ted wee
Tere] eieeebeay" reas
12 SbF Obese
Pe BA cea eae te]
3.
Te ET pe Se ee Pere tz
eg pn tere
saith:
t
Rr 4
t5. BAG ia aes NA Py
yumm ee hs ae
n
Pa et fae
orty $4 &
ae peGaty s2i¢i2
;¢
ts
ns
: ese ee Ries eet Benny Sterne
is} a
hee Fon tohcoes hee A We gp ane ae wots
‘Mew gie) =
ae
nape
*
Beat ba oir Se
Be
fa ted eee ie ts ig be letat abe ve'seiie Lat
«
ale.
slain oie
Pec L eorae
Sp ia News Tae:
feisists
sf
Spiceas
PDA ID Bey
ANCIENT KHOTAN
DETAILED REPORT OF ARCHAEOLOGICAL EXPLORATIONS IN CHINESE TURKESTAN
CARRIED OUT AND DESCRIBED UNDER THE ORDERS OF H.M. INDIAN GOVERNMENT BY
M. AUREL STEIN
INDIAN EDUCATIONAL SERVICE
VOL. «hE
WITH DESCRIPTIVE LIST OF ANTIQUES BY F. H. ANDREWS SEVENTY-TWO ILLUSTRATIONS IN THE TEXT, AND APPENDICES BY L. D. BARNETT, S. W. BUSHELL, E. CHAVANNES A. H. CHURCH, A. H. FRANCKE, L. DE LOCZY D. S. MARGOLIOUTH, E. J. RAPSON, F. W. THOMAS
OXFORD AT THE CLARENDON PRESS
ae 7
, EDINBURGH
Sew? ninen | end
i!
‘TO THE UNIVERSITY OF OXFORD
Z zZ a
} : sl a & Ss . nf . a z wo 4 S Za
HENRY FROWDE, M.A.
‘
PUBLISHER
TO THE MEMORY OF
COLONEL SIR HENRY YULE BiB fOBiyo Reid.
THE GREAT ELUCIDATOR OF EARLY TRAVEL AND A PIONEER IN THE HISTORICAL GEOGRAPHY OF CENTRAL ASIA
THIS ACCOUNT OF THE ANTIQUARIAN RESULTS OF MY JOURNEY ON WHICH HIS WORKS WERE MY BEST COMPANIONS IS DEDICATED WITH DEEP RESPECT AND ADMIRATION
FOR THE SCHOLAR, THE WRITER, AND THE MAN
iE
“YsOMaM 3aT
“AqUY YAVGH Ale
a if 3 “g
cies
+: \SSVART YARSA FO BOTAG!
=
at BaF:
—s Ha
YVHYAROGORD
FRAT BO
TREUGCLDA
Ta RUDE Ye. GO a
*
WOIVAIMOD 4
4
VOITAAIMGL Gi ~
INTRODUCTION
Tue present work is intended to describe in detail the archaeological results of the explorations which I carried out in 1900-01, under the orders of the Government of India, in the southern portion of Chinese Turkestan, and particularly in the region of Khotan. It has been my endeavour to make this account as exact and exhaustive as seemed justified by the importance of the discoveries which rewarded my journey, and by the interest of the observations then collected regarding the early geography and civilization of the country. Considering the length to which the work has grown, and also that its conclusion coincides with the commence- ment of a fresh and absorbing task in the same field, I may be allowed to restrict my introductory remarks to a brief record of the circumstances which led me to form my plan and enabled me to carry it into execution, and to some needful explanations concerning the conditions in which the results were finally elaborated.
Ever since 1891, when the famous birch-bark codex acquired by Colonel Bower from Kucha became known to Indologists, my eyes had been turned towards Eastern Turkestan as a field for archaeological enterprise. But the difficulties with which then, and for long years thereafter, I had to contend, in order to assure the completion of my labours on Kalhana’s Chronicle of Kashmir, while the leisure left to me by onerous administrative duties was of the scantiest, precluded any actual steps in that direction. In the spring of 1897, when the conclusion of those labours was drawing within sight, | became acquainted, through personal communications from the late Professor Biihler, with the important find of ancient birch-bark leaves, containing a Buddhist text in early Prakrit and in Kharosthi writing, of which M. Dutreuil de Rhins had acquired a portion during his stay at Khotan. These remarkable fragments, of which others had found their way to St. Petersburg, were on their publication by M. Senart at once recognized as the oldest Indian MS. then known, and they decided me to fix upon Khotan as my goal. |
Other acquisitions from the same region, which had reached Calcutta, helped to assure me still further of the importance of this particular field. Dr. A. F. R. Hoernle, c.1.5, to whose long-continued efforts we owe the decipherment and elucidation of those first discoveries of ancient Sanskrit texts from Kucha, had induced the Government of India to issue instructions to its political representatives in Kashgar, Ladak and Kashmir for the acquisition of antiques from Chinese Turkestan. The most notable additions which the collection thus formed in Calcutta under Dr. Hoernle’s charge received during the years 1895-7, consisted of MSS. and other remains which were said to have been obtained by native ‘treasure-seekers’ from sites in the Khotan oasis and adjoining parts of the Taklamakan desert. But no reliable information was ever forthcoming as to the exact origin of these finds, or the character of the sites which were supposed to have furnished them. The many questions of historical interest which from the absence of such data had to be left unanswered—and the doubts which soon arose as to the genuineness of a large portion of these acquisitions, consisting of MSS. and block-prints in
vi INTRODUCTION
a variety of ‘unknown characters’—were the best proof of the urgent need for proper explorations on the spot.
I had discussed the plan of such explorations with Professor Biihler before I left Europe in 1897, and often have I thought since of the words with which that lamented scholar, during the last hour I was destined to spend in his ever stimulating company, expressed his keen interest in my plan as well as his appreciation of its practical difficulties. Not the least of these was the question how to secure the needful freedom and means. The exacting duties of the post then held by me under the Panjab University did not allow me to take preparatory steps towards the execution of my plan until the summer of 1898, when, after having obtained in Kashmir the necessary practical data, I asked Dr. Hoernle to help in securing for my project the support of the Indian Government. I owe it largely to Dr. Hoernle’s generous interest in my proposals that the formal application which I submitted in August, 1898, met with favourable consideration. Regard for the exigencies of the post then held by me had obliged me to restrict the first scheme to a journey of only six months, to be undertaken in 1899. The total cost was estimated at Rs. 6,800, and this sum the Governments of India and the Panjab agreed to provide by joint contributions of two-thirds and one-third, respectively.
Final sanction had scarcely reached me in the spring of 1899 when my appointment to the Indian Educational Service as Principal of the Calcutta Madrasah necessitated a post- ponement of the journey, and a reconsideration of the official arrangements concerning it. Fortunately the Government of Bengal, under the late Sir Jonn Woopsurn, a zealous friend of Oriental research, proved as ready to recognize the importance of my proposed explorations as had Sir Mackwortu Younc’s administration in the Panjab. The changed conditions of my appointment permitted an expansion of the time limit and also of the scope of my journey. The modified scheme to which the Government of India accorded their final sanction in July, 1899, provided for my deputation to Chinese Turkestan for a period of one year from June, 1900. A grant of Rs. 11,000 (£733) was placed at my disposal to meet the estimated cost of the expedition, the contribution of the Government of India being increased to Rs. 6,500, while the remaining sum, as far as it was not covered by the grant which the Panjab Govern- ment kindly allowed to stand, was provided from provincial funds of Bengal.
It affords me gratification that, in spite of the difficulties arising from great distances and physical obstacles, and of the uncertainties inseparable from an enterprise in a new field, the whole of my programme was accomplished strictly within the sanctioned estimates of time and expense. I should probably not have been able to keep the expenses within the relatively modest limits here indicated had not the Survey of India Department offered its assistance for an important adjunct of my researches. The necessity of fixing accurately the position of ancient sites and generally elucidating the historical geography of the country brought surveys of the regions visited into close connexion with my immediate task. But I also realized from the first that only by utilizing every opportunity the journey might offer for geographical work of a more general character, could I gain that practical familiarity with the physical conditions of the country which is needed for the right interpretation of its past.
Colonel St. GrorGE GorRE, R.E., C.S.L, late Surveyor General of India, readily offered for this object the aid of his Department. He kindly agreed to depute with me one of its experienced native surveyors, Babi’ (now Rai Sahib) Ram Singh, and to provide the necessary equipment of surveying instruments, together with a grant of Rs. 2,000 (£133) intended to cover the additional expenses of survey operations. With the assistance of the Surveyor, to whose excellent services I have had occasion to refer elsewhere, it became possible to carry on a continuous system of
INTRODUCTION Vil
surveys by plane table, astronomical observations, and triangulation during the whole of my travels. The results of these surveys, which in the mountains I supplemented by photogrammetric survey work of my own, and to the direction and supervision of which I devoted a good deal of time and attention, will be found illustrated by the ‘Map of the territory of Khotan and adjoining regions’, on the scale of 8 miles to 1 inch, which accompanies this Report. Its use will, I hope, considerably facilitate a correct comprehension of the numerous questions, equally interesting for the historical student and the geographer, which I have had to discuss in relation to the ancient sites now abandoned to the desert.
For the generous consideration and aid which Lord Curzon’s Government and the local administrations of Bengal and the Panjab accorded to me, I wish to record here my deep sense of gratitude. To it I owed the long sought for chance of serving the interests of Oriental research in a new field, and with that measure of freedom which permits full concentration of scholarly efforts. But grateful, too, I must feel to Fate for having allowed me to use this chance to the full, and for having rewarded my labours on difficult ground with results which competent fellow-scholars have recognized as an ample return for the means and facilities so liberally granted to me. Of the serious difficulties and risks which had to be faced at various stages of my journey and which might easily have thwarted my aims, there is no need to speak here in detail. A perusal of my Personal Narrative! will suffice to show the character and extent of the obstacles which had to be reckoned with during my work in the Taklamakan.
I could not have overcome them successfully without the active co-operation of the Chinese administrators of the districts upon which I depended for whatever was needed during the winter campaign in the desert. In the Ambans Pan Darin and Huang-Daloi, then in charge of Khotan and Keriya, I was fortunate to find reliable friends, ever ready to help me with everything in their power. I look back with all the more gratitude to their unfailing attention and kindness, as it was shown at a time when the Empire was in conflict with the European powers, and as neither honours nor material rewards were to be expected for it. The true historical sense innate in educated Chinese, and their legendary remembrance of Hsiian-tsang, the great ‘monk of the T’ang period’ (Z“ang séng) whom I claimed as my guide, proved helpful in explaining the objects of my explorations. But the sympathetic attitude which the provincial administration from the first showed towards my work, must mainly be attributed to the effective support given to me by Mr. G. Macartney, ¢.1.£., the representative of the Indian Government at Kashgar. Apart from the advantages which his personal influence secured to me, I derived manifold benefits from the practical help and advice with which this accomplished friend favoured me during my stay at Kashgar. For these, and the watchful care with which he followed my explorations from afar, I must record here the expression of my lasting gratitude.
The value of the scientific results of my explorations may be left to be judged from the record presented in these volumes. Yet I feel I am justified in referring to the generous acknowledgement which those results have already received from fellow Orientalists. Their opinion, whether recorded in the formal resolution of the XIIIth International Congress of Orientalists 2, or expressed in reviews and other notices by such distinguished scholars as MM. Foucher, Sylvain Lévi, Senart, and others, have been a great encouragement to me and have helped me to bear the strain entailed by the elaboration of the results.
* Sand-buried Ruins of Khotan, Personal Narrative of a London, T. Fisher Unwin, 1903; second edition, Hurst and journey of archaeological and geographical exploration in Blackett, 1904. Chinese Turkestan. By M. Aurel Stein, First edition, > See Ruins of Khotan, p. xiii.
Vili INTRODUCTION
The difficulties to which I refer are accounted for largely by the extent of the archaeological operations carried out, the great variety of the finds and observations made, and not less by the scope and character which I felt obliged to give to my task of recording them. The extent of the field-work is sufficiently illustrated by the fact that, taking only the territory of ancient Khotan, the old sites and remains which I surveyed in detail or excavated are scattered over an area which in a direct line stretches for more than 300 miles from west to east. The wide range of the antiquarian remains is equally striking. Dating back to periods which may be estimated to extend over at least eleven centuries, they show remarkable diversity in nature and character. Among the sites surveyed there are represented the curious wind-eroded ‘Tatis’, typical of the great physical changes to which the soil of the tracts bordering on the desert is subject; ‘culture- strata’ left behind by thickly inhabited ancient settlements and deeply buried below the loess accumulations of centuries ; and in the area now overrun by the moving sands of the Taklamakan, sites with structural ruins of all kinds and in all stages of preservation, The remains discovered of temples, fortified posts, Sarais, private dwellings, &c., with the relics of ancient art-work and industrial products found in them, have been so plentiful and multifarious as to enable us to study most aspects of the early civilization which once flourished there.
The very abundance and novelty of the archaeological materials brought to light necessarily made the task of correctly recording and interpreting them one of considerable difficulty. But what greatly added to the difficulty—as it undoubtedly did also to the fascination and scientific interest of the work—was the remarkable diversity of cultural influences thus revealed in the arts of ancient Khotan. That the Buddhism which we knew from Chinese records to have been early established in the country had been imported, directly or indirectly, from India was, of course, always certain. But there was little to prepare us for such overwhelming evidence as the wealth of Kharosthi documents on wood and leather discovered at the Niya Site has furnished of the large place which Indian language and culture must have occupied in the administration and daily life of this region during the early centuries of our era®. That Sanskrit Buddhist literature was studied in Khotan down to the end of the eighth century a.p. has been proved beyond all doubt by the texts in Brahmi script which I excavated in the ruined shrines and monastic dwellings of Dandan-Uiliq and Endere. At the same time, the large series of ancient stucco sculptures, frescoes, painted tablets, and-relief carvings in wood, from ruins the dates of which, as determined by exact chronological evidence, range from the third to the eighth century, has conclusively demonstrated that the Graeco-Buddhist art of the extreme north-west of India found a new home in Khotan and enjoyed a long-continued local development.
The territory of Khotan had, like the rest of the Tarim basin, been under Chinese supremacy for considerable periods both under the Han and the T'ang dynasties. The interesting finds of Chinese records on wood and paper, Chinese sgraffiti, coins, articles of industrial art, &c., at the sites excavated by me have made it abundantly clear that Chinese political control was effective, and that the influence of Chinese civilization, too, must have strongly asserted itself. The fact that Khotan occupied an important position on the ancient trade route from China to the Oxus basin no doubt helped to increase this influence. To the same connecting link we must ascribe those early relations with the distant West which the remarkable series of classical seals found impressed in clay on a number of the ancient Kharosthi documents from the Niya Site attests in so striking a fashion*. The imitation of Persian art, of which we find unmistakable indications five centuries later in Buddhist paintings recovered from the shrines of Dandan-Uiliq, .
® See chap. x1, in particular pp. 363 sqq. * See below, pp. 354 sqq.
INTRODUCTION 1X
forms an instructive parallel, and, like the curious find of a Judaeo-Persian document at the same site, convincingly demonstrates the continuity of these relations. Finally, the finds of Tibetan MSS. and sgraffiti in the ruined fort and temple of Endere prove that we have to reckon also with influences from the side of Tibet. From the seventh century onwards these are accounted for by historically attested invasions; but there are ethnic connexions, at present less clearly recognizable, which suggest them also for earlier periods. '
But with all these varied foreign influences at work, there is ample evidence to show that Khotan, together with other territories of Eastern Turkestan, possessed during the Pre- Muhammadan epoch a well-defined civilization of its own. Neither the occasional references of the Chinese Annals nor the brief notices of Chinese Buddhist pilgrims could help us far to reconstruct, in its main aspects, that bygone phase of civilization of which all indigenous records and remains seemed to have vanished. Nor could acquisitions of chance finds, made from afar, of uncertain, if not suspected, origin and unaccompanied by precise archaeological evidence, be expected to throw true light on this cultural meu and the historical changes and developments it had witnessed.
In order to secure materials that would help us to recover this interesting chapter of lost history, and to interpret it rightly, it was not enough to conduct excavations and to arrange for the deposition of what official language styles the ‘archaeological proceeds’ in Museums. It was at least equally important that an exact and detailed record should be kept of all observations made on the ground, and that the evidence thus secured should be published with all possible care and fullness. It is needless to emphasize for fellow-scholars the impor- tance of this condition, which the canons of scientific archaeology impose upon any qualified worker. Yet for the benefit of others I cannot refrain from quoting the warning which an archaeological explorer of unequalled experience, Professor Flinders Petrie, has so eloquently recorded in his admirable handbook on Methods and Aims in Archaeology. ‘To leave a site merely plundered, without any attempt to work out its history, to see the meaning of the remains found, or to publish what may serve future students of the place or the subject, is to throw away the opportunities which have been snatched from those who might have used them properly’ ®. | :
I have often had to refer to the grave risks with which all ancient remains at the desert sites of Khotan are threatened by the destructive effect of wind-erosion, and even more at the hand of ‘treasure-seekers’, whose operations have in recent years been stimulated by the demands of European collectors. The thought of these risks and the knowledge that neither Khotan nor any other part of Eastern Turkestan, in spite of the conserving capacity of the climate, could ever have competed with countries like Egypt in wealth of antiquarian remains, always urged me on to unwearied exertion. The same thought and knowledge made me feel with equal keenness my responsibility in regard to the duty ‘to work with the fullest care and detail in recording, to publish everything fully’ *.
