs of the decorative styles of north-west Persia in the preceding five hundred years. The characteristics of the " Median " style in fine metalwork are still far from clear99 but they are unlikely to represent an entirely new tradition of metalworking. Historical circumstances suggest that " Median Art " was a blend of the NeoAssyrian tradition as it appears at Ziwiyeh and the native Iranian styles found in artefacts from the same site and its near neighbours which developed naturally into the better known Achaemenian style of the late sixth and fifth centuries B.C. Belt I represents an early stage in this process : the application of Neo-Assyrian motifs to Iranian artefacts. Belt II, comparable in form but distinct in decoration, exemplifies a local Iranian decorative style which was to be completely superseded. III The use of metal belts in Achaemenian Iran is not easy to establish. In representations both of the Medes and Persians,'o0 belts appear to be of leather or fabric without metal fittings of any kind. At Persepolis and on the Oxus gold plaques'o' the Medes wear a tunic belted at the waist with a girdle from which the akinakes hangs. Although the flowing robe of the Persians often conceals the belting of their undergarments, belts when shown do not appear to be Belts similar to those of metal-plated.'?2 the Medes were worn by the Scythian tribes of South Russia. They were of leather with a strap, run through a slit cut in the belt, to take a bow case,'03 or sometimes, on the evidence of Herodotus (IV. Io), a drinking cup. Since Sarmatian dress'04had much in common with that of the Medes and Scythians : tunic and baggy trousers, often of leather, they might be expected to have worn exactly the same type of belt. A significant change, however, is reflected in the appearance of a new form of belt fitting : the hook-clasp,105 often very elaborately made in metal. Present evidence suggests that such fittings were not part of Achaemenian or Scythian dress. Also, in at least one case,'I6 a Sarmatian is represented wearing a broad belt made of metal plaques. Despite the absence of surviving examples documentary evidence indicates that such belts may also have been worn on ceremonial occasions in Achaemenian Iran. Quintus Curtius,'o7 in describing the elaborate costume of Darius III, records that he wore a gold belt, knotted like a woman's. The Avesta, which probably covers the period from the seventh to the third centuries B.C.,'s8 also refers to gold belts.'09 Such a belt, or rather a close relation, is represented by an object recently received in the Ashmolean Museum, Oxford`" (Plates II and III). As circumstances of the belt's discovery and acquisition are not known, any attempt to identify the place and date of manufacture must rest entirely on internal evidence. The belt, just over a metre long (I *05 m.), is made of fourteen rectangular silver gilt plaques mounted on a single strip of leather, now broken in a number of places, but otherwise remarkably well preserved."' Each plaque is made from a single piece of beaten silver (approximately X -0o84 m. in size) with the upper and lower edges turned over (by about -oI3 m.) to form a flange ?075 which secures the plaque to its leather backing. Each plaque has a raised margin with milled surface and the upper face of the plaque, apart from the zoomorphic devices, is decorated with a series of traced semi-circles to form an irregular overall scale pattern. Raised, rectangular strips of metal, slightly angled in the centre, are soldered on to the surface of each plaque, one above and one below
99R. D. Barnett, IranicaAntiqua,2, 1962, pp. 77 ff. ; W. Culican, The Medes and Persians, 1965, PP. 46 ft. for recent discussions. For details of their dress and its antecedents see: A. Roes, Ioo Bib. Or., 8, 1951, pp. 137 ff. ; B. Goldman, IranicaAntiqua,4, 1964, pp. 133 ff. ; G. Thompson, Iran, III, 1965, pp. 121 if. E. Ior Schmidt, Persepolis,I, pl. I2o, 12I ; 0. M. Dalton, The Treasureof the Oxus, 1926 (hereafter " Oxus "), pl. XIV, 48, pl. XV, 7o. E. io2 Schmidt, Persepolis,I, pl. 70 C. 103For instance the Chertomlyk vase relief-E. Minns, Scythians and Greeks,1913, p. 162, fig. 49. 104 For bibliography see: M. Rostovtzeff, Iranians and Greeksin SouthRussia, 1922, p. 231, chapter VI, note 2.
105 M. Rostovtzeff, The AnimalStyle ..., 1923, P- 42. io6On fragment of a silver rhyton-N. Kondakof, et al., Antiquitis de la RussieMiriodionale,Paris, 1891, fig. 184, pp. 205 ff.

III, iii, 7. o07 R. o08 C. Zaehner, The Dawn and Twilight of Zoroastrianism, 1961, p. 26. 0o9 Tashts,XV, 57. nio was deposited on permanent loan in It 1963 by courtesy of the National Museum of Wales from the Howard de Walden Collection. Accessed as de Walden Loan No. 13. I am very grateful to Mr. M. L. Ryder for his report on this leather which will be found in an appendix.

92

JOURNAL

OF

PERSIAN

STUDIES

the main decorative motif. The scale pattern runs below them. They are made so as to form a are for series of cells or cloisons garnet inlays."2 The cloisons rectangular in shape on the three plaques at either end of the belt, otherwise triangular, with stones cut to the appropriate shape. The flat surface of each inlay is slightly set back from the face of the cloisonwhich serves to hold it in place. As far as it is possible to see the inlays are partly cemented into each cell. with The plaques at either end of the belt are decorated with a winged bull in repousse', the details of the rib cage, wing feathers, mouth, ear and nostril chased on the surface (Plate IIIa). An almond-shaped cloisonis soldered on to the lower part of its wing to take a garnet inlay with a domed face which rises above the edge of the cloison.The animal is represented in a " flying gallop " with tail in the air, its end turned over. with a pair of insects, set back to back, with an All the other plaques are decorated in repousse in the plaque between the back of their wings (Plate IIIc). In the centre of each insect's back opening an almond-shaped cloisonis soldered on to take a garnet inlay cut with a domed surface as on the winged bulls. Details of the insect's prominent eyes and abdomen are chased on the surface. The exact identity of this insect is not easy to establish."3It is almost certainly not a cicada as similar insects are sometimes called in archaeological literature. The most likely model is a bee-particularly if it is correct to recognize a protrudingsting in certain of the insects-with its large eyes and pointed abdomen carefully represented. The clasp is extremely simple (Plate IIIb). The end of the belt is turned over to form a much deeper metal flange than is usual on the upper and lower edges (approximately m. wide) and two hooks, "023 one now broken off, are fitted to it (01o2 m. apart). Each hook is made of a strand of silver, the lower end hammered out to form a thin flange of metal which is soldered"4to the surface of the flange. The head of the hook is treated zoomorphically. Delicately chased lines suggest the head of a swan, the curve of the hook its neck. Two rivets are,driven through both plaque and flange from the front, their flat heads appearing on the surface of the belt. They do not appear to have any function in relation to the hooks, but may be to secure the flange firmly to its leather backing. Two pear-shaped holes are pierced through the correspondingend-plaque in the appropriate position to take the hooks. A single rivet passes through both plaque and flange between the two holes. The belt has only one other feature (Plate IIa & d). Two holes are pierced through the second plaque from the right, one in each of the lower corners. Through each hole passes a small fragment of copper wire, presumably original; in one case coiled round to form a loop. They seem too slight to take anything except a very light weapon; but on the evidence of Herodotus already cited, they might well have served to suspend a cup. and Iconography Chronology The bulls with curved wings on the end plaques provide an important clue to identification of the belt's source and date of manufacture. They were a popular motif in the art of south-westAsia during the Achaemenian period. Though the winged bull is a creature of Assyrianorigin, it was always shown in Assyria, from the Middle Assyrian period onwards, with a straight wing,"5 as were all winged creatures on objects produced in local Assyrian workshops."6The curved wing appeared in Assyria on the ivories and bronze bowls made by Phoenician craftsmenworking in the Egyptian manner, but
it does not seem to have been adopted by local craftsmen."I7Mischwesenwith curved wings appear in the early first millennium B.C. on seals from Babylonia and Elam,"8 whence they were adopted and
II Mr. A. W. Kingsbury of the University Museum, Oxford, kindly commented on these stones for me. "3 I am very grateful to Mr. J. P. Doncaster, Keeper of Entomology, British Museum (Natural History), and his staff for advice on this problem. The following comments owe much to their suggestions. "4 1 use the term with caution; perhaps " welding " in the sense defined by Mrs. Davidson in H. Hoffmann and P. Davidson, Greek Gold, 1965, PP. 44-7 would be more accurate. H. Frankfort, CylinderSeals, 1939, pl. XXXIIc, XXXVh; 15s S. Smith, The Early History of Assyria, 1928, pl. XXIb89575; E. Porada and B. BMIo3313 ; pl. XXIc-BM Buchanan, CANES, I, No. 753. "i6 The same seems to be true of Urartu-R. D. Barnett, Iraq, XII, 1950, pl. VII, XVIII, XIX, XXI. 17 de Mertzenfeld, Ivoires Phiniciens, 1954, pl. CI-CIII; of pl. LXXXII ff., pl. CXIII ; A. H. Layard, Monuments Nineveh,II, 1853, pl. 58E, 59B, 63, 68. '8 Porada and B. Buchanan, CANES, I, Nos. 747, 749, 750; E. E. Porada, AncientIran, 1965, fig. 30, 49.

SOME

ANCIENT

METAL

BELTS:

THEIR

ANTECEDENTS

AND

RELATIVES

93

developed by the bronzeworkers of the central Zagros region."9 In north-west Persia there is some reason to think both forms were current in the pre-Achaemenian period. The straight wing appears on the fine metalwork of the Marlik culture120 and objects in a derivative Neo-Assyrian style from Ziwiyeh."' In the other objects reported from Ziwiyeh'22 and unprovenanced objects usually ascribed to Median workshops'23 the upper end of the wings already have a pronounced curve. The appearance of the curved wing in Achaemenian art is first documented at Persepolis ;"24 the winged " genius " erected by Cyrus at Pasargadae, perhaps under Syrian influence,'25 has two pairs of straight wings. After the late sixth century B.C. the curved wing, often with a very concave upper edge, became a hallmark of Achaemenian art. The lack of chronologically fixed points and the apparent absence of major iconographic developments in Achaemenian art over a period of two hundred years does not allow for precision in dating specific examples of a common motif. However, certain factors, apart from the garnet inlay, make clear that this particular treatment of the winged bull in a " flying gallop " belongs to the final phase of Achaemenian art in the fourth century B.C. Unfortunately the winged bull only appears on the sculptured friezes at Persepolis126 as the handle of a spouted amphora carried by one of the tribute bearers (Plate IVa), who was identified both by Schmidt and Herzfeld"7 as a Syrian ; Barnett2"8has argued that he is in fact a Lydian. Whatever the exact area of origin, his delegation certainly comes from the west of the Empire and brings as tribute vessels and bracelets now considered characteristic of Achaemenian workmanship. The only handle of this kind known to me'29 is reported to come from Mesopotamia, where it was found with a hoard of coins dating from the fifth and early fourth centuries B.C. Amandry'30 has discussed it at length and proposed a date in the first half of the fifth century for its manufacture. The hatched " beard " along the jaw and the chasing of the wings is well paralleled at Persepolis. Very much more detailed representations of winged bulls appeared on the gold overlay from one of the doors in the Apadana Palace at Persepolis'31 (Plate IVb). Both context and style suggest that they belong to the earlier phases of Achaemenian art, before about 470 B.C. The distinctive margins and panels of curled hair on the body and legs of the bulls are a legacy from Urartian even more perhaps than from Assyrian art.'32 With these bulls may be associated a similar one on an unprovenanced silver bowl 33 (Plate IVc). A virtual absence of such panels of hair and greater emphasis on musculation of the body in stylized geometric forms is more characteristic of developed Achaemenian art. Such devices are conventionally referred to as the " bow-and-dot " motif on the haunches'34 and the " pear-and-apple " motif on the shoulders.'35 In a rudimentary form they may first be seen on the reliefs at Persepolis on the ordinary bulls shown passant on the canopy of Darius, in reality probably gold appliques rather than embroidery.'36 A virtually identical treatment of the bull, this time winged, appears on a " goldsmith's " trial piece from Egypt'37 (Plate IVd). Though their tails flutter in the air, all these bulls are represented in the formal passant position, each leg shown, one behind the other. This is in marked contrast to the " gallop " of the bulls on the two belt plaques. During the period of Achaemenian supremacy in the Near East representations of animals running at full speed, in the same posture as the bulls on the belt, were a particularly common subject on the so-called " Graeco-Persian " gems.'38 The developed naturalistic style of these seal stones is closely
17 ; pl. XLIII, 17I. 12o O. Negahban, LL.N., 28th April, 1962, fig. B, G, H. E. I A. Godard, Le Tre'sor Ziwiye, 1950, fig. 15 de f. i2z W. Culican, The MedesandPersians, 1965, pl. 24. 123Sept Mille Ans, Paris, i96I, pl. LII. of I4 H. Frankfort, The Art and Architecture theAncientOrient,1958, pl. 8oB ; cf. pl. 77 top. 25 H. Frankfort, op. cit. fig. 116; for bibliography and discussion see E. Schmidt, Persepolis,I, p. 22. i6 Herzfeld, Iran in the Ancient East, 1941, pl. LXXXIV; winged human-headed bulls and ordinary bulls are much more common. 17 E. Schmidt, Persepolis,I, p. 85, n. 123. z8sR. D. Barnett, Iraq, XIX, 1957, pp. 68-9.
7

19 A. Godard, Les Bronzes du Luristan, 1931, pl. XLII,

E. 129 G. S. Robinson, Iraq, XII, 1950, pp. 44 ff. 130P. Amandry, Antike Kunst, 2, 1959, pp. 46 ff. I3 A. Sami, Persepolis, Shiraz, 1954, p. 82 (not numbered); Mostra d'ArteIranica, Rome, 1956, pl. XXX. E. 132 Herzfeld, Iran in theAncientEast, 1941, fig. 353, P- 249133 Sept Mille Ans, Paris, 1961, No. 678, pl. LXV; note musculation of thighs. 134 Salmony, GazettedesBeauxArts, XXXV, 1949, p. 6. A. 135 Roes, ArtibusAsiae, 15, 1952, P. 18, fig. I. A. 136Godard, L'Art de L'Iran, pl. 67. 137 Frankfort, J.N.E.S., IX, 1950, pp. III ff., pl. III. H. 138 Richter, Hesperia,supp. VIII, 1949, PP- 291 ff. ; ArchaeG. ologicaOrientalia-E.HerzfeldFestschrift, 1952, pp. 189 ff. and an article by H. Seyrig in the same volume, pp. 195 ff.

94

JOURNAL

OF

PERSIAN

STUDIES

reflected in the hunt scenes on a gold sheath and a silver disk'39from the Oxus Treasure. Significantly it is in this very hoard that the closest parallel to the winged bulls on this belt may be found. Among the rich variety of objects reported with this hoard was a metal plaque,'40 said to be of gold, exactly like is the belt plaques in form, but not inlaid. On it, in repouss6, represented a winged bull in every detail like those on the belt under discussion (Plate IVe). Its date, as with all pieces in this hoard, can only be established on external evidence.'4' Some indication is provided by a pair of very similar belt plaques, in silver, from the second kurgan at Pazyryk'42 decorated in repoussewith an animal combat: between lion and ibex. The overlapping animals and their muscular stylization is as typical of Achaemenian art as is the motif. The distinctive treatment of the ibex horns and its " bow-and-dot " body markings appear on a magnificent ibex vessel handle included in the Oxus hoard.'43 These two plaques are objects of definite Achaemenian inspiration, if not actually of Iranian manufacture. The relative chronology of the five large kurgans at Pazyryk, extending over approximately fifty years, has been established on the basis of tree-ring counts-kurgans I and 2 are the oldest;'44 but the question of their absolute chronology is more complex. Rudenko has argued for a date in the fifth century B.C.'45 but Griaznov'46 argued for a fourth to third century date and Kiselev'47 has suggested something even lower. Azarpay,'48 after a detailed examination of the Greek and Near Eastern motifs in the art of Pazyryk, gave the beginning of the fourth century B.C. as the terminus post quemfor all the burials. This argument is supported by the conventional date for a broken Chinese mirror Pazyryk found in the sixth kurgan; a type ascribed by Karlgren to the fourth century B.C.'49 It also agrees with the generally accepted view that the definite Achaemenian imports at Pazyryk must have reached the area before c. 330 B.C. when Alexander's conflicts with the central Asiatic tribes terminated trade between the Oxus region and Iran. Owing to a fairly wide margin of error, recent Carbon 14 determinations do not add significantly to the discussion, though happily they confirm the main chronological limits set by conventional archaeological arguments. A beam from the covering of kurgan II yielded a date of 2350 ? 140 B.P. (1963) ; part of the covering of kurgan V gave a date of 2440 - 50 B.P. (1963)."5o The Pazyryk belt plaques may then be ascribed to the early fourth century which will also be taken here as the terminus post quemfor the winged bull motif used on the belt B.c., under discussion. The closest parallel to the swan hook on this belt is to be found on a gold object, perhaps a perfume bottle, now in the Hermitage, Leningrad, as part of Peter the Great's collection of Siberian goldwork.'5' It is made in the shape of a falcon standing on a swan or goose, whose body is represented by an openwork oval with the neck and head cast solid in exactly the form found on the belt hook.: Scale-shaped cloisonscover the falcon's neck and chest, larger cells are set on its wings and tail, two rectangular cells on its thighs. On the grounds of its technical similarity to the polychrome jewellery in the Oxus hoard Salmony dated this object to the fifth or fourth centuries B.c. and attributed it to the first phase of Sarmatian art.'52 Similar bird terminals were very popular on spoons and ladles'53 and on bracelets'54 in the Achaemenian Empire.
'39O. M. Dalton, Oxus,pl. IX, 22, X, 24-. '4oN. Kondakof, et al., AntiquitIsde la Russie Miridionale,Paris, 1891, fig. 251. 141 Franfaise Schlumberger : Mimoiresde la Diligation Archdologique

en Afghanistan,XIV, 1953, PP- 46-9, has presented a cogent argument for accepting the integrity of the hoard in contrast to the scepticism of Dalton-Oxus, p. XVI. Even so, as the most recent examination of the numismatic evidence makes clear-A. R. Bellinger, The American Numismatic Society, is still a chronological MuseumNotes, X, 1962, pp. 5 Iff.-there range of at least three hundred years, c. 500-185 B.c., or perhaps as low as c. 125 B.c. 142M. Griaznov, L'Art anciende l'Altai, Leningrad, 1958, pl. 41 ; for the motif and parallels at Pazyryk see : S. I. Rudenko, Kulturanaseleniya Tsentralnogo Altaya..., Moscow, 1960, fig. 151. '430. M. Dalton, Oxus, pl. V. 1o. '44I. M. Zamotorin, Sov. Arkh., I, 1959, PP- 2I ft.

1958, p. 104146L'Art Anciende l'Altai, Leningrad, 1958, p. 15. '47Drevnia istoria iuzhnoi Sibiri, Moscow, 1951, p. 373 (not available to me). 148 ArtibusAsiae, 22, 1959, PP. 313 ff. '49Ibid., p. 339, n. 215. 7, '5oRadiocarbon, 1965, P- 223. A. Is5 Salmony, Gaz. des Beaux-Arts,XXXI, 1947, p. 9, fig. I. '52 Ibid., pp. 9 if. 153 Materialsfor the Archaeology Russia, 1894, pl. VI, figs. 2, 3 ; of P. Amandry, AntikeKunst, I, 1958, pp. 13 ff. ; pl.I10, 1, 16 ; H. Luschey, Berliner Museen, LIX (4), 1938, p. 79, pl. 4; P. Amandry, Collection Hdline Stathatos, III, pp. 268, 269 no. I, pl. XXXVIII No. 181 ; D. Stronach, Iran, III, 1965, pl. XIIa, b.
'54 Oxus, pl. XIX,

45Sov. Arkh., 27, 1957, PP- 30I ff. ; Artibus Asiae, XXI,

142.

SOME

ANCIENT

METAL

BELTS:

THEIR

ANTECEDENTS

AND

RELATIVES

95

The insect motif, which I take to be addorsed bees, is not easy to parallel when, in a schematic bird's eye view, a number of insects can be strikingly alike in conventionalized representations. Indeed Zeuner has suggested'55 that many of Mrs. van Buren's identifications of flies in Mesopotamian art should be regarded as representations of bees.'56 In view of the inherent difficulties I will merely cite a number of formal parallels from Iran and the Asiatic periphery of the Achaemenian Empire, where bees were common.'s7 I do not know an example of this or related insects on metalwork of the Achaemenian period in Iran, though they occur earlier in Western Iran on a dirk'1s and a fragment of bronze,'59 both without secure provenance. The bee appears as a bead in Greek jewellery from South Russia'60 and was a very popular motif in Greece itself,'6' but until the arrival of the Sarmatians in the steppe region of South Russia neither flies, bees nor cicadas were a common decorative motif. Kithn'6' in fact ascribed the introduction of the " cicada " motif to the Sarmatians, who probably derived it from further east, where it appears early and persistently in Chinese art.163 The lineal descendants of the insects on this belt may thus be found among the " cicada " fibulae of the early first millennium A.D. Indeed examples from Kertch of the fourth century A.D. closely resemble them in form.'64 The motif was diffused across Europe into Merovingian France where bees or cicadas were sewn on to the purple shroud or placed in the coffin of Childeric I (died A.D. 482) at Tournai.'65 Techniqueand Chronology Close identity in the form of the two winged bulls and between the twelve winged insects indicate that the maker of the belt worked with a set of patterns, one for each bull and one for the bees. As the belt seems to have been worked mainly from the front it would appear to have been hammered over a pair of low relief cores, of stone or bronze.'66 Such a method would greatly facilitate the production of an object like this, in which a number of identical patterns were constantly repeated. Although the cores have usually not survived there is evidence that in the sixth century B.c. both Greek and Scythian jewellery was made in South Russia by hammering the sheet metal over a wooden core with punches of horn or bone.'67 Cores in bronze have survived in a Ptolemaic jeweller's hoard from GaljiAb in the Egyptian Delta'68 and a magnificent Scythian example is reported from Garchinovo in Bulgaria.'69 At present most of the evidence for the technology of fine metal work in the Achaemenian period comes from Egypt during the Persian occupation. Frankfort'70 identified a small-scale stone relief from Egypt of the Achaemenian period as a goldsmith's trial piece. It is known from the finds at Galjfib that Ptolemaic goldsmiths often worked in close collaboration with sculptors. In his excavations at Memphis Petrie found some small lead "trays ", which he suggested " were made for the silversmiths to carry with them, both for taking orders and for scaling out their work ".7"' One of these is definitely of the Achaemenian period,'72 the other is more likely to be Ptolemaic.'73 Also from Memphis in the Ptolemaic period come stone models of helmets and shields74 which almost certainly served as goldsmiths' models in a manner recently explained by Mis. Davidson.'75 In an important series of reliefs from the fourth century tomb of Pedusiri at Tfina the production of zoomorphic rhyta is shown in
'55A History of Domesticated Animals, 1963, p. 498. 156The Fauna of AncientMesopotamia, 1938, pp. Io8 ff.
'57 E. 158A.

and Greeks,1913, p. 6. Minns, Scythians Godard, Bronzes du Luristan (Graeffe Collection), pl. 18, No. 44. 159R. Ghirshman, ArtibusAsiae, XIII, 1950, fig. 4, p. 184. o60 Minns, op. cit., p. 399, fig. 318. E. 161 Jacobsthal, Greek P. Pins, 1956, pp. 73 ff. 162 Ipek, 10, 1935, pp. 85 ff. 163 Salmony, The Connoisseur, A. 1933, PP. 174 ff. ; E. Waterbury, Early Chinese Symbols and Literature..., New York, 1942, pp. 83 ff. 164Kiihn, op. cit., pl. 24. 52, 56. 165 Abbe Cochet, Le Tombeau Childe'ric, Paris, 1859, PP. 177 de I, ft;f. he reviews two centuries of discussion on the symbolic significance of these insects, which were re-instated on the ceremonial

garments and throne of France by Napoleon I and III. See also E. Salin, La Civilisation Merovingienne, Paris I959, IV, pp. 18o ff. See H. Hoffmann and P. F. Davidson, Greek Gold,1965, pp. 28-9 166 for the method. 167 A. Higgins, Greek R. andRomanJewellery, 1961, p. I I, n. I. A. 168 Ippel, Galjt2b, 1922, pl. 5-9. j69B. Filov, E.S.A., IX, 1934, pp. 197 ff. I7 J.N.E.S., IX, 1950, pp. II 1If., pl. III. and Memphis,III, 1910, p. 441'7 F. Petrie, Meydum 12. 7'2 Ibid., pl. XXXIII, 173 Petrie, The Palace of Apries (Memphis II), 1909, pl. XV. F. 174 M. Rostovtzeff, The Social and Economic Historyof the Hellenistic World, I, 1941, p. 391, pl. XLVIII, I, 2-helmets ; von Bissing, E.S.A., IX, 1934, PP. 221 ff., fig. I, 2-shield. '75 H. Hoffmann and P. F. Davidson, Greek Gold, 1965, p. 28.

96

JOURNAL OF PERSIAN STUDIES

some detail, but unfortunately not the earliest stages in the process.'76 The methods of chasing and polishing shown here were no doubt those universally used at the time. As far as it is now possible to judge, for much of it has worn off, the surface of each plaque in this belt was completely fire-gilded. In the Achaemenian period such gilding is more often used to pick out specific details in the design,'77 than as the overall treatment of an object's surface. In fire-gilding an amalgam of mercury and powdered gold is spread over the surface of the silver, which is then heated to vaporize the mercury, leaving the gold behind in contact with the silver. The early history of this process is still very obscure. Mercury was known to the Assyrians, but, as far as textual evidence goes, only for medicinal purposes and perhaps to colour small figures.1'7 Cinnabar, native mercury sulphide, was available both to the Assyrians and the Persians near Kirkuk in Iraq,'79 at a number of places in Persia'8s and perhaps also from Anatolia.'8s As far as I am aware no artefacts, fire-gilded rather than gilded by applying gold leaf to a base, with or without an adhesive, have been reported from the Near East before the Achaemenian period. In Egypt fire-gilding does not appear before the fifth century B.c. at the earliest,'8' and may well have been introduced into the country by Persian metalsmiths. In Greece's3 mercury is first mentioned by writers of the fourth century B.c. Theophrastus is the first to describe its preparation from cinnabar by treatment with vinegar and mentions that it has some practical uses, though he does not describe them. It was Vitruvius, in the first century B.C., who first recorded that " neither silver nor gold can be properly gilded without it (mercury) ",'84 and there are no unequivocal examples of fire-gilding in the Graeco-Roman world before the Roman period.'85 In discussing the belt's iconography it became clear that it was unlikely to be the product of a workshop under direct Achaemenian influence; the bees and the swan hook suggest the Asiatic periphery of their Empire. Examination of the belt's cloisonnedecoration endorses this conclusion and defines more clearly the time and place of manufacture. Knowledge of Achaemenian polychrome jewellery is still restricted. Few of the surviving examples are from good archaeological contexts, and it is difficult to date within less than fifty years those that are. Moreover, even where sockets or cloisons remain, the actual inlay has more often than not disappeared. Yet there may be no doubt of its popularity at the time. The Achaemenian love of colour and inlay was not confined to jewellery; it extended to all the arts. Herodotus'86 records that the seven circuits of the city wall of Ecbatana had battlements, each a different colour. Sculptures at Pasargadae and Persepolis were originally painted in blue, scarlet, green, purple and yellow ; details were covered with gold foil and sometimes inlaid.'87 At Susa the glazed brick friezes reflect the same fashion and also reveal the richness of colour used in the textiles and fabrics for which the area had long been famous,'88 only known otherwise from the textiles of Iranian inspiration, if not actually of Iranian manufacture, found at Pazyryk. Cups inlaid with precious stones are mentioned by Parmenion in his list of booty taken from the Persians in the time of Alexander. 89 The history of inlaid and cloisonng jewellery in Persia is slowly becoming apparent. It appears in Sialk IV and some time later at Susa,'90probably under Sumerian influence.'9' If Tukris'isto be located in Western Persia, the late Bronze Age inventories from Qatna suggest'9' an intimate association between this area and a particular type of inlaid jewellery. By the early first millennium it was being
de Lefebvre, Le Tombeau Petosiris,Paris, 1924, III, pl. VIIVIII. '77E. L. B. Terrace, Antike Kunst, VI, 1963, PP. 72-3, lists examples. and Geology, 178R. C. Thompson, A Dictionary AssyrianChemistry of 1936, pp. 29-30. '79R. C. Thompson, op. cit., p. 29. of 18oJ. R. Partington, Origins and Development AppliedChemistry, 1935, P- 415. on 181 Theophrastus Stones,ed. E. R. Caley and J. F. C. Richards, 1956, p. 195. D. Cooney, A.J.A., 68, 1964, p. 75. i8J. E. s183 R. Caley and J. F. C. Richards, op. cit., for following information on Classical sources; also D. E. Eichholz, de Theophrastus Lapidus,Oxford, 1965.
I76 G. 184VIII,
185

8, 4I am most grateful to Dr. R. A. Higgins for advice on this aspect of the problem. Uncertainty still attaches to the technology of the gilded relief of the Greek Orientalizing period in New York: G.M.A. Richter, Handbookof the Greek Collection,1953, p. 32, pl. 20oB. 186 I, 98. 187 E. Herzfeld, Iran in the Ancient East, 1941, Pp. 255 ff. i88M.D.P., IX, 1907, passim ; Herodotus, IX, 8o. i89Athenaeus : Deipnosophistae, XI, 782. 190 M.D.P., XXV, 1934, p. 21o, fig. 53-3 ; R. Ghirshman,

Sialk,I, pl. XXX, p. 69.

'9'

Cf. C. L. Woolley, U.E., II, pl. 138 ; A. Parrot, Tello, 1948, p. io8, fig. 26c. I92J.Bottero, R.A., 43, 1949, p. 22, I43-

SOME

ANCIENT

METAL

BELTS:

THEIR

ANTECEDENTS

AND

RELATIVES

97

produced in north-west Persia,'93perhaps initially under Assyrian influence. The technique was probably commoner in Syria and Assyria than the scanty existing evidence allows. The range of inlaid ivories, first revealed in nineteenth-century excavations94 has been greatly increased by Mallowan's work at Nimrud.'95 Many were gilded and inlaid with red and green frits, glass and lapis lazuli. Cloisonne work was used by the jewellers of Tell Halaf'96 and a cloisonnibreast plate is worn by the amber statuette of an Assyrian king, now in Boston.197 It may be that the Scythians adopted the fashion after their contacts with the Medes in north Persia, for it appears, if rarely, amongst the earliest examples of fine metalwork ascribed to the Scythians in South Russia. Simple cells of triangular or rectangular shape were used on objects from Kelermes'98 and a horse's head from the U1 barrow'99 in the late seventh to sixth century B.c. The almond-shaped cloison also appears for the first time in this period on a cast bronze lion from the " Golden barrow " near Simferopol.o00 The form probably has an Iranian origin, for this shape is cut for inlay on the lion handles of a gold vessel, now in the Kevorkian Collection20' which may be ascribed to the Median period, c. late seventh to sixth century B.C. As far as present evidence goes it seems that blue stones - lazulites and turquoises alone or combined with yellow ones, were preferred by Achaemenian and earlier workshops in Western Asia.'o2 The appearance of red stones, garnets or zircons, was a fashion introduced from the Asiatic periphery of the Achaemenian Empire by the Parthians who had inherited it from their predecessors in that area. Garnets came from India and Ceylon along a trade route which ran through Bactria, down the Oxus to the Caspian. Other sources existed in Siberia. It is the use of garnets and their placing on this belt which most convincingly associate it with this area in the very late or immediately post-Achaemenian period. It is a characteristic of Achaemenian polychrome jewellery that the inlay or cloisonsare placed to emphasize the stylized musculation of animals or monsters. By contrast the inlay on the Siberian goldwork203 is often lavish and indiscriminate, in many instances concealing rather than accentuating the organic structure of the creature to which it is applied. The garnet inlays on this belt have nothing to do with the anatomy of the animals or insects to which they are fixed, but they are applied with restraint. At Pazyryk almond-shaped inlays in felt were used on the neck of an elk and on the mane and ribs of a griffin, usually without any direct relevance to anatomical forms.204This use of inlay and cloisonne work for pure ornament seems to be a distinctive trait of Asiatic " nomad " art. It is generally agreed that the great development of polychrome decoration from the fourth century B.C. onwards in the artistic tradition conventionally attributed to the Sarmatians was stimulated by close contact with the jewellery of Achaemenian Persia.205 Speiser'26 argued that the well-known pair of polychrome bracelets in the Oxus hoard,'07 from the final phase of Achaemenian art,'"o provide the crucial link between it and " Sarmatian " art. To this transitional phase I would also attribute this belt. In almost every respect, technically and iconographically, it belongs to a new artistic tradition, but the winged bull is so characteristically Achaemenian and so like the one on the gold plaque from the Oxus hoard, that it seems unlikely that this belt was manufactured more than fifty years after the end of the Achaemenian Empire. Such distinctively Achaemenian motifs are not found in developed " Sarmatian " art. Indeed it could have been made in the half century before c. 330 B.C., but this seems less likely. It is far more difficult to say whether it was made in a workshop west or east of the Caspian.
Porada, AncientIran, 1965, fig. p. 119 ; Ziwiyeh A. Godard, Le Trisor de Ziwiyd, 1950, p. I04, fig. 90. P. Amiet, Revuedu Louvre,1966, pp. 93 ff., fig. 1-2. A vase of " Marlik " type with chariots like those on the Hasanlu Gold Bowl and a frieze of animals in silver inlaid with coloured pastes. 194R. D. Barnett, A Catalogueof the Nimrud ivories in the British Museum,1957, pp. I56 ff. s95M. E. L. Mallowan, Nimrud and its Remains,1966, II, ffPP. 554 i96B. Hrouda, Tell Halaf, IV, 1962, pl. I, I.8a, p. 19. '97 Olmstead, Bulletin of the Museumof Fine Arts, Boston, 36, 1938, pp. 78 ff. ,98M. Rostovtzeff, Iraniansand Greeks S. Russia, 1922, pl. IX. in
193 Hasanlu-E. '99 Ebert, Real. der Vorg.,XIV, pl. IB.
200

Borovka, Scythian Art, I928, pl. 15A. Ans, Paris, i961, No. 159, pl. LI ; see also R. D. Barnett, A Catalogue theNimrudIvories..., pl. CIX, S.389a, b. of 202 O. M. Dalton, Oxus, p. XLIV. 203 E. Minns, Art of the NorthernNomads, pl. XVII B. 204M. Griaznov, L'Art anciende l'Altai, 1958, pl. 48, 49. 205 Salmony, E.S.A., XI, 1937, p. 98. 2o6 Vorderasiatische Kunst, 1952, P-. 132-3. 207 British Museum, Oxus, pl. I No. 116 ; Victoria and Albert Museum, E. Porada, AncientIran, I965, top plate on p. 173This dating seems more acceptable than the earlier one given 208 by Dalton and others, see P. Amandry, AntikeKunst, I, 1958, p. 22.
20o Sept Mille

98

JOURNAL

OF

PERSIAN

STUDIES

In view of the date I have indicated and the links with the Oxus hoard I would tentatively suggest a source in Bactria or Sogdiana. This belt is clearly a luxury object for ceremonial rather than routine use, for a man rather than a woman. Although it is known that various awards for valour and service were made in the ancient Near East they do not appear to have included belts. But since daggers are the most common award it may perhaps be assumed that such a gift included a belt or baldric. Though the Egyptians usually awarded gold flies as military decorations,209daggers, sometimes bearing the name of the royal donor, were also presented.210 In Assyria Sennacherib rewarded the builders of his canal with gold daggers and rings21' and Ashurbanipal, in his own words : " wrote my name (phonetically) upon an iron dagger (to be worn in) the girdle the mounting of which was golden, and gave it to him (Necho)."212 Kings of the second Isin Dynasty, and more rarely their officers, presented daggers to mercenaries from Western Persia for acts of valour.2,3 Among the peoples who migrated from central Asia into Europe across the steppe the warrior's belt alone was a reward for valour. All the available evidence is considerably later than the belt under discussion, but the general idea may have been current much earlier.214Among the Avars and the Franks the power of a defeated foe passed to his conqueror with his belt and later among the Mongols alliances were concluded by the exchange of belts captured from enemies. It is possible, by analogy with Avar practice, that this belt is the second of two belonging to a warrior. The first, of simpler form, would have carried his arms, the second, more elaborately decorated with totemic emblems-perhaps the bees-his cup.225 APPENDIX
REPORT ON THE LEATHER

By M. L. RYDER
The thickness of the specimen, combined with the coarseness of the collagen fibres seen in sections, suggests that the material is ox-hide, but no hair remains that would confirm such an identification (cf. Ryder, 1964)216 were found in any of the sections cut from two separate fragments. The specimens were appreciably degraded, and so there is no reason to suppose that it is not the original mounting. An overall maroon staining reaction was obtained with the method of Ryder (I 963),217 which suggests vegetable tanning, but it is possible that a similar reaction could have been brought about by long exposure to the earth through burial. It is of interest to record that the sections had an outer zone with a pale brown colour, which had presumably arisen by the removal of the basophilic (maroon staining) elements, possibly by the alkali in the softening fluid used prior to sectioning. NOTE ON PLATE IV (a) (b) (c) (d) (e)
209 W.

After Herzfeld, Iran in the AncientEast, P1. LXXXIV. After Mostra d'ArteIranica, Rome, 1956, P1. XXX. After Sept Mille Ans d'Art en Iran, Paris, 1961, P1. LXV. After H. Frankfort in J.N.E.S., IX, 1950, P1. III. After N. Kondakof, Antiquitis de la Russie MIridionale, Paris, 1891, Fig. 251.
Gallery,Baltimore, 1961, p. I I, n.I3 lists the evidence ; the following comments draw heavily on his remarks. 215 G. Laszl6, Archaeologia Hungarica,New Series, XXXIV, 1955, p. 163, fig. 47. 216 Ryder, M. L. (1964), in Coles, J. M., Coutts, J. and Ryder, M. L. " Report on the skin " in A Late Bronze Age Find from Gold, Cloth, Leatherand Angus, Scotlandwith associated Pyotdykes, Proc.Prehist.Soc., 30, 186-198. remains. Wood 217Ryder, M. L. (1963). " Remains derived from skin " in Brothwell, D. R. and Higgs, E. S. (Eds.) Sciencein Archaeology (Thames and Hudson).

von Bissing, Praehistorische Zeitschrift,XXXIV-V, 1949-50, p. 217. W. C. Hayes, The Scepter 210o ofEgj)pt, II, 1959, P- 77, fig. 40. z21 A. R. Luckenbill, AncientRecords Assyria and Babylonia, II, of 1927, No. 337, P. 151. 212 J. B. Pritchard, ed. A.N.E. T., 1950, p. 295. 213 Nagel, AfO, XIX, 1959-60, pp. 95 ff. ; Dossin, Iranica W. Antiqua, II, 1962, pp. 149 ff.-I follow here the more widely accepted interpretation of these weapons, without denying that some were definitely votive. 214 P. Verdier in : Arts of the Migration Period in the WaltersArt

99

QAL'EH-I YAZDIGIRD*
A Sasanian Palace Stronghold in Persian Kurdistan By E. J. Keall
The site of Qal'eh-i Yazdigird' occupies a naturally fortified position in the Zagros range of mountains in western Iran, near Qasr-i Shirin. It is bordered on the western and northern sides by a sheer escarpment that falls away to the plain of ZuhTb, and on the eastern side by steep cliffs that

build up into the higher reaches of the Dilahfi mountain. A broad shoulder overhanging the Diyala plains forms a saucerlikeplateau, which has created a catchment area for the winter streamsthat have
worn a precipitous ravine through the scarp, plunging down to join the main stream of the Halwin

River on the Plain of Zuhab (v. Fig. I).

Zuhab, which is generally accepted as being the site of the medieval city of Halwan.2 This involves a
journey of four farsakhs, including the ascent of the formidable escarpment. An even more difficult ascent to the tableland could be made at the northern end of the plateau.3 The easiest approach to the site-despite the fact that it entails crossing the stream of RijTb, which in time of flood is quite unford-

It is across this plain that the approach to the fortress can be made from the direction of Sar-i Pul-i

able-is that which leads over a distance of fourfarsakhsfrom Taq-i Girrah, a solitary arch of lime-

it meets the foot-hills of the Zagrosjust beyond Sar-i Pul. From this point it is a steep ascent to the top of the pass at Pa-i Taq-i Balatar (gendarme post), which marks the last difficult rise before the long gradual climb towards the plain of Kirind and beyond to Kirmdnshdh. Seleucid period, but more probably to the Sasanian. It has been variously described as a toll house marking the boundary of Media ; as a hunting lodge ; or as a royal resting station beside the great highway. " Madharfistdn, a place on the road to Khurasan from Baghdad, situated two days' march from Halwdn towards Hamadan ; and Marj-al-Qal'eh is one march from it. At this place there is a large
* The Editors apologize for discrepancies between the transliteration of place-names on the maps and figures and that in the text. The site is mentioned under this name, together with a short description, by Major Rawlinson, " Notes on a march from Zohab, at the foot of the Zagros, along the mountains to Khuzistan, and thence through the province of Luristan to Kermanshah ". Journal of the Royal GeographicalSociety, vol. IX, 1839, p. 33: Colonel E. I. Tchirikov, "Journal de voyage du Commissaire arbitre Russe pour la ddlimitation de la frontiere Turco-Persane en 1849-52 " (in Russian). Mimoiresde la sectionCaucasienne la Socidtd de Russe, Gdographique vol. IX, 1879 : and P. Schwarz, Iran im Mittelalternach dem arabischen vol. Geographen, VI, Leipzig, 1929, p. 683. I am indebted to S. C. R. Weightman for my original information regarding the existence of the site. I made my first visit, acting on his instructions, in the spring of 1964" March from 2 Rawlinson, ZohTb ", J.R.G.S., IX, p. 35: G. le Strange, Landsof the EasternCaliphate,Cambridge, 1930, Reise p. 19' : and Fr. Sarre and E. Herzfeld, Archdologische im Euphratund Tigris Gebiet,vol. II, Berlin, 1920, p. 833 Rawlinson, in examining the possibilities of an alternative route in antiquity from Mesopotamia to Kirmanshah, remarked that this was not a feasible route for general communications. " March from Zohab ", J.R.G.S., IX, p. 35. 4 Idem,p. 345 E. Flandin and P. Coste, Voyageen Perse, 184o-41, vol. I, Paris, 1843-54, P- 465 ; vol. IV, p. 172-3, 215 : J. de Morgan, " Recherches Archdologiques ", Mission Scientifique en Perse, vol. IV, Paris, 1896, p. 335-9 : Fr. Sarre and E. Herzfeld, IranischeFelsreliefs, Berlin 1910o, p. 232-5, pl. XLVII : E. Herzfeld, Archdologische aus Mitteilungen Iran, vol. II, 1930, p. 8o: and O. Reuther, " Sasanian Architecture ", Surveyof PersianArt, vol. I, p. 493, 509, 570, fig. 136 ; vol. IV, pl. 153C. 6 Rawlinson, " March from Zohib ", J.R.G.S., IX, p. 34; and cf. Le Strange, Lands of the Eastern Caliphate,p. 191 : and Schwarz, Iran im Mittelalter,vol. VI, p. 693-

stone masonry, which marks the physical division between Babylonia and Media,4 and lies just short of the summit of the pass termed the " Zagros Gates ", the ancient line of communication between these two provinces. The highroad from Baghdad rises gradually over the low eroded dunes from Qasr-i Shirin until

The Tiq-i Girrah has already been studied exhaustively, though the conclusions drawn have been almost as numerous as the studies made.5 It is now generally accepted that it does not date to the

Rawlinson suggested that the Taq-i Girrah may have formed part of the palace of Midharfistin which is recorded by the Arab geographers who described this region.6 Both Yaqilt and Qazvini state :

A\

47
. x.. . .
,, BA
r

.".~~~~\ .
?
Oasr
e~~

"?? ." .... "

*
...
.

~ ".~ .."..B
Oos
e?.4'
-

~
,',

~ APYazdgir~d .ab.:. Zoab S Put 'e1 e., Sar"" "E ".'RIJAB"

. " "e.n. 0 D... S A . "rP
.= r 'DOKKAN GE

~
.' . ... e "...-
e

~~1

,"] S ii AJIQaleh ...,. CHAAR

..

S hirin .

? "

/.,Khosrov
?. ..

.%? 4,

oo
... , "", 41
? ,,...

06 f
0S; 0

TAQ//7

.. .: 10 0 .. ., '. " . r : " . .:... -: .

: . o
"
-

??

.'' '?.:.

" .. ..' '. ?0-....":.Q 0 ?.

.a" '0/. .i
:

V~ ~~~.~r?u r

" ."

a 0 46 ? ...

\.

so? .DKAN .... QALE E 8, YADGRDE :;.

E 0 o & 00 00 0 P a ePI s 0l ...7000-9 .cs

:.

E GR
AV
Z4r

"~~~~~~~.? ..
...

10

... ... .20k"7 /

?

" .. ... ". QALEH ..c \\<>

,

" :.
___1"**

,i+

YAZDGIRD

r -:
SALE i
0,00_____000.0

'--'ov
a

AI

,~
______________0/

SH

Fig. I. Qal'eh-i razdigird: Relief Map.

QAL'EH-I

YAZDIGIRD

101

aivan and a great terrace in front of it, and the traces of a ruined garden built by Bahrdm Gfir. They declare that the snow falls in the quarter of the hilly district ; but never falls on the side which faces Iraq."7 However, the arch of Taq-i Girrah does not fit the description of the aivdn at Madharfistan. It is an isolated monument by the highway. There are neither signs of there having been water courses and terraced gardens nor traces of any subsidiary building, which would have been essential if there had been a palace. Indeed, the hillside presents a most rugged boulder-strewn terrain. Furthermore, the total distance between Sar-i Pul and Kirind (identified with Marj al-Qal'eh) is no more than fifty kilometres, or eightfarsakhs, which can hardly have represented three days' march, unless a considerable detour was made from the highroad. It must be remembered that in Yaqfit's time Midhariistan had long been abandoned, and there may no longer have remained an accurate record of the distances. On the other hand, the earlier records of Ibn Khurdadbih and Quddma indicate that the distance of Mddharfistan from Halwin is only four farsakhs, while Muqaddasi gives " two postal stages ".8 These are the distances favoured by Le Strange and Schwarz in their respective works.9 The shorter distances may mean that it would be possible to correlate Qal'eh-i Yazdigird with Mddharfistdn, particularly as the description of the site seems appropriate. The north-western portion of the plateau overlooks the Iraqi plain, and it is unlikely that the snow would lie for long in this area. The citadel, by contrast, is considerably more elevated, and even in summer is noted for the coolness of its breezes. In winter the massive hills behind carry a perpetual blanket of snow, while the area of the orchards is subject only to falls that very soon melt away. It would be difficult to find a more appropriate situation to fit the description of variations in climate within such a limited distance from Halwin. Moreover, the reference to the gardens and a great arched building may also seem fitting when the subject of the ruined pavilion and enclosure wall of Gach Gunbad is discussed. (v. infra.) Divdr-i Gach'0(v. Fig. 2) An impressive line of defences stretches for two and a half kilometres across the southern and open neck of the plateau, and thus completes the total circumvallation of the tableland, which on its other sides relies more on the strength of natural protection. The defensive wall of field stone and mortar follows an irregular course, hugging the bank of a deep gully, and climbing up towards the cliffs that mark the back edge of the tableland (v. P1. Ib). The only weak point is at the lower end where the ground is reasonably level and open to approach from the direction of Thq-i Girrah. In the lower reaches the wall is approximately 4 m. wide. In other places the single thickness is in the region of 2 5o m. At the upper end, where the defences are carried up the cliffs, the width is reduced to 2 m. In this section the stones employed are relatively small, owing to the steepness of the terrain which has made the importation of materials difficult. Vertical bands of plaster mark clearly the emplacement of formwork to construct a solid mass entirely dependent upon the bonding powers of gypsum mortar. Except for some brick walling found in Gach Gunbad, the standard building material throughout the site is this stone and mortar masonry. Reuther states that "in this type of construction the stones are not really built up like masonry, but are simply piled on top of each other, without any bond, the walls being modelled, as it were, in the rapidly setting mortar, and the stones packed in only as a filling. Were it not for the mortar, the walls would immediately tumble into a shapeless heap."" This method of construction is not possible with lime mortar which hardens more slowly, binding the bricks or stones by forming a chemical compound, in contrast to the mechanical bond of gypsum mortar.'2 It is a technique found in both early and late Sasanian structures in Fars, where "... the method was
7 Zakariya b. M. b. Mabmfid al-Qazvini, " Atharu-l-Balad", BibliothecaGeographorum ed. Arabicorum, B. J. de Goeje, 1906, 2, 302, 23 : and Yaqfit ibn 'Abd Allah al-Hamawi, " Mu'jam 9 Schwarz, Iran im Mittelalter, vol. VI, p. 693 : Le Strange, io The local names provide convenient labels to refer to the

Landsof the EasternCaliphate,p. 191.

al-Bulddn ", Geographisches ed. Wdrterbuch, Wustenfeld, Leipzig, I866, C7, 354, 7. I am indebted to Dr. A. D. H. Bivar for the translation of the passage from Qazvini. 8Quddma ibn Ja'far, B.G.A., de Goeje, vol. VI, I98, 4 : and Muhammad ibn Ahmad al-Muqaddasi, B.G.A., de Goeje,
vol. III, 135, 6.

different parts of the site. The use of the word "gach" (gypsum) is frequently found to apply to ruins of a Sasanian date in Iran, where this material abounds. Here it means " gypsum wall " " Reuther, " Sasanian Architecture ", Survey,vol. I, p. 498. SReuther, " Parthian Architecture " Survey,vol. I, p. 427.

102

JOURNAL

OF

PERSIAN

STUDIES

Ashiaba

s z

B&abohb

Yadga

shr1e_.IIIMfi

"".,-,,,.'
I-

orch-rds .ar.......

'>l(i~d

-,

Nctqqareh

,\,1,lit"

,! it9/ .....

it,
path to Shah Neshin J
S (sond.age

. , *..

" . .Diva

Gac

a ..orchards ,,- -',."

QSkLehhBata

eh r

ftIf> 91

Tepe0

Rashrw

Sketch
I
EIlA II

Map

Gcach
Gumbad

QLHEYZGR OA

IfDivar

e Gach

june ma".-north
1965" '

G" ""

?"-hpa ",

.
,r

jeep track

Rjab
to,,,--

e:
Sca"

o 100

500m - .-

1km

pt o Scarp &\ P
S..". . . "--Drwazeh

... -

Fig. 2. Qal'eh-i Tazdigird:SketchMap.

obviously indigenous, and spread from there to the frontier districts between 'Iriq-i 'Arab and Ardiln, where similar conditions prevail."' Even assuming that the architects of the early Sasanian palaces were employing for the first time in major construction a material that had previously been used only in domestic architecture in Fars, it can be argued that the technique was already known under the Parthians. However, it was exploited most fully by their successors. Once established as the mean for palace architecture, it was readily adopted in the western portion of Iranian plateau, as at Qasr-i Shirin, when this area became popular for the development of royal residences.'4 Its use there, and at Qal'eh-i Yazdigird, must be considered typically Sasanian.'5 Throughout the main length of the Divir-i Gach, the wall is strengthened on the outer flank by a series of rounded towers, of a typically Sasanian type: a stilted semi-circular tower, in which the length
" Sasanian Architecture ", vol. I, p. 498. Survey, 4 The best general survey of the ruins of Qasr-i Shirin is in G. L. Bell, The Palace and Mosqueat Ukhaidar,Oxford, 194 . 15 It is noteworthy that the buildings of Qasr-i Shirin are constructed in a crude style and lack any refinements of structural detail. The Chahar Qiipi, though impressive from Reuther, '13 its sheer size and dome span which measured 18 m. across, relies entirely on the massiveness of its solid piers for the load-bearing support. Though later in date than the monuments of Fars, the palaces of Qasr-i Shirin represent a deterioration of architectural principle and the blind adoption Sof a material without realizing its full properties. cf. Reuther, " Sasanian Architecture ", Survey,vol. I, p. 553.

QAL'EH-I

YAZDIGIRD

IO3

Coleh e Yozdgird
Scale: 0 5 10m

COMPARATIVE WALL DIMENSIONS

JA Y DAR

IBAL

GACH GUMBAD

wall dimensions. Fig. 3. Qal'eh-i Tazdigird: comparative

extending from the face of the wallis equal to its own diameter, and the sides are initially straight (v. Fig. 3). A similar type of tower occurs at Ctesiphon, Takht-i Sulaiman, Qal'eh-i Gabri, and Dastigird (the latter in baked brick).'6 It is notable that when Khusrau II ordered the rebuilding of Antioch in A.D. 540, at its new site near Ctesiphon, the buttress towers were constructed with a rectangular form that was more typical of the West.'7 Indeed, although the rounded tower was known and employed in Mesopotamia as early as the Assyrian period, it did not really come into vogue until post-Seleucid times. Unfortunately our knowledge of the Parthians is too limited to make dogmatic statements on the prevalence of either architectural form in that period. There is, for example, a mixture of types in northern Mesopotamia. But as far as the Sasanians were concerned, the rounded tower was employed almost without exception in stone construction. It is concluded, therefore, that these walls are of Sasanian date. The towers seem to vary in size according to their position and strategic importance, extending from between 4 m. to 8 m., with a similar variation in breadth. On the cliff section they are reduced to 3 m. by 3 m. Accurate measurements cannot be recorded until the foundations can be exposed in excavation ; because of stone robbing and decay, it is impossible to record the intervals between each
i6 At Ctesiphon the es-Sur city wall (mud brick) is io m. thick,

with towers 9 30 m. in diameter ; the turrets of the al-Tuwaibah wall are smaller. O. Reuther, Die Ausgrabungender deutschen im Ktesiphon-Expedition Winter1928/9, Berlin, 1929. (English Translation in "The German Excavations at Ctesiphon", Antiquity, vol. III, 1929, p. 451). At Takht-i Sulaiman the towers of the stone perimeter wall are similarly stilted. H. H. von der Osten and R. Naumann, Tahkt-i

'7

Suleiman, Vorlaufiger Bericht iiber die Ausgrabungen 1959, Berlin, 1961, p. 39-53. At Qal'eh-i Gabri the stilted turrets flank a wall of rubble masonry 2 m. thick. A. Stein, " An Archaeological Tour in Ancient Persis ", Iraq, vol. III, 1936, p. 142. At Dastigird towers of baked brick, measuring i i-6o x 0-20 m., flank a wall 16-6o m. thick. Sarre and Herzfeld, Archiologische Reise, vol. II, p. 91, fig. 176. vol. I, p. 574. Reuther, " Sasanian Architecture ", Survey,

104

JOURNAL

OF

PERSIAN

STUDIES

tower. Heaps of rubble which seem to indicate their remains are generally spaced from 15 m. to 25 m. apart. In one stretch where the wall is better preserved, the curtain measures 23 m. and is pierced by a series of loopholes (v. Pl. IIIa). Each embrasure cuts through the full 2 -50 m. thickness of the wall, narrowing from 6o cm. on the inner face to a gap of o20 cm. on the outer face. The loophole is i '20 m. deep, with its base 50 cm. up from the foundations. There is an interval of 2 m. between each embrasure. One loophole is blocked by an adjoining tower, suggesting that though contemporary, the towers were constructed in a secondary stage of the building process. In addition, there are traces of chambers abutting on to the inner face of the wall, from which refuge point there was a clear view to the outside.