To the obligation here indicated the peculiar conditions of the field in which my work lay added another closely allied one. I mean the necessity of recording in some detail my observations on the modern physical conditions and ethnography of the region. To the critical student of history, whether he deals with the written records or the antiquarian relics of the past, the powerful influence exercised by geographical conditions and surroundings must ever be present. Nor ought he to ignore the useful guidance which ethnic inheritance affords for
* See Methods and Aims in Archaeology, p. 179. * Comp. Flinders Petrie, loc. cit., p. 175. STEIN b
x INTRODUCTION
the proper comprehension of the past of a people and its civilization. Eastern Turkestan is not a part of the earth as easily accessible as Greece or Egypt or India. If in recent years much has been achieved, particularly through Dr. Hedin’s labours, towards the elucidation of its topography, yet there are great tracts where the detailed knowledge of local conditions, routes, &c., must still be gained primarily on the spot. The same observation applies still more forcibly to the ethnography of the country. These considerations will explain why I often felt obliged to record modern physical and anthropological facts when discussing questions connected with earlier economic and ethnic conditions. They also justify me in looking upon the publication of my Personal Narrative, in which observations of geographical interest together with others illustrating the life and character of the population could be suitably detailed for a wider public, as a useful and even necessary preliminary to the present work.
Just as it was important for the proper investigation of antiquarian questions to keep in view the geographical and ethnographical background, so too my task presupposed constant references to whatever written records have come down to us of the early history of these regions. The great majority of these records have to be gleaned from scattered passages in the Chinese dynastic Annals and the itineraries of Chinese Buddhist pilgrims to India. The limitations of my philological knowledge, which, as I have often bitterly regretted, does not extend to Chinese, would not allow me to consult these sources in the original. Nevertheless I have thought it right to attempt the succinct sketches of the history of Eastern Turkestan, Kashgar and Khotan which are embodied in chapters m1 and vi1?. Fortunately M. Chavannes’ recent publication Documents sur les Tou-k'iue (Turcs) occidentaux has opened to us a rich storehouse of precise and valuable information on the history of Central Asia during the period of Chinese ascendancy under the T‘angs, including a great portion of those notices in critically revised and annotated translations. For the Chinese notices on Khotan I could avail myself of the full extracts which Abel Rémusat, in a small but very meritorious publication, had rendered accessible from the Encyclopaedia of the Pzen z tzen*. I hope those historical sketches will prove useful by the endeavour made in them to apply to the recorded historical data the light of geographical and antiquarian knowledge acquired on the spot.
The preparation of my detailed report on the lines indicated was bound to prove an exacting task; but the difficulty of securing adequate leisure made it more difficult still and has greatly delayed its completion. The period of three months’ special duty which the Secretary of State had been pleased to sanction after I had, early in July, 1901, deposited my archaeological finds in the British Museum as a temporary measure, barely sufficed for the provisional arrangement of the collection and the preparation of the Preliminary Report published the same autumn®. At its expiration I was obliged to return to India, where during the following half-year my heavy official duties as Inspector of Schools in the Panjab precluded the possibility of any scholarly occupation. Fortunately, however, the Government of India, on the proposal of the Panjab Government, and influenced, no doubt, by the interest which H. E. Lord Curzon was kind enough to show in the results of my expedition, sanctioned, with the concurrence of H.M.’s Secretary of State, my deputation to England in May, 1902, in order that I might be enabled to elaborate those results and to make arrangements for their publication with the help of expert scholars in Europe.
7 See below, pp. 52-72, 166-84. topographical exploration in Chinese Turkestan. By M. A. 8 See Histoire de la Ville de Khotan, Paris, 1820. Stein. India Office, London, rgor. ° Preliminary Report on a journey of archaeological and
INTRODUCTION xi
I had every reason to feel deeply grateful to the Viceroy and the Indian Government for the generous consideration thus shown to me and my work. But the period of seven months’ stay in England, to which my deputation was at first limited, proved, as I had feared, wholly insufficient for the task. Two successive extensions which the Secretary of State, after reference to the Government of India, kindly sanctioned, added to the above period an aggregate of ten months. By dint of great exertions I could thus complete while in England those important portions of my task for which it was indispensable to have ready access to my archaeological finds deposited in the British Museum, as well as to the manifold expert help I needed for their detailed study and reproduction. The difficulty of these labours had been considerably increased by the uncertainty as to the length of time available for them. Yet before I left England I had managed to conclude the arrangements for the decipherment and publication of all the varied finds of MSS. and documents; to prepare and pass through the press the numerous plates of antiques; to supervise the drawing, &c., of the detailed map from original surveys which accompanies this work, and, last but not least, to bring out my Sand-buried Ruins of Khotan, which in many respects has formed a necessary preparation and complement to the present work.
After my return to India at the close of 1903 I devoted every moment of leisure I could spare from my duties as Inspector-General of Education and Archaeological Surveyor, North-west Frontier Province and Baluchistan, to continued work on my report. But in spite of all efforts its publication might have been delayed for years, had not the Government of India, in order to enable me to resume my explorations in Chinese Turkestan in the spring of 1906, been induced to set me free once more by the grant of six months’ special duty from October of last year. It was during these busy winter months in Kashmir, while many practical preparations for my new journey claimed attention, that the greater part of my text was written.
From the reasons which explain why this final report could not be presented earlier, I now turn with pleasure to the duty of recording my acknowledgements to those scholars and friends without whose valuable collaboration and kind help this work could not have been completed in its present form. First among them I must mention my artist friend, Mr. Frep. H. ANpREws, late Principal of the Mayo School of Art, Lahore, and now Director of the Art Department, Battersea Polytechnic, whose devoted exertions have effectively furthered my labours ever since my return from Chinese Turkestan. His wide knowledge of ancient Indian art, especially of the Graeco-Buddhist style, his familiarity with the products of many old Eastern industries; and his own high artistic abilities, exceptionally qualified him to assist me in the arrangement, description, and illustration of my collection-of antiques. The detailed descriptive list of these included in the report is wholly his work, though, having myself checked and revised it, I must accept full responsibility for its contents, and in particular for the data as to the origin of individual antiques. Without the painstaking accuracy and fullness of Mr. Andrews’ descriptions it would probably have been impossible for me to complete my report at a time when some six thousand miles separated me from my collection. The value of Mr. Andrews’ descriptive list will be all the more appreciated by students of the early arts and industries of Central Asia since, according to a decision of the Indian Government, the whole of the archaeological finds will now be distributed between the museums of Calcutta and Lahore and the British Museum. As comparison of the original objects will thus be difficult, their use for future researches will mainly depend on the details furnished by the descriptive list and on the illustrations given in the plates. For the preparation of the latter and for the investigation. of all questions affecting the technical and art aspects of my finds, Mr. Andrews’ expert help and advice have been of the utmost value to me. His unwearied assistance did not cease after I left England, and
| b 2
Xli INTRODUCTION
has greatly facilitated the passing through the press of these volumes. Besides the sketches and diagrams reproduced among the illustrations, I owe to Mr. Andrews’ artistic skill the black and white drawing which has furnished the vignette for the title-page. My volumes could scarcely issue under a more felicitous emblem than this faithfully restored enlargement of the figure of Pallas Promachos as seen in several of the ancient seal impressions in clay excavated by me.
I shall always associate the remembrance of the labours given to the elucidation of my archaeological finds with the British Museum, where they found their first resting-place. I may hence fittingly record in this place my obligations to its authorities who readily accorded them shelter. Both in the Coin Department and in that of Mediaeval Antiquities, where my collection was successively accommodated, I enjoyed exceptional facilities for its arrangement, &c., Mr. C. H. Reap, Keeper of the latter Department, proving in particular a most helpful adviser.
I owed these facilities in the first place to the kind help of my friend Prof. E. J. Rapson, then Assistant Keeper of the Coin Department, who during my absence in India charged himself also with the care of my collection. It was with a feeling of keen relief that I received the same friend’s offer to undertake what from the first I had recognized as the most difficult of the philological tasks resulting from my explorations,—the decipherment and publication of the mass of Kharosthi documents on wood and leather recovered from the Niya Site. No scholar could have been better qualified for this arduous work. Prof. Rapson’s preliminary papers show the advance he has already made in the task?*°. I owe it to Prof. Rapson’s kindness that in chap. x1 I have been able to indicate some results of his labours possessing special antiquarian bearings. The full publication of these documents, for which Prof. Rapson has been fortunate enough to secure the assistance of two distinguished savants, M. E. Srnart, Membre de I’'Institut, the decipherer of so many important Indian epigraphical records, and M. lAbbé A. Boyer, had to be reserved for a separate volume which is to follow this report at a date that cannot yet be specified.
For the MSS. and documents written in Brahmi characters, and showing texts partly in Sanskrit, partly in two non-Indian languages, Dr. A. F. R. Hoernte, c.1.£., was from the first marked out as the most competent editor. The eminent Indologist has completed his edition of these finds. It would have been difficult to provide for their publication in connexion with this report except by a separate volume; and as Dr. Hoernle has recently undertaken, under the auspices of the India Office, the editing of a series in which all previous MS. acquisitions from Eastern Turkestan entrusted to his care are to be published, the inclusion in it of those Brahmi texts has appeared more convenient. In the meantime Dr. Hoernle has placed at my disposal a valuable analysis of the Brahmi documents and inscriptions, which has been of great help in describing these finds in chapters 1x and xm, and for which I wish to offer here my special thanks.
Among the Appendices contributed to my report there is none for which I feel more deeply grateful than that in which M. Ep. Cuavannes, Membre de I’'Institut, has with masterly clearness analysed my finds of Chinese documents and inscriptions. It is not for me to appraise the exceptional qualifications which have rendered M. Chavannes the leading authority on all Chinese sources of information concerning the history and geography of Central Asia. The great help I have derived from his annotated translations of those finds, as well as from
© See below, pp. 326 sqq., 364 sqq. 1! See below, pp. 257 sqq:, 295 Sqq-, 439 Sq.
‘INTRODUCTION Xili
the long series of his earlier publications, will be evident to my readers from every chapter. But in addition I owe a debt of gratitude to M. Chavannes for a service of friendship which, I know, must have cost heavy sacrifices of time and labour. In order to secure in my work that uniformity in the transcription (according to Wade’s system) of Chinese names and terms, the want of which in the various publications consulted for Chinese records had often been a sore difficulty to me, and with a view to assuring general correctness in Chinese references, he kindly charged himself with a detailed revision of the latter, both in my MS. and in a final proof.
For my finds of Tibetan MSS. and sgraffiti (Appendix &) I had the good fortune to secure as joint editors two scholars of critical thoroughness, Dr. L. D. Barnett, of the British Museum, and Rev. A. H. Francxe, of the Moravian Mission, Ladak. Dr. Barnett’s preliminary notice will have shown to Orientalists how well qualified he was to elucidate the importance of those MS, finds for the history of Tibetan language and literature 1°. By his exceptional knowledge of Tibetan rock-carvings and of the colloquial language, the Rev. Mr. Francke has been able to deal successfully with the sgraffiti which would have baffled any Tibetan scholar in Europe. To Mr. F. W. Tuomas, the learned librarian of the India Office, my sincere thanks are due for having kindly supplemented the Tibetan notices of Khotan otherwise accessible to me by the valuable extracts from the Kanjur reproduced in Appendix &. From another and widely distant field of Oriental research has valuable help come to me in the learned analysis which Professor D. S. Marcotiouta has given of the curious Judaeo-Persian document from Dandan-UViliq reproduced in Appendix C.
The classification of my collection of coins could not have been carried out by more competent scholars than Dr..S. W. BusuHELL, c.m.c., the distinguished Sinologist, and Prof. E. J. Rapson. Though I had endeavoured, as far as my means of reference would permit, to determine on the spot all coins actually found by me, yet the inventory list in Appendix D, and the plates illustrating it were prepared mainly from the materials which those two friends were kind enough to place at my disposal. To Dr. S. W. Bushell I am still further indebted for the liberality with which he has allowed me to draw upon his store of antiquarian and historical knowledge in questions connected with Chinese lore. To Dr. Percy Garpner, Professor of Archaeology at Oxford, I must offer my best thanks for having accorded to me the same favour in regard to objects of classical and semi-classical art contained in my collection.
The two interesting Appendices /, G, in which Professors A. H. Cuurcn, F.R.s., and L. pe Léczy have recorded the results of their respective analyses of the stucco materials, and the loess and sand specimens brought back by me from Khotan sites, deserve all the more my grateful acknowledgements, as they clear up questions in which the archaeologist can derive guidance solely from natural science. The same is the case with the important investigations which Prof. J. Wiesner, Director of the Institute for Plant-Physiology, Vienna University, has devoted to the ancient paper materials represented among my antiquarian finds!*, Nor ought I to omit grateful mention here of the valuable assistance I have received in the anthropological field, linking as it were the domains of historical and natural science, from the instructive analysis to which my friend Mr. T. A. Joycg, of the British Museum, Secretary of the Anthropological Institute of Great Britain, has subjected the anthropometric materials I collected on my journey ™.
# See /.R.A.S., 1903, pp. 109 sqq.; below, pp. 426 sq. 307 Sq., 426.
* See J. Wiesner, Ein neuer Betirag zur Geschichte des | See On the physical anthropology of the oases of Khotan Papieres, in Stteungsberichte of the Imperial Academy, and Kertya, By T. A. Joyce, in Journal of the Anthropol. Vienna, histor.-philos. Klasse, 1904; comp. below, pp. Institute, 1903, pp. 304~24 ; comp. below, pp. 144 sqq.
Xiv INTRODUCTION
A reference to chapter vi, where I have discussed the racial origin of the Khotan population, will show how illuminating these investigations are to the historical student.
For the publication of my report, the liberal arrangements sanctioned by the Secretary of State have proved of great help, and for them I may be allowed to record my special gratitude. At my request the preparation of the collotype plates illustrating antiques was entrusted to Mr. W. Gricecs, whose name has been made familiar by the labours of a lifetime to every student of Indian art. I appreciate all the more gratefully the personal care and attention which Mr. Griggs has bestowed on the work as I know that much of this, as in the case of other scholarly publications to which he has lent his skill, must have been a labour of love. The drawing of the map of Khotan and the adjoining regions was entrusted to Mr. J. W. Appison, map draughtsman of the Royal Geographical Society, and effected in a fashion which has been commended by competent judges.
I had every reason to feel gratified when the Delegates of the Clarendon Press, with a grant from the India Office, agreed to undertake the printing and publication of the work, together with the preparation of the remaining plates. My hope that the resources of that great officina would facilitate the early issue of the work, and in a fitting form, has been justified by experience. For the typographical care bestowed on these volumes and for much kind attention besides, I wish to express here my thanks to Mr. C. Cannan, Secretary to the Delegates, and Mr. H. Hart, the University Printer. |
Owing to the great distance separating me from the Press, it was impossible for me to see more than one proof, and even for this the available time was very scanty. I could not have hoped to assure the requisite degree of correctness, had not the self-sacrificing kindness of friends come to my help. Mr. J. S. Corron, editor of the /mpertal Gazetteer of India, kindly revised my manuscript before it went to the press, with regard to those literary requirements of which he is so experienced a judge. Mr. Vincent Smiru, the distinguished Indologist, was good enough to read a proof and favoured me with a number of useful corrections. Mr. T. W. Arwnoxp, Professor of Arabic at University College, London, charged himself with seeing the final revises, a task for which I could not have found a friend of wider Oriental attainments nor one more willing to help. To Mr. Arnold also is due the elaborate Index, which I was unable to compile myself, but which I should not have liked to entrust to any other hands. But to none are my obligations greater than to my friend Mr. P. S. Aten, of Corpus Christi College, Oxford, who from a scholarly task of absorbing magnitude, in a field wholly Western, spared time to correct my proofs with most painstaking thoroughness. He also took entire ‘charge’ of the book when I had passed beyond the reach of regular postal communications. I look upon these proofs of friendly devotion as the best compensation for the difficulties I had to overcome while preparing my work.
During that happy year of travel through the mountains and the deserts which once saw the passage of Hsiian-tsang, the great pilgrim, of Marco Polo, and of him ‘who sought lost Cathay and found Heaven,’ the thought of the great scholar who had first with true critical intuition traced their tracks and those of many another early traveller through Central Asia, was ever with me. From his immortal volumes, which have accompanied me everywhere, I never failed to derive guidance and encouragement, whether I turned to them in camp after long hours of rough travel or in my improvised study after desk labour yet more tiring. In dedicating this work to the memory of Sir Henry Yuxe I claim no small privilege. But if the interest of researches on ground that was cherished by the Master, and the endeavour to
INTRODUCTION xv
carry them on in his spirit can atone for whatever there is of defects and dryness in the presentation of their results, I hope that this tribute will not be deemed inappropriate.
Fate has willed that I should write these closing lines at the foot of the mighty Kun-lun, and yet within sight as it were of the desert, on my way back to Khotan which I have so often longed to see again, and at the commencement of an enterprise which is to take me yet further afield. Grateful do I feel to Fate—and to those who dispense its decrees—for the opportunity which enables me to proceed to a fresh task at the very time when the old one
has reached its end.
Gladly do I now forget all the efforts that were needed to assure me
this freedom, and gladly do I see in its achievement a reward for the labours recorded in
these volumes.