In all probability there were a number of these apartments attached to the defences. The best preserved examples are a group towards the lower end of the Divir-i Gach, in the area known locally m. wide, they are roofed by a barrel vault which is as the Ashpaz Gdh. Measuring 5 m. long by 3.30 the standard form of ceiling in Sasanian buildings.'9 Such vaults, formed of the same rubble and mortar composition as the walls below, can be produced either with or without centring.'0 In the former case only light framework is necessary to produce a soffit in rapidly setting mortar, over which the rubble infill can be added. The removal of the centring leaves a slight recession where the springing of the archivolt is set back on the impost. The effect could be enhanced in stucco to create a cornice, as in the Qal'eh-i Dukhtar at Firfizdbhd." That such an interval was attractive to the Sasanian architects is known from the arch at Taq-i Bustain, where it is executed in solid rock as a purely stylistic device and not as a structural necessity.2 Alternatively, a barrel vault can be constructed without centring by the gradual shift of masonry from the extremities of the impost to the central point of the arch. To achieve this, the springing may often begin quite low down ; the resulting parabolic or half elliptical section, which is not a consideration of strength, is a direct result of this structural method. It is this type of construction that appears at Qal'eh-i Yazdigird. In some cases the workmanship may be faulty, producing a slightly pointed arch; such discrepancies were remedied by the application of plaster. In these cases the diameter of the intrados would be less than the full width of the interval spanned, in the same way that the corbelling at Sarvistdn reduced the interval spanned by the vault.23 The Sasanian builders seem to have deliberately avoided the visual effect of the pointed arch. For no Sasanian building has yet been discovered in which this form is revealed, whether in corbelled construction or true vaulting. There are a few examples of the pointed arch illustrated in arcades on metal-work of the period, in which the form tends to be more akin to the ogee or keel shapes of India.'4 The chambers in this part of the defences must have formed part of the garrison barracks in conjunction with an isolated but adjacent building that takes advantage of a spur at the end of a natural ridge. This seems to represent the quarters of the officers of the garrison. A series of small chambers run on either side of the entrance ; the walls have been pierced by loopholes similar to those in the main defences. The compound is extremely denuded (one wall has been eroded away down the steep incline), but can be seen to measure 145 m. by 35 m. The walls, 2'20 m. thick, were flanked with towers circa 3 m. by 3 m., and roughly 20 m. apart. Darwdzeh At a point where the defensive wall crosses the head of the ravine (already mentioned as the outlet for the streams from the tableland), and follows the brink of the cliffs until their steepness prevents any access, a large shapeless block of masonry survives in its position of guarding what may have been the
" kitchen " of the great palace. '9 e.g. A. Stein, Old routesin Western Iran, London, 1940, p. 68 ; S18 : and " Ancient Persis ", Iraq, vol. III, p. I42 sqq. 20Reuther, " Sasanian Architecture ", Survey,vol. I, p. 499 ; fig. 128: and K. A. C. Creswell, Early Muslim Architecture,
vol. 11, Oxford, 1940, p. 61. x8 These ruins are interpreted locally as having been the

Gdh,'8 A.shpaz

21M. Dieulafoy, L'Art Antiquede la Perse, vol. IV, Paris, 1885, fig. 24-26; 43. 22 E. Herzfeld, Am Tor von Asien, Berlin, 1920, pl. XXXI, XXXIII. 23 Dieulafoy, L'Art Antique,vol. IV, p. 12 sqq.; fig. 14-19; pl. I. 24 Reuther, " Sasanian Architecture ", Survey,vol. I, p. 512-14.

QAL'EH-I

YAZDIGIRD

105

gateway area. It is in this region alone that there is a reasonably level track ; and there is no evidence
of an entrance further up the wall. stretches of walling still cling to the actual rim of the precipice. Wherever a gulley offers access to the top, the fissure has been blocked off with masonry, so that the complete circumference of the enclosed area amounts to approximately twenty kilometres. Similar attempts to secure an elevated tract, though

Great attention has been paid by the builders to the total encirclement of the plateau-isolated

on a lesser scale, may be seen in the defences of Qal'eh-i Dukhtar at Firfizabdd and Qal'eh-i Gabri
near Fasd."5There too the walls have been carried up to towering pinnacles and can be traced along the edges of formidable escarpments. The local name of "gateway" has been applied to this segment of the plateau. It is from this point that the weekly caravan from the village of Ban Zardeh leaves to visit the bdzdr town of Sar-i Pul, following a tortuous path down to the plain. Slightly to the north of this track there are signs that a pathway had once been engineered-its purpose being to bypass a gully that can be dangerous in winter. These aids to navigation have long since disintegrated, but even in its pristine condition the original route must have presented a fearsome ascent.

Qal'eh-iYazdigird (Upper Castle) The cliffs that form the back-cloth to the plateau are in their turn surmounted by eroded pinnacles, and finally by the massiverange itself. It is upon one of these pinnacles that the defences culminate in a lofty citadel which has given its name to the whole site (and is referredto as Qal'eh-i B~ld on the map to avoid confusion). It perches above the cliffs, with steep slopes of scree on all sides, except for a narrow neck of land which links the fort to the higher ground behind (v. P1. Ia). The path from the lower tableland follows the stream of Ab-i Ghuslan round in a steady rise past its source near the shrine of Bdbd until it reaches the further side of the castle which it approaches by crossingthe neck of land Yddgar,26 already mentioned. Beyond this point the track climbs up to the summer grazing grounds, and eventually passes on to G5wdreh. In its lower stages before the path reaches the Upper Castle, it has to pass through the jaws of a narrow gorge through which flows the stream of Ab-i Ghusldn. The jagged cliffs of these narrows have been made accessible to movement along their tops by the addition of masonry, to give a completely commanding control of the pathway. As a last line of defence it would present an overwhelming obstacle, and properly manned, would be impervious to direct onslaught from either direction. However, the theory of Rawlinson that this was the ultimate refuge for Yazdigird can be discounted in view of the difficulties of maintaining supplies of food and water if it were once cut off from the rest of the stronghold. The function of the castle can best be interpreted as a sort of barbican against attack from the high ground. Any invaders who forced their way down to the jaws of the gorge, the last obstacle before reaching the tableland itself, would by their very action have turned their backs on the defenders of the castle, and become entrapped within the narrows. The entrance to the fort is marked by an extended length of walling that curved round like an arm and protected the gateway. It also served as a stabling quarter, if the small divisions on the inner face of the wall can be interpreted as stalls. The main wall is obliged to cross some very steep ground, so that the towers are really supporting buttresses, and in some cases project outwards at the base to accommodate the incline. Set at close intervals of between 6 m. and I6 m., they vary in size between
2 m. and 6 m. in diameter (v. Fig. 3 and P1. IIIb). A broad tower at the north-eastern corner measures I I m. across and includes a room within its span. Within the fort, the barrel-vaulted ceilings of some of the small chambers span an interval of 3 m. A deep cellar beneath the gateway area measures 9 m. long, m. wide, and has a vault standing m. above the floor level. A large proportion of the perimeter 3"70 4.50 is standing to a height of 5 m., and in one section there is an intact face Io m. high. walling Almost half the castle, on the southern side, is at a much lower elevation than the rest. In this quarter a few fragments of polychrome glazed pottery indicate that the buildings were occupied during
25 Stein, " Ancient Persis ", Iraq, vol. III, p. 123; I42 sqq.
z6

For the significance of the shrine and a full bibliography see V. Minorsky, Notes sur la Sectedes Ahle Haqq, Paris, 1921.

io6

JOURNAL

OF

PERSIAN

STUDIES

the early years of Islam. The masonry does not, however, suggest that they were constructed after the building of the site as a whole. The castle forms a logical part of the defensive network. These potsherds probably represent squatter occupation, since it is unlikely that a garrison was housed here by the conquering Arab army. The position is a defensive rather than a commanding one, and is too difficult of access to provide a satisfactory home for any other than the militia. It would, however, have afforded the ideal refuge for an organized band of robbers intent upon plundering the caravans of the great highroad. As part of the defensive network, the view that the castle enjoys is supplemented by two adjacent and smaller structures, which surmount similarly isolated pinnacles above the line of the cliffs. From the combination of these three it is possible to command a view over the whole of the tableland and to trace the highroad running through Sar-i Pul, some 15 kilometres away as the crow flies, and three thousand feet below. The more northerly post is called "dshiabd ", and the other " naqqdreh khdneh 2.7

Jd-i Ddr (v. P1. IIb)

The watered gardens of Bdn Zardeh, which are limited in extent by the volume of water that the Ab-i Ghuslin can provide, partly cover the ruins of a heavily fortified palace. The entrance to the Ja-i Dar compound is provided with the same shielding arm to the gateway as found in the Upper Castle. Flanked by huge buttress towers, measuring 8 m. in diameter on a stilted semi-circular plan, the ruins reveal traces of an internal communication system within the defences-narrow corridors and small chambers running the whole length of the wall, which is itself 6 m. thick. The best example is on the north-eastern side, where a tower accommodates a chamber 8 m. long and 210o m. wide ; m. width and a sloping roof indicate that there was a passage leading at one end, the doorway of I o30 from this point. A window with embrasure 60 cm. wide penetrates the remaining 2 m. thickupwards ness of the wall through to its inner face. On the other side of the chamber, two approaches, I -20 m. wide and at different levels, form the connection from the main corridor system into the tower itself (v. Fig. 3).Asimilar pattern was probably repeated at all the other towers. The north-western and south-western sides of the enclosure each have five towers as well as the corner bastions along a face measuring 16o m. The south-eastern stretch splays out at an angle, so that the whole forms an irregular square. The m. north-eastern reaches continue for another 100oo beyond the 16o m. mark to form part of the gateway with a further 60 m. at right angles to this to complete the shielding arm, which faces down extension, towards the Darwizeh area-the suggested entry-point to the tableland. The centre of the enclosure, of which the whole is subject to the considerable slope of the terrain, is distinguished by a raised mound whose surface is liberally sprinkled with potsherds, suggesting that it is composed of occupational debris. But the prospect of excavation is rendered extremely difficult by the fact that the slopes have been terraced to accommodate ploughing ; the area has been exposed to the attention of treasure seekers ; and the clearance of rubble from such a collapse of walling presents what may be an overwhelming obstacle. The question now arises as to who might be expected to have been the aristocratic residents of the palace stronghold. The region of Qal'eh-i Yazdigird is by no means subject to severe winters and the summers are sufficiently warm to warrant the migration of sheep, while herds of goat remain on the lower ground. The trees represented in the orchards which cover the site include apricot, pear, pomegranate, fig, walnut, apple, almond, and grape vines. In Yaquit's day the figs of Rijib were "not to be equalled anywhere in the world".28 The most perfect season is autumn, which in conjunction with the fruit harvest is particularly pleasant. This is in contrast with Qasr-i Shirin which is more favourable in winter, and Hamadan, which provides an ideal summer residence. Qal'eh-i Yazdigird might have provided an ideal autumnal residence for the king. There is in addition the reference of Yaqfit to the gardens of Bahrim (probably Gach Gunbad). But " it was natural for later Arab writers to pick up local attributions of the type, and we may increasingly have to
27

d'Orient Kurde,Correspondence According to J. Blau, Dictionnaire no. 9, Bruxelles, 1965, ase bayi can be translated " wind mill ", and dar means " tree " (in this case, " Orchard Place ").

Naqqjreh Khdneh can be translated " Drumstand ": see S. Haim, New Persian-EnglishDictionary,Tehran, 1960. 28Rawlinson, " March from Zohdb ", J.R.G.S., IX, p. 34.

QAL'EH-I

YAZDIGIRD

107

deal in the field with the constructions of great noblemen.'"29 There was, for instance, at the time of Bahrim, an intensive scheme of development undertaken in Fars by his First Minister, Mihr NarsE, who according to Tabari3o founded (in the region of Firfizabdd) villages, plantations, and fire-temples.3' The fact that such exhaustive efforts have been made to secure the area at Qal'eh-i Yazdigird rather suggests that it was not the supreme monarch who was involved. A permanent residence would possibly have embarrassed the resources of the district, and it would be quite unnatural for a city not to have grown up around a royal residence. Nor does its explanation as a mountain retreat for hunting expeditions seem convincing. There would be no need for extreme fortification in this case, for any king who felt so insecure would hardly venture even beyond his own city walls. Furthermore, it does not seem to have been the habit of Sasanian monarchs to indulge in safaris. If they can be compared with the Safavid kings, it can rather be expected that game was driven into a walled enclosure for the pleasure of the nobles to hunt without danger to themselves.32 There would be no need for the king to travel to such remote parts to enjoy this kind of sport. The mountain stronghold does seem more appropriate for a great noble. It may be that he actually had designs upon the throne himself, in the same way that Ardashir must have secured his position at FirfizTbtd in preparation for the overthrow of the Parthian monarchy.33 The numerous occasions when the royal authority was subsequently challenged would perhaps lend support to this theory. Conversely, the noble may have been established there by the king in order to prevent such disorders from arising. Situated on the fringe of a wild mountain tract, the stronghold acts as a police post, a fortress to deter local attempts to secure autonomy and resist the collection of taxes. It would also reduce the possibilities of the caravan traffic being disrupted by ambush on the great highway at its hazardous point of entry on to the Iranian plateau. Tepe Rash (v. P1. IIa) To the west of the palace, a number of low hillocks can be traced roughly in the shape of a horseshoe. The more easterly of these have been terraced and ploughed in the same way as the Ja-i Dar mound, while on the surface there appears the same scatter of sherds-mainly of a coarse quality and reddish or reddish-orange in colour. It seemed likely that these slopes represented the occupational debris of the quarters of the artisans and servants dependent upon the palace. But trial excavation showed that bedrock was no more than one metre below the ploughed surface, which suggests that the occupation was relatively limited. Such material as might have offered dating evidence has been scattered over the slopes and ploughed away. It had been hoped that the work at Qal'eh-i Yazdigird with stratified evidence would contribute towards remedying the lack of knowledge about Sasanian pottery in western Iran. The excavation was, therefore, disappointing in this respect. However, it may be profitable to discuss briefly the nature of the sherds recovered from the surface. It is misleading to think in terms of a standard ware throughout the empire, with its vastly different cultural traditions and variations in local material. There are, however, sufficient characteristics to link the local pottery of this period with that of other sites on the Iranian plateau. There is too a marked similarity between this collection of surface sherds and those gathered from the remains of the town that lies immediately to the east of the arch at Tiq-i Busthn,34 and which might also be dated as Sasanian. The use of a burnished slip is not nearly so apparent here as at Qasr-i Abii Nasr in Fars," but the fine grit-tempered red clay with reddish-orange slip is reminiscent of the unglazed wares from Sarvistan.36
29 I am grateful to Mrs. Deborah Thompson for this suggestion.
30

32I am indebted to Mr. Ronald Ferrier for this comparison.

33 E. Herzfeld, An Archaeological History of Iran, London, 1935, 34E. F. Schmidt, Flights OverAncient Citiesof Iran, Chicago, 1940,

der Muhammad ibn Jarir al-Tabari, Geschichte Perserund Araber aus des Chronicle Tabari, ed. zur Zeit der Sassaniden der arabischen Th. N61deke, Leyden, 1879: and Chroniquede Tabarf, ed. M. Hermann Zotenberg, Paris, 1867. These constructions have been shown by Vanden Berghe to be situated in the plain of Farrishband. L. Vanden Berghe, " R6centes D6couvertes de Monuments Sassanides dans le Fars", IranicaAntiqua,vol. I, 1961, p. 187.

P. 95.

3'

p. 8o; pl. 95. W. Hauser, " The Persian Expedition ", Bulletin of the Metropolitan Museum of Art, November, 1933, section ii: and " The Persian Expedition 1933-34 ", B.M.M.A., December, 1934, section ii (pottery by J. M. Upton). 36 Survey of the author 1965.
35

108

JOURNAL

OF

PERSIAN

STUDIES

There is the same tendency towards incised wavy lines,37 both combed and single grooving, and an extensive use of raised ribs or " pie-crust " moulding. Punched holes in continuous rows are also a feature. The rims of heavy storage jars are characterized by thick nail-head or bulbous profile, often with a sharply defined shoulder just below the rim. The excavations at Qasr-i Abfi Nasr, Istakhr, and Damghan in the 1930's produced evidence to show that glazed wares were not much in existence at these sites before the Abbasid period. 8 It was concluded that Iran had no share in the development of glazed wares, and that the mountain ranges between the Persian highlands and the plains of Iraq were a dividing line between two ceramic cultures. It is particularly interesting to note that at Qal'eh-i Yazdigird there is also a marked lack of glazed wares. It is unfortunate that little surface pottery can be recovered for comparative purposes at Qasr-i Shirin which topographically belongs more to the plains, and which might be expected to produce a greater proportion of glazed wares. A further point of interest is that there were amongst the surface sherds at Qal'eh-i Yazdigird two examples of a fine, hard ware-grey core, fired red on the outside and burnished-comparable perhaps to the " pieces of unusually fine pottery, almost as hard and smooth as polished stone " from Qasr-i Ab-i Nasr, and dated as Parthian or Seleucid.39 This type of pottery has recently been found at a number of sites in the Kirmanshah province, and, termed " clinky ware ", has been tentatively dated as Parthian.4? The fineness of the clay so distinguishes it from the coarse red wares that it must be assumed that it was a luxury product and possibly the substitute or equivalent of a glazed ware during this period on the plateau.

Gach Gunbad
Separated from the Jd-i Dir palace by the mounds of Tepe Rash, and assuming a central position surrounded by a ring of hills, the traces of a walled enclosure can be followed for a distance of 535 m. stretching down in the direction of the Darwizeh. The eastern reaches are no longer evident on the surface, but the whole area can be seen to have formed a rectangular layout measuring 265 m. along the southern end, and possibly slightly less at the northern limit where the hills encroach upon its regular, rectangular shape. At the lower end, two piers of masonry standing at right angles to the curtain wall, and 8 m. apart, mark the entrance to the enclosure at a point 210 m. from the south-west corner. Near this corner a stretch of wall stands to a height of 3 m. and reveals that there was a regular series of slender buttresses on the inner face, comparable with those found in the al-Dhabai compound immediately adjacent to the Taq-i Kisrd at Ctesiphon, which the excavators interpreted as an arena4' (v. Fig. 3). The large enclosure at Tdq-i Bustan has been associated with a hunting park.4 It seems that at Qal'eh-i Yazdigird the concept is the same-an irrigated garden or paradise. In the northern section of Gach Gunbad, there survives a huge block of shapeless masonry which Rawlinson had assumed was the base for a pavilion or temporary superstructure43 (v. P1. HIa). Its function is not at all clear, except that any building imposed above would tend to resemble a tower. A deeply cut stream-bed has broken through the ruins at this point, but it is still possible to trace the lines of small chambers, constructed of the same rubble masonry as the perimeter walling. Immediately adjacent to the huge block (Gach-i Buzurg) on the eastern side, and covering an area 8o m. square, the ground is coloured red by powdered fragments of baked brick, amongst which are pieces of stucco plaster. This stretch of land has been cleared to facilitate cultivation, with the result that there is a slightly sunken area, bounded by heaps of field stone. It is here too that the villagers recover fragments
37There is no reason to suppose that these features are not Sasanian, though in a different context, in the northwest of Iran, they could easily be confused with second millennium B.C. material, cf. R. H. Dyson, Jr., " Problems of Protohistoric Iran as seen from Hasanlu ", J.N.E.S., XXIV, p. 193.38Qasr-i Abfi Nasr, see n. 35: Istakhr. E. F. Schmidt, The Treasuryof Persepolisand Other Discoveriesin the Homelandof the Achaemenians,O.I.C. 21, Chicago, 1939, p. 101: and Damghan, see also Treasuryof Persepolis,p. o101.
39 Hauser, " Persian Expedition ", B.M.M.A., I934, section ii.
40

I am indebted to Mr. David Stronach and Dr. T. Cuyler YoungJr. for this term and the information. 4' Reuther, Die Ausgrabungen, 1928/9; and "The German Excavations ", Antiquity,vol. III, p. 444. 42 Reuther, " Sasanian Archaeology ", Survey,vol. I, p. 569. 43 Rawlinson, " March from Zohab ", J.R.G.S., IX, p. 33.

QAL'EH-I

YAZDIGIRD

109

of gypsum plaster, which they resmelt to provide the facing for the dome of the BTbh Yddgar shrine.44 Two trenches were opened up where the red colouring was strongest on the surface.45 The edges of baked brick appeared only 20 cm. below the topsoil, scarcely damaged by the wooden plough (v. P1. IVd). The impression was as if a wall had gently toppled over intact. The prospect of removing heavy brick tumble did not seem justifiable in view of the limited time available, and the work was halted. An adjacent area produced the same results, whereupon close examination revealed that the bricks were actually in their original position ; these same bricks were found to be standing to a height of 4 m. above the level of the floor (v. Fig 4).The bricks are square, their measurements varying from 30 to 31 cm. along the sides, and from 6 - 5 to 8 cm. in thickness. From the great variation in one building it is evident how misleading a comparison of brick sizes can be. A number of recognized Sasanian sites also show marked inconsistencies in brick dimensions. At Aivan-i Karkheh46 they measure 32 X 37 x 8 cm.; at Ctesiphon,4' 39 X 39 X 12 cm. ; at Sarvistdn48 27 X 27 x 8 cm. ; at Damghdn49 37 x 37 x 8 cm. ; at Gumush Tepe5? in Khurasan 35"5 X 35"5 x 8 cm. ; and 35-2 cm. square at Dastigird.5' The unconditional acceptance of size as the criterion for dating can be seen to be misleading in the mud brick walls of the Qal'eh-i Dukhtar in Kirman.5 The upper levels with bricks measuring 20o X 20o X 5 cm. were clearly Islamic. But the original foundations, in which three different sizes had been ased in the same single building process, reveal two courses of baked brick 30 X 30 x 6 cm. laid over a rubble pack on the cliff face, followed by fourteen courses of mud brick 52 X 52 x 15 cm., and topped by standard courses 38 x 38 X 8 cm. The excavators at Barghuthiat assumed that the 36 X 36 x 9 cm. size could be attributed to the Sasanians, while the 40 X 40 X 12 cm. and 44 X 44 X 12 cm. sizes were Parthian (but only from the observation that they were larger)." In view of these discrepancies it would be fair to state that the brick sizes at Qal'eh-i Yazdigird tend to corroborate a Sasanian date but are not conclusive proof on their own. The individual bricks have been marked before firing by a crude finger scratch. Three types have been recovered-a plain cross, a small crescent, and the impression of three finger points. In construction, they are laid in vertical courses turned alternatively through 90 degrees, and reinforced at intervals with horizontal lays, and bonded with gypsum mortar.54 Vertical lay construction occurs at Assur ; at Tell 'Umar (A.D. first century) ; in the foundations at Tdq-i Kisra ; in the Sasanian palace at Ddmghdn ; and in the eighth-century Tnrik Khaneh there as well. It cannot, therefore, be specifically pin-pointed to either the Parthian or Sasanian period. Reuther suggests that it may relate to an unknown form of building material, but his assertion that this technique was associated with the plano-convex bricks of Mesopotamia is unsatisfactory, and the reason for its employment must remain obscure." A length of walling measuring 8 -35 m. was exposed, which made a right angle turn, and seemed to continue beyond the three metres actually exposed. A shallow pilaster marks the corner, extending m. 42 cm. along each face, and projecting 8 cm. out from the wall (P1. IVa). From the corner, 3"95 along the main length, another pilaster I -8o m. wide also projects outwards for 8 cm. (P1. IVc). Both pilasters run to the full height of the wall, the entire surface of which carries a double layer of gypsum plaster. Each layer is 4 cm. thick, plain for the most part, except for the upper 6o cm. of the wall's
44Hence the name GachGunbad(gypsum dome) does not reflect the survival of an ancient tradition of a magnificent domed building, but indicates the abundance of stucco fragments to be found in the topsoil. It is a corroboration too of the ease with which stucco can be handled. Simply by heating in a kiln, gypsum is reduced to a powder, and by the addition of water is rendered in a plaster form, cf. N. C. Debevoise, " The Origins of Decorative Stucco ", A.J.A. 45, 1941, p. 48. 45The trial excavation and survey of the site, which lasted for three weeks in June 1965, was conducted by myself with the help of Mr. Rahnamoun from the Department of Antiquities, as the Representative of the Ministry of Culture. I would also like to acknowledge the support of the British Institute of Persian Studies, and in partcular the continual encouragement and guidance of the Dirtxtor, Mr. David Stronach. 8A
46 Survey of the author, 1965. 47 S. Langdon, " The Excavations at Kish and Barghuthiat in 1933 ", Iraq, vol. I, 1934, p. II17. 48Dieulafoy, L'Art Antique,vol. IV, p. 2. 49 F. Kimball, "A Sasanian Building ", Excavations at Tepe

MuseumBulletin, 1932. Hissar, Pennsylvania C. and Sistan, p. 272-73. 50o E. Yate, Khurasan 5' Creswell, Early Muslim Architecture, vol. II, p. 22: and Sarre and Herzfeld, Archdologische Reise, II, p. 90. Survey of the author, 1963. 53Langdon, " Excavations at Kish ", Iraq, vol I, p. I17, cf. note47. 54 cf. Reuther, " Parthian Architecture ", Survey,vol. I, p. 423, fig. 99. Idem, p. 422-23. s55
52

110

JOURNAL

OF

PERSIAN

STUDIES

Qaleh

Yaoz

dgird

WALL PLAN of EXCAVATED

GUMBAD GACH
verticallays

---

cap plaster
0
12m

___

._

.._

._ ,,_

-stubble S---- top soil " -eroded fragments

SECTION & DETAIL

mixed broken Adebris

stucco panetlling & pilaster

falen
masonry

baked

brick in vertical
lays

plaster
facing

.=---

plaster
fine

crumble
deposit
scae
___.,,,
,__

floor

,
jtne

I9s"

trench. Fig. 4. Gach Gunbad:plan and sectionof excavated

QAL'EH-I

YAZDIGIRD

111

total height of 4 -20 m. Two flat horizontal ribs or fillets mark the limits of a frieze which contains a series of panels in three bands, displaying a repetitive design of interlocking key patterning and intermediate medallions (Fig. 5 and P1. IVb). The upper rib or cornice fillet, marking the top of the wall, indicates that it is a single storey structure, perhaps the curtain wall of a courtyard, as found in the temple precincts at Assur.56 The cement cover to the top of the wall still survives in places. A large enough area of this cement surface has been cleared to show that it extends back too far to be considered a step-back on the impost before the springing of the archivolt. There is no evidence for such a vault, but the section reveals that a considerable number of stucco fragments, both of panelwork and sculptured figures, have fallen down together with a heavy collapse of masonry from a height greater than the four metres of the wall itself. It would be wrong to think in terms of a flat roof. " All known Sasanian buildings are vaulted, whether they are built in the rubble technique indigenous to Fars, or in brick ".11 Perhaps it would be more feasible to suggest that rather than being capped by a barrel vault, the wall was surmounted by a colonnaded balustrade or parapet.58 The close proximity of many of the fragments recovered, as well as the fact that there are duplicates in some cases, points to their emplacement in horizontal bands. None of the pieces excavated were curved in section, in a way that would have seemed natural had they been connected with a barrel vault. In view of the fact that the excavated trench lies to one side of the ruined complex, it would be possible to accommodate the theory of a single-storey wall flanking the main aivdn, which was a strong feature of Oriental architecture from Parthian times. On the other hand, the aivdn had tended to become subordinate to the domed structure, particularly in religious architecture.59 But the aivdn still survived, and perhaps enjoyed a revival towards the end of the Sasanian period. There was also a tendency to combine long linking corridors with a central domed structure. The section shows that the building was exposed in an abandoned state for some length of time before its total collapse. There is no sign of burning, but the floor is covered with a thick layer of greenish clay deposit-such as one might associate with a building exposed to rain and the accumulation of wind-blown particles on its deserted floors. Following this there was a sudden collapse, and one large block of masonry still has stucco decoration intact upon its face. It may have come from an adjacent double-storey structure, for the wall has been preserved to its standing height by the accumulation of this debris, which is too extensive to be explained away simply as the collapse of a parapet. Pending further excavation it can only be stated categorically that there has been a fall of stucco fragments, together with brick masonry, which has originated from a height greater than that of the wall itself. The StuccoDecoration The history of stucco has already been thoroughly investigated by Debevoise, though not all his interpretations are universally accepted.60 He suggests that the introduction of stucco, particularly with elaborate and painted designs, coincides with the appearance of the aivdn, which reflected a general reaction against Hellenism in the first century A.D. Brick glazing was no longer practised, and the stucco provided a convenient base, often with an additional coating of fine plaster wash, for large-scale painting. It is sufficient to stress here that, as a malleable form which covered the building like a garment, the stucco lent itself readily to the application of mould impressions of continuous and interlocking floral and geometric designs, and the abandonment of the architectural details of Hellenistic tradition. Goldman classifies the different stucco patterns from Seleucia-on-Tigris as follows :61 (a) stucco pattern composed of a single element either carved or moulded into the stucco, e.g. the circular medallion with turning wheel design.
56 W.