Camp KO6xK-YAR, Fuly 24, 1906.
ADDENDA ET
P. 10,1. 41. The identification of Kao Hsien-chih’s route with that leading over the Bardghil and Darkét has been confirmed in all particulars by the observations I made on my visit to these passes in May, 1906. From Sarhad to the Yasin side of the Darkdt Pass three marches are reckoned, Chikmar-robat and Vedinkot forming the recognized inter- mediate halting-places.
P. 14, 1. 9. For Kashkar read here and elsewhere Kashkar.
P. 16, 1. 29. The name reproduced in Chinese transcription by A-shé-yii-shth-to probably survives in that of the modern village of Shuist, on the upper Mastij river, near which I traced, in May, 1906, extensive areas of former cultiva- tion, said to have been abandoned for a long time owing to the advance of glaciers. Their re-occupation is now slowly proceeding.
P. 35, 1.1. The remains of K7z-kurghan have since been surveyed by me, on May 30, 1906. They were found to consist of massive bastioned walls of stone and brick, manifestly of ancient construction, which crown the west rim of an isolated rocky spur rising circ. 600 ft. above the Taghdumbash river. The top of the spur is very difficult of access, even on that side, and protected on all others by wholly unscalable precipices.- The position is one of exceptional natural strength, and the ruins surveyed prove its occupation as a place of refuge at an early period.
P, 39, 1. 29. Crossing the Chichiklik plateau on June 4, 1906, I traced near its centre the foundation walls of a large stone-built quadrangle, now partly occupied by decayed graves. The recent construction, by the Chinese authori- ties, of some shelter-huts in the immediate vicinity proves the need of a hospice in this locality. The physical aspects of this inhospitable plateau agree accurately with Hsiian-tsang’s description.
P. 62, 1. 33 marg. For Hsiian Tsung read Hsiian-tsung.
P. 91. 1. 9. The anthropometric data collected by me in July, 1906, concerning the Pakhpo people support the
M. AUREL STEIN.
CORRIGENDA
ethnic affinity here indicated. I have not been able to trace the use of a separate language among them.
P. 91,1. 19. During my stay at Kék-yar, in July, 1906, I received information about a number of caves which are venerated as Mazars in the mountains drained by the Tiznaf river.
P. 103, 1. 14. The new district established since 1903, com- prising Giima, Moji, Sanju, &c., has received the official designation of P‘z-shan, showing that the identification here assumed is accepted by the Chinese authorities.
P. 111, last line, and p. 285, 1. 3, a. 4. or Pi-mo read P‘i-mo.
P. 113,]. 12. For fig. 3 read fig. 31.
M. oor. k. For LI read IL.
P. 114. T.M. 003. j. Add See Plate XLII.
P. 130, note 23. On revisiting Karanghu-tagh in August, 1906, I ascertained that the story of Aba Bakr’s riches thrown into the Yurung-kash on his flight to this valley still lingers in local tradition. It is also popularly known in Khotan.
P, 131, 1.5. For Ujat read Ujat.
P. 136, 1. 7. Hor Cherchen read Charchan, and so elsewhere.
P, 166, n. 1. For 150 read 151.
P. 172, 1. 26. For Si-shan read Hsi-shan.
P. 188, 1. 31. Read of M. Petrowsky, late Russian Consul General, &c.
P. 190, note 2. Read on M. Petrowsky’s large collection.
P. 208, 1.25. Mor Y. o11 read Y.oo11; 1. 28, for B. oor. 1. read B. oor. |.
P. 213. Y. oo11. i. For XLIV read XLVII.
Y. oo11. j. Add See Plate XLVI. P. 218 (col. a), 1. 31. For Plate 36 read Plate XLVII.
P. 230, 1. 2, For D.tv. § read D. x. 4; n. 31, for sec. v. read sec. iv.
Xvi ADDENDA ET CORRIGENDA
236,n.1. ForPlateLV read PlatesLVI-LVII, LXXXVIII. 238, 1. 2. For Yamen read Ya-mén and so elsewhere. 246,1.3,a.f. For R. iv read R. ii.
252, 1. 36. While this chapter was passing through the
press Mr, F. H. Andrews informed me of his having dis-
covered, among the objects brought back from Tibet by the late expedition, and now deposited in the British
Museum, a suit of mail, evidently old, which shows the
plates or scales arranged in the reversed order exactly as
on the Kubera image of the Dandan-Uiliq shrine and in the Gandhara example. The plates are not rivetted but ingeniously threaded, by a method which is illustrated in the accompanying diagram of Mr. Andrews. The dispo- sition of the holes, one at the rounded end and three on each of the longer sides, is exactly the same as in the
piece of hard leather, N. xv. 005 (see diagram, p. 41 t),
and thus strikingly confirms the suggestion made regarding
the latter (p. 374) that it belonged to a scale armour.
P. 257, 1. 16. For king read kind.
P. 269, n. 15. For D. m1. 13 read D. ut. 12; for D. vi. 4, 6
read D. v1, 2, 6.
288. D.1. 11. Add See Plate LV.
297. D. v1.3. Add See Plate LXVI.
298. D. vm. 3. d. Add See Plate CXVI,.
302. D.(T.) 05. Add See Plate LV.
335, 1.26. Nx. 2. c is not shown in Plate XCVII.
343, n. 4, 1.2. For 104804190 read 10+ 86+ 190.
374, ]. 26. For the discovery of a Tibetan coat of mail
which conclusively proves that the piece of leather must
have formed part of a scale armour, see above, Add. to
p. 252, with illustration.
P, 376, 1. 20. For N. xix.1 read N. xix. t.
1.5, af. For N. xx. 1 read N. xx. 1.
P. 377, n. 6. For p. 335 read p. 334-
P. 378, 1. 13. On my recent passage through Chitral and Mastiij I repeatedly traced the same flower ornament in the architectural wood-carving of those valleys.
P. 393. N. iv. 80. Add See Plate LXXI.
P. 394. N. iv. 133. Add See Plate XCIII.
P. 395. N.v. 6. Add See Plate CII.
P. 398. Fora reproduction of N. x.1 in fuller detail see p. xxiv.
N. x. 2. c is not shown in Plate XCVII.
P. 399. N. xii. 3. Add See Plate LXX.
P. 402. N. xv. 59. b. Add See Plate CXIII.
rd td
Lee Se
Ofructure of Scale Armour
Obverse
Loop relaxed fo show method of threading straight lace .
P. 404. N. xv. 122. Add See Plate XCVII. P. 407. N. xv. 207. Add See Plate CXIII.
P. 410. N. xv. oor. a. or (interior) read (exterior), for (exterior) read (interior).
. 411. N. xv. oo1. c, f, g, h, j. Add See Plate LXXIV. 413. N. xx. 03. Add See Plate LXIX.
. 425, 1. 11, a.f. For 31 read 314.
438. E.i. 4. Add See Plate CIX.
. 439. E.i. 39. Add See Plate CIX.
441. E. i, 025 does not appear on Pl. LXXVII.
Mid dH
Section
INTRODUCTION
TABLE OF CONTENTS
ADDENDA AND CORRIGENDA
LIST OF ILLUSTRATIONS
LIsT OF ABBREVIATED TITLES
CHAPTER I.—FROM KASHMIR TO THE PAMIRS
Section i.
ye aR
The Gilgit Road and Kisanganga Valley
Chinese Historical Relations with Gilgit
The Chinese Occupation of Gilgit and the Route to shin Ancient Remains in Gilgit and Hunza
CHAPTER II.—SARIKOL AND THE ROUTE TO KASHGAR
Section i. pene ale
The Geographical Position and Ethnography of Sarikol Early Chinese Accounts of Sarikol ° ; Historical Sites of Sarikol
From Sarikol to Kashgar
CHAPTER III.—HISTORICAL NOTICES OF KASHGAR
Section i, oy ie
The Old Names of Kashgar
Kashgar during the Han Epoch
Eastern Turkestan under the T‘angs
Notices of Kashgar during the T'ang Period .
PAGES V—-XV
XV-XVi. XX=-Xxi
XXii-xxiv
22—46
47-72 47 52 57 65
CHAPTER IV.—THE ANCIENT REMAINS OF KASHGAR AND THE OASES OF YARKAND AND KARGHALIK
Section i.
oe in Se
CHAPTER V.—THE ROUTE FROM KARGHALIK TO KHOTAN:
The Stipas of Kurghan-Tim and Kizil-Debe . The Ruins near Khan-ui y The Oasis and City of Yarkand Karghalik in Chinese Records .
REMAINS
Section i.
STEIN
By the Desert Edge to Khotan The Oasis of Gima
The Tati of Kakshal .
Ancient Remains at Moji . From Moji to the Khotan Oasis
73-93 73
79 86
89
ITS ANCIENT TOPOGRAPHY AND
94 99 103 110
115
XViil CONTENTS
CHAPTER VI.—THE KHOTAN OASIS: ITS GEOGRAPHY AND PEOPLE. Section i, The Oasis in its Geographical Features » ii. Agriculture and Industries in Khotan ; » iii The Population of Khotan: its Distribution and Chseceter » iv. The Population of Khotan: its Physical Characteristics and Racial Origin
CHAPTER VII.—HISTORICAL NOTICES OF KHOTAN
Section i. The Early Records and Names of Khotan » ii. The Legendary Traditions of Khotan » wi. Khotan in Chinese Records, from the Han to the Sui Dynasty » iv. Khotan during the T‘ang Period » v» Later Chinese Records of Khotan
CHAPTER VIII—ANCIENT SITES OF THE KHOTAN OASIS Section i. The Hill of Gogrnga . » i. The Culture-strata of Yotkan » iii. The Site of the Ancient Capital » iv. Antiques acquired from Yotkan and in Khotan » v. Buddhist Sites described by Hsiian-tsang
CHAPTER IX.—THE RUINS OF DANDAN-UILIQ
Section i. The Desert March to the Site. : ; , 3 » di. First Excavation of Buddhist Shrines » wi. Art Relics of Shrine D., II. » iv. First Finds of Ancient Manuscripts . » Vv. Discovery of Dated Documents 5» vi. Records from the Hu-kuo Convent . , » vii. Other Ruins of Dandan-Uiliq and General Ohaeitions on Site », viii. List of Objects excavated or found at Dandan-Uiliq
CHAPTER X.—FROM DANDAN-UILIQ TO THE NIYA RIVER . Section i. The Rawak Site . » ii. A Judaeo-Persian Document » ii. Keriya, Niya, and Imam Ja‘far Sadiq
CHAPTER XI.—THE ANCIENT SITE BEYOND THE NIYA RIVER Section i. The Ruin N. 1 and the first Finds of Inscribed Tablets
di. Excavation of Ancient Residences, N. 11, N. 1, N. Iv
» lii. Discoveries in an Ancient Rubbish-heap, N. xv.
» iv. Ancient Documents on Wood and Leather :
» v. Chinese Documents from N. xv. and the Writing on Wood vi. Decipherment of Ancient Documents, Kharosthi and Chinese » vii. Exploration of Ruins N. vI-xm and General Observations on Site , viii. List of Antiques excavated at Niya Site
CHAPTER XII.—THE ENDERE RUINS Section i. The March to the Endere River » ii. Excavation of the Endere Temple » iii, The Ruined Fort and Stipa of the Endere Site : iv. List of Antiques from the Endere Ruins
PAGES
» 123-150
123 130
136
143 151-184 151
156
166
172
177
185-235 185 1go 199 206 22%
236-303 236 242 249 256 264 273 281 288
+ 304-315
34 306
so
. 316-416
316 328 338 344 358 363 374 385 417-442 417 421 430 438
CHAPTER XIII.—KARA-DONG AND THE SEARCH FOR HSUAN-TSANG’S P‘I-MO Section i. ae
CONTENTS
Expedition to the Kara-dong Ruins ; Hsiian-tsang’s Notice of P‘i-mo and Marco Polo's Pein The Sites of Uzun-Tati and Uligh-Ziarat
The Tiiga-dong Mounds and the Keriya—Khotan iiotite
CHAPTER XIV. THE RUINS OF AK-SIPIL AND RAWAK
Section i. Hanguya Tati and the Site of Tam-Oghil » di, The Remains of Ak-sipil and Kighillik » iii. The Rawak Stipa » iv. The Sculptures of the Rawak Vihara » v. The Date of the Rawak Remains; the Jumbe-kum Site
CHAPTER XV.—DEPARTURE FROM KHOTAN
Section i. ” li. ane
APPENDIX A.—CHINESE DOCUMENTS FROM THE SITES OF DANDAN-UILIOQ, NIYA AND ENDERE. » 521-547
Islam Akhiin and his Forgeries Last Days in the Khotan Oasis From Khotan to London
APPENDICES
Translated and annotated by EDOUARD CHAVANNES Part I. Les Documents Chinois de Dandan-UViliq
i es » IIL.
APPENDIX B.—TIBETAN MANUSCRIPTS AND SGRAFFITI DISCOVERED AT ENDERE.
Les Documents sur Bois de Niya 2 Les Sgraffiti et Documents Chinois du Fort D’ Suidecs
L. D. BARNETT and A. H. FRANCKE Part I. Fragments of the Salistamba-siitra », II. Miscellaneous Fragments of Religious Works : » III. Two Religious Poems » IV. Sgraffiti of Endere
APPENDIX C.—THE JUDAEO-PERSIAN DOCUMENT FROM DANDAN-UILIQ. Edited by D. S.
MARGOLIOUTH
APPENDIX D.—INVENTORY LisT OF COINS FOUND OR PURCHASED. Prepared from Notes of - 575-580
APPENDIX E.—EXTRACTS FROM TIBETAN ACCOUNTS OF KHOTAN. Communicated and annotated
S. W. BUSHELL and E. J. RAPSON
by F. W. THOMAS
Part I. Notes on Rockhill’s Life of = Buddha, pp. 230 sqq., and qh. Works there quoted
II. Notes from the Gosraga Vyadkarana
+P
APPENDIX F.—NOTES ON SPECIMENS OF ANCIENT STUCCO FROM KHOTAN SITES. By A. H.
CHURCH
APPENDIX G. NOTES ON SAND AND LOESS SPECIMENS FROM THE REGION OF KHOTAN. By . 588-590
L. DE Loczy
INDEX OF OBJECTS FOUND, PURCHASED, ETC. PLATES DESCRIBED, OR REFERRED TO, IN THE TEXT GENERAL INDEX
xix PAGES 443-469 443
45%
457
465 470-56 470
474
482
488
500
507-520 507 514 517
521 537 546
- 548-569
549 557 561 566
579-574
581-585 581 584
586-587
591-596
- 597-598
599-621
LAST OF: BEAST RACE NS
Rock-carved Relief near Naupir, Gilgit : ‘ ‘ ‘ : . To face p. 18 Castle of Mir of Hunza at Baltit ‘ : ; ; : ; % 18 Mount Rakipdshi, seen from Aliabad, Hunza : : 4 ; . ‘6 20 Ruined Stiipa at Thol, Hunza : , . j : ; P 20 Wakhis settled at Khaibar, Hunza . ; ; : , , 3 4 26 Kanjitis at Misgar . ‘ ; : . ‘ 3 26 Glaciers at source of Oxus, seen oC approach to . Wakhjir Pass F , ‘ g 32 Kilik Pass, looking south from Khush-bél spur. : Fe oe Interior of ruined town, Tash-kurghan, looking north-east Te Shindi defile ; : i 36 Chinese fort within ruined town, Tash-kurghan . ; : , . i 36 Muztagh-Ata Peak seen from Shamalda ridge : f i , ’ 2 44 Muztagh-Ata Peak seen from Lake Little Kara-kul j i ‘ : Se 44 Ruined Stipa and Mound of Mauri-Tim, seen from south-west . ; ' : ‘ 74 Ruined Stipa of Kurghan-Tim, seen from north . ‘ , , 3 i 74 Mecca Pilgrims from Kashgar F ; ‘ , ; ; 3 go Mosque and Tank at Yetimlukum, Karghalik , , . : ‘ 5 go. Yetimlukum Mazar with cemetery, near Karghalik ; ree ; ‘ ; 7 go Kaptar-Khana ruin near Khan-ui, Bésh-karim : ‘ : ; ns ae Kepek-gholuk Mazar near Moji ; : , ‘ ; ‘ : as eee Ruined Mound of Karakir-Tim, near Pialma : ; ‘ 112 Eroded Ranges towards Kara-kash river gorge, seen from ities yaoi Dawan : : % 128 Kun-lun Range, with Glaciers of Nissa Valley, seen from above Brinjak Pass : 3 128° Tawakkél labourers brought to Dandan-Uiliq : ; ‘ ; a 43 144 Taghliks and exiled Khotan criminals at Karanghu-tagh . : : ; P 144 Mazar of Kohmari , ; ‘ ‘ , ; y 186 Entrance to sacred cave of Kohmari Hill ; . , : 186 North-west corner of excavated area at Yotkan, with entrance to Yotkan-Yar é ae South-west banks of excavated area, Yotkan, near Allama hamlet. * 192 Ruins of Buddhist shrine, D. 11, Dandan-Uiliq, seen from south before excavation . ’ re 246 Ruins of Buddhist shrine D. 11, Dandan-Uiliq, seen from north-east after excavation } 3 246 Fresco and relief sculpture in small cella of shrine D. 11, Dandan-Uiliq . : » age Small cella of Buddhist shrine D. 11, Dandan-Uiliq, seen from west after excavation. , 252 Ruined dwelling D. 111, Dandan-Uiliq, in course of excavation . ; . » 956 Interior of ruined dwelling D. 11, Dandan-Uiligq, after excavation. ‘ : > aa8 Remains of ruined shrine D. Iv, Dandan-Uiliq ; ‘ : 7 e 264 Remains of eastern portion of ruined dwelling D. v, Dandan- vila , ; : ‘ 264 Mazar of Burhanuddin-Padshahim on Keriya River. : : , Kee Ruined house N. 1, Niya Site, seen from north-west after excavation , , eee. *)
Ruined Stipa, Niya Site, seen from south-east ; ; ; ; ; ‘ 328
LIST OF ILLUSTRATIONS
Group of ruined dwellings N. 11, seen from south-east after excavation
Ruined house N. Iv, with arbour, Niya Site, seen from north-east
Ruins of house N. I11, with fallen garden trees, Niya Site, seen from west
Western portion of ruined residence N. 111, Niya Site, before excavation .
Ruins of residence N. 111, Niya Site, seen from south-east before excavation
Ruined dwelling N. v, Niya Site, seen from south before excavation
Ancient rubbish-heap N. xv, Niya Site, seen from west in course of excavation
Inner surfaces of double wedge, showing fastening at string-hole end
String twisted into loop for fastening double wedge
Reverse of under-tablet of double wedge, with string-hole and folds of sue held by grooves
Obverse of covering-tablet of double wedge, with string-hole and string fastened in seal socket
Remains of dwelling completely eroded, near N. XI, Niya Site
Remains of ancient trees and tank, near dwelling N. vil, Niya Site
Fragments of woollen carpet
Oblong piece of hard ‘green’ leather : :
Ruined building E. 111, within Endere Fort, seen from south-west j
Interior of ruined temple cella E. i, Endere Fort, seen from east after excavation
Ruined Stipa, Endere Site, seen from south-west .
Ruined rampart near gateway, Endere Fort, seen from interior
Interior of ruined quadrangle, Kara-dong, seen from north-east after partial pear :
Wooden gateway of ruined quadrangle, Kara-dong, seen from east after excavation
Remains of ruined Stipa, Arka-kuduk-Tim, Hanguya Tati, seen from south
Remains of deserted modern dwelling, at Kobzagiram-shah, Domoko
Lachin-ata Mazar, seen from west
Portion of north segment of ruined rampart, Ak-sipil, seen hon eee ;
Portion of north segment of ruined rampart, Ak-sipil, seen from interior .
Ruined Stiipa of Rawak, seen from south part of court
Ruined Stipa of Rawak, with south corner of court, seen from east at commencement of excavation . . °
Relief sculptures, R. i—R. v, on inner south-west wall, Rawak Stapa Court
Relief sculptures, R. viii—R. xii, on inner south-west wall, Rawak Stipa Court
Statue of seated Buddha, R. xi, between colossal reliefs, in inner south corner, Rawak Stipa Court
Colossal statue, R. xiii, with relief ee in inner Eee corner, faye Stipa Ciea
Relief sculptures, R. xlix—lii, on outer south-east wall, Rawak Stiipa Court
Colossal statues, R. Ixvi-lxxiv, or outer walls of south corner, Rawak Stipa Court, with remains of outer passage in foreground
Relief sculptures, R. xxviii, xxix, with Dvarapdlas on right; at gate of nadae Stiipa Court
Relief sculptures, R. lvi-lxi, on outer south-east wall, Rawak Stipa Court
Relief statues, R. i-iv, on inner south-east wall, Rawak Stipa Court .
Turdi Khwaja, of Tawakkél, treasure-seeker .
Islam Akhin, forger of antiques
To face p.
»
To face p.
Page
9?
To face p.
xxi 328 330 330 332 332 340 340 348 348 349 349 378 378 398 411 422 422 434 434 446 446 471 471 471 474 474 484
484 490 490
492 492 496
496 498 498 508 508 508
Bist. ‘OR ABRREY intra? thigeo
Brat, Si-yu-k7.
Bipputpn, Hindoo Koosh
BrRETSCHNEIDER, Mediaeval Researches
Busuet., Early History of Tibet.
Cuavannes, Les Livres chinois . Cuavannes, Zurcs oceid. .
CHAVANNES, Voyage de Song Yun Cuavannes-Livi, L’Liinératre d’ Ou-Kong Curzon, The Pamirs
Euras and Ross, Zarikh-t-Rashidi
Forsyts, Varkand Mission Report .
Foucner, Géographie du Gandhéra .
Foucner, L’Ari du Gandhdra .
Foucurr, LZ’ Jconographie bouddhique .
Si-yu-ki. Buddhist records of the Western world, translated from the Chinese of Hiuen Tsiang (a.p. 629). By Samuel Beal, in two volumes. London, Triibner & Co., 1884.
Tribes of the Hindoo Koosh, by Major J. Biddulph, BS.C., Political Officer at Gilgit. Calcutta, Office of the Superintendent of Government Printing, 1880.
Mediaeval Researches from Eastern Asiatic sources. Fragments towards the knowledge of the geography and history of Central and Western Asia from the 13th to the 17th century, &c. By E. Bretschneider. 2 vols. London, Triibner, 1888. —
The early history of Tibet, from Chinese sources. By S. W. Bushell, M.D., Physician to H.B.M. Legation, Peking. Reprinted from the /.R.A.S., N.S., vol. xii.
Les livres chinois avant Vinvention du papier, par Edouard Chavannes. Extrait du Journal asiatique, numéro de Janvier—Février 1905. Paris, Imprimerie Nationale. 1905. '
Documents sur les Tou-kiue (Turcs) occidentaux. Recueillis et commentés par Edouard Chavannes, Membre de I’Institut, Professeur au Collége de France, avec une carte. Saint-Pétersbourg, Académie Impériale des Sciences, 1903.
Voyage de Song Yun dans l’'Udyana et le Gandhara (518-522 p.c.). Traduit par
M. E, Chavannes, Membre de l'Institut. Extrait du Bulletin de l’Ecole Frangaise d’Extréme-Orient, Juillet--Septembre, 1903. Hanoi, F. H. Schneider, 1903.
Voyages des pélerins Bouddhistes. _ L’itinéraire d’Ou-k‘ong (751-790). Traduit et annoté par MM. Sylvain Lévi et Ed. Chavannes. Extrait du Journal asiatique (Sept-Oct. 1895). Paris, Imprimerie Nationale, 1895.
The Pamirs and the Source of the Oxus. By the Right Hon. George N. Curzon, M.P., with map and illustrations. London, the Royal Geographical Society, 1899.
The Tarikh-i-Rashidi of Mirza Muhammad Haidar, Dughlat. A history of the Moghuls of Central Asia. An English version edited with commentary, notes, and map, by N. Elias. The translation by E. Denison Ross, (First edition.) London, Sampson Low, Marston & Co., 1895.
Report of a Mission to Yarkund in 1873, under command of Sir T. D. Forsyth, K.C.S.1., C.B., Bengal Civil Service. With historical and geographical information regarding the possessions of the Ameer of Yarkund. Calcutta, Foreign Department Press, 1875.
Notes sur la géographie ancienne du Gandhara (commentaire 4 un chapitre de Hiuen-Tsang), par A. Foucher, maftre de conférences 4 I’Ecole des Hautes-Etudes. Extrait du Bulletin de ?Ecole Frangaise d’Extréme-Orient. Hanoi, F. H. Schneider, 1902.
L’art gréco-bouddhique du Gandhara. Etude sur les origines de linfluence classique dans l’art bouddhique de l’Inde et de I’Extréme-Orient, par A. Foucher, Docteur-és-lettres. Tome Premier. Paris, Imprimerie Nationale. Ernest Leroux, éditeur, 1905.
Etude sur l’iconographie bouddhique de I’Inde d’aprés des documents nouveaux. Par A. Foucher. Vol. I, Paris, Ernest Leroux, éditeur, r9g00. Vol. II, 1905.
LIST OF ABBREVIATED TITLES XXili
FrankE, Zur Kenntniss der Turkvilker Beitrage aus chinesischen Quellen zur Kenntniss der Tiirk-Vélker und Skythen
GRENARD, Mission D. de Rhins
GriwnweveL, Mythologie .
GRUNWEDEL-Burcess, Buddhist ari .
Henin, Reisen in Z.-A. .
Hepin, Through Asia HoErnie, Report on C.-A, ant.
Jutien, Mémoires
Jutien, Méthode pour déchiffrer
Jurien, Vie de Hiouen-Thsang .
Leeor, Fa-hien
Livi, Notes chinorses sur Inde
Marquart, Lrankahr
Rapson, Specimens .
Rémusat, Ville de Khotan
RICHTHOFEN, China.
RitTER, Asien
Central-Asiens. Von QO. Franke. Aus dem Anhang zu den Abhandlungen der kGnigl. Preuss. Akademie der Wissenschaften von 1904. Berlin, Verlag der kénig). Akademie der Wissenschaften, 1904.
Dutreuil de Rhins (J. L.). Mission scientifique dans la haute Asie, 1890-1895. Vols. I-III. Deuxiéme et troisitme parties, par F. Grenard. Paris, Ernest Leroux, 1897-8.
Mythologie des Buddhismus in Tibet und der Mongolei. Fiihrer durch die lamaistische Sammlung des Fiirsten E. Uchtomskij, von Albert Griinwedel, Dr. phil., Leipzig, F. A. Brockhaus, 1900.
Buddhist Art in India. Translated from the ‘Handbuch’ of Prof. Albert Griin- wedel, by Agnes C. Gibson, Revised and enlarged by Jas. Burgess, C.I.E., LL.D. London, Bernard Quaritch, rgor.
Die geographisch-wissenschaftlichen Ergebnisse meiner Reisen in Zentralasien, 1894-1897, von Dr. Sven Hedin. (Erganzungsheft No. 131 zu ‘ Petermanns Mitteilungen.’) Gotha, Justus Perthes, 1900.
Through Asia. By Sven Hedin, in two volumes. London, Methuen & Co., 1898.
A Report on the British collection of Antiquities from Central Asia. By A. R. Rudolf Hoernle, C.LE., Ph.D. Part I. (Extra-Number to the Journal of the Asiatic Society of Bengal for 1899.) Calcutta, Baptist Mission Press, 1899.
Part II. (Extra~-Number 1 to the Journal of the Asiatic Society of Bengal, Vol. LXX, Part I, 1901.) Calcutta, rgo2.
Mémoires sur les contrées occidentales, traduits du sanscrit en chinois, en l’an 648, par Hiouen-Thsang, et du chinois en frangais par M. Stanislas Julien. Vols. I, II. Paris, Imprimerie Impériale, 1857-58.
Méthode pour déchiffrer et transcrire les noms sanscrits qui se rencontrent dans les livres chinois, inventée et démontrée par M. Stanislas Julien. Paris, Imprimerie Impériale, 1861.
Histoire de la vie de Hiouen-Thsang et de ses voyages dans I’Inde, depuis I’an 629 jusqu’en 645, par Hoei-li Yen-Thsong; suivie de documents et d’éclaircissements géographiques tirés de la relation originale de Hiouen-Thsang ; traduite du chinois par Stanislas Julien. Paris, Imprimerie Impériale, 1853.
A record of Buddhistic kingdoms, being an account by the Chinese monk F4-hien of his travels in India and Ceylon (a. p. 399-414) in search of the Buddhist books of discipline. Translated and annotated, with a Corean recension of the Chinese text, by James Legge, M.A., LL.D. Oxford, at the Clarendon Press, 1886.
Notes chinoises sur Inde. Par M. Sylvain Lévi, Professeur au Collége de France. IV.—Le pays de Kharostra et l’écriture Kharostri. Extrait du Bulletin de I’Ecole Frangaise d’Extréme-Orient. (Juillet-Septembre, 1904.) Hanoi, F. H. Schneider, 1904.
EranSahr nach der Geographie des PS. Moses Xorenacii. Mit historisch-kri- tischem Kommentar und historischen und topographischen Excursen von Dr. J. Marquart. Abhandlungen der kéniglichen Gesellschaft der Wissenschaften zu Géttingen, philologisch-historische Klasse, Neue Folge, Band III, Nro. 2. Berlin, Weidmannsche Buchhandlung, 1gor.
Specimens of the Kharosthi inscriptions discovered by Dr. Stein at Niya in Chinese Turkestan. Tentative transcriptions by E.J. Rapson, M.A. [Printed for presentation at the XIVth Congress of Orientalists, 1905. ]
Histoire de la ville de Khotan, tirée des annales de la Chine et traduite du chinois ; suivie de recherches sur la substance minérale appelée par les Chinois preRRE DE 1U, et sur le Jaspe des anciens. Par M. Abel Rémusat. Paris, imprimerie de Doublet, 1820.
China. Ergebnisse eigener Reisen und darauf gegriindeter Studien. Von Ferdi- nand Freiherrn von Richthofen. Erster Band. Berlin, Verlag von Dietrich Reimer, 1877.
Die Erdkunde von Asien, von Carl Ritter. Band V. Berlin, G. Reimer, 1837.
XXIV
LIST OF ABBREVIATED TITLES
Rocknity, Life of the Buddha .
Stein, Prelim. Report
Stein, Rajat.
Stein, Ruins of Khotan
Watters, Yuan Chwang —
Wiesner, Weuer Beitrag zur Geschichte
des Papieres.
Yutz, Cathay
Yur, Marco Polo ,
The Life of the Buddha and the early history of his order. Derived from Tibetan works in the Bkah-hgyur and Bstan-hgyur. Followed by notices on the early history of Tibet and Khotan. Translated by W. Woodville Rockhill. London, Triibner & Co., 1884.
Preliminary Report on a journey of archaeological and topographical exploration in Chinese Turkestan. By M.A. Stein. Published under the authority of H.M.’s Secretary of State for India in Council. London, rgor.
Kalhana’s Rajatarangini, a Chronicle of the kings of Kasmir. Translated, with an introduction, commentary, and appendices, by M. A. Stein. Vols. I, IL London, A. Constable & Co., 1900.
Sand-buried ruins of Khotan. Personal Narrative of a journey of archaeological and geographical exploration in Chinese Turkestan. By M. Aurel Stein. (First edition.) London, T. Fisher Unwin, 1903.
On Yuan Chwang’s travels in India, by Thomas Watters, M.R.AS. Edited, after his death, by T. W. Rhys Davids and S. W. Bushell. Two vols., with two maps and an itinerary by Vincent A. Smith. London, Royal Asiatic Society, 1905.
Ein neuer Beitrag zur Geschichte dés Papieres, von J. Wiesner. Sitzungsberichte der kais. Akademie der Wissenschaften in Wien, philosophisch-historische Klasse, Band CXLVIII. Wien, 1904, in Kommission bei Carl Gerold’s Sohn.
Cathay and the way thither; being a collection of medieval notices of China. Translated and edited by Col. Henry Yule, C.B. In two vols. London, Hakluyt Society, 1866,
The book of Ser Marco Polo the Venetian concerning the kingdoms and marvels of the East. Translated and edited, with notes, by Colonel Sir Henry Yule, R.E., C.B., K.C.S,L., corr. inst. France, Third edition, revised throughout in the light of recent discoveries by Henri Cordier (of Paris); with a memoir of Henry Yule by his daughter Amy Francis Yule. In two volumes. London, John Murray, 1903.
REPRODUCTION IN FULLER DETAIL OF THE FRAGMENT oF CarpszT, N. x. 1, PAGE 398.
AT
bid f
Blue & Light Yellow
Blue & Dark Yellow
. Blue ie Light Vellow
Red & Light Yellow?
tf
=,
=: ralew =~ me = =
a
ma,
Coll git dee Bd FROM KASHMIR TO THE PAMIRS Section I—THE GILGIT ROAD AND KISANGANGA VALLEY
ASHMIR, where in 1898 the first plan of my explorations had been formed, and where Kashmir,
two years later, in May 1900, I completed the practical preparations of my journey, was gia al by reason of historical associations and geographical position alike a most appropriate starting- ake point for my archaeological enterprise. I may therefore fitly commence this detailed account of my antiquarian observations from the same ground.
The labours which I devoted during the preceding ten years to the elucidation of Kalhana’s
Chronicle of Kashmir, and the manifold historical researches necessitated by this task, had made me eagerly look out for whatever traces survived of ancient relations between Kashmir and the Central-Asian regions lying beyond its great mountain barriers. The information to be gleaned from the sources accessible to me was scanty, even in regard to the tracts immediately adjoining Kashmir, which were undoubtedly subject from early times to its cultural influence. But the results of the close study I bestowed on the ancient topography of Kashmir, assured me of at least one interesting fact, which at the time of my start (May 31, 1900) appeared like an auspicious omen: there could be no doubt from the first that the route I was to follow through the mountains on leaving the Kashmir Valley for the north had a claim to considerable antiquity and historical importance. |
Practical advantages as well as geographical interests had made me decide for the route Antiquity of through Gilgit and Hunza, and over the Taghdumbash Pamir, as the most suitable line of approach ha to Kashgar, my immediate goal in Chinese Turkestan. The ‘Gilgit Transport Road’, which I was authorized to use for the first portion of my journey, from Kashmir to Gilgit, dates in its present form only from the years 1890-92, when the placing of an Imperial garrison in Gilgit and in the adjoining valleys leading towards the Eastern Hindukush necessitated the construction of a military road fit for laden transport during at least a portion of the year. But the route which it follows, between and over the high ranges separating the Kashmir Valley from that of the Indus, is marked out by nature as the most accessible line of communication from Kashmir to the Dard territories northward, and there is historical evidence to prove its use at an early period.