Andrae and H. Lenzen, Die Partherstadt Assur, W.V.D.O.G. 57, p. 86, abb 42. 57Reuther, " Sasanian Architecture ", Survey, vol. I, p. 499. 58cf. the al-Dhabai structure at Ctesiphon. Reuther, Die Ausgrabungen,1928/9, and " The German Excavations ", Antiquity,III, p. 444.

59 cf. Debevoise, " Origins of Stucco ", A.J.A., 45, p. 60. 60Debevoise, "Origins of Stucco ", A.J.A., 45, see also J. Baltrusaitis, " Sasanian Stucco, Ornamental ", Survey, vol. I. 61B. Goldman, "The Allover Pattern in Mesopotamian Stuccowork ", Berytus,vol. X, 1952-53.

112

JOURNAL OF PERSIAN STUDIES

(b) stucco pattern composed of a series of repeated elements arranged in a band or ribbon, e.g. the gamma cross meander. (c) stucco pattern composed of interlocking and intersecting repeated elements over an area, e.g. the rosettes composed of overlapping arcs of circles. (d) miscellaneous patterns, e.g. the multiple saw-toothed flutings on attached columns. The patterns have in common a "light and dark" effect gained by the sharp differentiation of the face of the design from the background, and the result of deep narrow cutting.

Fig. 5. Stucco panelling.

At Qal'eh-i Yazdigird the decoration includes type (b)-the gamma cross meander or running key pattern, which had already been liberally employed during the Parthian period, for example in the stucco cornice of the west aivdn of the palace at Assur.6' Here it echoes closely the Classical fret design which was a popular motif on the cornice corona of Greek monuments. The Sasanian artists had added their unmistakable touch by transferring the key pattern from a purely linear concept to forming part of an overall design covering the entire wall face. But the swastika cross cannot be ascribed to any origin in particular, since it is an obvious variation of a plain cross, and in fact was used freely from the earliest times in Mesopotamia and Anatolia.63 It has been noted that the key pattern can be interpreted in two basically different ways-either as a repetition of successive swastika crosses, or with the emphasis placed on the long linking axes.64 In the case of Qal'eh-i Yazdigird the latter explanation is
62W. Andrae and H. Lenzen, Die Partherstadt Assur, 1930; and E. D. Van Buren, "Symbols of the Gods in Mesopotamian Art ", AnalectaOrientalia,no. 23, 1945. 64 Baltrusaitis, " Sasanian Stucco ", Survey, vol. I, p. 602-6 fig. I8o.

W.V.D.O.G. 57, Taf. 14.

63 cf. F. Petrie, Decorative Patternsof the Ancient World, London,

QAL'EH-I

YAZDIGIRD

113

more satisfactory, since each line of the design is constituted by a double raised bar divided by a narrow groove. The double bar and groove effect is a fairly sophisticated rendering of this form of decoration. The visual effect here is that the arms interlock as a gamma cross, while the swastika impression is obscured. It is perhaps closer to the Classical version than many Oriental examples, being restricted to a horizontal frieze, with three bands of decoration. At Ctesiphon the design covers the entire surface, and is set diagonally instead of on a regular horizontal plane.65 In every case a series of rosettes alternates in each row with the key design. At Ctesiphon these are positioned so that they are free-standing, whereas the Qal'eh-i Yazdigird examples form the centrepiece of a medallion, comparable with those from Kish. There is a variety of six different types of rosette: a quatrefoil with heart-shaped leaf; a combined plain quatrefoil and heart-shaped leaf; a lozenge-shaped sexfoil with independent triangles; a lozenge-shaped quatrefoil with independent bows; a turning whorl; and a sexfoil of triangles with slightly concave sides (which is really the reverse negative of circles with overlapping arcs).66 In addition to these purely repetitive features of decoration which remain in situ, there are a number of sculptured figures, fragmented pieces of friezework, and relics of a balustrade, which were recovered in excavation from the main sondage. These will be discussed individually.67
Figure61 Hermaphrodite From a study of the three naked figures, which seem to have been identical in their pristine condition, the problem arises of deciding whether the artist had intended to portray a hermaphrodite or simply an effeminate youth. Certainly there is a clear indication of the male genitalia. The difficulty is not one of acknowledging the male qualities, but of deciding whether the figure is sufficiently feminine to deserve the term "hermaphrodite".69 It is not very much more feminine than the youthful picture of Hermes in a terracotta statuette excavated at Dura Europos,7o and does not really match the sophisticated image of a mature woman in the figurines from Seleucia.71 But it must be remembered that Hermaphroditus was still a youth when he bathed in a fountain of Caria, and by his union with Salmacis became a dual entity. The Iranian artists did in fact find it difficult or undesirable at this time to express the subtleties of the female form. It is noteworthy that the nude did not appear in Sasanian art until about the sixth century. Ghirshman notes that its appearance was certainly due to foreign influence; and even though the makers of the mosaics at Bishapur had already shown an appreciation of the beauty of the human form, the nude found its full flowering only in silverware.72 It is true that they occasionally portray "languorous attitudes, the strong yet supple bodies of young women with swaying movements and frankly sensuous appeal".73 The most voluptuous rendering is perhaps that of a nude dancing-girl depicted on a boat-shaped silver bowl of the sixth or seventh century.74 But even here the artist fails to realize the true nature of a supple breast, and resorts to a more typical button shape, which appears almost without exception on both silverware and rock reliefs.75The bull-necked version of Anfhita, the
65 Survey,vol. IV, pl. 172 E; and E. Ktihnel, Die Ausgrabungen

der zweiten Ktesiphon-Expedition, Berlin, 1933, 25. 66 The rosette has been used as a decorative fig. motif from the earliest times, and the heart-shaped leaf and swastika have been frequently employed in a variety of contexts. Goldman analyses the geometric all-over pattern, in particular the interlocking circle (with special reference to Seleucia-on-Tigris). He suggests that it appears to have been based on a fabric antecedent, perhaps of Assyrian origin, which was translated by the Parthians into stucco grill-work, and that the mode of carving points to a perforated, wooden screen ancestry. Goldman, " The Allover Pattern ", Berytus,vol. X, 1952-53. 67 It is unfortunate that thorough cleaning of the objects was not possible during the 1965 trial excavation. Proper treatment promises to aid the interpretation significantly. 68 The statuette measures 46 cm. high x 18 cm. 69 In the case of some figurines from Seleucia-on-Tigris, there is

simply an exaggeration of the female pudenda, rather than there being a fully hermaphroditic quality. W. Van Ingen, Figurinesfrom Seleucia-onTigris, 1927-32, Ann Arbor, 1939, p. 43. 70 C. Hopkins, " The Season 1934-35 at Dura ", A.J.A. 39, no. 3, fig.-57' Van Ingen, Figurines from Seleucia-onTigris, P1. LVI, no. 868. 72R. Ghirshman, Iran: Parthiansand Sassanians,1962, p. 214. It is also significant that these figures at Bishapur were executed under very strong Classical influence, and possibly even by western craftsmen. bowl:Walters Art Gallery, Baltimore. Ghirshman, Iran, fig. 258. 75 e.g. Kalar Dasht jug: Iran Bastan Museum, Teheran. Ghirshman, Iran, fig. 256.
73 Idem. 74 Boat-shaped

114

JOURNAL

OF

PERSIAN

STUDIES

goddess of fertility, at Naqsh-i Rustam76 is equalled in sturdiness by the robust queen of Bahrfm II at Sar Mashhad.7" The small wings that are clearly visible behind the shoulders of one of the statuettes suggest that it may portray a cupid, although in a very effeminate rendering. The face is moulded in a distinctly Romano-Classical style, rather moonlike, with the same full roundness that is shown in the twin Victories adorning the spandrils of the main arch at Taq-i Bustan.7'"Certainly the appearance of the cupid in Sasanian art is quite frequent, being included on silver bowls, and shown in particular accompanying royal personages on rock reliefs, in the form of a victory. They are obviously western in inspiration, and usually, as at Qal'eh-i Yazdigird, the Classical qualities survive. The artist here has positioned the head and hands in such a way that they express an attitude of coyness-the head is tilted to one side, with the cheek resting on the right hand as it comes up to cover the left shoulder. A close parallel in the facial features is the head of a lunar goddess on a marble relief from Hatra. Ingholt compares the subtle smile of this figure with the charm of the ancient Greek "korai".9 But in view of the traditional reluctance to acknowledge the real essence of feminine proportions, the meagre rendering is no reason on its own to doubt that it is a female figure. If the conclusions of Lensen should be taken as a standard guide,so the distinction between the female breast and the overdeveloped is that the female breast is pendulous in male chest depends upon the shape of the cleavage-that and assumes a round and separated position frontally, whereas in the male they meet in front to profile form a letter Y cleavage. On this score again the figure is probably feminine. From the combination of these female characteristics with their undoubtable male qualities, it is reasonable to conclude that the Qal'eh-i Yazdigird figures, in which the western artistic influence predominates, also reflects a western tradition: the hermaphrodite.

Reclining Figure
The same Classical influence can be recognized in the several pieces that form part of a continuous and repetitive scene in a horizontal frieze,8' which shows a reclining male form and a boyish figure, who with wings extended behind faces away from the divan and takes up a stance to grasp the tail of an animal. The slight bulge at the end of the tail seems to imply that the low-slung, smooth hindquarters belong to a lion, only the rump of which survives. It would be difficult to ascribe the reclining form to any one prototype, for the attitude is merely a repetition of the normal western reclining position. The Oriental is usually shown to be seated crosslegged, so much so that the king, except when viewed frontally mounted on a throne, is shown as maintaining a decorous position on a divan by crossing one leg over the other. There is however very little difference between the western attitude of the reclining figures at Palmyra-in the triclinium of Maqqai and the tomb of Varkai82-and that of a Sasanianking supported bycushionsat a royal banquet depicted on a silver bowl of the sixth century." There is a similarly oriental attitude at Tang-i Sarvak.84 The Seleucid statue of Hercules at Bisitfin portrays the god in a horizontal position, repeating a Hellenistic tradition which often shows him lying on a lionskin cover. It would be a temptation to assume that the lion and the reclining form indicated that this was in fact a portrayal of Hercules.

78Vanden Berghe, Archdologie, I28 C: and Herzfeld, Am Tor pl. vonAsien, pl. XXXVI. 79H. Ingholt, " Parthian Sculpture from Hatra ", Orient and Hellas in Art and Religion, Memoirsof the Connecticut Academy of Art and Science,no. 12, New Haven, 1954, pl. III. " 80V. F. Lensen, The Triumph of Dionysus on Textiles of Late Egypt ", Universityof CaliforniaPublicationsin Classical Archaeology, no. 5 (i), 1960. I am indebted to Mrs. Deborah Thompson for this reference. I would also like to acknowledge

76 L. Vanden Berghe, Archiologie l'Iran Ancien,Leiden, de 1959, pl. 30 C: and Ghirshman, Iran, fig. 2 8. 77 Ghirshman, Iran, fig. 217.

her valuable criticism of my suggestions regarding the dating and styling of the stucco. 8, The frieze is 20 cm. deep, of which a length of 55 cm. can be reconstructed. 82Ghirshman, Iran, fig. 83; 90. 83Silver bowl: Walters Art Gallery, Baltimore. Ghirshman, Iran, der fig- 259. cf. also Kiihnel, Ausgrabungen Zweiten KtesiphonExpedition, I933, fig. 36. 84 W. B. Henning, " The Monuments and Inscriptions of Tang-i Sarvak ", Asia Major, N.S., 1951-52, pl. II and III: Ghirshman, Iran, fig. 67; and Vanden Berghe, Archdologie, 88 A. pl.

QAL'EH-I

YAZDIGIRD

115

EntwinedBeasts (Fig. 6) Perhaps the most interesting piece of all is a square block, of which the narrow sides are plain, while the two faces display on the one side a simple rosette pattern, and on the other a scene of two entwined beasts.85 The initial impression is of a griffin-like creature-streamlined, with low-slung hind quarters, long lean bodies, and sharp pointed snouts; goatee beards and peaked ears; and ferret tails just brushing the ground. The creatures have assumed symmetrically an "X" position, linked to each other by a single turn about the midriffs, so that their heads and short fore-legs create an impression of straining away from each other in antagonism. The positioning of the half raised limbs is consistent with the Sasanian concept of the beast rampant.86 The form is recognizable in many animals executed in both stone and silverwork, particularly in versions of the simurgh8'and in scenes depicting horses captured pictorially in mid-flight."

Fig. 6. Stuccocapital.

The short wings, confirming a dragon-like appearance, are again typical of those borne by griffins and simurghsin Sasanian art.89 Being foreshortened with blunt base and volute scroll tips they are in distinct contrast with the streamlined wings of Classical griffins. There was already a tendency for the wing tips to splay outwards in Achaemenian art (continued from the Assyrian tradition),90 and this was still apparent in Parthian times, with a drake's tail curl evident in the winged sphinx from Nysa9' (second century B.C.). It became the standard shape for the wings of hybrid beasts in the Sasanian period, and the curl was emphasized even more until it approached the nature of a volute scroll.92 The wings of the horses that support the throne on the "Cup of Solomon"'93 are the closest example to the highly stylized form of the Qal'eh-i Yazdigird animals. The beasts are flanked on the one side by a stylized tree of life. Ghirshman refers to the abundance in Sasanian metal and silkwork of confronted and addorsed animals and notes that according to the Avesta their function was often to guard a spring marked by a sacred tree.94 Touching upon the popularity in Sasanian art of all manner of exotic and startling creatures, he offers the explanation
scene measures 24 cm. square. The block is 37 cm. high, tapering from 37 cm. along the top edge to 27 cm. along the base. 86e.g. the crossed lions on a sixth-century ewer in the Bibliothbque Nationale, Paris. Ghirshman, Iran, fig. 404. 87e.g. Ghirshman, Iran, fig. 271-76; and Survey, vol. IV, pl. 177 F. 88e.g. Ghirshman, Iran, fig. 247-52.
85 The

89I am indebted to Mr. Charles Wilkinson for this comparison. 90 e.g. the sphinx in the Tripylon and the bull in the Gateway at Persepolis. 9, Ghirshman, Iran, fig. 40 B. 92 e.g. R. Pfister, "Gobelins Sassanides du Musee de Lyon ", RevuedesArtsAsiatiques,Paris, 1930, pl. II. 93 Ghirshman, Iran, fig. 401; and Survey,vol. IV, pl. 203. 94 Ghirshman, Iran, p. 232.

116

JOURNAL

OF

PERSIAN

STUDIES

that "the fierce, demonic aspect of many of these animals has been interpreted as an expression of their supernatural powers. Probably they were also meant to satisfy the taste for tales of travel in fabulous lands, for romances, for all that was mysterious-even terrifying. The Iranian craftsmen did not hesitate to cater for this taste by depicting legendary creatures calculated to startle and intrigue."95 The purpose of the block is enigmatic. What is clear is that two faces were intended to be exposed to view, though possibly the side with the rosette was destined for a rather obscure position. One vertical edge of the block tapers inwards, with a slightly rounded corner, which may indicate that it was a pilaster capital. The pillars and discs from the al-Dhabai structure at Ctesiphon were conceived as belonging to a balustrade,96 and a stucco pilaster capital was recorded at Tel 'Umar.97 A comparable parallel may perhaps be found in the frame of a stucco grill recovered at Qasr-i Abfi Nasr.98 Except when deliberately attempting to echo Greek capitals, Sasanian versions lack well-defined mouldings or contours, relying on pictorial decoration for effect. If the concept of a balustrade surmounting a single storey wall should be acceptable, this block may have formed the capital of a pillar in the parapet.

Fig. 7. Stucco plaque.

Male Portrait (Fig. 7) Several fragments of identical circular plaques can be used to reconstruct a picture of the original work,99 which consisted of a male head in medium relief, encircled within a border of egg and dart moulding. The head bears the characteristics of a type that is depicted on many of the rock reliefs and statues executed during the Parthian period. The finest example of the trappings and gear that accompany these personages is displayed on the bronze statue recovered from the temple of Shimi in 1936 by Stein."? Apart from the details of the accoutrements, the style of this figure is under strong Hellenistic influence. Perhaps more typical are the heads depicted on the reliefs at Mdlamir,'o' Shimbdr,'0o and Tang-i Sarvak.'03 A close parallel can be seen in one of the marble heads excavated at Shami.'o4
95 Idem, p. 219. " 96Reuther, Die Ausgrabungen 1928/9 and The German Excavations ", Antiquity,III, p. 444; and see E. Porada, AncientIran, London, 1965, p. 211. 97L. Waterman, SecondPreliminary Reportupon the Excavationsat Tel 'Umar, Ann Arbor, 1933, pl. II. 98W. Hauser, "The Persian Expedition, I932-33 ", Bulletin Museum,New York,November, 1933, sect. ii, of the Metropolitan fig. 799The reconstructed disc measures 40 cm. in diameter.

Stein, Old Routes,p. 130 sqq., pl. 46-7: A. Godard, L'Art de l'Iran, Paris, 1962, fig. 159-61; and Ghirshman, Iran, fig. 99. " L. 1o0 Vanden Berghe, Le Relief Parthe de Hung-i Nauruzi ", IranicaAntiqua,vol. III, Leiden, 1963, pl. 53. " 1o0A. D. H. Bivar and S. Shaked, The Inscriptions at Shimbar ", B.S.O.A.S. XXVII, I964, pl. II. " 1o3Henning, Tang-i Sarvak ", Asia Major, N.S., I952, pl. II. Marble head: Iran Bastan Museum. Stein, Old Routes,pl. 49; 1o4 and Ghirshman, Iran, fig. 107 A.
100oo

Pl. Ia. The Upper Castle: view north-west,showing the enclosed plateau area beyond.

Pl. Ib. Divar-i Gach: view south-west from the cliffs towardsDarwazeh.

Pl. IIa. Gach Gunbad: view south-east, from west of the enclosure.Tepe Rash beyondto the left.

Pl. IIb. Ja-i Dar: view east, showingthe UpperCastle beyondthe gorge.

P1. IIIa. Divar-i Gach: loopholesseenfrom the innerface of the wall.

face. P1. IIIb. The UpperCastle: buttresstowersalong the north

Pl. IVa. GachGunbad:corner pilasterandpanelling.

Pl. IVb. GachGunbad:stucco panelling.

P1. IVc. GachGunbad:pilaster,panelling,andplaster cap.

Pl. IVd. GachGunbad:verticalbricklay and collapse.

QAL'EH-I

YAZDIGIRD

117

The same characteristics can be distinguished in all the examples: bobbed hair, drawn into thick bunches on either side of the neck by a broad head-band just above the temples. This is in distinct contrast with the closely cropped "Caesar" look of Mithridates I on the Hung-i Naurfizi relief at Mdlamir.I05 The features are variable, not least owing to very crude workmanship, but there is a general tendency towards gauntness, and the beard is usually closely shaped and slightly pointed, without the "ring and knot" extravagance favoured by Sasanian monarchs.'o6 Long moustaches complete a very striking facial treatment. There is also adequate evidence to suggest that the Qal'eh-i Yazdigird bust was decorated with a necklace, of a simple banded nature. This tends to distinguish it from the figures of a later period, for a beaded necklace is invariably found on Sasanian portraits.'07 Although the lack of detail may in some cases be the result of poor craftsmanship, from the evidence of coinage and the Shami statues it can be assumed that a banded form was more normal in Parthian times. It is true that the bobbed style of hair enjoys a long history, and can be distinguished in some of the Achaemenian figures at Persepolis. But in these cases there is an almost overworked stress on the curled effect. The nearest approach to the Parthian style is that worn by the Bactrians. Most of Achaemenian figures are shown wearing some form of head gear, which is replaced in the Parthian period by a head fillet. Under the Sasanians there is tendency towards a loosely curled style, and the side bunches are much less reserved, often being shown to be floating in the breeze. In short, this style enjoys a continuous tradition right through Persian history. In this example the pieces are nearest in inspiration to the Parthian version. As such there is some difficulty in dating them, since they are associated with other pieces in a Sasanian context. The only satisfactory explanation is that the artists were indulging in deliberately archaic practices or were simply very conservative in their taste.

Engaged Half-Column"'s(Fig. 8) Equally Parthian in inspiration-if the comparison be made with the stucco finds of Warka,"o9 Assur," and Kfih-i Khwaijeh",-is a semi-circular half-column, which is decorated with a limited number of motifs, divided by runs of reel and bead moulding. These are quite exceptional in that, by dividing up a diaper pattern, they have lost all allegiance to the continuous linear concept of their origin. Of the individual motifs, the stepped crenellation can be traced at least as far back as Assyrian architecture."2 The crenellated battlement first appears as a defensive feature in conjunction with loopholes below. The latter were sometimes replaced by decorative medallions. Later the battlement is repeated as a purely decorative cresting, with the loophole reduced to a slit between the merlons. It has already reached this stage by the time of its employment in Achaemenian architecture. At Assur, in the of the Parthian palace, it is produced as a decorative form in a stucco frieze, while at Kfih-i facade Khwaijeh and Warka it abandons the linear tradition and becomes an isolated motif, being employed in repetitive grill work. The ultimate in this trend is achieved at Qal'eh-i Yazdigird, where it forms a part of varied diaper pattern. The earlier tradition also survives elsewhere in this period, for example in the battlement of a pavilion illustrated on a silver plate of the sixth century."3 The other surviving designs decorating the half-column are a plain quatrefoil, a quatrefoil with heart-shaped leaf, and a gamma cross swastika. They echo the patterning of the wall frieze.
105os Vanden

Berghe, "Le Relief Parthe ", IranicaAntiqua,vol. III,

pl. 55. 1o6 e.g. Ghirshman, Iran, fig. 216 (Bahrdm II at Sar Meshed) ; fig. 157 (Ardashir I at Naqsh-i Rustam); fig. 196 (Shapfir I at Bishipfir). 107Ghirshman, Iran, fig. 164, 209, 212. cf. figs. 135-155 with 304-326 : and Godard, L'Art de l'Iran, pl. Ioo-Io4-. 1o8 length of 6o cm. was recovered; the diameter measures A 28 cm. "o9W. K. Loftus, Travels and Researches Chaldea and Susiana, in New York, 1857, p. 225.

Andrae and Lenzen, Die Partherstadt Assur, p. 17, i. m Herzfeld, Archaeological History,p. 66. " Reuther, " Sasanian Architecture ", Survey, vol. I, p. 418; Debevoise, "Origins of Stucco ", A.J.A., 45, P- 51; and W. Andrae, " Die Festungswerke von Assur ", W.V.D.O.G., 23, Leipzig, 1913. "3 Silver plate: Iran Bdstdn Museum, Tehran. Ghirshman, Iran, fig. 246: The arch at Tdq-i Bfistdn is similarly decorated with a battlement cresting. Vanden Berghe, Arch6ologie, pl. 128: and Survey,vol. IV, pl. 159 B.

1o

118

JOURNAL

OF

PERSIAN

STUDIES

column. stucco Fig. 8. Engaged

AnahitaStatuette"4
The much broken torso of a small female figure gives a clear indication of its original significance by the positioning of its arms "represented in the typical posture of the great goddess of fertility, thousands of times reproduced in late Babylonian terracottas and alabaster statuettes, i.e. supporting her breasts with her own hands.""'5 This stance enjoys a seemingly uninterrupted history, being particularly apparent in figurines, of which Zaehner illustrates a fine example."6 Van Ingen states that "the type in which the woman presses or supports the breasts, which was also used from the earliest times, continued in popularity during the later periods.""7 Although in this example the stance and style are reminiscent of a figurine, the statuette formed part of a relief panel, like all the stucco figures at the site, being moulded in high relief, but with the back attached to a flat surface in the manner of a rock-carving.

Classical (Fig. 9d) Moulding"8

A perfect example of the degeneration of Classical motif and moulding is to be seen in the several fragments of a continuous cornice decoration, which is strictly western in inspiration. The lower fillet bears a perpetual wave or rope scroll pattern; the ovolo carries an egg and tongue enrichment (though the tongue has become a vertical reel); and the ultimate cavetto moulding shows a simple and very stylized honeysuckle anthemion. It is conceivable that this cornice was a top member in a balustrade which included the frieze with the "reclining figure" already described.

Miscellaneous
There are several pieces of grill-work, or rather panelling-for they do not constitute open latticework, but simply recessed patterning. The merlon and loophole motif appears here too: as a separate
London, 1961, pl. 29: cf. also Loftus, Travels; and Vanden Berghe, Archiologie,pl. ioi A, B. " The 15 M. Rostovtzeff, Squatting Gods in Babylonia and at "7 Van Ingen, Figurines,p. I8. Dura ", Iraq, IV, p. 20. The height is 14 cm.; and the largest piece measures 33 cm. xx8 n6 R. C. Zaehner, The Dawn and Twilight of Zoroastrianism, long.
,4 Existing height: 23 cm. tall x x 3 cm. wide.

QAL'EH-I

YAZDIGIRD

119

stucco Fig. 9. Miscellaneous fragments.