In the Memoir on the Ancient Geography of Kashmir, accompanying my annotated translation Routes of Kalhana’s Rajataranginz, 1 had occasion to point out the historical interest of the mountain pee be: routes leading from the north shore of the Volur Lake into the part of the Upper Kisanganga Valley. Valley which is known as Guréz!, Only through the valley of the Burzil stream, here joining the Kiganganga from the north, is it possible to gain the high passes which cross the snowy watershed between the Kisanganga and the Indus and give access to Astor and thence to the other Dard Valleys. The Kisanganga itself, though draining an extensive mountain area, flows
: Kalhana’s Riaatarangini, translated, with an Intro- London, 1900, II. pp. 406 sqq.; also note on Rajat. duction, Commentary, and Appendices, by M.A. Stein, vii. 1171. STEIN B
Passes of Trag*bal and Dud?- khut.
‘Northern Gate’ of Kashmir.
2 FROM KASHMIR TO THE PAMIRS [Chap. I
throughout its upper course in an extremely confined valley. To the west, some twenty miles below the junction of the Burzil stream, this valley for a considerable distance becomes an uninhabited and almost inaccessible gorge; it cannot be crossed nor is it reached there by any route from Kashmir. Eastwards again the great massif of the hoary Haramukh (Skr. Haramukuta) peaks, and the glacier-crowned range continuing it towards the head of the Sind Valley, raise an effective barrier between Kashmir and the Kisanganga. It is crossed only by a few difficult mountain tracks, over which even: the hardy Dard hillmen cannot take their ponies, except during brief periods in particularly favourable seasons and then unladen.
The geographical facts here briefly indicated make it quite certain that in ancient times the line of communication between. Kashmir and the Dard territories must have led through Guréz, and over that part of the watershed which separates Guréz from the north shore of the Volur Lake. Starting from a point due south of the main village of Guréz, the summit ridge for about twenty miles westward shows a height of from 12,000 to 13,000 feet above the sea. The route which is followed by the present ‘Gilgit Transport Road’, and which has been in general use for military purposes since Sikh times, crosses the range by the Trag*bal or Razdiangan Pass, at an elevation of close on 12,000 feet. Another route which leads over the Dudekhut Pass, about eight miles further to the east, is distinctly referred to in the Rijatarangini in connexion with an invasion of the Dards from the Kisanganga Valley towards the close of the eleventh century a.p. In my comments on this, and in some other references, I have shown that the position of the hill fort of Dugdhaghata, which the Dards had occupied and which King Harsa of Kashmir vainly attempted to recover, in order to close the route to their inroads, can still be traced on the summit of the Dud@khut Pass to which it has left its name*. Though somewhat less exposed than the Trag*bal route, and evidently preferred in ancient times on this account and as being more direct, the Dud*khut Pass is equally liable to heavy and early snowfall. Like the corresponding portion of the present ‘Gilgit Road’, it can scarcely have been kept open for regular use by laden animals or large bodies of men during more than four months of the year.
Both the Trag*bal and Dud*khut routes ascend the southern slopes of the range from the stream of the Band@por Nullah, which flows into the Volur Lake from the north, and which, among the Brahman population of Kashmir, is still known by its ancient name of Madhumatz. Close to the point in the valley where the two routes diverge lies the village of Band*koth, once the site of a castle (40th, Skr. Aotta) occupied by the ‘Malik’ or feudal chief who in Muhammadan times watched over the passes giving approach to Kashmir from the Kisanganga. It is very probable that in the same neighbourhood there also existed one of those fortified frontier watch-stations which in the time of the Hindu rulers closed all routes leading into the Kashmir Valley, and which, under the designations of dvara, ‘gate’, or dranga, figure so prominently in the Sanskrit Chronicles of Kashmir ®.
It is true that the watch-station of this particular locality is not specifically mentioned by either Kalhana or the later Sanskrit Chronicles. But there can be no doubt that it was known to the Chinese pilgrim Wu-k‘ung, who visited Kashmir from Gandhara, the present Peshawar District, and stayed there during the years 759-763 4.p. The topographical data
2 Compare note on Rajal. vii. 1171. Rajat. i. 122; iii. 227, &c.; J.A.S.B., 1895, Part I, pp. 382 3 The character and historical importance of these sqq. The system of rahdari maintained in connexion with ancient watch-stations or ‘Gates’ of Kashmir have been de- these watch-stations, by which persons leaving or entering monstrated by me in my Memoir on the Ancient Geography the Valley were required to produce special permits or
of Kashmir, § 40 (Raat., Il. pp. 391 sqq.); see also notes passes, survived almost to our own days.
Sec. i] CHINESE HISTORICAL RELATIONS WITH GILGIT 3
regarding Kashmir contained in his itinerary have been fully discussed by me elsewhere‘. Wu-k‘ung correctly described the kingdom of Kashmir as enclosed on all sides by mountains forming its natural ramparts. Only three roads had been opened through them, and these were secured by gates. The road in the east, leading to T‘u-fan or Tibet, corresponds plainly to the present route over the Zoji-La to Ladak and thence to Tibet. By another road, which started from ‘the western gate’ and went to Ch'ien-t‘o-lo or Gandhiara, the route following the Jehlam river in its course to the west is clearly meant. The third road, in the north, is described as reaching into P‘o-lii. Its identity with the route represented in its main direction by the present ‘Gilgit Road’ was beyond all doubt, even before Professor Chavannes’ lucid analysis of the original Chinese records had conclusively proved that the term P‘o-lii or Pu-lu applied to modern Gilgit as well as to Baltistan.
Of the territory into which that route first takes us, one important historical fact can still be established from Kalhana’s Chronicle. Several passages relating to events on this side of the Kashmir frontier show plainly that the hill tract formed by the drainage area of the Upper Kisanganga was occupied, in ancient times as now, by a population belonging to the Dard race®. The tenacity with which the ethnographic watershed has been here maintained during long centuries creates a reasonable presumption that the occupation by Dard tribes of the mountain territories extending north-westwards of Kashmir up to Chitral and the Hindukush likewise dates back to a very early period.
We must attach the more value to this evidence because the indications which Sanskrit texts and the works of classical geographers have preserved for us as to the early seats of the Dards (Skr. Darad, Darada; the Aapddpa of Ptolemy, Dardae of Pliny, &c.), are by no means as exact as we should wish. From the account given by Kalhana of events among the Dards of the Kisanganga Valley near his own time, it may be concluded that this hill tract then formed a small chiefship. The division of the territory inhabited by the Dards into a considerable number of more or less independent principalities has prevailed also during recent periods of their history, and is amply accounted for by geographical features. Mountain ranges of exceptional ruggedness, covered by eternal snows on their summit lines and culminating in some of the highest peaks of the world, render communication between the narrow valleys difficult even for the hardy hillmen who have succeeded during so many centuries in maintaining here their struggle with a harsh climate and a barren soil. Frequently the confined valleys themselves present almost equal difficulties, where they narrow to rock-bound gorges filled completely for considerable periods of the year by snow-fed rivers. The formidable nature of the mountain-barriers in this region is reflected in the notice which Ptolemy records of ‘the Daradrai in whose country the mountains are of surpassing height’*. Nor is there need to search far in the accounts of modern travellers in order to find ample illustration of the difficulties and dangers so graphically described by the old Chinese pilgrims who made their way along the precipitous rock walls of the Dard Valleys’.
I have fully discussed elsewhere the results of that remarkable seclusion which Kashmir enjoyed during a long period of its history, owing to the natural strength of its mountain
* See ‘ Notes on Ou-k‘ong’s account of Kagmir’, in Pro- PP- 431, 435- ceedings of the Imperial Academy, Vienna, 1896, vol. cxxxv. ° See Ptolemy, Geogr., VII. i. 42. pp. 228q.; also Raat, II. pp. 357 sq. ” Compare Fa-hien’s Zravels, translated by Legge, p. 26; * Compare, regarding the Dards and their relations Voyage de Song Yun, transl. by Chavannes, p. 28; Hsiian- with Kashmir, my notes on Réjaé. i. 312; vii. 9113 also II. tsiang’s Sz-yw-hz, transl. by Beal, i. p. 134 sq.
Bi
The Dards of the Kisanganga Valley.
Kalhana’s Daraddesa.
Alberini’s notice of Dard valleys.
4 FROM KASHMIR TO THE PAMIRS [Chap. I
ramparts and the rigorous system of guarding the frontier-passes*. In view of this jealously maintained isolation it cannot surprise us to find that Kalhana’s horizon northward was practically limited by the petty hill state on the Upper Kisangaiga and its tributaries, which he designates as Daraddefa, ‘the Darad land.’ The seat of its chiefs, referred to by the name of Daratfuri7, ‘the town of the Darads’, may have stood at the present Guréz*. This, the main settlement of the valley, was the residence of the Nawabs who ruled the tract until the Sikh conquest. On account of the advantages offered by its position in the widest and most central portion of the valley, Guréz was probably a place of some importance in earlier - times, though among its rude log-huts and near the rubble-built fort no remains of antiquity could be traced by me.
Once beyond this point the guidance of the Chronicler fails us completely for the historical topography of the route. It is, indeed, significant, as I have pointed out elsewhere, that ‘Kalhana, when describing. the home of the “Mleccha” chiefs from the north who, in his own time, invaded Kashmir together with the Darads of the Ki§anganga Valley (viii. 2762-4), can treat us only to details of the mythical geography of the Himalaya regions’’°. Judging from Kalhana’s usual accuracy in matters topographical we may feel convinced that he would not have withheld from us the old Sanskrit designations of Astor, Gilgit, and the other Dard valleys from which the northern allies were undoubtedly drawn, if those territories had been as familiar to him as they are now to the educated Kashmiri.
Considering this restricted knowledge prevailing in the Kashmir of Hindu times, and the complete inaccessibility of Kashmir itself to all Muhammadans, it is remarkable that Albérini, more than a century before Kalhana, should have succeeded in obtaining any information, however scanty, about that northern region. In my Memoir on the Ancient Geography of Kashmir, 1 have shown the value attaching to the chapter of Albériini’s /zaza, which contains his description of Kashmir derived from indigenous sources™. He refers there to the mountains of ‘Bolor and Shamilan’ as visible for two marches on the left of the traveller who enters the Kashmir Valley from the western ‘Gate’, i.e. through the gorge of Baramtla. At present we can trace neither the name Stamilan nor the designation Bhattavaryan given by Albérini to the tribes which inhabit those mountains, and whose ‘king has the title of Bhatta-~Shah’, But the position indicated, and the use of the term Bolor, which has been applied to Gilgit and Baltistan for centuries, plainly show that Albérini’s informant meant the mountain-region of the Dards, forming the confines of Kashmir to the north and north-west. This is confirmed by the subsequent mention of ‘Gilgit, Aswira, and Shiltas’ as the chief places of those tribes ; for here it is impossible to mistake a reference to the modern Gilgit, Hasora, and Chilas ™.
Section II.—CHINESE HISTORICAL RELATIONS WITH GILGIT
Nowhere among the Dards do we meet with written records of any antiquity or other evidence which could throw light on the early history of the region traversed by the ‘Gilgit Road’, Nor are there archaeological remains to offer us guidance; at least I did not succeed in tracing any going back to pre-Muhammadan times until Gilgit itself was reached. But
8 See my introduction to Raaz., I. pp. 30 sq. 1 See Rajat, II. pp. 359 sqq., §§ 12-14.
® Compare my note to Raat. vii. 911; regarding Guréz 2 See Albériini, /zdza, transl. Sachau, i. p. 20; for a and the neighbouring portions of the valley, see Ruzns possible explanation of the term Bhata, see my note, Raat., of Khotan, pp. 15 sqq- II, pp. 363-4.
0 See Rajat. I. p. 31, note 6.
Sec. ii] CHINESE HISTORICAL RELATIONS WITH GILGIT 5
a remarkable course of historical events, for a short period about the middle of the eighth century A.D., invested this barren mountain-land at the extreme north-west confines of India with considerable political and strategical importance to the Chinese Empire. The record preserved of those events in the official Annals of the Chinese T‘ang dynasty gives us a glimpse of Gilgit history at that interesting phase, and strikingly illustrates the value which has always attached to the route connecting Gilgit with Kashmir.
It is only quite recently, and solely through the brilliant researches of Professor Edouard Chavannes, that the Chinese records containing these curious references to Gilgit and the neighbouring mountain-territories have been rendered accessible to the Western student. They have been translated and also elucidated with rare critical acumen in M. Chavannes’ Documents sur les Tou-kiue occidentaux. In the ninth chapter of the last and purely historical part of his work M. Chavannes has given a lucid exposition of the main events which occurred in the territories between the Oxus and the Indus during the period that witnessed the furthest
Gilgit in the Chinese Annals.
extension of Chinese power westwards, from the middle of the seventh to the middle of the ©
eighth century of our era. <A few facts gleaned from this account will suffice to make clear how such a remote region as Gilgit and the adjoining valleys, even as far as Kashmir, was brought within the sphere of Chinese activity.
After the destruction of the empire of the Western Turks (658-659 a.p.) China secured suzerainty over a vast region, which comprised both Eastern and Western Turkestan and extended across the Oxus Valley and to territories south of the Hindukush and within the ancient confines of India*. The rising power of the Tibetans, who from 670 to 692 a.p. occupied Kashgar and the Tarim Basin, combined with dynastic troubles in China itself, reduced Chinese dominion over those distant territories in the West to a more or less nominal condition by the end of the eighth century. But the victorious advance of the Arabs, resumed with fresh vigour during 705-715, under the leadership of Qutayba ibn Muslim, obliged the small states along the Oxus and in Sogdiana to turn to the Imperial Court for protection.
The danger threatening from Muslim conquest was increased by the Tibetans, who, after having been driven back from the Tarim Basin, were now endeavouring to join hands with the Arabs, their allies, across the Pamir region. Against these aggressions from west and south, China during the reign of the energetic Emperor Hsiian-tsung (713-755 A.D.) directed a long series of well-sustained efforts, both diplomatic and military. In the course of the struggle with the Tibetans Gilgit acquired strategic importance; for through it led, then as now, the nearest and most practicable route giving access from the Upper Indus, via Yasin and the Baroghil Pass, to the central Pamirs and the Oxus Valley. It thus became an object of constant and special care for the Imperial Government to keep this gate of invasion firmly closed against the Tibetans, who probably from very early times were in possession of Ladak, and had thence pushed down along the Indus into Baltistan or Skardo, the immediate neighbour of Gilgit on the south-west. |
On examining the notice ‘on the Great and the Little P‘o-lii’ as now presented by M. Chavannes from the T‘ang Annals, we cannot fail to recognize, with the learned translator, that it is this territory of Baltistan which is meant by the term ‘Great P‘o-lii’® It is correctly
* See Chavannes, Zurcs occid., pp. 287 sqq. tification of ‘Great P’o-lii (Po-liu)’ with Ladak, suggested by * Compare the detailed analysis of the Chinese administra- Sir H. Yule, Cathay, p. lxx, and previously also accepted by tive divisions established between the Oxus and Indus, Zurcs me (faaz., II. p. 435), was based on an imperfect extract occtd., pp. 277 Sqq-, 280; also S, Lévi, Journal astat., 1900, of this notice from the T’ang Annals, given by Rémusat, XV. Pp. 305, 315. Nouveaux mélanges asiat., i. p. 194; it cannot any longer be
* See Chavannes, Zurcs occid., pp. 149 sq. The iden- maintained.
Gilgit between Chinese and Tibetans.
‘Great Po- li’ of T‘ang Annals.
‘Little P‘o- li’ of the T’ang Annals.
6 FROM KASHMIR TO THE PAMIRS [Chap. I
described as being right to the west of the Tibetans and adjoining ‘the Little P'o-lii’, i.e. Gilgit, while westwards it had as its neighbour ‘the territory of Wu-ch'ang (Udyana) which belongs to Northern India’. In regard to the latter indication it must be remembered that the ancient Udyana, as shown by the accounts of the Chinese pilgrims and other evidence, comprised not only the whole of the Swat Valley but also the mountains on the right bank of the Indus, even beyond the river’s great bend to the south.
At the time of Hsiian-tsang’s visit (circa 631 a.D.) the frontiers of Udyana in the north-east evidently extended to the Valley of Darél (Za-&-/o), almost opposite Chilas, and possibly even further up the Indus‘. On the other hand, we find to the present day the half-Tibetan Baltis, who have given their name to Baltistan, settled along the Indus to within a short distance above the point where the Gilgit river joins it. Thence toward Chilas and Darél the rocky gorge of the Indus is even more confined and barren than elsewhere, and must in early times have held a very small population. Thus we can fully account for Udyana being named in the Chinese description as the western neighbour of ‘Great P’o-lii’®
We learn from the notice in the Annals that ‘the Great P*‘o-lii’ had become subject to the Tibetans some time before the middle of the eighth century. But both the Annals and the official records preserved in the encyclopaedia 7s@ fu yiian kuec show that previously, in 717 and 721, two successive rulers of Great P*o-lii had acknowledged the suzerainty of China, and received the decree of royal investiture from the Imperial court. We have no means of ascertaining to what extent the ethnic character of the present Balti population, Tibetan in language and largely also in race, is due to this Tibetan occupation. Possibly the close ethnic connexion of Baltistan with Ladak is of far earlier date, and may have facilitated the extension of Tibetan power along the Indus Valley.
The ancient local term reproduced by the P'o-lii of the Chinese records still survives in the designation Paloyo which, as I ascertained on my passage through Gilgit, is applied by the Dards of those parts to their eastern neighbours, the Baltis.
The territory of ‘Little P‘o-li’ has been identified by M. Chavannes with the modern Gilgit *. A short analysis of the topographical and historical data furnished by the Annals proves that this identification is undoubtedly correct, if the term Gilgit is extended so as to include the valleys drained by the Gilgit river in its upper course, and in particular the important mountain tract to the north-west known as Yasin. ‘Little P‘o-lii’ is described as having ‘Great P*o-lii’ or Baltistan on its south-east, at a distance of three hundred li; Ku-shih-mi or Kashmir lay five hundred li to the south of it; while at the same distance to the north there was the town
* See Sz-yu-k7, transl. Beal, i. p. 134. Also Fa-hien’s to west and narrow from north to south,’ as rich in gold and
Travels, transl. Legge, p. 24, speaks of Zo-d¢ (Z"o-leth), which is plainly shown by the context to be identical with Hsiian-tsang’s Za-i-lo, or Darél, as being immediately within the borders of Northern India. This evidently indicates dependence on, or ethnic connexion with, Udyana, to which country the traveller next proceeded and which he further on specifies as the commencement of North India.
5’ The country of Po-lu-lo, which Hsiian-tsang reached after leaving Za-i-lo, by going to the east up the course of the Indus, for a distance of about 500 li, or circ. 100 miles, was manifestly Baltistan. The direction and character of the route (‘by the help of flying bridges and footways made of wood across the chasms and precipices’) distinctly point to this, Also the description of the territory as ‘long from east
silver and possessing a climate continually cold, would apply more strictly to Baltistan than to Gilgit. But it is well to remember that the term ‘Bolor’ or ‘Bolur,’ which Hsiian- tsang’s Po-Ju-lo is undoubtedly meant to represent, had a much wider application, including at one time or other all the mountain tracts on the southern slopes of the eastern Hindu- kush, from Kafiristan to Skardo. Compare, regarding this somewhat loose and once much-disputed term, Sir H. Yule’s note, Marco Polo, i. pp. 178 sq., giving a lucid summary of exhaustive researches, and the judicious remarks of Elias, Tarikh-t-Rashidi, p. 385, note. The origin of the term is doubtful; and so, too, its philological relation to P‘o-li. 6 See Zurcs occid., pp. 150 sq.
Sec. ii CHINESE HISTORICAL RELATIONS WITH GILGIT 7
So-lez, belonging to the territory of Mu-mi, which has been recognized with certainty as the present Wakhan on the Upper Oxus’. A glance at the map shows that the bearings, as well as the relative distances, here indicated are in remarkable agreement with the geographical position of that central and most productive part of the main valley to which the name of Gilgit properly applies.
The historical notices regarding ‘Little P‘o-lii’, rendered accessible by M. Chavannes’ researches, afford proof of the political importance which the Chinese attached to the Gilgit Valley as the main line of communication between the Upper Indus and the Oxus, and of the persistent efforts made by them during the reign of the Emperor Hsiian-tsung (713-755 a. D.) to close this route against Tibetan invasion. At the commencement of that reign the king of Little P'o-lii, Mo-chin-mang, came to pay homage at the court of Hsiian-tsung, who accorded him protection and constituted his country into a military territory called Swé-yiian. ‘Owing to its proximity to the Tibetans his country suffered much from them; the Tibetans declared to him: “It is not your kingdom which we covet, but we wish to use your route in order to attack the Four Garrisons (i.e. the present Chinese Turkestan) ”’ ®,
In 722 a.p. Mo-chin-mang, having been deprived of nine townships by the Tibetans, applied for help. The Imperial Commissioner resident at Pei-ting thereupon declared: ‘ P*‘o-lii is the western gate of [the territories of] the T‘ang dynasty; if P*‘o-lii is lost [to us] then the western countries will all become Tibetan’, and he dispatched a force of four thousand chosen troops, under the prefect of Kashgar (Su-lé), to his help. Mo-chin-mang thus succoured defeated the Tibetans and recovered the nine townships. The same historical work, the Zz#% chch “ung chien, which has preserved these details, records a fresh attack upon P‘o-lii by the Tibetans fifteen years later. On that occasion, 737 a.p., Chinese action took the form of a diversion from the centre of the empire, which resulted in a great defeat of the Tibetans west of the Kuku-Nor and relieved P‘o-lii®.
After Mo-chin-mang’s death ‘Little P‘o-lii’ was ruled in succession by his sons Nan-ni and Ma-hao-lai (Ma-lai-hi). The Imperial edict concerning the latter’s investiture, in 741 A.D., is still extant among the records extracted by M. Chavannes’. He too seems to have died early, and his successor Su-shih-li-chih was won over by the Tibetans, who induced him to marry a Tibetan princess and thus secured a footing in his territory. In consequence ‘more than twenty kingdoms (i.e. little hill states) to the north-west’ of Little P‘o-lii are said by the Annalist to have become subject to the Tibetans. Their customary tribute no longer reached the Imperial court.
7 See Turcs occid., p. 279, also pp. 152, 154, where good reasons have been advanced for the assumption that the name which the Chinese represent by So-/ed was really that of the main branch of the Oxus, now known as Panja. For another mention of this river made in connexion with Lien- yiin, a stronghold which in all probability occupied the position of the present Sarhad at the point where the route from Yasin over the Bardghil Pass strikes the Ab-i-Panja, compare below, p. 8. Whether the town So-lei stood in
the same place seems doubtful.
° See Zurcs occid., pp. 1508q. By the term ‘Four Garrisons’, the territories of Kashgar, Khotan, Kucha, and Kara-shahr, then occupied by Chinese forces, are meant; comp. ibid., p. 113 note.
* See extract given Zurcs occid., p. 151, note 2.
*° See for a translation of this document Zurcs occid.,
pp. 211 sq.
Gilgit in the eighth cen- tury.
Kao Hsien- chih’s ex- pedition into Yasin, 747 A.D.
His march across the Pamirs.
8 FROM KASHMIR TO THE PAMIRS [Chap. I
Section III.—THE CHINESE OCCUPATION OF GILGIT AND THE ROUTE TO KASHMIR
The situation thus created obliged the Chinese Government to make special efforts for the recovery of their lost hold upon the Gilgit Valley. Three expeditions led against ‘Little P'o-lii’ by the Protector of ‘the Four Garrisons’ proved fruitless. But a fourth, entrusted in 747 a.p. by Imperial decree to his Deputy, Kao Hsien-chih, a general of Korean origin, was crowned with complete success. The military operations connected with this expedition deserve to be briefly considered here, in view both of the historical importance of the result and of their interest for the ancient topography of these regions’.
Kao Hsien-chih was specially appointed by the Emperor Hsiian-tsung to take charge of the campaign against the Tibetans in Little P‘o-li, under the title of ‘director of camps’ with a force of ten thousand cavalry and infantry? Starting from An-hsi, the present Kucha, he reached Su-lé or Kashgar in thirty-five days, apparently via Maralbashi, Twenty days later his force arrived at the military post of the Ts‘ung-ling mountains, which occupied the position of the present Tash-kurghan in Sarikol®. The march thence, through the valley of Po-mi or the Pamirs into Shighndn, called by the Chinese ‘the kingdom of the five Shzh-nz’, occupied forty days. |
Kao Hsien-chih then divided his troops into three columns, which were to march by different routes to the attack of the fortress of Lien-yiin held by the Tibetans. The localities by which the routes for two of these columns are indicated (Pe-ku, ‘the northern gorge’, and Chith-fo-t'ang, ‘the hall of the red Buddha’) cannot be traced at present; but the route of the — third column, under Kao Hsien-chih himself, which is described as leading through the kingdom of Hu-mi or ‘Wakhan, undoubtedly lay up the valley of the Oxus. Lien-yiin itself, where the several columns effected their junction on the appointed day, the thirteenth of the seventh month, had the river of P‘o-lei or So-lei in front of its walls, M. Chavannes has shown good reason for assuming that this name designates the main branch of the’Oxus now known as the Ab-i-Panja, and that Lien-yiin occupied a position corresponding to the present village of Sarhad, but on the opposite or southern bank of the river. From Sarhad starts the well- known route which leads southwards over the Bardghil Pass to the headwaters of the Mastij river, to this day representing the easiest line of access from the Upper Oxus to Chitral as well as to Gilgit.
The position of this locality also readily accounts for a concentration of the attacking forces from three different directions, such as the Chinese record indicates. Besides the main column ascending the Ab-i-Panja Valley from Kala Panja and the lower part of Wakhan, another might with advantage have co-operated from the opposite direction by descending from the headwaters of the Ab-i-Panja. These could be reached without serious difficulty from Tash-kurghan over the Taghdumbash Pamir and the Wakhjir Pass, or via the Naiza-tash Pass
1 The details of Kao Hsien-chih’s memorable march over south of, the Hindukush. These, therefore, have required
the Pamirs and across the Hindukush are contained in his biography, which has been reproduced from the T‘ang Annals and annotated by M. Chavannes in his Zurcs occid., pp. 152 8qq. The learned translator deserves special credit for having recognized that Kao Hsien-chih’s route lay over the Bardghil and Darkét Passes. He did not, however, attempt to identify the several localities mentioned along the route across, and
special discussion here.
2 The biography notes particularly that ‘at that period the foot-soldiers all kept horses (i.e. ponies) on their own account’, The use of these animals for the transport of food supplies must have increased the mobility of the Chinese force.
8 Compare Zurcs occid., p. 125; also below, pp. 28, 36.
Sec. iii] CHINESE OCCUPATION OF GILGIT AND ROUTE TO KASHMIR 9
and the Little Pamir. The third column, which used ‘the northern gorge’, may possibly have followed the Shakh-Dara, or Ghund-Dara Valley of Shighnan, to the Oxus branch coming from the Great Pamir or Victoria Lake, and may, by subsequently crossing the watershed towards the Ab-i-Panja, have descended upon Sarhad from the north. It is highly probable that the division of the Chinese forces, as well as Kao Hsien-chih’s move into the distant Shighnan, were intended to reduce the difficulties about supplies and transport which for so large a body of troops must have been as serious then, in the inhospitable Pamir region, as they would be now.
The topographical indications contained in the description of the battle by which the Chinese general rendered himself master of Lien-yiin and of the route it was intended to guard, fully confirm these identifications. The stronghold itself was occupied by a thousand men, and the river in front being in flood offered a serious obstacle. Kao Hsien-chih, however, with a picked body of mounted men, succeeded in crossing unopposed and without loss. He at once led his troops, whom this success had filled with confidence, to the attack of the main force of the enemy, eight or nine thousand in number. They were posted fifteen li, or about three miles, to the south, where advantage had been taken of the mountainous ground to erect palisades. The Chinese general, having gained the heights, i.e. turned the fortified line, engaged the defenders in a fight which ended in their complete defeat with heavy loss and the precipitate flight of the survivors during the night. It is evident that this battle was fought at the entrance of the defile which leads to the Baroghil Pass +.
Kao Hsien-chih decided to leave behind in his camp certain high officers who were opposed to a further advance, together with three thousand men worn out by the previous hardships, and ordered them to guard Lien-yiin. With the rest of his troops he pushed on, and after a three days’ march arrived on the summit of ‘Mount T‘an-chii’; ‘from there downwards there were precipices for over forty li (circ. eight miles) in a straight line.’ By a trick Kao Hsien-chih prevailed upon his wavering troops to effect their descent into the valley, and after three more marches was met by ‘the barbarians of the town of A-nu-yiieh’ offering their submission. The next day he occupied A-nu-yiieh, where the commander of his advance guard had previously entrapped the five or six chief dignitaries of the king of Little P‘o-lii who were supporting the Tibetans. These he executed, and then hastened to break the bridge which led over the Sp-yi river at a distance of sixty li (circ. twelve miles) from A-nu-yiieh. ‘Scarcely had the bridge been destroyed in the evening when the Tibetans, on foot and mounted, arrived in great numbers, but it was then too late for them to attain their object. This bridge had the length of an arrow shot; it had taken a whole year to construct it. It had been built at the time when the Tibetans, under the pretext of using its route, had by deceit possessed themselves of (Little) P‘o-lii.’ !
It is easy to trace on the map the successive stages of the Chinese general’s progress. All details recorded of it agree accurately with the route that leads over the Bardghil saddle (12,460 feet above the sea) to the sources of the Mastiij river, and then, crossing south- eastwards the far higher Darkot Pass (15,200 feet), descends along the Yasin river to its junction with the main river of Gilgit. Three days are by no means too large an allowance of time for a military force accompanied by baggage animals to effect the march from the Oxus
* The account of this encounter vividly recalls the attempts of the valley, leaving it free to their opponents to occupy the made by Tibetan strategy in 1904 to bar the advance of the commanding heights on either side. Considering how scanty Tibet Mission Force near Guru and again on the Karo-La. timber must at all times have been about Sarhad, it appears In each case the main Tibetan force was content to await probable that by the ‘ palisades’ mentioned in the translation the attack behind stone walls erected across the open ground walls or sangars constructed of loose stones are really meant.
STEIN C
Capture of Lien-yiin and the Barodghil Pass.
Chinese invasion of | Yasin.
Crossing of Darkot Pass (Mt. Z*an- chit).
Bridge over the Gilgit
river.
Return of Kao Hsien- chih.
10 FROM KASHMIR TO THE PAMIRS [Chap. I
to the summit of the Darkot Pass, considering that the ascent to the latter lies partly over the moraines and ice of a great glacier. The Darkot Pass corresponds exactly in position to the ‘Mount T‘an-chii’ of the Annals, and possibly preserves the modern form of the name which the Chinese transcription, with its usual phonetic imperfection, has endeavoured to reproduce. The steep southern face of the pass, where the track descends close on 6,000 feet between the summit and the hamlet of Darkot, over a distance of about five or six miles, manifestly represents ‘the precipices for over forty li in a straight line’ which dismayed the Chinese soldiers on looking down from the heights of Mount T‘an-chii.
From the foot of the pass at Darkot a march of about twenty-five miles brings us to the village of Yasin®, the political centre of the valley, in which we can safely recognize the town of A-nu-yiieh. Not only does the time indicated for Kao Hsien-chih’s march to A-nu-yiieh fully agree with the three marches which are reckoned at the present day for the journey between Yasin and the watershed, but it is also evident that we have in A-nu-yiieh a fairly accurate rendering of the name Avniah or Arniya by which Yasin is known to the Dards of the Gilgit Valley *
The correctness of this identification is confirmed by the reference to the bridge over the river So-yz, which is mentioned in the above account as sixty li beyond A-nu-yiieh. This river So-yi is named in the Annals’ notice of Little P'‘o-lii as the one on which Yeh-to, the capital of the kingdom, stood, and there can be no doubt that the main Gilgit river is meant. Now a glance at the map shows that, descending the valley from Yasin, we reach the Gilgit river at a distance of about twelve miles, which agrees exactly with the sixty li of the Chinese record’. It is equally clear that as the further route towards Gilgit proper and Baltistan leads along the right or southern bank of the Gilgit river, the Tibetan reinforcements hurrying up from that direction could not arrive at Yasin without first crossing the main river. Hence the importance to the Chinese invaders of destroying the bridge ®.
Once in possession of the Yasin Valley, the Chinese general induced the king of Little P‘o-lii to surrender, and subsequently pacified the whole territory. Leaving behind a garrison to hold it, and taking along his royal prisoner he returned within two months to Lien-yiin. Then he regained the Pamir, whence he dispatched news to the Imperial Court announcing his victory. That the fame of Kao Hsien-chih’s remarkable expedition must have spread far and created a deep impression, is shown by a closing remark of the Annals, the historical significance of which M. Chavannes has duly noted : ‘Then the Fu-din (Syria), the Za-shzh (i. e. Tazi or Arabs) and seventy-two kingdoms of divers barbarian peoples were all seized with fear and made their submission ®.’
It was, no doubt, the greatness of the natural obstacles overcome in the course of the victorious march across the Pamirs and the Hindukush which gave to this success of the Chinese
5 See Biddulph, Z7ibes of the Hindoo Koosh, p. 55.
® See Biddulph, Zribes of the Hindoo Koosh, p. 62 (where ‘Arinah’ is a misprint for Arniah). The omission of an equivalent of 7 in the Chinese rendering of this name has its parallel in the Chinese A-Ast-Jan, ta-kan, which reproduce the Turki names Arslan, farkan; comp. Turcs occid., pp. 239, 317-
7 Compare for this distance also Zrebes of the Hindoo Koosh, p. 55-
§ All communication between Yasin and Gilgit is now carried on over the bridge which spans the Gilgit river near Giipis at the mouth of the Yasin Valley. A wire suspension
bridge, constructed by Col. Aylmer, R.E., has, since the Chitral campaign of 1895, replaced the difficult rope bridge which previously existed here. Like all similar rope bridges between Kashmir and the Hindukush, the old bridge con- sisted of twigs twisted into ropes, and required frequent repairs to retain even a moderate degree of safety. — It is very probable that the Chinese record, by its ‘ pont de rotin,’ intends to designate a bridge of this peculiar description. The modern fort of Giipis marks the strategic importance of the position. 9 See Zurcs occid., pp. 151, 154, 296.