120

JOURNAL

OF

PERSIAN

STUDIES

feature, removed in concept from its original function as a linear cresting; and in conjunction with other designs, including a square set diagonally within a square. The development of the merlon and a comparison with the material from Assur and Warka has already been discussedin the case of the halfcolumn, q.v. (Fig. 9c, e, f, g). A medallion,"9 roundel with convex face is decorated with a geometric flower composed of two or sets of twelve petals, so arranged that the tips of the inner ring meet the interstices of the outer layer. The resulting effect is somewhat peony-like (Fig. 9b). A fragment of wing20 with volute scroll tip is reminiscent of the pair of eagle wings frequently employed in Sasanian art to frame a subject.'2' It appears in this way in some stucco panel-work from Chah Tarkhan,"' and is a persistentembellishment on the crowns of the kings of the latter part of the empire.123 The curl of the feather tips is heavily exaggerated in the manner of the "entwined beasts" q.v. (Fig. 9a). Two broad bands of stucco,"4 plain apart from a series of recessed squares with a central cross pattern, suggest an architrave-whichis perhapsfeasible in view of the theory of a wall parapet (Fig 9). Conclusion The overall impression of the site is unmistakably Sasanian: in the use of rubble and mortar masonry, with a stilted semi-circulartower; and in the parabolic arch construction without centring. There are no known examples of monumental Parthian buildings constructedentirely of this material. This negative condition is not, however, a positive proof, for the technique must have been known in domestic architecture in Fars. But generally speaking, on the basis of archaeological evidence, the Parthians are not known to have favoured monumental fortifications,and the intense work undertaken at Qal'eh-i Yazdigird was rather atypical of their tradition. Brick construction cannot on the other hand be associated with either period in particular. But it may be noted that the earliest example is dated to the first century A.D. The paradise enclosure, with Gach Gunbad representing a pavilion within the gardens, is perhaps more Sasanian in concept, and has parallels at Thq-i Bustan, and especially the al-Dhabai enclosure at Ctesiphon. The historical record indicates that the site may be Mddharfistin, since there is no trace of another site that fits the description so well. Its association with Bahrim may be the result of romanticism on the part of the Arab geographers, particularly as those who attribute the site to such a founder were writing five hundred years after the fall of the Sasanian empire. The legendary association with Yazdigird III, which Rawlinson favoured, can be discounted too."5 For although it was in this region that the all but last battles were fought against the Arab armies, it is hardly likely that the Sasanian monarch contemplated the necessity for an isolated retreat before the time when flight was the only escape. The Upper Castle can in no way be regarded as an ultimate refuge. It is rather a barbican against attack from the higher ground. Qal'eh-i Yazdigird could not support a large defendinggarrison indefinitely, and in fact the purpose behind the system of fortificationsseems more appropriatefor the residenceof a great nobleman than the retreatof the monarch of an empire. Pope has pointed out that "with our present knowledge (i.e. 1938) stucco does not justify the division of the Sasanian period into sub-periods each representing an artistic phase, but implies rather a great variation in the historical course of different types of elements within an ornamental repertoire"."16The
relief figures show elements of Classical influence, but more in the original inspiration than the actual rendering, emphasizing the strong orientalizing of western traditions. An anti-Hellenistic phase had
"9 Diameter I7 cm. 12oHeight 33 cm.

24Breadth 20 cm. in view of the last scenes of an epoch which were enacted in this area, has been reinforced by the doctrines of the Ahl-i Haqq sect, which states that Babd Yddgir, whose tomb lies at the source of the Ab-i Ghuslin, was a reincarnation of Imam Husein. The latter became the son-in-law of Yazdigird by marrying his daughter, Shahrbanfi. See Minorsky, Notes sur la Sectedes Ahle Haqq. vol. I, p. 645. x26A. U. Pope," Sassanian Stucco; Figural ", Survey,

"s The retention of the name of Yazdigird, though quite natural

cf. Creswell, Early Muslim Architecture, vol. I, p. 198/9; and Debevoise, " Origins of Stucco ", A.J.A., 45, P- 53. 22 Panel-work: Philadelphia Museum of Art. Ghirshman, Iran, fig. 229. 123 Coin of Yazdigird III from Susa: Bibliotheque Nationale, e.g. Paris. Ghirshman, Iran, fig. 329: see also Porado, AncientIran, p. 212, fig. I I4.
121

QAL'EH-I

YAZDIGIRD

121

already set in during the first century, only to be reversed for a short time by the influence of the invasions of Trajan and Hadrian. If this site is of Sasanian date there is an exceptional western flavour about some of the pieces, though the mouldings show that they are sufficiently removed from their original tradition to make their appearance at this date feasible.'27 Certain aspects are remarkably Parthian in inspiration, though the very fact that the crenellated battlement was employed continuously as a decorative motif from the first millennium B.C. shows how conservative the native artists could be. If the stucco was produced by Sasanian craftsmen, there are some peculiarly archaic elements, including the merlon design and the bobbed hair style. But equally typical of their period are the Sasanian wings of the entwined beasts, the volute scrolled wing, and the running key pattern and medallion frieze-work. There can be no compromise over the question of dating. The stucco fragments are connected with a building that suffered decay following what seems to have been a single period of occupation; and that building forms a normal and logical part of the residential layout of the site, protected by a thorough and well-integrated system of defensive fortifications. From the aggregate of various indications, the weight of evidence falls most heavily on the Sasanian side. The use of rubble masonry, without any of the inspired structural techniques that were shown in Fars during the early Sasanian period, suggests that the site may be close in date to the monuments of Qasr-i Shirin. But the absence of involved arabesques and heavily intricate stucco decoration are a sign that it is earlier than Ctesiphon. In view of the literary attribution to Bahram Gfir, it is conceivable that Qal'eh-i Yazdigird, or Mddharfistdn, was constructed in the fifth century A.D., around the time of the reign of Bahram, possibly by one of his nobles.

"7

The dating of the stucco at Ctesiphon was complicated by the presence of Hellenistic elements as well as more schematic decoration showing all the characteristics of the late Sasanian epoch. M. S. Dimand, " The Second Expedition to Ctesiphon der Berlin 1933, 1931-32 ", Summary in English, Die Ausgrabungen (weiten Ktesiphon-Expedition, P- 34.

123

CERAMIQUES PEINTES DE TURENG TEPE

ByJ. Deshayes
La steppe turcomane n'avait rivdl6 jusqu'a prdsent, en fait de c6ramique peinte, que celle de Shah Tep6 III: elle est tres caractdristiquede toute la plaine de Gorgan, et lors d'une expedition effectude en janvier 1960 en compagnie de D. Stronach et deJ. Christophe,nous avons pu en ramasserdes fragments sur la plupart des sites que nous avons explordsentre Gorgan et Gombad-i-Qabous.' Elle est cependant loin d'etre unique, et les fouilles de Tureng Tep6 ont eu l'avantage de reveler l'existence de plusieurs autres categories de ceramique peinte, dont il importe d'6tablir les traits essentiels et la chronologie. La plus ancienne, celle de Tureng Tep6 I (P1. Ia et d), n'a nulle part 6tdtrouvdeen place, bien que nous ayons maintenant men6 jusqu'a son terme le plus important de nos sondages (A). Sans doute avons-nous etd arretes par la nappe d'eau souterraine ddja rencontr6e par Wulsin;2 mais la presence dans les couches les plus profondes que nous ayons atteintes (sols 25 et 26) d'6pais remblais d'argile destindsvraisemblablementa lutter contre les infiltrations montre que nous ne saurions avoir manqu6 que de peu le sol vierge. Or dans ces couches les plus anciennes de notre sondage la ceramique en question 6tait totalement absente, a l'exception d'un unique tesson au niveau 26. En fait elle fut trouvee exclusivement dans les briques crues dont taient faits les b timents des sols 13 ? 16, correspondant chronologiquement a la pdriode Hissar II. L'argile qui servit a fabriquer ces briques fut 6videmment puisde en un point du site au moins en partie deserte a cette 6poque, mais occup6 ant6rieurement. On ne peut naturellement preciser a l'heure actuelle oil se trouve situ6 cet etablissement primitif. L' dtude de cette c6ramique-dont nous n'avons malheureusementtrouv6 que d'infimesfragmentsconfirme son anciennet6. La technique en est en effet tres rudimentaire: l'argile est rose et pleine de paille servant de d6graissant; la pate est assez friable; la surface est recouverte d'un engobe creme probablement lissd; le d6cor est effectu6 a l'aide d'une peinture brune, assez claire, l6gerement rougeatre, qui s'efface au moindre frottement. La nature des motifs est difficile a apprecier 6tant donn6 les faibles dimensions des tessons conserves; on distingue ndanmoins d'6troites bandes horizontales riguli'rement espac6es, des lignes verticales, parfois coupees de quelques lignes horizontales moins serr6es, des lignes onduldes horizontales, des triangles (?) hachurds de lignes obliques, des losanges reserv6s entre des motifs pleins l'gerement curvilignes, enfin des lignes obliques ' peine onduldes coupdes d'une bande horizontale. En plusieurscas le decor orne les deux faces du tesson. Les formes sont encore plus indiscernables;cependant un tesson prdsente un profil carend extremement anguleux; de plus nous avons trouve plusieurs fragments de bols profonds ' paroi l6gerement ouverte et a 1 vre parfois 6vasde (Fig. ia-c). Ces quelques donnees sont 6videmment bien peu parlantes. Cependant un certain nombre de rapprochements avec d'autres ceramiques de l'Iran prdhistoriquesont possibles. La technique meme de ces vases est caract6ristique,car l'emploi d'un degraissantv6getal est surtout attestden Mesopotamie du nord et en Iran; on le trouve representdentre autres"a Djeitun, dans la grotte de Hotu,4 a Sialk I,5 a Dalma Tep6,6 PisdeliTepd,7aGeoyTpd, a Tep6 Guran,9enfin, beaucoupplusausud, hTall-i-DjariiB.'o

Dans cette cat6gorie les bols profonds avec ou sans lRvre 6vas6e se trouvent ~ peu pros sur tous les sites, mais la panse anguleuse de l'un de nos fragments rappelle surtout une jarre cardnde A ddcouverte
Cf. Bylin-Althin, "Keramische Funde von den, Tep6's der Tiurkmenensteppe ", Sdrtryckur Orientsiillskapets Arsbok, 1937, p. 26-38; Arne, Shah Tepe,p. 5-30. 2 Suppl. to the Bull. of the Amer. Inst. for Persian Art and Arch., vol. II. no. I bis (mars 1932), p. 7. 3 V. M. Masson, Antiquity,XXXV (1961), p. 2044 C. Coon, Proc. of theAmer.Philos. Soc., XCVI (1952), p. 242-43 (pas d'illustration).
5 R. Ghirshman, Fouillesde Sialk, I, p. I-I -2. 6 Expedition, (2), I962, p. 38 et V fig. p. 39. 7 R. Dyson, Antiquity,XXXIV (1960), p. 21, fig. I et 2. 8 B. Brown, Excavations in Azerbaijan1948, p. 20, fig. 4 et 5. 9 J. Meldgaard, Acta Archaeologica, XXXIV (1963), p. -I14- 116, fig. 15 et fig. 17-18. " Lux ", no. XII (19511o Vanden Berghe, Jaarbericht Ex Oriente de 52), p. 212-13, fig. 28; Archdologie l'Iran Ancien,pl. 49, a.

Fig. I.

A, vigital (i/4). (a) Sondage sol 14, bol digraissant

A, vigital (I/4). (b) Sondage sol 14, bol ddigraissant

A, vigetal (1/4). (c) Sondage sol 16, bol digraissant

A, (1/6). (d) Sondage sol 2o,fragmentdegrandejarre

B, (e) Sondage sol 4,fragmentdegrandejarre(1/6).

D2, sol 4,fragment de grande jarre (i/4). (f) Sondage

A, (g) Sondage sol. i6, fragmentdejarre (I/4).

A, (h) Sondage sol 9g, vaseglobulaire(1/2).

A, (j) Sondage sol 20, vaseglobulairea surfacelissie (1/2).

A, (i) Sondage sol 14,fragmentde coupe(1/2).

CERAMIQUES

PEINTES

DE

TURENG

TEPE

125

Djeitun." Quant aux motifs decoratifs, ils ne sauraient &trecompards a ceux de Tall-i-Djarii, de Tep6 Guran ou de Dalma Tep6, qui sont de nature fort diff6rente, ni h ceux de Pisdeli Tepe, qui sont beaucoup plus dvoluds. En revanche ils rappellent un peu les vases de Geoy Tepd M, qui sont souvent ornds de lignes onduldes horizontales; mais ce sont tres videmment les sites les plus proches du n6tre, Sialk et Djeitun, qui offrent les rapprochements les plus probables. A Sialk I, du moins A partir du niveau I, 3, le decor orne souvent l'int6rieur et l'exterieur du vase simultandment,"2 comme c'est parfois le cas a Tureng Tepd; mais il semble n'etre jamais confind a la seule face exterieure, comme ' nous en avons maints exemples Tureng Tepd; d'autre part le motif le plus frdquent a Sialk, les groupes de lignes horizontales ou verticales rdunies deux par deux par de petites hachures transver' sales,'3 n'est represent6 Tureng Tep6 que d'une maniere tres incertaine, par un tesson dont le d cor est presque effac6. Enfin l'aspect des tessons de Sialk est fort diff6rent: ils sont d'un gris sale, sans engobe, et la peinture en est gris fonced. " ' Au contraire Djeitun nous avons affaire comme Tureng Tdpe a une peinture de couleur brun rouge sur fond creme. A en juger par les quelques fragments publids jusqu'ici, la nature meme des motifs semble confirmer ce rapprochement: les lignes onduldes paralldles y jouent notamment un ' Sialk. grand r6le. II est vrai que celles-ci se retrouvent egalement se pourrait done que les trois sites, Sialk I, Tureng Tdpd I et Djeitun, representent du point de II vue de la c6ramique trois cultures assez voisines, bien que Masson considere Djeitun comme anterieur ' Sialk I ;14 a vrai dire nous manquons d'informations suffisantes sur les d6couvertes du Turkmenistan sovidtique pour juger du bien fonde d'une telle chronologie. Un hiatus assez prolong6 separe certainement la culture de Tureng T6pe I de celle que caracterise la c6ramique peinte de type Shah Tep6 III. Du moins n'avons-nous trouv6 aucun intermediaire entre ces deux poteries peintes, dont la premiere remonte sans doute au Ve milldnaire, la seconde datant probablement, comme nous le verrons, du milieu du IVe mill6naire. Est-ce a dire que ces interm6diaires n'existent pas? Il serait d'autant plus aventureux de le pretendre que 1'exploration archdologique de la steppe turcomane ne fait que commencer et qu'a Tureng Tep6 meme les couches les plus profondes n'ont jusqu'i prdsent 6t6 atteintes qu'au fond d'un sondage de dimensions forcdment restreintes. La c6ramique peinte de type Shah T6p6 III y a 6td trouvee en place depuis le sol 20 jusqu'au sol 26, le plus basque nous ayons pu toucher: ces sept sols constituent notre p6riode Tureng Tdp6 II A.'5 Mais de nombreux fragments de plus faibles dimensions ont it6 ramassds Apeu pros i tous les niveaux de nos sondages A et B, tous deux localishs sur le tcpd ouest (le" Mound C " de Wulsin), et jusque sur la surface de ce tipd. En revanche cette poterie est tr~s rare dans les autres secteurs (le tip6 principal exceptd): nous n'en avons relevi qu'une poign6e dans nos sondages C et D au sud et E au pfriode est beaucoup moins richement reprasent6e dans ces parties sud-ouest. Est-ce dire que cette du site, ou bien qu'elle y fut recouverte par une couche d'occupation beaucoup plus importante pendant la p6riode Tureng T pd III (correspondant en gros L Tip6 Hissar III), comme nos sondages semblent le laisser supposer ? Ou bien faut-il invoquer 1'absence des dnormes cavitgs qui ont boulevers6 en profondeur les couches du tipd occidental et ont fait remonter jusqu'h la surface des fragments des p6riodes les plus anciennes? Cette dernipre explication nous semble la plus probable, sans toutefois 6ter toute valeur aux pricidentes. L'argile de ces vases, assez bien raffinde et bien cuite, est de couleur rouge brique; la surface enest A revstue, l'extprieur et sur la partie visible de la face intdrieure, d'un badigeon g6ndralement rouge foncd, presque lie-de-vin, parfois plut6t ocre, 6tal en traindes horizontales; le d6cor peint est d'un noir plus ou moins franc; aucune trace de lissage n'a it6 observ e dans cette s6rie. On peut se demander comment cette c6ramique 6tait fabriquie, car de fines lignes horizontales L l'int6rieur pourraient faire a croire l'usage du tour de potier. Nous partageons ndanmoins l'opinion d'Arne, selon qui ces vases zi 6taient en rdalit6 faits la main'e6 ou la rigueur au tour lent; pourtant la c6ramique lissue qui fut
Masson, o.c., fig. 2 = Sov. Arkh., 196!2, no. 3, p. 16o, fig. 3. Cependant un vase de profil caren6, en argile a d6graissant Hassuna Ia 6galement, mais n'est pas v6g6tal, a 6t6 d6couvert tA d6cor6: cf. JNES, IV (1945), p. 276-77, pl. XIII, I, en haut. Z2Sialk, I, p. 13.
31

'3 '5

'4 Antiquity,XXXV (1961), p. 206-207.

Ibid., pl. XXXVIII.

Jusqu'd present aucun vase de cette cat6gorie n'a vert dans une tombe. 16 Shah Tepe,p. 164-65.

t6 d6cou-

9A

126

JOURNAL

OF

PERSIAN

STUDIES

trouvee c6te-h-c6te avec la c6ramique peinte dans les memes couches de Tureng Tep II A 6tait le plus souvent tournee. Les formes et motifs ne sont guere varies. 11 s'agit surtout de grandes jarres-dont aucune n'a pu ' la partie sup6rieure, h paroi lge reetre reconstitude-d'un type frequent 6galement Shah Tep ment convexe, est sensiblement cylindrique ou un peu conique;"' la panse dessine une ployure plus ou moins marquee a proximite du fond, qui est beaucoup plus etroit que l'embouchure. Celle-ci est pourvue d'un rebord tres ouvert, quelquefois large et assez mince (Fig. Id et e), parfois au contraire 6troit et 6pais (Fig. If et P1. Ib). En d'autres cas la partie sup6rieure est franchement tronconique, et le rebord tombant (Fig. Ig). Nous avons trouv6 aussi des fragments de jarres a peu pres globulaires, a embouchure &troite et rebord 6vase, plus ou moins developp6 (Fig. Ih et P1. Ic, en bas)." Une autre jarre globulaire, dont malheureusement la lvre est trop incomplete pour que le profil puisse en etre dessin6 avec certitude, est pourvue non d'un rebord, mais d'un bourrelet anguleux (P1. le). ' Enfin un seul fragment de coupe a 6t6 d6couvert, dont la paroi peine concave est fortement 6vasee, et la lkvre simplement amincie (Fig. I i). Ces deux derniers vases sont les seuls qui constituent une ' nouveaut6 veritable par rapport au repertoire des formes trouvees Shah Tepe. La m eme remarque s'applique aux motifs et a leur disposition a la surface du vase.'9 Celle-ci est g6neralement fondee sur un arrangement vertical de motifs s6pare's couvrant toute la hauteur. Ces motifs sont parfois tous identiques et regulierement disposes: il s'agit en ce cas de bandes entierement noires et plus ou moins paisses, ou, sur un tesson a paroi tres mince qui parait assez exceptionnel, de 2 lignes fines et serries (P1. HIa, en bas au milieu). Mais le plus souvent une certaine alternance est dans le choix des themes et leur 6cartement: des series de lignes verticales encloses entre des observee bandes l6ge'rement plus 6paisses peuvent alterner avec des zigzags (P1. IIb); des bandes entierement noires peuvent etre s6parees par des lignes serpentines aux ondulations larges (P1. IIc), ou bien tres rapprochees et irregulieres (P1. IId), ou encore par des lignes ondulees qui se croisent en dessinant une succession de motifs biconvexes (P1. IIIa, en bas 'a droite); de larges bandes remplies de fins zigzags verticaux alternent elles aussi avec des lignes ondul6es tres serrees, et en certains cas ces zigzags sont si rapproch6s qu'ils se touchent et dessinent un pseudo-quadrillage (P1. IIIb); ce motif alterne quelquefois non avec des lignes ondulees, mais avec d' troites bandes verticales (P1. Ib); on trouve aussi des bandes d6corees d'une arete de poisson alternant avec des groupes de zigzags serres (P1. IIa, en haut a droite); des s6ries de lignes largement ondulees, parfois groupies deux par deux entre des bandes verticales (P1. IIa, au milieu); des zigzags aux dents tres profondes et tres rapprochees, soit isolks (P1. IIa, en bas a gauche), soit butant contre des bandes noires verticales (P1. IIIa, en haut a droite); enfin de larges zones d6cordes, entre lignes verticales, de zigzags horizontaux superposes (P1. IIIc). Dans cette s6rie le motif le plus original est constitud par un rameau vertical qui alterne avec des bandes d6corees de croisillons obliques; ces themes verticaux sont en outre separes par des rangles de gros disques noirs tres espaces (P1. Ie), que l'on avait deja rencontres dans une situation analogue a Shah Tep6." A c6t6 de ces motifs organisis selon un systeme purement vertical, d'autres fragments comportent un decor couvrant apparemment toute la surface du vase. Les croisillons obliques delimitant soit des carris (Fig. Ig), soit des losanges, sont ddji bien connus i Shah T~pe," de meme que le quadrillage de lignes horizontales et verticales (P1. Ic, en bas et P1. IIId, a gauche)."3 En revanche un theme nouveau doit etre signal6, un damier aux cases rectangulaires 6troites et allongees. Le d6cor des jarres est confin6 t la surface extirieure du vase, h l'exception de quelques groupes de lignes verticales sur le rebord; mais le fragment de coupe Aparoi 6vas6e est d6core h l'intdrieur, et ne comporte h l'extirieur qu'une bande horizontale sur le bord (Fig. I i). McCown a tres sommairement 6tudi6 la chronologie relative de Shah Tepe III, pour dater cette couche de l'6poque Hissar IIA;'4 il n'en a pas moins not6 les affinites qui unissent cette c6ramique t
22 Ibid., pl. XLI, fig. 294. Ibid., p. 166, fig. 278. Comparative Stratigraphy, 54. Ce point de vue a 6t6 repris sans p. autre discussion du problkme par R. Dyson, in Ehrich, Chronp. ologiesin Old WorldArchaeology, 239-40.

'7

Ibid., p. 168, fig. 282, 283, 286, 287.

1s Ibid., p. 165, fig.-276. '9 Ibid., p. 167-69.

23 24

21 Ibid., p. 170o, fig. 295, a etc.

o Ibid., p. I65, fig. 276.

CERAMIQUES

PEINTES

DE

TURENG

TEPE

127

celle des periodes ant6rieures."5 La d6couverte a Shah Tep6 d'un fragment de type Hissar IC, orn6 de cornes de bouquetin enfermant une 6toile A branches termindes par des boules,26 constitue un argument tr s important en ce sens. Une analyse plus poussde de la ceramique peinte de Tureng Tep6 IIA aboutit a un rdsultat analogue: des rapprochements significatifs, fond6s seulement sur les formes et les motifs les plus typiques, nous obligent en effet a remonter la date de cette ceramique A une 6poque ant6rieure a celle d'Hissar IIA. C'est ainsi que les tres grandes jarres a large ouverture, dont le diametre maximum se trouve au bas de la panse, dont I'dpaule est bombee et la partie infdrieure l6gerement concave,27 rappellent tres 6videmment une jarre de Sialk III, 1.2' Quant aux motifs, les zigzags serres dessinant une sorte de pseudo-quadrillage sont caractdristiques de la ceramique de Hagi Mohammed29 et, dans un style plus 6volud, se retrouvent peut-&tre a Giyan VA;30 les bandes verticales orndes de zigzags horizontaux superposes dvoquent la c6ramique de Cheshmeh Ali;3 les damiers aux rectangles allonges se rencontrent frdquemment en M6sopotamie depuis les couches profondes d'Eridu2 jusqu' a l'dpoque Ur-Ubaid 11, mais 6galement a Sialk III, du niveau 4 au niveau 7 ; les lignes onduldes croisdes dessinant une succession de motifs biconvexes sont utilisdes en Mesopotamie dans la civilisation de Halaf r6cent35 comme dans celle d'Ur-Ubaid 136, mais aussi a Giyan VB ;37 les gros disques noirs sont courants dans le sud misopotamien pendant toute la periode d'Ubaid, 38 mais egalement pendant la meme pdriode dans le nord de la M6sopotamie,9 en Susiane4o et a Giyan VC,4 et semblent disparaitre ensuite du rdpertoire de la cdramique; enfin le theme du rameau, que l'on trouve dans le nord mesopotamien a Gawra XVIII,4 est surtout frequent en Iran a une epoque assez voisine,43 mais persiste jusqu'a Giyan VD.44 Aucune de ces analogies ne suffirait A elle seule a dater notre p riode Tureng T p6 IIA; mais l'ensemble de ces rapprochements constitue, nous semble-t-il, une preuve suffisamment parlante, d'autant ea plus qu'aucun de ces motifs, au surplus, n'est reprdsentd Hissar IIA. De le me^mefagon la composition en bandes verticales qui caracterise la c6ramique peinte de Tureng Tepd IIA et Shah Tepd III est A peu pres totalement &trangere a Hissar IIA, qui prdfdre une composition en bandes horizontales. Elle est au contraire beaucoup plus frdquente, comme l'a ddji indiqu6 McCown, a Hissar IC; 4 l'on y retrouve aussi une meme alternance de motifs verticaux diff6rents et, mieux encore, comme sur un fragment de Tureng T pd, des thames verticaux s6pards par une rangde verticale d'61kments isolds, par exemple des 6toiles a boules.46 La seule diff6rence importante entre les deux ceramiques, du point de vue de la composition du moins, reside dans le fait qu'a Tepd Hissar ces themes verticaux s'interrompent sur une bande horizontale, alors qu'a Tureng Tepd et Shah Tepe ils descendent jusqu'au pied du vase. Par ailleurs plusieurs de ces motifs eux-m mes sont identiques a Tureng Tep6 IIA et Hissar IC: groupes de lignes onduldes, groupes de zigzags, croisillons, aretes de poisson;47 on trouve meme a Hissar IC un rameau analogue a celui de Tureng Tepd, mais renvers6.48 La c6ramique de Tureng T p6 IIA presente donc, quant au decor, des ressemblances certaines avec celle d'Hissar IC et en est sans doute en grande partie contemporaine. Toutefois les deux produc38 Ziegler, motif 135, pl. 6, I, m (de H. Mohammed); UVB, O.c., p. 54. o.c., Shah Tepe,p. 171, fig. 299. IV, pl. 16, B, n et pl. i16, D, u, v, d'Uruk XVIII et XVII; 27 Ibid., pl. XC, fig. 280. Ur Excavations,IV, pl. 52, aU 32 et aU 34 (d'Ur-Ubaid II, 28 Sialk, I, pl. LXII, S 1693. o0 il s'agit 6videmment d'un h6ritage de l'6poque antdrieure). 29 Ziegler, Die Keramik von . . . motif 6o, 39 Tobler, TepeGawra,II, pl. LXIX, 21, de Gawra XIX. Haggi Mohammed, 40 MMAI, XXX pl. 9, h. (1947), p. 166, fig. 28, I et 2, de T6p6 Djowi. 30 G. Contenau et R. Ghirshman, Tipe Giyan,pl. 41, en haut au 41 Tipl Giyan,pl. 57, en bas, 3e a partir de la gauche. 42 milieu. Tepe Gawra, II, pl. LXX, 8. 3' Trisorsde l'AncienIran (Geneve, 1966), pl. i. 43 Tall-i BakunA, pl. 45, 2 (niveau incertain); Tip Gi?yan, 48, pl. 32 Sumer,IV 2e rangde droite (VC); MDP, XIII (1912), pl. V, 2; pl. VII, e (1948), pl. X, 3e rangee, 3e Apartir de la droite. 33 Ur Excavations, pl. 50, en haut a gauche. IV, 4; pl. VIII, 5; pl. XII, 4, de Suse A; MMAI, XXX (1947), 34 Sialk, I, pl. LXV, S I8I5 et pl. LXXIII, S 87; les rares sp6cip. 204, fig. 48, 4, de Suse A 6galement. mens de ce motif trouves ' Sialk IV ne sont guere que des 44 Tip6 Giyan,pl. 59, 2e rangee, au milieu. 45 Par exemple Tepe Hissar, pl. IX, H survivances locales. 3478 et H 4747; pl. X, H 35 Iraq, II (1935), p. 127, fig. 59, 2 et pl. XXII, i, d'Arpachiyah 4637 et H 4378. 46 Comparer Tepe Hissar, pl. VIII, H 4383 et notre pl. le. TT 6. ' 36 Ur Excavations, pl. 49, 2e rangee au centre. 47 Cependant IV, T6p6 Hissar ces motifs sont moins souvent qu'a 37 Tipd Giyan, pl. 43, 2e rangee a gauche, et pl. 44, en haut A Tureng T6p6 enserr6s entre des bandes verticales. 48 TepeHissar, pl. X, H droite. 4637.
25