Sec. iii] CHINESE OCCUPATION OF GILGIT AND ROUTE TO KASHMIR 11
arms its peculiar lustre. And it is from more than one point of view interesting to find that other serious difficulties, which from the physical character and position of the country must always beset military operations in the Gilgit region, are reflected in a contemporary notice of the Chinese records.
Among the important diplomatic documents embodied in the great encyclopaedia 7s‘é fu ytian kuet, and now rendered accessible by M. Chavannes’ researches, there is found the record of a representation which Shih-li-mang-ch‘ieh-lo, the ruler (jadgu) of T‘u-ho-lo or Tokharistan on the Oxus, addressed in 749 A.D. to the Imperial Court. Its contents throw light on the situation prevailing in the Gilgit Valley immediately after Kao Hsien-chih’s expedition”. The prince, whose territory comprised the present Badakhshan, with some tracts on the northern bank of the Oxus", complains of a neighbouring hill-state called Chieh-shuai', which, relying on the protection of its high mountains, had allied itself with the Tibetans. Its chief ‘knows that the territory of P*o-lii is limited, its population dense; that the area available for cultivation is small, and consequently when garrison troops are placed there, the supplies fail. It then becomes necessary to purchase salt and rice in Kashmir (Ku-shth-mz) and it is thus that the difficulty is met. The traders’ caravans must, on going and coming back, all pass by the kingdom of Chieh-shuai; its king has therefore accepted the presents offered by the Tibetans, who claimed to establish a stronghold in his territory with a view to getting possession of the important route that leads into P'o-lii. Since Kao Hsien-chih opened up P*o-lii, there are 3,000 more troops there, and P‘o-lii has been crushed by it. The King of Chieh-shuai, in alliance with the Tibetans, has taken advantage of the exhausted condition of P‘o-lii and decided to invade it’
Chinese garrison in Little P‘o-lii.
In order to meet this danger, the ruler of Tokharistan then proposes a bold plan of Expedition
operations, which, if supported by the Imperial authorities, would allow him to conquer ‘the Great P’o-lii’ and the countries east of it. This would bring him in a straight line to Khotan, Kara-shahr and beyond, where the Tibetans could then no longer maintain themselves. Besides the dispatch of Chinese troops into Little P‘o-lii, he solicits that the King of Kashmir, as a loyal ally of China, be encouraged by an Imperial edict and special honours to lend to the enterprise the aid of his great resources in troops and supplies. The Emperor is said to have responded to this appeal of the T‘u-ho-lo prince, and an edict preserved by the same source records the investiture in 750 A.D. of Su-chia as king of Chieh-shuai in place of his rebel brother P‘o-t*é-mo ¥.
A briefer notice of these events, extracted by M. Chavannes from another work, the Lz chih tung chien, also mentions the defeat, in 750 A.p., of Chieh-shih by Kao Hsien-chih, the capture of its chief P‘o-t’é-mo, and the installation of Su-chia in his place ™.
Chinese intervention succeeded on this occasion in relieving Gilgit and Tokharistan from the threatening invasion of the Tibetans. But it is certain that the complete defeat of Kao Hsien-chih by the Arabs in the year following (751 a. p.) and the consequent rapid decline of the Imperial power in the ‘Western Countries’ must have forced the Chinese to abandon
See Turcs occid., pp. 214 8q.; comp. also p. 296 on the 2 The name thus transcribed, o ail » appears in a historical import of this document. : : sah Lens . ‘
4 T"u-ho-lo, a great and important territory, then largely a Te eae hee #6) Bi eroticy inhabited by the descendants of the Hephthalites or White Huns, has long been identified with the Tokharistan of early Muhammadan writers; compare, e.g., Sir H. Yule, 7.R.A.S., N.S., vi. pp. 94 sqq.; also my note on Réja/. iv. 166. The valuable data concerning T‘u-ho-lo, contained in the Annals and other Chinese records of the T‘ang period, have been fully discussed by M. Chavannes, Zurcs occid., pp. 155 sqq-
shu; see Zurcs occid., pp. 158 sq., and below, p. 13.
8 See Chavannes, Zurcs occid., pp. 215 sq.; also the passage of the T‘ang Annals, ibid. p. 158, mentioning the defeat of the Chieh-shih people who were endeavouring to bring about an attack on Tokharistan by the Tibetans,
™ See Zurcs occid., p. 214, note 2.
against Chieh-shuai (750 A.D.).
Supplies for Gilgit garrison,
Transport of supplies from Kashmir.
12
FROM KASHMIR TO THE PAMIRS [Chap. I
their position in Gilgit, and the relations maintained through it with such distant territories as Kashmir, Kabul, and Udyana™.
The record just summarized is the last glimpse we obtain of Gilgit from Chinese sources. It is hence particularly gratifying to find so clear a reference to that great difficulty attending any military occupation of the Gilgit Valley, its dependence for supplies on Kashmir. The situation there described as resulting from the presence of a Chinese garrison must, to those acquainted with the modern history of Gilgit, vividly recall the efforts and sacrifices which the maintenance of a military force from the Kashmir side has entailed ever since the Sikhs, soon followed by the Dogras, established their first footing in the Dard mountains across the Indus’. The Chinese record supplies us, in fact, with striking evidence how little the lapse of a thousand years and momentous political changes can affect the conditions of life and communication prevailing in these mountain regions, or the aspects of military activity which equally depend on unchanging geographical features.
The difficulties which the letter of the T‘u-ho-lo ruler so graphically represents, and of which the Tibetans were scheming to take advantage, are exactly those with which the Kashmir rulers, and in more recent years the military authorities of the Indian Government, have had to contend in their occupation of Gilgit. Though the Dogra force employed in Gilgit and on the line of communication leading to it, probably never much exceeded 3,000 men, their maintenance, being wholly dependent on food supplies brought from Kashmir, involved a severe strain on the Maharaja’s resources, and often led to serious suffering and loss of life among his Kashmirian subjects!’. The very restricted area available for cultivation in the Gilgit Valley renders it practically impossible to feed a garrison on local produce; nor can surplus supplies be secured from any base nearer than Kashmir. The barrenness of the mountain region which the route from Kashmir traverses for a distance of over 200 miles would alone suffice to make the regular transport of supplies a matter of great difficulty ; for whether pack animals or men are employed, the food needed by them in transit has likewise to be provided from Kashmir. In addition, account must be taken of the natural obstacles presented by the two high ranges which have to be crossed en vowde, and the passes of which, owing to the heavy snowfall, rarely remain open for laden traffic during more than four months of each year.
The placing of a small Imperial garrison in Gilgit in 1890 was rapidly followed by the construction of the ‘Gilgit Transport Road’, with all the resources of modern European engineering, and by the introduction of systematic transport arrangements under the control of the Indian Commissariat Department’, These changes have made it possible to dispense altogether with human labour for transport purposes, and have greatly reduced the risk with which the use of this route, whether for the annual provisioning of Gilgit, or on occasion of
the adjoining Dard valleys, and to extend its political control to the Hindukush watershed, publications on the Hunza
4% Compare for these diplomatic relations maintained during the first half of the eighth century, M. Chavannes’
lucid exposition, Zurcs occ#d., pp. 295 Sq.; as to the events connected with Kao Hsien-chih’s defeat and the collapse of Chinese rule in Western Turkestan, ibid. pp. 297 sqq., also pp. 142 sqq.
16 An excellent sketch of the operations of the Sikhs and Dégras which brought Gilgit and the adjoining Dard tracts into dependence on Kashmir, and thus within the sphere of Indian political interests, is given, up to the year 1875, by Drew, /ummoo and Kashmir Territories, pp. 437 sqq. For the events which have since led the Indian government to assume a more direct share in the occupation of Gilgit and
(1891) and Chitral (1895) campaigns may be consulted.
1 For characteristic facts concerning the hardships which the requisitions of carriage for transport to Gilgit inflicted upon the village population of Kashmir until the construction of the ‘Gilgit Road,’ compare Sir Walter Lawrence, Valley of Kashmir, p. 413.
18 As to the working of these arrangements and the curious manner in which they recall to the historical student certain features of the watch kept in ancient times over these mountain-routes of Kashmir, compare my Ruins of Kho/an,
pp. 12 sq.
Sec. iii] CHINESE OCCUPATION OF GILGIT AND ROUTE TO KASHMIR 13
military emergencies, was always beset from breaks of the road, premature snowfall, and similar incidents of Alpine travel. Yet a journey along ‘the Gilgit Road’, or the perusal of an account of its marches as given in my Personal Narrative, will suffice to demonstrate how serious even now are the physical difficulties presented by the task of keeping a military force in Gilgit supplied from so distant a base and over such mountains.
In the facts connected with the present occupation of the Gilgit Valley we have the best illustration of the reasons which rendered it vital for the Chinese hold over ‘Little P*‘o-lii’ that the route towards Kashmir should be kept absolutely safe and open for food convoys", We find the same fact attested with equal clearness in the record which the T'ang Annals, in their notice on Kwu-shth-mi or Kashmir have preserved of the letter addressed to the ‘ Celestial Kagan’, the Chinese Emperor, by King Mu-to-pi, i.e. Muktapida, on his succession to the Kashmir throne (733 A.D.)*°. Muktapida, who requests his investiture by Imperial decree, as accorded before in 720 a.D. to his brother and predecessor Candrapida (Chén-t'o-lo-pi-li), particularly claims for his kingdom the merit of having, on the occasion when the troops of ‘the imperial Kagan’ arrived in Little P‘o-lii, come to their assistance by the dispatch of convoys of supplies, though their number amounted to two hundred thousand men*. It is not certain which Chinese expedition into the Gilgit region is alluded to here; but it is evident that the situation in which that service was rendered must have closely resembled the one prevailing in the Gilgit Valley immediately after Kao Hsien-chih’s success. Perhaps Muktapida’s reference, as suggested by M. Chavannes, is to the operations by which the Chinese in 722 a. p. relieved Mo-chin-mang, king of Little P*o-lii*.
It remains to locate that hill-state of Chieh-shuai, which by its alliance with the Tibetans so seriously endangered the Chinese position in Gilgit. M. Chavannes has not attempted to identify it, nor does the letter of the Tokh§ristan ruler which reports that alliance, furnish evidence as to the situation of ‘Chieh-shuai, except such as is conveyed in the statement that it could block the route of supplies from Kashmir to the Gilgit Valley. But some further indications, I believe, may be gleaned from another Chinese record which M. Chavannes’ researches have rendered accessible. The notice on the T‘u-ho-lo country, or Tokharistan, extracted by him from the T‘ang Annals, mentions an attack planned on that territory by the Chieh-shih people in alliance with the Tibetans, and its frustration through the help of Chinese troops?*, As the latter is said to have been secured upon the application of the Jabgu Shih-li-mang-ch‘ieh-lo, the same whose letter of the year 749 a.D. we have already considered, it is certain that the Chieh-shuai of that document and Chieh-shih mentioned here are identical **,
Now the territory of Chieh-shih is distinctly described as bordering on Tokharistan, and
this makes it highly probable that M. Chavannes is right in recognizing it in the mountain tract which a subsequent passage of the notice of Tokharistan mentions under an abbreviated
«Considérant que le roi de Kou-che-mi (Cachemire) »® See Zurcs occid., pp. 166sq. The date of Muktapida’s a été fidéle et loyal envers les Chinois, qu'il a en outre investiture is supplied by the imperial decree reproduced by beaucoup de soldats et de cavaliers, que son territoire est M. Chavannes, ibid. p. 209, from the 75 fu yiian kuet, vaste et que la population y est dense, que les vivres y sont *1 This passage had been erroneously rendered in A, Ré- en abondance, j’espére tout spécialement que la bonté musat’s MWélanges asiai., i. p. 196, where the notice of the impériale conférera au roi de Kou-che-mi (Cachemire) un T‘ang Annals concerning Kashmir was first translated. My édit écrit pour l’encourager, qu’on lui donnera des vétements reference to the Chinese data about Muktapida, in Raa. et des présents, ainsi que des ornements précieux et des iv. 126, note, should be rectified accordingly. ceintures pour faire que (ce roi) soit touché et reconnaissant 2 See Zurcs occid., p. 167, note 2; and above, p. 7. de la bonté sainte (de l’empereur) et qu’il redouble de fidélité 8 See Zurcs occid., p. 158. et de loyauté.’ From the letter of the Tokharistan chief, * For this very slight graphic difference between the two
quoted p. 11; see Chavannes, Zurcs occid., p. 215. transcriptions, see above, p. 11, note 12,
Assistance of Kashmir rulers.
Position of Chieh-shuai (Chieh- shih).
Shé-mi identified with Kafiristan.
14 FROM KASHMIR TO THE PAMIRS [Chap. I
form of the name, Chieh 2%, This tract of Chieh (vecte Chieh-shih) lay ‘in the middle of the Ts‘ung-ling mountains; to the west and south it is bordered by (the territory of) Shé-mi; to the north-west are the I-ta or Hephthalites’. As the T’ang Annals explicitly state of Tokharistan that it was the seat of the Hephthalites**, and as the portion of Tokh§aristan south of the Oxus is undoubtedly represented by the province of Badakhshan, it is plain that Chieh-shih or Chieh must be looked for to the south-east of the latter, i.e. in the direction of the valleys drained by the Chitral river.
The reasons which make me inclined to identify Chieh-shih (Chieh-shuai) with the portion of the main Chitral Valley, properly known as Kashkar or Chitral, are briefly these. The territory of Shé-mi f4¢ i}, which is said to border Chieh-shih on the west and south, is also mentioned under the identical name in the account of the Buddhist pilgrim Sung Yiin, and in the corre- sponding notices which the Wei Annals have preserved from the record of his fellow-pilgrim Hui-shéng. After a stay in the country of the Yeh-tas or Hephthalites, they passed (519 A. D.) into the small mountain tract of Po-chih, and hence into the territory of Shé-mi. There they gradually emerged from the Ts‘ung-ling mountains, and proceeded to Udyana which lay south of Shé-mi®.. The position indicated for Po-chih (to the south-west of Wakhan) and certain features mentioned of its mountains show clearly that it comprised the headwaters of the Varduj or Kokcha river, south of Zebak and towards the Hindukush watershed. From there it is possible to reach by a number of passes the cluster of valleys to the south of the great snowy range which since mediaeval times has been known by the general name of Kafiristan. Owing to the constant war waged against its inhabitants by the Muhammadan hill-states around, this great alpine region has remained more or less a derra zucognita until quite recent times. But there is nothing to prove that those valleys were similarly closed to traffic in earlier periods, while it is certain that a route leading down the easternmost of them to the Kinar or Chitral river and hence across Dir into the Swat Valley would form a shorter and in all probability an easier line of communication than a route crossing one of the high passes east of Lake Dufferin
and thence descending to Kashkar or Chitral proper **.
*° See Zurcs occid., p.159, with note 3 concerning the varying Chinese representations of the local name.
*° See Turcs occid., pp. 157 sq.
7 For an accurate translation of the passages mentioning Shé-mi in Sung Yiin’s account and in the Wei Annals see now Chavannes, Voyage de Song Yun, p. 28. Dr. Marquart, EranSahr, pp. 244 sq., had already rectified some serious mistakes in Beal’s version, Sz-yu-k7, pp. xcii sq. Compare also M. Chavannes’ abstract, Zurcs occid., p. 159, note 4.
*8 It appears to me that a route such as I suppose Sung Yiin to have followed on his journey from Badakhshan to Udyana or Swat would help to solve a geographical difficulty in Marco Polo’s narrative which has puzzled no less an authority than Sir Henry Yule. In Chapter xxx, Marco Polo tells us ‘that ten days’ journey to the south of Badashan there is a province called Pashaz, the people of which have a peculiar language, and are Idolaters of a brown complexion. They are great adepts in sorceries and the diabolic arts. They are a pestilent people and a crafty; and they live upon flesh and rice. Their country is very hot.’ The traveller then proceeds to tell us ‘of another country which is seven days’ journey from this one towards
the south-east, and the name of which is Keshimur (Kashmir)’ ; see Yule, Marco Polo, i. p. 164.
Sir Henry Yule was undoubtedly right in assuming that Marco Polo had never personally visited these countries and that his account of them, brief as it is, was derived from hearsay information about the tracts which the Mongol partisan leader Nigtdar had traversed, about 1260 a.p., on an adventurous incursion from Badakhshan towards Kashmir and the Punjab. In chapter xviii, where the Venetian relates that exploit (see Yule, Marco Polo, i. p- 98, with note, p. 104), the name of Pashai is linked with D7r, the territory on the Upper PanjkGra river, which an invader wishing to make his way from Badakhshan into Kashmir by the most direct route would necessarily have to pass through.