26

128

JOURNAL OF PERSIAN STUDIES

tions ne sauraient 8tre confondues. Il est inutile d'insister a nouveau sur les analogies qu'Hissar IC presente avec Sialk III, 4-6, dues probablement a des influences venues du sud-ouest; celles-ci se sont ' 6tendues jusqu'au Turkmenistan sovietique, puisqu'on les retrouve Namazga Tepe;49 elles ont en revanche curieusement 6pargn6 la plaine de Gorgan. La poterie peinte de la steppe turcomane n'est donc certainement pas une simple emanation de celle qu'a la meme epoque on fabriquait sur le plateau iranien. On songerait plut6t, comme le suggere McCown,so a une tradition h6ritee des antiques ciramiques iraniennes a d6cor noir sur fond rouge, qui ' T6p6 Hissar disparaissent seulement a partir du niveau IC: raison suppl6mentaire, nous semble-t-il, pour ne pas abaisser davantage la chronologie relative de Tureng Tepe IIA. I1 n'en reste pas moins vrai que notre c6ramique ne d6rive directement d'aucune poterie peinte jusqu'a prdsent connue. On souhaiterait pouvoir d6couvrir dans la steppe turcomane une 6volution ' conduisant de Tureng Tdpi I Tureng Tep6 IIA, paralldle a celle qui sur le plateau mane par ' de Sialk I Sialk III. exemple En verit6 McCown avait une raison tres pr6cise de considerer Shah Tep6 III comme contemporain " d'Hissar IIA, en dipit des difficultes auxquelles se heurte cette thdorie. C'est en effet Shah T6p6 III comme a Hissar IIA que la ceramique grise liss6e fait son apparition au nord et au sud de la chaine de l'Elbourz: il etait donc tentant d'6tablir un 6troit paralldlisme entre ces deux civilisations, notamment du point de vue chronologique. Mais &tait-ce bien une obligation absolue? ' A vrai dire la ceramique peinte que nous venons de decrire n'est pas la seule avoir 3t6 trouvie aux " Shah T6pe: aucune autre niveaux 20 a%26 de Tureng Tep6. Nous ignorons s'il en est de meme n'a 6t6 mentionn6e dans le rapport final. poterie-peinte En fait, entre nos deux categories, la diff6rence n'est pas immediatement perceptible, puisque dans les deux cas le decor est peint en couleur sombre sur fond rouge. Mais l'argile est ici d'un rouge l6gerement diff6rent, plut6t ocr6 que brique; surtout, aucun badigeon ne recouvre la surface, qui est toujours soigneusement liss6e; la peinture est beaucoup moins foncee, plut6t brune que noire, et est ' parfois tres claire, et en tout cas tres fluide; le d6cor est parfois peine distinct (P1. IIId, a droite). sont composes de fagon beaucoup De mime les motifs, bien qu'apparentis aux pricidents, plus lLiche; par exemple on ne retrouve pas les zigzags contigus imitant un riseau de croisillons; d'une faeon g6ndrale les zigzags sont plus larges, occupent plus de place et sont plus isolns (P1. IIIa, en haut 'agauche). Si l'on rencontre encore des rangees de disques sombres, ceux-ci sont moins rdguliers que dans la catigorie precedente, plut6t elliptiques que circulaires, et de contour assez flou (P1. IIa, en haut au milieu; P1. IIIa, en bas a gauche et au milieu). Parfois de grands quadrillages delimitant de larges losanges couvrent probablement toute la surface (P1. Ic, en haut); il arrive aussi, ce qui est nouveau, que le quadrillage soit compose de lignes verticales et de lignes obliques (P1. IIc, en haut a gauche). Un motif original doit en outre etre signal6, dont 1'importance chronologique est grande: il s'agit, d6corant la partie infdrieure d'une jarre dont l'int'rieur n'est pas travaill6, de s6ries verticales et contigues de lignes obliques alternees, figurant une succession ininterrompue d'aretes de poisson " (P1. IIa, en haut a gauche): ce motif se trouve en M6sopotamie5' et en Iran5' a des 6poques ant rieures Hissar IIA, puis il disparait. Dans les formes, une certaine parent6 unit les deux c6ramiques peintes de Tureng Tep6i IIA: on trouve dans les deux s6ries des jarres globulaires quasiment identiques: les deux vases peints les plus r6cents que nous ayons d6couverts L Tureng Tdp6 sont deux exemplaires de ce type trouvis c6te-L-c6te sur le sol 20, 1'un L badigeon rouge (P1. Ic, en bas), 1'autre L surface liss6e (P1. Ic, en haut). Mais ces deux cat6gories ne sont-elles que deux variantes d'une mime c6ramique ? On le croirait d'autant plus volontiers que le badigeon de l'une, nous l'avons ddji notd, prdsente parfois, bien que rarement, la couleur ocrie si caractdristique de l'autre. Une telle conclusion cependant ne tiendrait pas compte d'un fait essentiel: ces memes couches de Tureng Tfp6 IIA nous ont r~v61l en assez grande
49 M. E. Masson et V. M. Masson. Cahiers d'HistoireMondiale,V

(1959), p. 26-27.
50o Comparative Stratigraphy, 54. p.
5'

Ziegler, Die Keramik von . . . Haki

Mohammed,motif 98,

pl. I, e et pl. 27, b; Ur Excavations,IV, pl. 48, 2e rang6e A gauche (d'Ur-Ubaid I). 52 Tall-i BakunA, pl. 52, 13 (niveau incertain); Tdpd Giyan,pl. 55, 2e rang6e, 3e A partir de la droite, de Giyan VC; T&h6ran, Musee Iran Bastan, no. 578, de Cheshmeh Ali.

CERAMIQUES

PEINTES

DE

TURENG

TEPE

129

quantit6 une ceramique rouge ocree, a surface soigneusement lissee, mais cette fois sans decor peint. Elle " n'6tait connue jusqu'ici que par un bol oreilles d6couvert jadis par Wulsin53 et par un bol de Shah Tdp III.*54 Or cette ceramique rouge n'est certainement qu'une variante de la c6ramique grise tous les liss"e: interm6diaires existent entre les deux couleurs, et les formes sont identqiues. C'est le cas notamment d'une jarre d'un type tres particulier, et inesthetique, a fond large, ouverture rdtrecie et tres large rebord 'vas6 presque horizontalement (Fig. 2a, en argile rouge lissde, et Fig. 2b, en argile grise lissde). La c6ramique rouge lissde disparait elle aussi apres le niveau 20, a la seule exception de quelques vases ddcords d'incisions, identiques aux vases gris de la periode correspondante (Tureng Tep6 IIB). A Tepe Hissar aucune trace de cette poterie rouge n'a 6td decouverte, comme si la diffusion de la ceramique lissde n'avait atteint ce site qu'h une epoque oh sa variante rouge de la plaine de Gorgan avait d6ja disparu. Nous retrouvons alors le paralldlisme Tureng Tep6 IIA-Hissar IC que l'itude de la ceramique peinte nous avait amend a reconnaitre. Les rares formes de la c6ramique grise d'Hissar IIA publides par Schmidt 55 ne nous permettent malheureusement pas d' dtablir une correspondance pr6cise entre cette periode et l'une de celles de Tureng Tepe (c'est-a-dire Tureng T'p6 IIA ou IIB). En effet la ceramique grise d'Hissar II (Hissar IIA et IIB ne constituant en fait qu'une seule culture qu'il semble vain de pretendre subdiviser) diff6re sensiblement de celle de Tureng Tip6 II. Elle imite en formes de la ceramique peinte d'Hissar IC, encore presentes a Hissar II. fait, pour l'essentiel, les ' En serait-il de meme Tureng Tep6? On pourrait en effet imaginer que la c'ramique rouge liss6e de la plaine de Gorgan ne constitue qu'une imitation de la ceramique peinte locale. On ne pourrait en ce cas tirer aucune conclusion chronologique de son absence ' Tdpi Hissar. Ce problkme pose 6videmment celui, beaucoup plus vaste, des rapports entre ceramique peinte et c6ramique lissee dans la steppe turcomane, ainsi d'ailleurs qu'a' Tep6e Hissar. S'est-il simplement produit une evolution progressive de l'une Al'autre de part et d'autre de l'Elbourz, sans aucune intervention extirieure? On pourrait expliquer ainsi la similitude des formes, si caracteristique de Tepe Hissar, comme la prdsence a Tureng Tep6 de formules intermediaires entre ciramique peinte a engobe et c6ramique grise liss6e, c'est-a-dire d'une c6ramique rouge lissde a d6cor peint et d'une ceramique rouge lissee sans dicor. Cependant rien n'indique que ces quatre productions diff6rentes se soient rdellement succed6 dans le temps. Assurdment toute reponse ddfinitive nous est interdite, puisque les couches les plus profondes n'ont pas pu &treexplories. Cependant il est remarquable qu'au niveau 26 la c'ramique grise soit plus abondante que le ceramique rouge, alors qu'au niveau 21 la proportion se trouve plut6t inversee. Les deux productions ont donc vecu longtemps c6te-a-c6te, pendant dans doute plusieurs siecles; peut-on admettre en ce cas que l'une est d6riv e de l'autre ? On hesite d'autant plus a le croire que les formes des deux ciramiques sont, a Tureng Tepd, tres diffhrentes dans l'ensemble, a la seule exception des jarres globulaires qui en fait constituent un type trop banal pour ' bord rentr6, ni 8tre significatif. On ne trouve dans la ceramique peinte ni les coupes pied, ni les bols "a " les vases oreilles, ni les "alabastres ", ni les jarres a large rebord de la ciramique liss6e. Celle-ci pr6sente en fait un repertoire de formes beaucoup plus riche, meme si certains types ont pu 8tre empruntes a la c6ramique peinte, par exemple les grandes jarres sans col, a ouverture large et panse anguleuse. De plus ces deux poteries pourtant contemporaines sont fabriquees selon des techniques complktement diff6rentes, puisque les vases lissis sont pour la plupart tournes, alors que les vases peints h engobe ne le sont jamais. D'ailleurs on voit mal pourquoi deux ciramiques peintes tris distinctes, celle de la plaine turcomane et celle de Tip6 Hissar, auraient toutes deux 6volu6 simultandment et dans un sens identique, pour donner une mime ckramique grise liss6e, et pourquoi par ailleurs celle de Namazga Tip6 III, tris proche par certains c6tis d'Hissar IC, fut suivie au contraire de toute une sdrie d'autres poteries peintes. En fait, nous semble-t-il, on ne peut pas ne pas considdrer la ciramique lissde comme un dl6ment &tranger, venu probablement des steppes septentrionales. Mais il s'agit l1 d'un sujet qui ddborde du cadre de cet article.
53 Wulsin, o.c., p. 9, pl. XIII, fig. 2. 54 Shah Tepe, p. I76; non illustre. 55 Tepe Hissar, pl. XXIII;

MuseumsJournal, XXIII, 4 (1933),

pl. XCVII-CI.

Fig. 2.

A, (a) Sondage sol o20, vaserougelissi TT 278 (i/6).

A, (b) Sondage sol 18, vasegris lisse TT 220 (I/6).

(c) SondageE2, entresols I et

peint (i/I). <,tesson

E2, entresols I et 2, tesson (d) Sondage peint (i/i).

E2, entresols I et 2, tesson (e) Sondage peint (i/i).

CERAMIQUES

PEINTES

DE

TURENG

TEPE

131

iraniens. Ce fait jetterait un jour intdressant sur les relations commerciales de Tureng Teip6

Toujours est-il qu'a partir de Tureng Tepe IIB la ceramique lissde regne seule: aucune poterie peinte ne fut plus fabriquee sur ce site jusqu'a l' poque islamique. Pourtant trois tessons d'une c6ramique peinte jusqu'ici inconnue ont dt d6couverts dans notre sondage E2, entre les sols I et 2, datant tous deux de la periode Tureng Tep6 IIIC, correspondant " Hissar IIIC. Ces fragments semblent appartenir a des vases diff6rents, mais de technique identique: l'argile, d'un rouge vif, est trbsfine et bien cuite, les parois sont minces, la surface, sans engobe, est tres soigneusement lissde a l'exterieur; le d6cor est effectu6 au moyen d'une peinture gris tres fonce ou brun l'intirieur comme "a noira^tre suivant les cas (Fig. 2c-e). Un fond plat d'un vase tres 6vas6, nous a ete conserv6, ainsi qu'un rebord de bol l1gerement convexe, a lIvre simplement amincie, trop incomplete pour que le profil en pflt &tredessind. Le decor, uniquement exterieur, comprend des groupes de bandes horizontales, notamment sur la lRvre, des bandes verticales, une rosette constitude de plusieurs petits points, enfin un motif malheureusement tres incomplet, qui pourrait tre une tate de gazelle tres stylisee: on songe ividemment a certaines gazelles ou autres animaux a cornes reprisentes sur des vases peints de Kara Tepe 2,56d'Hissar IC,5 de Giyan VC58ou de Bakun A.59Mais ces ressemblances sont fallacieuses. Comment admettre que des tessons de plusieurs vases diff6rentsimportis du plateau iranien au IVe millinaire puissent etre simultanement remont6sjusqu'a un meme niveau datant du d6but du IIe millinaire? Au surplus aucune trace de telles importations n'a et6 d6couverte dans les couches explories de Tureng Tp6 hIIA.Surtout la technique est tr's diff6rente, qu'il s'agisse de la couleur de l'argile, de l'absence d'engobe ou du lissage tres soigne de la surface. II faut donc bien supposer que ces vases ont 6t6 importis au d6but du IIe millinaire d'une rigion apparemment 6loignie, puisque ni sur le plateau iranien ni au Turkestan aucune c6ramique de ce type ne semble avoir eite fabriquee a cette date. C'est en direction du sud-est qu'il faut apparemment chercher les affinitis les plus proches, sans pour autant pritendre en d6couvrir l'origine exacte: les confins sud-orientaux de l'Iran sont jusqu'a present trop mal explords. On sait du moins que la decor noir, ornee notamment d'animaux a c6ramique dite" chalcolithique ", en argile rouge liss6e et "a cornes tres stylises, demeura a la mode dans ces rigions, ainsi qu'en Inde, jusqu'a une epoque extremement tardive, probablementjusqu'au debut du ler millinaire. Les fouilles recentes de Rangpur, pres de Lothal, ont par exemple rev1ld une ceramique de ce type, d6corde d'antilopes et de bovides, aux niveaux IIC et III;6o on la trouve igalement a Navdatoli III moyen et r6cent.6'A Nevasa III, des rosettes composdesde plusieurs points et considerees comme des marques de potier ornent 6galement une ceramique a peinture noire sur fond rouge, souvent decoree d'animaux.6 Les niveaux IIC et III de Rangpur sont dates par S. R. Rao respectivement de I Ioo--ooo et iooo-8oo avant notre "re;63 ces dates extremement bassesont etd remontees par B. B. Lal, sur la foi d'dvaluationsobtenues par analyse du CI4: la ceramique rouge lissde 'adcor noir de Rangpur et Navdatoli apparaitrait peu apres le milieu du IIe millenaire6gpour se poursuivre sans doute jusqu'au debut du ler millinaire. Rien ne comme importes, dans s'opposedonc a ce que nous considdrionsles tessonspeints de Tureng Tp' IIHIC les premierssi cles du IIe millenaire, d'un site encore inconnu, voisin probablement des confins indo-

6poque, et l'on 6voqueraitvolontiers a ce sujet les nombreux objets de type Hissar IIIC decouverts dans les couches superieuresde la civilisation de l'Indus et au Baluchistan.61

a cette

V. M. Masson, EneolitIoujnykh OblasteiSredniei Azii, II, pl. VIII, 43, 6o, 61. 57 TepeHissar, pl. XIII, DH 44, 14b. 58 Tep6Giyan,pl. 54, 3e rangee, 2e a partir de la droite; cf. aussi pl. 63, en bas. 59 Tall-i BakunA, pl. 74, I (Bakun A III). 60 Ancient India, XVIII-XIX (1962-63), p. 103 et p. 114-16; pl. XXII et pl. XXIII, B (en couleurs), fig. 39, nos. 49 52. a Cette c6ramique est meme orn6e parfois d'6toiles ik branches
56

terminees par des boules qui indiquent bien de lointaines affinites iraniennes (ibid., p. 10xo9, 37, 2, de la p'riode III). fig. 61 Ibid., p. 198; Sankalia et autres, Excavationsat Maheshivanand Navdatoli, p. 103 et fig. 33, no. 37d. 62 H. Sankalia et autres, From History to Prehistory at Nevasa (1954-56), p. 250 sqq, fig. i o. 63 AncientIndia, XVIII-XIX, p. 204. 64 Ibid., p. 215 et p. 220. 65 Cf. Mode, DasfriuheIndien,p. 107.

a Pl. Ia. Sondage sol 14, tessons degraissant A, ve'getal.

P1. Ic. SondageA, sol 20, jarres globulaires.

A, ve'gtal. degraissant Pl. Id. Sondage sol 15, tessona&

Pl. Ib. Sondage sol 25,fragmentdegrandejarre. A,

sur Pl. le. Sondage tesson A, peint trouve' le sol 2 ( /112).

Pl. Ha. Sondage sols 24 et 25, tessonspeints. A,

peints. PI. IIc. SondageA, sol 20, tessons

Pl. JIb. Sondage sol 25,fragmentdegrandejarre. A,

de Pl. IId. SondageEI, tesson peint trouve'pris la surface.

peints. P1. IIIa. SondageA, sol 24, tessons

Pl. IIIb. Sondage sol 25, fragments degrande A, jarre.

jarre. Pl. IlIIc. SondageA, sol 25, fragmentde grande

Pl. IIId. Sondage sol 21, tessons A, peints.

133

SURVEY OF EXCAVATIONS IN IRAN DURING 1965-66

The Director's Report in Iran IV intimated that in Volume V it was hoped to include a survey of the archaeological work undertaken in Iran in 1965-66. Thanks to the generous co-operation of the Directors, a number of excavation reports have been assembled and are presented here, arranged alphabetically. It is hoped that continued support will be given to this new feature. Reports, accompanied by photographswhere appropriate,should be sent direct to the Editor, c/o The BritishAcademy, Burlington Gardens, London, W.i, or, if more convenient, to the Director of the British Institute of Persian Studies in Tehran. BdbdJdn Tepe Excavations at BdbdJan, Lfiristan, sponsored by the Institute of Archaeology, London, started on September 3rd and continued for two months. The site is about 250 yards acrossand consistsof two main mounds, about ten and twelve metres high respectively. A trench was opened on the flat summit of the

S
I

. .

'U

"

A CAS 5 . \C

P P/AN A l

. .
/ .
s

*TABRIZ

Takht- LSuleiman

Dlnkpooe Tepe

h O-asanlu

Tureng Tcpe

MESHED

>

.,
("

TEHRAN

Ghar- Khar & * HAMADAN Tet-Gnj-i S* Dcreh"< Godin Tepe Gcj * Baba Jan Tepe <4

L Tep Sialk

"* Tep's & War Kabud Kaolwali

o ISFAHAN

Susoa

rd-e- Nechandeh
/

ZABOL D ODahan.L.Ghulaman h

oKERMAN * Dorudzan * Persepolis

SHIRAZ

/ ZAHEDAN

Tall blis

* Bampur
*Tall__NIS

IN

AU

D I

APABIA
APABIA o

134

JOURNAL

OF

PERSIAN

STUDIES

largest mound, and by the end of the season had revealed the foundations of a large fortified manor about thirty yards across with walls up to 11 metres thick. The plan comprised a large central hall with a double line of columns down the centre and a stone paved area at one end. It was surrounded by long corridor-like rooms and may have had an open portico along the front. The plan is most closely paralleled at Hasanlu IV (Burned Building III), but also shows possible connections with contemporary Assyrian palace architecture, as well as foreshadowing later Achaemenid constructions. An earlier version of this building, equally large but with corner towers, was found directly underneath. Red on cream painted pottery from these levels, of the type known as " genre Luristan ", suggests that these buildings probably date to the ninth or eighth centuries B.c. They presumably formed the seat of a local tribal chieftain. A step trench cut into the side of the smaller mound produced a sequence extending from the painted wares described above through to pottery equating very closely with wares from Hasanlu III. Finds consisted of domestic architecture and artifacts, including baked wall tiles painted with designs recalling those on the pottery of Sialk, Necropolis B. C. L. GovFF.

Bdmpir
Excavations were undertaken at Bampfir in 1966 in order to establish the stratigraphy of this district of southeastern Iran, first surveyed by Sir Aurel Stein in 1932. This research was made possible by grants from the British Academy, the Leverhulme Trust and other bodies, and was carried out under my direction in collaboration with the Archaeological Museum, Tehran. Work began in March on a site (Z) to the northwest of the fort. The intended area of the excavation was restricted by the need to avoid Islamic burials but a trench 6 X 2 metres was sunk to natural at a depth of 7 metres. Site Z provided evidence of six periods of prehistoric occupation. The earliest levels contained no structures but sun-dried mudbrick buildings occurred in Period II. These were rebuilt in the subsequent period when five successive floors were laid. The area within the trench was levelled in Period IV and new structures erected on an entirely different plan. The site was then abandoned until the arrival of newcomers in Period V. The building of this period was short-lived and was succeeded in Period VI by an occupation layer containing much pottery. The site was used as a cemetery in early Islamic or later times and the upper levels of Period VI were much disturbed. A second trench (Y) was opened across the line of Stein's excavations. This showed that the original trench dug "to the level of the surrounding plain" had in fact only cut through a rubbish dump of Period VI. Excavations in this trench confirmed the sequence obtained on Site Z and revealed a wellpreserved kiln in a Period IV context. Other objects of interest in this horizon included a fragment of carved steatite, a compartmented metal stamp seal, an animal figurine and numerous oval clay slingstones. The decorated pottery from Bdmpfir is of importance since it can be related, during Periods III to V, to the wares of other countries. A cream-slipped, wheel-made red ware was current during Periods IIV and a thin grey ware was used for small flat-based bowls and cups with black friezes. Forms were rather angular during Period I but were succeeded by flattened spherical shapes. Cordons, striped and outlined in black, were applied to the thicker jars in either straight or undulating lines and wavy comb incising was used on some plain cream-slipped jars. A wide range of geometric and zoomorphic motifs was current throughout the Bampfir sequence. They included, during Period II, a repertory reminiscent of Susa I sigmas, broad multiple chevrons or angular patterns and fringed M's. New designs appeared during Period III which suggest contact with southeastern Afghanistan during the period of Mundigak IV, I. This connection lasted into Period IV, the context to which Mundigak IV, 2-3, sites in the Gardan Reg, Seistan, and Khurab burial, Bii, are best related. Fragments of incised grey ware " hut urns " appeared in this period and similarities between both the painted and incised wares suggest contemporaneity between Bampfir IV, the precemetery levels of Sh~hi Tump, Pakistan, and sites on Umm-an-Ntr off the Oman coast. The ceramics of Period V were inferior to those of the earlier periods from which they differed both stylistically and in the preference for a red instead of a cream slip. Incised and painted grey wares

SURVEY

OF

EXCAVATIONS

135

continued and streak-burnishing was introduced. New forms included ring-based bowls and canisters comparable to those of the Kulli culture. Kulli motifs included spiky trees surmounted by an arrowshaped object, palm trees with adjacent caprids-not bulls-fish, triangles point to base and wavy hatched bands. These designs were, however, invariably executed in black without a secondary colour. Characteristic motifs of this period not found on Kulli-ware were striped M's used as panel dividers and a chevron band set between straight lines. The latter is a design common to the probably contemporary Shugha culture of Fars. Period VI produced much decorated pottery but added little to the range of designs already published.' Streak-burnishing became increasingly popular but was used mainly on shallow platters and bowls. While the excavations were limited in scope, the results obtained are of considerable importance. Erosion has destroyed the stratigraphy of many of the prehistoric sites in Seistdn and a cultural sequence has yet to be established for the Makran in Pakistan. The relation of the Bampfir sequence to these regions and even further afield, may help to fill these gaps in our knowledge. A report will be appearing in Iran VI. B. DE CARDI. Bard e Nechandeh A short illustrated account of Professor Roman Ghirshman's discoveries of the Parthian period at Bard e Nechandeh was given in the IllustratedLondonNews ofJuly 16th 1966. Bisitawn2 The most recent programme of field research at the ancient site of Bisitfin has been that conducted for the past four years by the German Archaeological Institute in Tehran. The great rock of Bisitfin BAGISTANUS in the account of Diodorus and Bdgdstdna, the place of the gods, in Old Persian-stands directly above an abundant spring that irrigates a wide strip of ground to the east. This point, not far from Kermdnshah, became a resting place for countless caravans, tribes of nomads and marching armies passing from Babylon to Ecbatana or returning from the Plateau to Mesopotamia. The site owes its fame to a series of historical monuments that stand close to the old highway. These include the elevated relief and inscription of Darius the Great, which remained an enigma to all later travellers until Sir Henry Rawlinson-whose perilous climbs were epic feats in themselves-first published his pioneer decipherment in 1846;3 a Seleucid Herakles Kallinikos, erected in the year 148 B.C.; and three Parthian reliefs (belonging to Mithradates II, Gotarzes and an unknown regent), each of which were studied by Ernst Herzfeld in his famous book Am Tor vonAsien.To a lesser degree too, the site has become known in archaeological literature for three richly worked Sasanian capitals. These are known to have been moved to Taq-i-Bfistdn in the past few hundred years and have been variously attributed, together with the chief iwan at Tiq-i-Bfistdn, to either the reign of Khusrau II (590-628) or to that of Peroz (457-483). The first excavations at Bisitfin were those carried out in 1949 by Carleton S. Coon, who restricted his soundings to a small cave deposit near the spring-fed pool.4 In contrast to this specific study of Bisitiin's first occupation, the object of our own excavations was to gain a fuller understanding of the site as a whole, with special reference to any architectural appointments that might be related either to the relief of Darius or to the three displaced capitals of Sasanian date. From the results of our excavations and other studies in the pre-Islamic period I may list: (i) The recovery of an ancient terrace in the small gorge under the relief of Darius. As at least one indication of its date, this structure was found to be covered with small stone chips that had evidently fallen upon it during the construction of the relief above.
Reconnaissances north-western in India and Stein, Archaeological south-eastern (1i937), P1. VII. For designs typical of Bampfir Iran see Pls. XIII and XVI (lower 2 VI, report and those on Choga plate). Haft Tepe and Siraf This Mish, represent late additions to the Survey and do not appear on the map. 3 H. Rawlinson, JRAS X, 1846. 4 Carleton S. Coon, CaveExplorationsin Iran 1949, 1951, p. 6f.

136

JOURNAL

OF

PERSIAN

STUDIES

(2) A close study of the relief of Darius. In particular, a whole series of fresh photographs of the relief were obtained from a tall observation platform that was erected in December 1963, when His Imperial Majesty the Shdhanshdhcame to Bisitfin to open a sugar factory in the neighbourhood. (3) The partial excavation of a sizeable fortification of Achaemenian or pre-Achaemenian date. Embracing several terraces, this construction lies to the north of the main cliff face. (4) The discovery of an isolated column base of possible Seleucid date. (5) The recovery of impressive evidence for a major architecturalprogramme in the late Sasanian period. Starting with a carefulstudy of a terracewall that lies parallel to a hitherto undated rock cutting some 200 m. long, it was found that the mason marks not only matched those attested in a supporting wall attached to the rock carving but also those found on both the stones of a long wall situated beside the river and the blocks of a bridge that goes by the name of Pul-i-Khusrau.What is more, local tradition still calls the raised ground before the huge unfinishedcarving "Farhad'sTerrace", while one of the illustrations to Nizami's twelfth-century work shows Farhad himself carving the rock of Bisitfin, watched by Queen Shirin. Thus, from strictly technical indications (the mason marks), the evident legend and the unfinished characterof this ambitious work, we can almost certainly attribute the whole programme to the later years of Khusrau II, possibly between his conquest of Jerusalem in 615 and his fall in 628. Furthermorethe three sculptured capitals from Bisitfin can be seen to be influenced by Byzantine types of the sixth century-again suggestingthat they should be ascribed to the same period. Finally, within the present village, we have studied an early Islamic caravanserai, which may have been constructed in part with re-used blocks of Sasanian date. It appears to have been rebuilt in the Mongol period, when we know from Hamdullah Mustaufi of Qazvin (1340) that yet another royal abode was established at the foot of the rock. Also from the Mongol period, we should mention a kiosk built directly on the Sasanian river wall, which was decorated partly with tiles similar to those used at Kdshdn and Sultdndibd and partly with elaborate cut bricks of a type known from at least Seljuq times onwards. H. LUSCHEY.

Dahdn-i-Ghuldmdn The first of two reports of the Italian expedition's work at this site was published in the Illustrated London News of October 29th 1966, Archaeological Section no. 2255. Dr. Scerrato has discovered an Achaemenian city at Dahdn-i-Ghuldmdn, which he identifies as Zarin, capital of the province of Drangiana. DinkhaTepe During July and August of 1966 the Hasanlu Project of the University Museum of Philadelphia and the Metropolitan Museum of New York City carried out exploratory excavations at the site of Dinkha Tepe. Dinkha Tepe is located in the Ushnu valley a few miles southwest of Hasanlu in the southwestern corner of Western Azerbaijan. The site had been previously visited and tested by Sir Aurel Stein in I936 and was selected for investigation at the present time in view of the extensive second millennium B.c. deposits known to be there. The excavations included a number of test pits put down at various points both on and off the mound to determine something of the periods represented and the growth pattern of the mound. In addition four Io metre square area excavations were carried out on the northern edge of the site next the area, now washed away by the Qadar River, where Stein originally dug. From this area a trench was extended upwards to the highest crest of the site where two additional trenches were opened to ascertain the nature of the topmost deposits. An additional area was opened near the edge of the higher central portion of the site. Although loose sherds of Dalma Ware (c. 4200 B.C.) were found in late fill none of the test excavations encountered the original site of the Dalma occupation. The major portion of the site dated to the second millennium B.c. This second millennium settlement can be seen to include three broad stages: limestone outcrop next to the Qadar River which quite possibly ran (I) the original settlement on a at that time on the south rather than north side of settlement; (2) the gradual extension of the settlement beyond the walls of the original town to the limits of the presently visible tepe (and an additional

SURVEY

OF

EXCAVATIONS

137

area now washed away by the river); and (3) the abandonment of the original town wall and the subsequent construction of buildings over its site. It seems likely on preliminary evidence that stage 3, and possibly 2 as well, date to the Late Bronze Age of the second millennium-about 1500 or 16oo to B.C. The earliest stage (I) has not yet been penetrated by excavation. Following stage 3, which in 1250 at least one area ended with a conflagration, the site was occupied by the Iron Age I culture already known as Hasanlu V (c. 1250-1000 B.C.). This occupation was represented as far as presently known by deep deposits of ash and trash into which a number of graves or tombs were sunk. These graves and tombs are characterized by the presence of jars with unbridged horizontal spouts in burnished greyblack ware, polished grey-black goblets, and bowls with flaring sides and inverted crescent-shaped lugs and two holes (presumably for a string for hanging the bowls up). No vessels with bridged spouts were encountered anywhere in this period, as is also the case at Hasanlu. In addition to the greyblack ware, red and buff wares also occur along with a polychrome of black and red on cream bands set around the shoulders of jars or bowls. At least one sherd of grey-black pattern-burnished ware was found along with several sherds of white-filled incised ware. Into the top of this Iron Age I stratum were sunk brick and stone tombs of Iron Age II, or Hasanlu IV type. At Hasanlu burials have been primarily simple inhumations in contrast to these Dinkha tombs, although the contents of both are virtually identical. A total of some 8o tombs of Iron I and II were excavated. The Iron II tombs are invariably characterized by the inclusion of jars with bridged horizontal spouts, and carinated bowls in grey, red or buff ware. In contrast to these two later levels, the Late Bronze Age materials are characterized by fine wheelmade vessels of buff-to-yellow colour often painted with red, red-brown, brown, or dark brown paint in either simple parallel horizontal bands, or with a single zone decorated with cross-hatched triangles, circles, opposed solid-colour triangles, birds, and ladder patterns in various combinations. Larger jars were painted primarily with parallel bands. Much of the pottery was a plain ware ranging in colour from orange to buff-brown. Much of this was handmade. Quite characteristic were large storage jars fired red or buff, with zones of incised wavy lines or discontinuous "teeth" marks made with a comb-like tool. Often these wavy line groups are separated by raised ribs. Various other patterns of incised criss-cross lines sometimes alternate with the wavy lines, and sometimes are used with impressed circles. Sherds of this decorative type have frequently previously been called " Islamic " in various surface surveys since they closely resemble later decorative patterns. In addition, a few sherds of a Late Bronze Age fine, wheelmade, burnished grey ware were found. This latter appears to be quite distinct when compared with Iron I sherds. Two rock tombs belonging to the Late Bronze Age were found. One contained at least nine individuals, the other two adults and a baby. Plain ware pots and personal jewellery made up the major offerings with the addition of a bronze arrowhead and a bronze sword blade as weapons. Toggle pins were common. Two pendants of gold, a crescent and a disc with embossed star, were found. In general the parallels seem to be with Nuzi and Giyan III-II. At the end of Iron Age II (Hasanlu IV) the site, like Hasanlu and other large citadels in the area, was abandoned and was not reoccupied, it would seem, until late in the Islamic period. R. H. DYSON, JR. Dorudzan At the close of the 1935 season at Persepolis, the German architect Karl Bergner surveyed and reported the existance of several Achaemenian sites along the Rfid-i Kur river.5 These sites are located about 5o km. northwest of Persepolis. Excavation at the sites had not been undertaken until May, 1965, when Pahlavi University decided to field a rescue expedition due to the projected Dorudzan Dam project which threatens to inundate several monuments. The primary concern of the expedition, so far, has been Site A-a canal head called Sang-i Dokhtar6 -located 3 km. to the west of Dorudzan. This canal head, measuring approximately 8 x 3 m., is
5 Karl Bergner, Archaeologische Mitteilungenaus Iran VIII, 1937, pp. I-4. 6 Ibid., p. I and Tafel IV.

138

JOURNAL OF PERSIAN STUDIES

constructed of worked blocks irregularly set into and upon a base cut out of bedrock. Two arches have "shutter-type " sluice gates. During the course of excavation,

been cut through this base. The flow of water through the arches and into a canal was controlled by encountered. Carbon and soil samples have been extracted from layers laid down prior to and
contemporary with the canal head construction. Through C'4 and pollen analyses of these samples an of the monument's lower part will be cut away and subsequently attached to a prepared concrete core twenty superimposed strata were

absolute date for the canal head as well as a descriptionof the flora in the area should be obtained. Before this monument is removed to a safe spot the worked blocks will be numbered and the fagade

at a new site. A complete record consisting of sections, photographs, plans, and elevations is being prepared by the expedition architect. Two other sites, an Islamic dam and an Achaemenian dam,7 will be investigated in the next few months. M. B. NICOL. and Ghar-i-Khar Ganj-iDarek the summer of 1965 the writer made a prehistoric reconnaissancein Western Iran, as the During basis for a programme of future research in that area. The survey extended from northern Luristan through Kurdistan to Azerbaijan. The emphasis was on locating cave, rock-shelter and open sites which might yield data on the Palaeolithic, epi-Palaeolithic and early Neolithic occupations of the Zagros range. Most of the survey was carried out in close associationwith Dr. T. Cuyler Young of the Royal Ontario Museum, since my researchin the earlier prehistorywas to a large degree complementary to his on the later sequences.8Mr. Adad Darbani of the Tabriz Museum representedthe Archaeological Service of Iran. Some of this area had been examined rapidly in 1949-51 by C. S. Coon, who had excavated several Palaeolithic sites (Bisitfin Cave in the KermdnshThdistrict and Tamtama Cave in Azerbaijan). In such a large region our 1965 survey, like Coon's earlier one, could not attempt a thorough examination of each valley; this task is still to be done. Of the sites investigated in 1965, many contained no archaeological materials or had only shallow deposits from historic and late prehistorictimes. These will not be discussed here. However, in the Kermdnshdh district two important sites were investigated which are most promising for future work. Khar. Ghar-i Just outside the village of Bisitfinis a long (27 m.) and deep cave with archaeologicaldefrom at least Middle Palaeolithic (Mousterian) times until the present day. The site posits ranging ("Donkey Cave") is located about half-way up a face of Bisit-in Mountain facing the valley floor of the Gamas Ab. A sounding 2 m. X I m. was made in late August through the deposits near the mouth to a depth of 5- 15 m. Neither bedrocknor the base of the depositswasreachedby the time workwas suspended. The wealth of stone and bone artifacts, faunal materials, charcoal, etc., which were recovered in this brief and restricted sounding indicate the value of the site in shedding new light on the Palaeolithic occupation of the Zagros. The Mousterian implements (side-scrapers,thick blades) at the base of our sounding are followed by levels containing a stone industry with burins, end-scrapersand backed blades reminiscent of the Baradostian (Upper Palaeolithic). Above this are artifacts, often
microlithic and geometric in form, of epi-Palaeolithic or "Mesolithic" character, probably related to the Zarzian industry. Near the top are levels containing a possible aceramic Neolithic, as well as coarse red straw-tempered " software" pottery. These are followed by scanty traces of Sasanian (?) occupation and by later historic materials. A time span of at least 35,ooo years is attested for occupation of this site, with the Palaeolithic horizons making up the largest and most significant part of the deposits. Thus, Ghar-i Khar presents an excellent opportunity to study not only the cultural materials but also the changes in climate, flora and fauna of the region over this long period from Last Glacial to modern times.
7 Ibid., p. I if. and Tafeln VI-VII. 8 See T. C. Young and P. E. L. Smith, "Research in the prehistory of central Western Iran", Science, 153 (1966), pp. 386-91.

SURVEY

OF

EXCAVATIONS

139

Tepe Ganj-i Dareh. Some 15 km. from Bisitfin, in the Gamas Ab valley near the town of Harsin, we sounded a small mound (" Mound of the Treasure Valley ") for which the abundance of stone artifacts and the absence of pottery on the surface suggested a fairly early Holocene date. The deposits are c. 7 m. deep and the mound is c. 40 m. in diameter at present. Our sounding, which was carried down to sterile soil, revealed a number of levels of what seems to be collapsed mud walling interspersed with ashy layers. These levels yielded abundant artifacts in flint and other stones (except obsidian) as well as faunal remains. The typology of the flint artifacts (fine cylindrical and pyramidal cores, backed bladelets, scrapers) plus the bone awls and fragments of probable stone dishes suggest the assemblages here belong to the early Neolithic tradition which is represented at Braidwood's site of Tepe Asiab and at other sites in the Central Zagros, i.e., to what is sometimes called the aceramic or Pre-Pottery Neolithic. This estimate appears to be confirmed by a radiocarbon assay of charcoal from the base of the mound which gave a date of 8450 + 150 B.C. (GaK-8o7). If correct, this would presumably be fairly close to the supposed beginnings of animal and plant domestication in this area (cf. dates of c. 8900 B.C. from Shanidar Valley, Iraq, for the earliest evidence of goat domestication). The duration of the occupation of the mound, which is probably composed of a series of small villages of mud-walled houses, is not yet known. Charcoal from a higher level has given a C'4 date of 6960 + 170 B.c. (GaK-994), which may indicate that the later levels of the site are contemporaneous with the basal Jarmo horizon in Iraqi Kurdistan and with part of Tepe Guran in Luristan. The absence of demonstrably pottery-bearing levels suggests that the mound was not occupied during the sixth millennium or later; we are still uncertain if the elusive eighth millennium is represented here. Altogether, Ganj-i Dareh seems a most promising site for further investigations for, in addition to the abundance of artifactual and other evidence, the absence of deposits later than the aceramic Neolithic will make it practical to uncover large areas, or even the entire living surfaces, of each occupation level. The site provides a very useful transitional sequence between the Upper Pleistocene and early Holocene series, as represented at Ghar-i Khar, and the succeeding archaeological horizons as revealed by Dr. Young at Godin Tepe in the same valley. The combined data from all three sites should go a long way in illuminating the events and processes of human occupations and natural environments in this area during several vital moments in prehistory, particularly if botanical remains can be recovered to supplement the abundant faunal materials. P. E. L. SMITH.

GodinTepe
Preliminary test excavations were conducted by the Iran Expedition of the Royal Ontario Museum of the University of Toronto at Godin Tepe in western Iran from mid-September to early November, 1965. Godin Tepe is a large site located approximately 12 kilometres southeast of the town of Kangavar on the ethnic border between Kurdistan and Luristan. A high central citadel mound of some 5 hectares in area rises to a height of 26 metres above a sprawling outer town covering approximately 5 more hectares and rising 4 metres above virgin soil. The north face of the site has been eroded away by the east branch of the Gamas Ab River and two test trenches placed here revealed a stratigraphic sample of the site from summit to virgin soil. Seven main cultural periods or phases have been defined which will, without question, be subject to considerable sub-division and re-evaluation as further excavations yield more data. Period VII (c. 5500-5000 B.C.) is characterized by a plain straw tempered pottery similar to some wares found at other western Iranian sites such as Hajji Firuz, Guran, Ali Kosh and Sarab. Little was recovered for this period beyond pottery except for a few long, parallel-sided flint blades and secondary evidence for considerable chineharchitecture. Period VI (c. 5000-3500 B.C.) represented by some 4"2 metres of well-stratified occupational and constructional debris yielded a pottery most closely comparable with that of Couche V at Tepe Giyan. Of particular significance was the discovery that the plain wares characteristic of Period VII remained in use in the early strata attributable to Period VI, indicating no sharp break between the sixth millennium horizon at the site and the ceramic cultures which followed in the fifth and early fourth millennia. This conclusion is so far supported by the nature

140

JOURNAL

OF

PERSIAN

STUDIES

of the stratification itself, which shows no evidence of any major abandonment of the site between

VII Periods andVI.

Period V (c. 3500-3000 B.C.) is typified by pottery closely comparable to that of the Uruk Period in Mesopotamia, though recent analysis suggests that this level may also have links with the earlier B.C.?) strata were ranges of the Early Dynastic Period in Mesopotamia. Period IV (c. 2700-2000 separated from those of upper Period V by a marked erosion surfacewhich may, of course, prove to be only local, but which also may suggest something of a chronological gap between the two periods. Certainly there was a major interruptionin the ceramic sequence of the site at this point, for the pottery of Period IV is mostcloselycomparableto the EarlyBronzeAge I and II grey wares heretoforeexcavated in Iran only at Yanik Tepe near Tabriz. Period III (c. 2oo000-16oo B.C.) is representedby over 7 metres of mixed occupational and construction strata, and can be compared generally with Couche III at Tepe Giyan on the basis of the painted pottery recovered. Some evidence suggests the possibility of a rather extensive fortification of the citadel at more than one level in this period. Period II (c. eighth to fourth centuries B.c.) is found only in a rather thin stratum of wash immediately underlying constructionaldebris of Period I, the remains of a recent Islamic fort or watch tower. The length and relative completenessof this stratigraphicsequence is the most promising feature of the site. Further excavation should yield considerableevidence on: (I) the exact nature of the relationship between the late Neolithic of western Iran and the succeeding periods which cover roughly the Halaf-Ubaid periods in terms of Mesopotamia; (2) the relationship between the Uruk of western Iran and of Mesopotamia; (3) the significance of the appearance so far to the south of a ceramic horizon closely linked to Yanik Tepe, the Caucasus and eastern Anatolia in Early Bronze Age times; and (4) the nature of the stratigraphic sequence in central western Iran during the period covered by Giyan III, which has heretoforebeen defined only by a tomb sequence. A thorough study of the local micro-environment by the appropriate natural scientists and the correlation of the Godin evidence with earlier materials in the Paleolithic and early Neolithic time ranges being excavated in the same valley systemby the expedition of the Universit6 de Montreal, under the direction of Philip E. L. Smith, will further enhance the significanceof the expedition's efforts. The chance find of a large Assyrian stele belonging to Sargon II on a nearby mound in the AsadhbddValley late in the 1965 season resulted from following up leads supplied by the Godin workmen (P1. I). The stele is approximately 165 centimetres long, 65 centimetreswide and 70 centimetres thick. One side reveals a badly weathered relief of the Assyrianking, who has the beginning of an inscription across the bottom of his folded robe. Two other sides of the stele have 75 lines each of NeoAssyrianinscription. The fourth side is blank. Though badly weathered and thereforenot susceptible to easy preliminary reading, enough of the text has so far been recovered to indicate that it is an account of an Assyrianmilitary campaign into the central Zagrosregion, including the recording of a number of new local place names and some fresh onomastic evidence. Prompt action by the Iranian Archaeological Service in transportingthe stele to Tehran has insured its preservationfrom furtherweathering. Hopefully it will be possible to continue with more extensive excavations at Godin Tepe during the summer of 1967 and for as many seasons after that as will be necessary to add significant horizontal clearance to the evidence so far obtained from the vertical testing of the site. T. CUYLER YOUNG, JR.

Haft Tepe The initial excavations of the Iranian Archaeological Service at the site of Haft Tepe in Khuzistan lasted for eight weeks during March and April i966. The team opened soundings in two separate mounds. The sounding in Tepe Abu Fandowa consisted of a stepped trench designed to test the complete stratigraphicsequence from a pre-Susa I context down to Neo-Elamite times. In the second sounding, which was located at the northern edge of the chief mound of the whole complex, the expedition was able to uncover a large tomb with 21 skeletons. The tomb had a vaulted baked brick roof while the skeletons were laid out on a large platform, different parts of which still

SURVEY

OF

EXCAVATIONS

141

showed traces of red ochre. Next to the tomb was a large structure-at present assumedto be a templewith heavy mud-brick walls and a series of halls and courtyards.This structureproduced many objects of Elamite date including an incomplete cuneiform inscription written in the Neo-Babylonian script. According to ProfessorErica Reiner the inscriptionrepresentsthe second known Babylonian inscription of the Elamite king Tepti-ahar who may have ruled either in the fourteenth century B.C. (Hinz, Das ReichElam, p. 53f.) or in the early first millennium B.C. The Haft Tepe inscription is thought to be part of a text referringto the maintenance duties of the guardians of a tomb. Mention is made of detailed sacrificesto be made before the chariot of the god and of Tepti-ahar. Altogether 55 lines of the original stela survive, the stone itself measuring 50 X Io8 cm. It is planned to resume excavations at Haft Tepe during the first quarter of 1967.
E.
NEGAHBAN.

Rashi A superb collection of gold and silver vessels of the Sasanian period was displayed in the exhibition "Trdsors de l'Ancien Iran ", shown at the MusdeRath, Geneva, fromJune 8th to September 25th, 1966. Some of these are catalogued as coming from Rashi, a site near Marlik Tepe. It is reported that a shrine and a cemetery have been found there. Siraf With the supportof the Calouste Gulbenkian Foundation, the BritishAcademy, the BritishMuseum and the British Institute of Persian Studies, the first excavations at Sirafwere carried out in the autumn of 1966. First mentioned in A.D. 851, Siraf ranked with Basra as the leading port on the Persian Gulf in the ninth and tenth centuries A.D. Damaged by an earthquake in 977 and affected by the political consequences of the fall of the Bfiyid dynasty in the following century, the city began to decline and Yaqiat who visited it in the early thirteenth century describes it as an all but abandoned port. The site of Siraf was first identified with the ruins near the present village of Taheri, some 150 miles south-east of Bushire, in 1835. It was not systematically explored until 1933 when Sir Aurel Stein visited the site and described its buildings, rock cut cemeteriesand abundant fragmentsof Islamic and Far Eastern pottery. Siraf lies in a bay about three miles across and divided by a spur on the summit of which stands the fort of the Shaikh of Taheri. To the east is the village of Taheri and to the west the site of Siraf. The first season of excavations was planned as a series of soundings to test the nature and thickness of the archaeological deposits. This plan was modified when it was learned that a road was about to be constructed over one of the major public buildings of Siraf, a large mosque. The following excavations were carried out: the mosque was partly uncovered, a deep sounding was made some ioo yards north-west of the mosque and two smaller excavations were carried out on the site of a house or warehouse and a pottery producing centre. noted by Stein is situated near the sea coast about a quarter of a mile west of the Fort The Mosque and could be recognized by its walls and pier bases visible among the debris. It is possible to recover only the ground plan; and this season's work has revealed at least three successive phases of building.

All construction was of rubble masonry with or without a facing of plaster. (I) Probably the original plan accords with that of the classical mosque of the Umayyad and 'Abbasid periods-an open court surrounded on each of the four sides by arcades which are greater in depth on the qibla, that is, the side facing the direction of Mecca. The Siraf mosque is a rectangular building measuring 44 m. on the qibla and facing side and 58 m. on the remaining two sides. The sanctuary consists of four aisles parallel with, and thirteen bays of equal width at right angles to, the qibla wall. These arcades are supported on 48 freestanding piers of varying shapes. On the remaining three sides were porticoes (riwaqs) of double arcades. On the north-east wall and slightly south of the main axis was a minaret. On the outside, a columned portico led up to an entrance in the south-east side. (2) A fifth aisle was added to the sanctuary by the construction of a new qibla wall and by replacing the earlier one by 14 piers.
10

142

JOURNAL

OF

PERSIAN

STUDIES

(3) Two buildings were successively introduced into the sanctuary without regard to the earlier disposition and therefore after the latter was abandoned. The Deep Sounding:this revealed an accumulation of occupation debris some 25 ft. thick. Fourfifths of this accumulation belongs to the period of Siraf's greatest prosperity. The remains of three superimposed buildings were noted; the last of these contained stucco capitals. Associated with the buildings was a large quantity of 9th and I oth century pottery, both Islamic and Far Eastern. The Islamic material included the coarse pottery with a blue-green glaze noted by Stein and tin glazed wares either plain or decorated with blue and green or green and brown. Among the Far Eastern pottery were fragments of stonewarejars and bowls with an olive green glaze and of porcelain and celadon vessels. Below the 9th and i oth century levels were a series of deposits lying on undisturbed sand. These contained none of the fine glazed wares found higher up. The only local glazed pottery was the coarse ware with a blue-green glaze, here almost exclusively green; and from the Far East only stoneware with an olive green glaze. Thus, although the earliest levels are clearly of the Islamic period they antedate the period of Siraf's greatest prosperity and are provisionally attributed to the late 8th and early 9th centuries. a The Houseor Warehouse: series of soundings were made in an area about 300 yards south-west of the mosque and on the sea coast which was thought to contain a single building complex, possibly a warehouse. At least three superimposedbuildings were discoveredbut their elucidation requiresfurther investigation. Kilns: these were noted in the north-westernarea of Siraf and clearly were part of a factory suburb close to the sea. The kilns were discovered where the sea has encroached on the cliff. One kiln was excavated and coarse pottery recovered. Much glass slag and waste vessels of the 9th and Ioth centuries nearby suggest the presence of a glass house and the area warrants extensive excavation for kilns and workshops. Plans: Future (I) Continuation of work on the mosque in order to recover the earliest building phase; further investigation of the warehouse and kilns. (2) A sounding in the area of the fort which would seem to have been the nucleus of the town. (3) Investigation and recording of the cemeteries cut in the steep slopes of the hills to the north of Siraf. (4) Area reconnaissanceto investigate related sites.

D. B. WHITEHOUSE.

Suse Au cours de l'hiver I965/1966, les travaux " Suse ont 6td poursuivis sur trois chantiers: sur la "Ville Royale ", dans le quartier des palais ach6m6nides et sur l'acropole. VilleRoyale.Dans le chantier stratigraphiquecommenc6 il y a vingt ans, sur une superficie d'un hectare, la Delegation a degag6, a la profondeurde pres de quinze metres, la quinzieme ville (P1.HIIa). En rasant les murs d'une vaste demeure d'un seigneur dlamite, qui avait et6 mise au jour 1'an passe dans notre niveau XIV, nous avons decouvert les archives de ce personnage, qui comprenaient plus
de cent documents, tous 6conomiques. Le professeur L. de Meyer eut le temps de prendre connaissance de ces textes qui demanderont encore une longue 6tude, mais d~s maintenant nous savons que le propridtaire de cette grande maison, qui comprenait pas moins de 51 chambres, s'appelait Temti-wartal, qu'il possddait de vastes terres que cultivaient ses hommes, que ses bergers surveillaient des milliers de tites de bdtail grand et petit, et qu'il 6tait en relations d'affaires meme avec des gens de Bahrein d'ois il recevait de l'argent. Un regu mentionne la livraison de I 7 mines d'argent. Les donations royales qui font partie de ces archives sont signdes et scellies par Kutir-Nahhunte I, ce qui place notre niveau XIV vers le XVIIIeXVIIe sidcle avant notre ere. Une observation d'ordre social et dconomique, concernant la socidtd 6lamite de Suse de cette 6poque, vient d' etre faite dans notre ville du niveau XV. Sous la grande maison de Temti-wartal du

SURVEY

OF

EXCAVATIONS

143

niveau XIV, nous avons mis au jour une sdrie de maisons sensiblement plus petites. Or, cette observation s'accorde parfaitement avec les indications qu'offraient les textes 6lamites trouves depuis des d6cades 'a Suse, et qui font savoir qu'au ddbut du IIe millenaire avant notre 're, en Elam la petite propridtd avait tendance 'adisparaitre en faveur de la grande. Temti-wartas racheta plusieurs maisons du quartier pour 6lever sur leur emplacement sa grande residence. Mais il ne put le faire d'emblke: la difficultd qu'il rencontra provenait du fait de la prdsence parmi les maisons de ce quartier, d'un petit sanctuaire, d'une " maison du culte ", qui finit a la longue par disparaitre sous la grande maison de Temti-wartav. ' Ce sanctuaire du niveau XV est le premier de cette 6poque trouv6 Suse, ainsi que le premier qui ait 6te mis au jour sur la " Ville Royale ". D'un plan trbs simple, il prdsente une antecella et une cella en enfilade et en longueur. Ce plan n'est pas de la M6sopotamie du Sud voisine; il s'apparente aux ' plans des temples decouverts a Gawra, et plus tard Assur, et laisse a penser qu'on se trouve probablement devant des traditions de l'architecture religieuse dlamite qui remonte au IMe millhnaire. Si la supposition est exacte, elle induit a penser que des rapports liaient les habitants 6lamites, probablement autochtones du sud-ouest de l'Iran, a ceux qui 6taient fixes dans la partie du nord-ouest de l'Iran, sur les contreforts et meme dans les chaines des Zagros. lors de la premiere phase de l'existence L'entree de la cella en longueur se faisait par le c6td otroit du sanctuaire, et par le c6td long lors de la seconde. Dans le fond de la cella et coll6 contre le mur (P1. IIb) 6troit, s' levait un podium portant un d6cor en terre crue en relief sur un fond noirci avec du bitume, qui representait une porte ferm6e de temple, flanquee sur les angles du podium de deux de'parts de vofites qui 6voquaient les niches voittes de la facade d'un temple (P1. IIc). Pres du podium et pratiquie dans d'6paisseur du mur long int6rieur, se trouvait une grande cheminee. On saisit l'importance de cette d6couverte qui est susceptible de projeter une lumi"re nouvelle sur l'origine du temple dans cette partie du Moyen Orient ancien des bords du Golfe Persique. Le professeur E. Heinrich exprima, il y a deji quelque temps, une hypothese comme quoi le temple de Mesopotamie d6rivait d'une habitation. Le podium prenait la place du chef de la maison et l'autel celui du foyer. Notre ddcouverte appuye et 6taie cette these puisque nous savons maintenant que le podium du petit sanctuaire se trouvait t l'emplacement r6serve au maitre de la maison. Les preuves sont nombreuses qui ont 6te mises aujour par nous: plusieurs maisons dlamites des diff6rents niveaux de notre fouille stratigraphique ont fait connaitre des salles oblongues dont le fond 6tait d6cor6 d'une niche peinte en vert, bleu ou rouge. Le sol devant cette niche 6tait dall6 de belles briques cuites dont une 6tait perc6e d'une ouverture dans laquelle pouvait passer une main. Sous cette brique perc6e &tait toujours enfouie une jarre qui, remplie d'eau, conservait celle-ci fraiche meme pendant les grandes chaleurs. Ce fond decor6 des salles 6tait la place du chef de la maison; cette place fut choisie dans le petit sanctuaire de Suse pour y 6lever le podium qui supportait, probablement, l'image sacree. La decouverte d'une " maison du culte " encastree parmi les maisons habitues, apporte une autre preuve comme quoi la grande metropole dlamite qu'6tait Suse ne possedait pas seulement un temenos(certainement sur l'acropole) avec de grands temples " imperiaux ". Il faut croire que chaque ' une divinit6 particulikrement v6neree. A quartier de cette ville devait avoir son sanctuaire dedi6 notre trbs vif regret, aucune brique inscrite n'est venue nous divoiler le nom du didicataire de notre " maison du culte ". Le secondchantierL Suse fut ouvert sur le mur d'enceinte des palais de Darius. Nous cherchons h y reconnaitre la grande porte d'entrie qui donnait aussi bien vers le palais que vers l'apadanaet les jardins royaux dont parle le livre d'Esther. Au cours de travaux prdliminaires, deux forts murs en briques cuites lides avec du bitume, contre lesquels s'appuyait un mur en briques crues, de pros de 17 mitres d' dpaisseur, viennent d'y etre reconnus. Le troisidme chantiera 6t6 poursuivi sur la surface de l'acropole, derriere le chgiteau de la Ddligation. nous avons touch6 les restes des 6curies des " lanciers du Bengale " dont une brigade avait occup6 LA,
9 Voir J. B. Youssifov, Elam. Histoiresocialeet dconomique. Thise de doctorat. Bacou 1965, P. 59 (en russe). Distribution privie.

144

JOURNAL

OF

PERSIAN

STUDIES

Suse pendant la guerre de 1914-1918 et dont l'dtat-major avec le g6ndral en tate occuperent le chateau pendant presque cinq ans. Sous les ruines des murs en briques crues, mont6s il y a un demi-siecle, furent mis au jour des fers a cheval, des boutons d'uniformes et meme le papier d'emballage d'un paquet de cigarettes indoues; puis nous avons touch 6 quelques ilots du niveau des ddbuts du protodynastique. Tout ce qui existait sur ce tell post6rieurement a cette date, a disparu au cours de pres d'un demi-siecle de travaux. Mais les installations de l'dpoque de Djemdet Nasr restent intactes sur une bonne superficie. Cette fouille prdhistorique, qui parait prometteuse par les bitiments a peine touch6s, est inscrite dans le programme de la prochaine campagne.
R. GHIRSHMAN.

Suleiman Takht-i
Nachdem in den vergangenen Jahren der sasanidische Feuertempel und die dazugehdrenden Gebaude freigelegt worden sind (Teher. Forsch. I; AA 1961, 28; 1962, 633; 1964, 1; 1965, IV, 619; ILN 16, Jan. 1965), wurde im Sommer 1966 an folgenden Plitzen gegraben: Zuweg vom n6rdlichen Haupttor zum Feuertempel, innere Umfassungsmauer, Hof E, Gebiet westlich neben dem Feuertempel, grosser Iwan westlich vom See. Auf der Innenseite des Tores wurde eine Torhalle teilweise freigelegt, neben der sich Reste einer alteren Lehmziegelanlage fanden. Von der Torhalle fhfirt rampenartig ein Aufweg zur zweiten, inneren Mauer. Diese hat halbrunde Bastionen auf der Aussenseite und enthalt einen gew6lbten Gang im Inneren. Sie umschliesst den Tempelbezirk auf der Nordseite des Sees U-f6rmig im Westen, Osten und Norden. Das Tor der inneren Mauer hatte, Thnlich wie das iussere Tor, eine zum Tempel hin offene Torhalle. Nord6stlich vom Feuertempel wurde ein Pfeilerhof (E) und die ihn umgebenden Portiken zur Hilfte freigelegt. Westlich vom Feuertempel, den Raum zwischen diesem und dem westlichen Fligel der inneren Mauer fillend, wurden spitsasanidische Gebaude aufgedeckt: ein offenbar zu Wohnzwecken dienender Komplex von Raumen, eine grosse, rechteckige Halle mit Nischen, ein Hof mit Podien, Becken und Feuerplatz und eine mit kleinen Rundbastionen besetzte Fassadenwand. In tieferen Schichten fanden sich Reste einer grossen Abortanlage. In Fussbodenaufftillungen und Mauerfundamenten verbaut waren Fragmente sasanidischer Stuckreliefs, darunter die Halbplastik eines galoppierenden Rindes, sowie Mtinzen (5.-7. Jhdt. n. Chr.), Tonbullen, Glasperlen und Bronzegerate, Nadeln, Ringe, Glocken und ein mit einer Palmette und Pferdekopf dekorierter Aufhanger. Unter dem bei frtiheren Grabungen untersuchten Komplex von Silen und Kuppelriumen (PB-PG) wurden weitere Teile eines im Vorjahr gefundenen alteren Gebaudes freigelegt. Es Thnelt dem sog. Feuertempel von Susa und dem altesten Zustand des Feuertempels von Kfih-i Khwaja. Das Gebaude ist mehrmals umgebaut worden, ehe es durch die daruberstehende Anlage ersetzt wurde. Einem solchen Umbau geh6rt ein aus Ziegeln gemauerter und sorgfaltig verputzter Tisch mit rechteckiger Platte und einwirts geschwungenem Unterbau an, der in einem der Seitenraume neben dem zentralen 4-Stitzensaal gefunden wurde. Die Untersuchungen am grossen Westiwan ergaben, dass hinter der Iwanhalle, nur durch einen relativ engen Durchgang mit ihr verbunden, ein schmaler, quergelagerter Raum mit breitem Tor nach aussen liegt. Er wird seitlich von zwei kleinen Kuppelriumen mit kreuzformig angeordneten Nischen flankiert, die ebenfalls durch Tore nach aussen ge6ffnet sind. Uber den bis zu 5 m hoch erhaltenen Quadermauern der sasanidischen Anlage liegen die Reste zweier oktogonaler Kuppelsale mit Nischenraumen an allen Seiten und eines grossen, ungefahr quadratischen Mittelraumes, die beim Wiederaufbau des Palastes durch Abaqa Khan im letzten Drittel des 13. Jhdts. errichtet wurden. Von der prachtvollen Ausstattung der Raume fanden sich viele Fragmente von Stuckreliefs, Wandmalereien, Luster- und Lajvardina-Wandplatten. D. HUFF. Tal-i Iblis In the fall of 1966 an international expedition, organized by the Illinois State Museum and supported by the National Science Foundation, began uncovering an ancient site in the south-eastern part

SURVEY

OF

EXCAVATIONS

145

of the Persian Plateau. This is Tal-i Iblis or Devil's Mound, in the Bard Sir Valley 7,000 feet above sea level, located 80 kilometres south-west of the provincial capital of Kerman. The excavations were exploratory and partly to verify a previous discovery of late 5th Millennium smelting ofcopper from ore.'0 Besides the writer, archaeologists were David Chase, Montgomery Museum of Fine Arts; Geza and Irene Fehervari, University of London; Mohammed Sarraf, Archaeological Service of Iran; and Daniel Evett, University of Chicago. Dorreh Mirheydar, geographer from the University of Tehran spent some days with us as did Gary and Valerie Hume of the University of Minnesota-Bdmpfir Expedition. There was also a metallurgical team headed by Cyril S. Smith, of the Massachusetts Institute of Technology, whose report follows this one. With these scientists on hand, the evidence for early copper smelting was abundantly confirmed, but after the metallurgists had departed new evidence was found which may again push back the date for smelting. Excavations were carried on in several sectors of the mound, which had been partially ruined by people digging the rich earth for fertilizer. In one area was a large pre-Islamic pottery kiln of mud brick, beautifully preserved, with a firing chamber, clay grate, a separate firebox below, and a sophisticated system of vents. Charcoal samples will help date this construction. In another area was a well-preserved round tower of baked and unbaked mould-made bricks. Fragments of pottery reminiscent of Sialk IV, including part of a bevelled rim bowl, suggest an early 3rd Millennium date and a time of strong interaction and diffusion of ideas over a large area including Kerman, Sialk, Susa, and Jemdet Nasr sites in Mesopotamia. A 5 m. square was dug through a dumping area of the preceding Ali Abad period, which has radiocarbon beginning and ending dates of 3645?59 B.c. (P-928) and 2869?57 B.C. (P-929). Here also were bevelled rim bowl fragments attesting still earlier connections with Mesopotamia. Ali Abad pottery was painted in a style difficult to place but with some resemblance to Shahi Tump in southern Baluchistan. A considerable amount is polychrome and most is hand-made. One undecorated variety is wheel-made with string-cut bases. A somewhat earlier period with a radiocarbon date of 3792-16o B.C. (P-927) produced a subterranean furnace lined with apparently mould-made bricks and containing a deposit of white powder, analysis of which may indicate the purpose of this construction. Painting was used sparingly on the pottery from this level; most sherds have a kind of scratched decoration. m. Underlying this was a dump 6o cm. thick and at least 100oo long. From this came most of our evidence for early metallurgy-hundreds of fragments of small earthenware crucibles showing copper stain and dross. Yet this was not an industrial dump but the refuse of domestic activities, with quantities of charcoal, animal bones and pottery fragments. The dump is assigned to the Iblis Period and radiocarbon dates-4o91 ?74 B.c. (P-925) at the bottom and (P-926A) at the topB.c. it accumulated in a short time. In this area we found no4083?-72 furnaces for reducing copper. or suggest pits Nowhere, moreover, have we found any areas or dumps pertaining solely to metallurgy. There is some evidence, on the other hand, of metallurgical activities going hand in hand with domestic ones. At the bottom of the mound we began finding mud-brick walls which were earlier than the dump with the crucibles. These belonged to carefully planned houses. Since most of the upper levels of the site had been carried away by fertilizer diggers, the walls could be exposed without much difficulty. Altogether we exposed large parts of four houses with a total of 90 rooms. Rooms were not built around a court; instead there was an outer perimeter of living rooms around a central core of storage and special purpose rooms. The mud-brick walls were plastered, the doorways carefully made; there were buttresses, corridors, and special rooms for living, food preparation and storage. Walls and floors of living rooms were painted red and floors were covered with mats. Roofing was by poles extended from wall to wall and plastered with mud. We await radiocarbon dates from two tentatively defined periods, both of which, in copper, bone, stone, and clay artifacts, including animal figurines, seem to have their closest connections with Sialk II. The Bard Sir period is clearly older than the Iblis Period dump and
o10

Joseph R. Caldwell and Sadegh Malek Shahmirzadi, "Tal-iIblis: the Kerman Range and the Beginnings of Smelting", Illinois State MuseumPreliminary Number7, Illinois State Reports Museum Society, Springfield, 1966. Ralph C. Dougherty and

Joseph R. Caldwell, "Evidence of Early Pyrometallurgy in the Kerman Range in Iran", Science,August 26, I966, Vol. 153, No. 3729, PP. 984-985-

146

JOURNAL

OF

PERSIAN

STUDIES

crucibles. Most of Bard Sir pottery is chaff tempered coarse ware not unlike the earliest coarse ware of Hassuna, but associated with a handsome buff ware reminiscent of the older levels of Tal-i Gap. Two houses and another dump can be assigned to this period: older still are two houses we associate with Lalehzar Period, again awaiting C'4determinations.Lalehzar pottery is about 99 per cent coarse ware, with painted buff ware sherds occurring occasionally. The houses of both these earlier periods have yielded copper artifactsand fragmentsof ore, and the Bard Sir dump contained a few crucible fragments. Apparently built into the ruins of a building but older than the Iblis Period dump was a small copper-smelting pit containing a few bricky fragments perhaps derived from a ruined superstructure.In the bottom were innumerable bits of copper oxide and one crucible fragment. In the corner of a well-preservedroom of a house of the Lalehzar Period was a clay box with a round bottom 6 cm. above a fireplace. In the bottom of the box was a mass of fused white friable metal lighter than lead. We await a metallurgical analysis, but I suspect it will be difficult to explain this construction as other than for sintering or cupellation. In the same room was a small deposit of grain (yet to be identified) and several upright pottery vessels. This evidence and the mixing in the dumps of crucible fragmentsand domestic refuse raises the question whether our earliest metallurgy is not a home or "cottage" industry. Additional excavation should bring a clear answer to this. Various writers on early metallurgy stress the idea that the metallurgist must be a professional. Childe writes that "Probably from the first metallurgy was a craft as well as a technique . . . The operations of mining, smelting and casting are too elaborate and demand too continuous attention to be normally conducted in the intervals of tilling fields and minding cattle"." The Iblis evidence may be telling us of an earlier home and/or part-time industrywhen basic processeswere being developed at a time before full professionalismwas achieved.
JOSEPH R. CALDWELL.

Tal-i Iblis: Metallurgical Archaeology Planned in connection with ProfessorJosephR. Caldwell'sexcavation at Tal-i Iblis, a small expedition travelled through the Kerman Mountains and along the fringes of the Dasht-i-Kavir and Dasht-iLilt, seeking information on what might be called the ecology of early metallurgy in Iran. It was financed by the United States National Science Foundation. The group consisted of Cyril Stanley Smith, metallurgistfrom the MassachusettsInstitute of Technology; Theodore A. Wertime, consultant Akademie to the Smithsonian ,Institution; Radomir Pleiner, archaeologist from the Ceskoslovenska v'd Archeologick- Ustav in Prague; and Gholam-Hossein Vossouqzadeh, a geologist loaned by the mining department of the Iranian Ministry of Economy. The Ministry proved extremely helpful in providing information on ancient sites, which have been sought for clues to ore deposits. The route followed was from Tehran through Kashan to Nain, then east through the important mining centres of Anarak and Nachklak, continuing acrossthe desertthrough Khur and Rabat to Tabas, then south via Naiband to Kerman and Mashis (to join ProfessorCaldwell's excavation for a few days) and return via Yazd and the iron workings at Haneshk. The (Sasanian?) gold mines at Muteh yielded rotary grindstones and a ceramic bowl, apparently used for ore washing. The old workingsat all mines visited had been started at outcropping veins of ore which were followed with all their irregularitiesuntil work was stopped by the termination of the vein,
by the ore becoming too hard to mine, or when ground water level was reached, or the oxidized ore yielded to sulphide which was not worked. Miners' lamps and other pottery found in the mines were of undatable types, but wooden props from two mines will be C'4-dated. Furnace ruins (usually incorporating refractory siliceous rocks not native to the site) and numerous piles of slag (containing charcoal) gave evidence of very extensive lead smelting at many places along the route. These were usually located in relatively flat ground I or 2 kilometres from the base of the mountains at sites supposedly selected for combined access to ore, fuel, and food, though there is little enough of any of these in the desert today. At Nachklak and near Tarz, residual lumps of fused litharge provided evidence for the extraction of
"

in V. Gordon Childe, What Happened History, Pelican,

1954.

SURVEY

OF

EXCAVATIONS

147

silver from the smelted lead, though no remains of cupellation furnaces were found. Instead there were enormous numbers of broken clay rods, slightly tapering and rounded at each end. When intact, these would have been about 50 cm. long and 3 to cm. diameter. They are unlike any known metal3"5 lurgical devices or tools, and clues to their function are being sought by laboratory examination of them. With the advice and assistanceof Mr. M. B. Saeedee, who has had several decades' experience in lead and copper smelting by both traditional and modern methods, smelting experiments were conducted at Tal-i Iblis in a reconstructed blast furnace and a hearth using goatskin bellows. Material found by Caldwell in level Iat Tal-i Iblis (Science, 1966, 153, 984-85) leaves no doubt that a large copper took place there around 4,000 B.C.The curious shape of the crucible fragments, the smelting operation copious residuesof charcoal, the sparsity of slag and the lack of any furnace debris all indicate that the smelting practice was quite different from later standard methods.
C. S. SMITH.

TepesKalwali and War Kabud A brief report of the excavations at Tepe Kalwali and Tepe War Kabud, under the direction of ProfessorL. Vanden Berghe and Madame Yolande Maleki, is given in the catalogue of the exhibition " Tresors de l'Ancien Iran ", shown at the Musde Rath, Geneva, from June 8th to September 25th Cemeteriesof the firstmillennium B.C. were found. 1966. TurengTipde' L'essentiel du travail de la mission fran?aise de Tureng Tepe a consiste en 1965 a approfondiret mener jusqu'a son terme le sondage principal situd sur le plus occidental des tdpes qui composent le site. Entre les sols 19 et 26 nous avons digag6 les restes de diffirents batiments de brique crue; les maisons des deux niveaux inf6rieurs 6taient 6difides sur d' pais remblais d'argile destinds a remedier aux infiltrations des eaux souterraines: de fait celles-ci nous ont obliges a interrompre notre travail aussit6t apres le sol 26. De plus un mur entourant le site primitif fut decouvert au niveau 25. Ces diff6rentssols constituent notre periode IIA, caractdrisdepar la prdsence de vases peints a fond rouge et de vases rouges lisses dont les formesrappellent celles de la c6ramique grise lissde, elle-meme ddja fort abondante. En revanche le decor liss6 n'existe pas encore durant cette periode. D'autres sondages ont 6tdentrepris, destinds surtout a explorer les couches les plus recentes-et les plus obscures--du site. Dans l'un d'eux nous avons pu observer une excellente stratigraphie de la ea periode IIIC, correspondanta Hissar IIIC et succedant sans solution de continuit6 la periode prdc'dente; quelques tessons peints manifestementimportes ont it6 trouves a ce niveau. Un autre sondage devait preciser la position stratigraphiquedu four de potier reperd en 1963, et qui appartient d6finitivement a la fin de la p6riode IIIB; le four lui-meme a 6td fouill6. II 6tait en grande partie souterrain, et la sole se trouvait au niveau du plancher; seule d6passait la coupole, dont ne subsistaientque des traces; un puits d'acces 16gerementoblique penetrait jusqu'au fond de la chambre de chauffe et s'ouvrait au niveau du plancher. De nombreux fragments de figurines ont etd trouvis dans les couches correspondantes. Dans le mime sondage nous avons d6couvert a proximit6 de la surface les vestiges d'un important
edifice que des tessons permettent de situer vers le XIIIe si~cle de notre are.

J. DESHAYES.
Choga Mish The site of Choga Mish is located about twenty-five kilometers south-east of Dezful on the east side of the Ab-d-Dez River. It is one of the largest sites on the edge of the Susiana plain and only a few miles from the foothills of the mountains. The tepe was known to the French excavators at Susa and appears on their topographic maps of the region,'2 but no excavations were undertaken by them on it. Dr. Robert M. Adams, during his survey of the ancient irrigation systems of northern Khuzestan in zzM.D.P.
XXIX (1943), Fig. lO6, p. 140, and ibid., XXX (1947), Fig. I, p. 121.

148

JOURNAL

OF

PERSIAN

STUDIES

1960, noticed the abundance of Protoliterate pottery sherds on the surface in addition to sherds of prehistoric painted wares. Excavation of this site seemed very promising, first as a source of new evidence concerning the rise of the Protoliterate civilization in south-west Iran (in which the writer has been interested for many years) and second as a single source for a stratified sequence of prehistoric Susiana cultures, one which could serve as comparison for that so ingeniously built up by le Breton and Vanden Berghe from several sites. Consequently, the writer proposed that the excavation of this site be included in the Oriental Institute's programme. After obtaining the necessary permit from the Iranian Government, small-scale excavations began on November 6th 1961, and lasted through December 18th of the same year. Two subsequent campaigns were carried out, the second from February 5th to June 27th 1963; the third from November 5th 1965 to April 5th 1966. Through all these campaigns Professor Helene J. Kantor participated with the writer in conducting the excavations. The site itself, roughly rectangular in shape, is oriented in length almost exactly north-south. Ancient remains can be located immediately below the surface in an area about 300 m. wide and 600 m. long, i.e. approximately 45 acres. However, there is little doubt that the surrounding fields, the surface soil of which consists largely of debris washed down from the higher parts of the site, cover similar remains and that the settlement must have been considerably larger, probably 60-70 acres. Topographically, Choga Mish consists of two parts. The northern part, the "Main Mound", comprises about one-fifth of the total and consists of a series of steep ridges and gullies, with the highest point more than 25 m. above the surrounding plain. The southern and considerably larger area is much less rugged and features five less prominent knolls reaching only about 8-10 m. above the surrounding fields. This area is known as the "Terrace ". The whole site is thickly covered with potsherds. In order to test the distribution and dates of the settlements on the site, more than thirty trenches, soundings, and pits were dug at strategic points on it. In most of these soundings virgin soil has been reached. During the third season we extended the excavations in the vicinity of certain trenches to form much larger areas. In summary, the results are as follows: at every point where we excavated, we found, usually immediately below the surface, building remains of the Protoliterate period. As a rule, these remains were denuded to such an extent that it was impossible to recover more than vestiges of the original buildings. However, there is sufficient evidence to show that the Protoliterate occupation lasted in most places for three or four consecutive rebuildings. There are also indications that even earlier buildings of the same period must have existed on the site because some of the kiln-fired bricks of the lowest levels were in secondary use. The architectural remains consist of private dwellings as well as of more imposing buildings. Worth mentioning are a circular building about eleven metres in diameter in Trench V; traces of a solid mud brick platform with projecting "towers" in the north-western terrace; and monumental walls over 4 m. thick in N9 which are capped by even thicker walls (more than 11 m.) of the Elamite period. The remains of the houses were largely disturbed by numerous "refuse pits ", varying in size and depth, which cut through many walls and floors. These pits can also be dated to the Protoliterate period since they contained no later remains. An abundance of drains constructed of kiln-baked bricks is a typical feature of the Protoliterate levels. They occur at different depths and the upper ones apparently belonged to buildings which have been completely denuded. All the indications are that Choga Mish must have been a city of considerable size during the Protoliterate period. Its fairly regular rectangular form suggests that it may have had a town enclosure. However, we have not yet located the town wall nor the actual limits of the city, since the outlying remains are probably deep below the washed-down material from the higher parts of the town of the Protoliterate and later periods. However, certain topographical features suggest the probable location of the town walls. By far the largest bulk of finds of the Protoliterate period consists of pottery. Pottery of this period was unearthed in such overwhelming quantities that we found it necessary to devise a special method

(Photo: J. Powell, Rome) Pl. I. Assyrianstele of SargonII, found nearGodinTepe.

ire. R. Pl. IIa. Suse.NiveauXV. DebutduIHemillinaireavantnotre Cliche' Ghirshman.

Pl. IIb. Suse. Sanctuaire. Cella. Cliche' Ghirshman. R.

Pl. IIc. Suse. Sanctuaire. Podium.ClichiR. Ghirshman.

SURVEY

OF

EXCAVATIONS

149

for recording it. This method aims at a total quantitative as well as qualitative analytical record. While the pottery has many parallels among previous finds, especially at Susa and Warka, large numbers of new types were discovered. The small finds are too numerous and diversified to be discussed within the limits of this report. Undoubtedly the single most significant class of such finds, and the one that may contribute more than any other group to our better understanding of Protoliterate civilization, is the great number of cylinder seal impressions. These vary from nearly complete impressions on large conical sealings or on hollow balls to minute fragments recovered by means of sifting. To piece together such fragments into complete or coherent seal designs and to draw them for publication is a difficult and time-consuming task. From the prehistoric periods we have obtained stratified series of pottery from deep trenches in various parts of the site. Trench II, consisting of three large "steps" which together represent an accumulation of about 21 m. of prehistoric remains, gave us the first of such stratified series. In several other trenches virgin soil has been reached at depths varying from about 5 to 11 m. The analysis of the masses of stratified prehistoric pottery will take time but some important results are already clear: one is that the entire area was not uniformly settled during all prehistoric phases; the other is the fact that some of the prehistoric settlements appear to antedate any phase of the Susiana sequence as established by le Breton and Vanden Berghe. Whether this earliest habitation is of a single period or contains several phases is as yet uncertain. Finally, remains of fairly elaborate architecture discovered in Trench XXV seem to be related to this earliest period of settlement.

P. DELOUGAZ.

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 El 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 ffir Orientforschung 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 Inscriptiorum Arabicorum Dl6fgation Archeologique frangaise en Afghanistan, memoires Encyclopaedia of Islam Eurasia Septentrionalis Antiqua E. E. Herzfeld, Iran in the Ancient East (194I) 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 Mimoires de la D6l1gation en Perse Museum Journal, Philadelphia Oriental Institute, Chicago, Communications Oriental Institute, Publications Orientalia Suecana Praehistorische Zeitschrift Revue d'Assyriologie Royal Central Asian Journal
Revue des tEtudes 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 Turk Tarih, Arkeologya ve Etnografya Dergisi Die Welt des Orients Wissenschaftliche Verbffentlichungen der Deutschen Orientgesellschaft Zeitschrift fur Assyriologie Zeitschrift der Deutschen Morgenlandischen Gesellschaft

FORM OF COVENANT

of ......................................................................................................................................................................................... hereby covenant with the British Institute of PersianStudies that for a period of seven years from the date of this Deed or during my lifetime (whichever shall be the shorter period) I will pay to the said Society from my general fund of taxed income such an amount annuallyas after deduction of income tax at the Standard Rate yields the net sum of. .............................................................................................................................................................................. of. The first paym is to be made on ent Date d t his the.........................................................................................................

day of 19 ................................................................. ................................................................. .......... Signed, sealed and delivered by the said: In the presence of:

NOTICE It is hoped that as many as possible of those applying for Membership of the Institute will undertake a seven-year covenant, thus making a further substantial contribution to the resources of the Institute.

BANKER'SORDER To..........BANTo BANK...BAN ..................................................................................................................................................................... ADDRESSADDRESS ..................................................................................................................................................................... Please pay to the British Bank of the Middle East, 7 King William Street, London E.C.4,for the credit of the British Institute of PersianStudies, the sum of ? : now and a similar sum on Ist Januaryof each year until further notice/for the next six years.* Signe d .................................................................................................................................. Date ..........................................................................................
* Delete whichever does not apply.

".

\
.....

250000V

'
R B l JAN0

-.o

e 40O~iinwtii
!i !li 4oo.Kilometres I i

.S

"' ome...res.

S.

R.

ARBAIJANS. zayeh,
L.'
e

Qv "
e

<:44.
8NDARAN

.Gorgan
.DamghnL

*Meshed
n

" .."

TEHRAN H.adaLK H UKHRAS A N ""

S .Kermansha

.Qum
.Kashiaw

Ss

Sus a.

AN

Isfahan
's

A
Yazc

N
\ sisTAN~ OA

__'

Ahwa -

: ?q

sr

''
/

TasargadaeKermaiL
...epo, E R M A

ibShiraz ?

..4
SA

ARABIA
J.A.

D'

"Bandar'Abbas

di

Ai

PRINTED 109

IN GREAT KINGSWAY

BRITAIN LONDON

BY WC2

LTD HEADLEY BROTHERS KENT AND ASHFORD