The name Pashaz is still borne to this day by a Muham- madanized tribe closely akin to the Siah-pdsh, settled in the Panjshir Valley and in the hills on the west and south of Kafiristan. It has been very fully discussed by Sir Henry Yule (ibid., i. p. 165), who shows ample grounds for the belief that this tribal name must have once been more widely spread over the southern slopes of the Hindukush
Sec. iii] CHINESE OCCUPATION OF GILGIT AND ROUTE TO KASHMIR 15
If the territory of Shé-mi be identified with Kafiristan, we obtain a satisfactory explanation why, in the notice on Chieh (Chieh-shih), Shé-mi is mentioned as bordering the latter both on the west and the south. The Siah-pdsh or ‘Kafir’ tribes, which have given to Kafiristan its mediaeval and modern designation, inhabit not only the valleys due west of Chitral, but supply also the bulk of the population in the Valley of the Kinar or Chitral river itself, along that little-known portion of its course from Kala Dardsh southwards to Asmar. These settlements, though Muhammadanized under Chitrali or Pathan rule, form a living proof of the fact that Kafiristan originally included a considerable portion of the main valley due south of Kashkar ”.
The open and fertile part of the main valley, containing the large villages which bear collectively the name of Chitral and form the political centre of the Kashkar or Chitral State, answers remarkably well to the description given in the T‘ang Annals of the mild climate and rich produce of Chieh (Chieh-shih). And in view of the topographical arguments already adduced for this identification, we need not hesitate to suggest also that it was the local name Kashkar, or an earlier form of it, which the Chinese endeavoured to reproduce by Chieh-shih or Chieh-shuai. The application of the term Kashkar to the territory of Chitral is well attested from Muhammadan sources, and its use is still current throughout those regions®, Chieh-shih, as an attempt to represent Kashkar by Chinese sounds, would have a parallel in the name Chiia-sha which Hsiian-tsang gives to the present city and oasis of Kashgar, in Chinese
Turkestan 2,
as far as they are comprised in the limits of Kafiristan. If the great commentator nevertheless records his inability to account for Marco Polo’s application of ‘the name Pashai to the country south-east of Badakhshan’, the reason of the difficulty seems to me to lie solely in Sir Henry Yule’s assumption that the route heard of by the traveller, led ‘ by the Dorah or the Nuks4n Pass, over the watershed of Hinduktsh into Chitr4l and so to Dir’.
Though such a route via Chitral would, no doubt, have been available in Marco Polo’s time as much as now, there is no indication whatever forcing us to believe that it was the one really meant by his informants. When Nigiidar ‘ with a great body of horsemen, cruel unscrupulous fellows’ went off from Badakhshan towards Kashmir, he may very well have made his way over the Hindukush by the more direct line that passes to Dir through the eastern part of Kafiristan. In fact, the description of the Pashai people and _ their country, as given by Marco Polo, distinctly points to such a route; for we have in it an unmistakable reflex of characteristic features with which the idolatrous Siah-pdsh Kafirs have always been credited by their Muhammadan neighbours.
It is much to be regretted that the Oriental records of the period, as far as they were accessible to Sir Henry Yule, seem to have retained only faint traces of the Mongol adventurer’s remarkable inroad. From the point of view of Indian history it was, no doubt, a mere passing episode. But some details regarding it would possess special interest as illustrating an instance of successful invasion by a route that so far has not received its due share of attention.
* Compare Biddulph, Aindoo Koosh, pp. 64 sq.
*° Compare Biddulph, Hindoo Koosh, pp. 89 sqq.; Raverty, Notes on Afghanisién, 1888, pp. 152 sqq.
*. In regard to the above proposed identification of Chieh- shih (Chieh-shuai) with Chitral, it is necessary to consider briefly some other Chinese geographical notices which have previously been assumed to refer to Chitral. Both Cunningham and V. de Saint-Martin had expressed the belief that the mountainous territory of Shang-mi is] aap which Hstian- tsang describes as situated beyond a great range to the south of Ta-mo-hsi-t‘ieh-ti or Wakhan, was identical with the Chitral Valley (see /.A.S.B., xiv. p. 433, and Saint-Martin, A/émorre analyt., p. 426). Their suggestion has received the weighty support of Sir Henry Yule, who observes that the yellow arsenic or orpiment mentioned in Hsiian-tsang’s account is still a characteristic product of Chitral (/.2.A.S., N.S., vi. 114). The further fact that Shang-mi, with the alternative name of Chii-wet, is mentioned in a brief notice of the T‘ang Annals first extracted by A. Rémusat, did not escape Sir Henry Yule’s attention. But without an exact and reliable rendering of the passage, as now supplied in M. Chavannes’ work (Zurcs occtd., p. 129, note 2), it was impossible to perceive that the name had in reality a much more restricted appli- cation.
The notice of the Annals tells us: ‘ Chd#-we7 1H. va is also called Shang-mz'; its capital is in the town of A-shé-yi- shth-to; it is situated amidst the great snowy mountains, north of the river of P‘o-lii. This country is cold; it produces the five cereals, wine and pomegranates. During the winter people live in caves. The inhabitants of this kingdom have always assisted the Little P‘o-lii in spying out the Middle Kingdom (China) ’.
The river of P’o-lii must be the Gilgit river, and a glance at the map shows that the territory meant by Ché-we7' or Shang-mt corresponds exactly to the present Mastiij and the
Chieh-shih identified with Chitral.
Chilis and the Kash-
mir—-Gilgit route,
16 FROM KASHMIR TO THE PAMIRS [Chap. I
It was, as we have seen, the danger threatening the route between Kashmir and Gilgit which induced the Chinese to lend their aid against the hill-state of Chieh-shih. Chitral seems distant from that route; yet it does not need elaborate argument to prove that the co-operation of the Chitral chief would have made it far easier to the Tibetans to close that route of supplies for Gilgit, and thus to render the Chinese position untenable. The route followed by the ‘Gilgit Road’ is well protected against inroads from the east by the height and heavy snowfall of the ranges that flank it on that side, and the uninhabitable nature of the elevated valleys and plateaus between them*, The great glacier-crowned spurs descending from the Nanga- Parbat massif and the inaccessible character of the gorges that drain them, form a similar bulwark westwards as far as Astor; but it is different beyond that place where the route descends to the Indus.
From the adjoining territory of Childs, lower down on the Indus, raids upon the route | are possible both along the river itself and by certain tracks leading over the northernmost spurs of Nanga-Parbat. Such irruptions into the Astor Valley were frequently made by the Dards of Childs, who have always been renowned for their bravery and love of plunder, and were sometimes extended as far as the Kisanganga Valley, Baltistan, and northward into Gilgit **. These raids continued up to the time of the Dogra occupation of Gilgit, and were only stopped when Maharaja Gulab Singh’s troops, about 1851, succeeded after great difficulties in invading
Childs and reducing its chief stronghold.
Yarkhin Valley. This valley, through which flows the main feeder of the Chitral river, lies due north of the headwaters of the Gilgit river near the Shandur Lake ; and the name K/d, by which it is known below Mastiij, its chief place, is manifestly the modern derivative of the name intended by the Chinese transcription C/i-wei (compare for the name Kho, which is also given to the largest section of the Dard population of Kashkar Bala, Biddulph, Hzmdoo Koosh, pp. 59, 62). Further local research may yet help us to trace also the name of the capital A-shé-yii-shth-to, as recorded in the Annals.
The fertile and thickly populated Mastij, with some adjoining parts of Kashkar Bala or Upper Kashkar, has, under the still ruling Khushwakte family, enjoyed independence from Chitral or Kashkar proper for a considerable period. This division, which is partly accounted for by geographical features, appears to me reflected in the distinction between Shang-mi (or Chii-wei) and Chieh-shih as indicated by the notices of the T‘ang Records. The name Shang-mi_ is, I believe, properly applied only to the valleys drained by the Mastiij branch of the Chitral river. But we could well under- stand, from analogous cases elsewhere, how Hsiian-tsang, who did not personally visit this region, could have been led by his informants to comprise under this designation the whole Chitral valley, including that lower part which formed the separate territory of Kashkar or Chieh-shih. In the first place, Mastiij or Shang-mi lay nearest to the route which the pilgrim followed through Wakhan. It must further be remembered that both Mastiij and Kashkar may at the time of his journey have been under one ruler, as Chitral traditions record for more recent periods (see for such traditions retain- ing curious traces of earlier historical events, Hindoo Koosh,
pp. 149 sqq-).
But the Childsis preserved their independence
Dr. J. Marquart, in his exhaustive monograph on the historical geography of Tokhiristan, has likewise discussed the Chinese notices concerning Shang-mi, with his usual — learning and thoroughness. But the fresh data contained in M. Chavannes’ publication were not then before him, and he could thus not give adequate attention to the geographical reasons against the identity of Shang-mi and Sung Yiin’s Shé-mi as assumed by him. On the other hand, he rightly emphasizes the fact that the pilgrim Wu-k‘ung correctly names Chii-wet immediately after Hu-mi or Wakhan. [Since the above note was written the ‘Errata supplémentaires’ (litho- graphed) which M. Chavannes has added to his Zures occtd. (1904), show that the distinguished Sinologist has also recognized the identity of Chii-wei or Shang-mi with ‘la région de Mastoudj et Tchitral’, He previously located Chii-wei in Yasin; see Zurcs occid., p. 129, note 2.|
It may be added that Wu-k‘ung, in 752 .D., travelled from Hu-mi or Wakhan to Chii-wei and thence, through two localities (Ho-an and Lan-so) not yet identified, to Veh-ho, in which M. Chavannes recognizes, with good reason, Veh-/o, the capital of Little P‘o-lii; comp. L’Linéraire a’ Ou-kong, p. 12, and Zurcs occrd., p. 129, note 2.
82 The extent of uninhabited ground on either side of the route, bare of all resources for the support of human life, and either actual glacier area or under heavy snow for the greater part of the year, is well illustrated by the snow and race maps in Drew’s Jummoo and Kashmir.
83 Compare, for an interesting account of the Chilasis and their predatory expeditions, which are still remembered in the adjoining regions, Biddulph, Hindoo Koosh, pp. 15 sqq.; see also Drew, /ummoo, pp. 398, 404.
Sec. iii] CHINESE OCCUPATION OF GILGIT AND ROUTE TO KASHMIR 17
notwithstanding a nominal tribute, and their turbulent disposition and the danger it represented for the newly-constructed ‘Gilgit Road’ ultimately led, in 1892, to the occupation of their territory by an Imperial garrison.
The Chilasis, in race, language, and ethnic peculiarities, are closely allied to the other Dard communities which, organized in small republics, extend far down both banks of the Indus, and collectively form the tract known as Shinkari or the Kohistan of the Indus Valley. Though this region is still inaccessible to the European traveller, it is certain that the several sections of the valley are not separated from each other by great natural barriers. Hence constant relations are kept up between these Dard communities which used to combine when threatened by an external foe. In fact, tradition at Childs distinctly asserts that in pre- Muhammadan times the whole of Shinkari was under the rule of one Raja*.
We have seen already that, according to the indications furnished by the T‘ang Annals and Hsiian-tsang, this tract in the Indus Valley was at one time politically dependent on the kingdom of Udyana. The supremacy exercised from the Swat Valley may, at the period when we hear of the alliance between the Tibetans and Chieh-shih, have been replaced by predominant influence from the side of Chitral. Both from Chitral proper and from Upper Kashkar or Mastij the Indus Valley can readily be reached by a number of routes leading across the headwaters of the Swat and Panjkora rivers, and the remarkable extension which in recent years the Khan of Dir’s power has taken in the direction of the Swat Kohistan and the Indus Valley presents a curious parallel. Chilas and the other Dard communities along the Indus, if left to themselves, could without great difficulty have been overawed by the Chinese garrison placed in Gilgit and Yasin; but when controlled and supported by a neighbouring hill- state of such resources as Chitral, they were bound to become a serious menace to the Kashmir— Gilgit route, which they flanked, and upon which the maintenance of that garrison depended.
Section IV.—ANCIENT REMAINS IN GILGIT AND HUNZA
The total absence of reliable records makes it impossible at present to trace more of the early history of Gilgit and the adjoining regions than is revealed for a brief period by the illuminating notices of the Chinese Annals. A patient study of local traditions and ethnology, and a systematic search for ancient remains would, no doubt, bring to light materials likely to help us in restoring some aspects of the life and culture that prevailed there during pre- Muhammadan times. But for such labours the series of rapid and often trying marches which, between June 11 and 28, carried me through the valleys of Gilgit and Hunza to the Hindukush watershed, as described in chapters 11 and 111 of my Personal Narrative, left no opportunity. —
Among the few relics of Buddhist worship which are extant above ground in the main valley of Gilgit, and to which Major J. Manners Smith very kindly drew my attention, I was able to visit only the great rock-carved relief at the entrance of the Kergah Nullah, some four miles above Gilgit Fort and not far from the village of Naupir. This relief, which appears to have been first described by Colonel Biddulph 1, occupies a conspicuous position, more than
** For much interesting information collected by Colonel attempted identification of this figure with the colossal Buddha Biddulph regarding the Indus Kohistan, see Hindoo Koosh, image seen by Fa-hsien in Darél, requires no serious considera: pp. 8 sqq. tion. A somewhat indistinct photographic reproduction of the
* See Hindoo Koosh, pp. 109 sqq., where a rather primi- rock-carving is given in the Pamir Boundary Commission tive sketch of the relief is reproduced in lithography. The Report, p. 32.
STEIN D
Rock- sculpture near Gilgit.
Pose of Buddha statue.
18 FROM KASHMIR TO THE PAMIRS [Chap. I
thirty feet above the ground, on a precipitous rock-face. It shows the figure of a colossal Buddha, about nine feet in height, carved in low relief within a shallow niche of trefoil shape (see Fig. 1).
Buddha is represented standing with the right hand and forearm raised across the breast, in the gesture which, in Buddhist convention, is known as the abhayamudra (‘the pose of assuring safety’), while the left hangs down grasping the edge of the robe. The robe is indicated only at the sides of the figure from the hip downwards, and leaves the limbs entirely bare, with the result that the statue at first sight recalls the representation of a naked Jina rather than a draped Buddha.
Yet, coarsely modelled and executed as the image is, examination of the photograph clearly shows that a fully-draped figure was intended. Broad pleats or bands appearing around the neck, and also over the right forearm, mark the folds which result from the classical draping of the robe, as invariably exhibited by Buddha statues in this posture both in Gandhara and Khotan sculpture. The type which the Gilgit sculptor endeavoured to reproduce, with all the imperfections of his art, is unmistakably the same as appears in numerous stucco reliefs of Dandan-Uiliq, specimens of which may be seen on Plates LIV and LV. Earlier representations of it are to be found in numbers, both among Graeco-Buddhist sculptures from Gandhara and the great stucco reliefs excavated by me at the Rawak Stipa near Khotan ’.
It is significant that nowhere among these early instances of the type do we find the right hand brought across the breast as in the Gilgit relief. While the outward turn of the palm, which is a characteristic feature of the Abhayamudra, could be obtained in this position only by an artificial twisting of the forearm, it is a perfectly natural gesture when the forearm is extended forward, with more or less of an upward slant, as seen in the sculptures, of which illustrations are quoted in note 2, The same strained pose of the right arm and hand is found, however, in a small relief representation of Buddha which occurs as part of the stucco wall decoration in one of the Dandan-Uiliq shrines, and a specimen of which, D. 1. 11, has been reproduced in Plate LV.
In either case the change in the position of the forearm is manifestly due to the artist’s inability to show in low relief the natural pose of the original model. The Dandan-Uiliq stucco figure certainly belongs to the second half of the eighth century, and the resemblance between it and the Gilgit rock-carving, in the clumsy device just discussed, must warn us against seeing in the crude modelling of the latter a mark of special antiquity. The pointed form of the trefoil arch enclosing the image seems to me, in fact, distinct evidence of a later date, judging from what a close study of that architectural element in Kashmirian monuments has enabled me to ascertain as to the successive development of its forms. We have another sign of late workmanship in the exaggerated prominence which is given to the edges of the drapery, particularly that falling from the proper left arm. Its appearance there might easily mislead the untrained observer into a belief that the left hand, instead of supporting the robe (as clearly seen in the Dandan-Uiliq and Rawak reliefs), was holding ‘a staff or some kind of weapon ’*.
2 Compare, e. g., Griinwedel, Buddhist Art, pp. 169, 170, ticular Buddha images, as contained in Nepalese MSS., and 174 (with the interesting Chinese Buddha, in wood, p. 177, among extant Buddhist sculptures of Northern India; for which is traditionally derived as a replica from Udayana’s illustrations comp. PI. II and Figs. 7, 8. famous statue); for corresponding relief statues at Rawak, 3 See Biddulph’s description, Hzndoo Koosh, p. 110.’ For see Figs. 68, 69, Plate XIV, &c. similar exaggerated treatment of the drapery in imitation of
M. Foucher, who has lucidly discussed the A dhayapanz- ancient forms, comp, Griinwedel, Buddhist Art, p. 171 and
mudra in his Iconographie bouddhique, pp. 68 sq., points out Fig. 125. its frequency both in the miniature representations of par-
LLILTVd LV VZNONH AO MIN AO ATLSVO ‘LIOTIO “UNdAVN UVAN AAITAU GAAUVI-NIOU
ia ay os
ae
Rae
Sec. iii] ANCIENT REMAINS IN GILGIT AND HUNZA | 19
The deeply-cut square holes, arranged outside the trefoil niche in the form of a pentagon, undoubtedly served to support a wooden framework which once protected the image.
Colonel Biddulph has already noticed the remains of an ancient irrigation work, which can be traced close by, along the right or eastern side of the mouth of Kergah-Nullah. The square sockets which are seen there, sunk into the steep rock face at a uniform level and for a con- siderable distance, were manifestly designed to hold stout pieces of timber, on which a wooden trough could be fixed to conduct water from the stream of the Nullah for the irrigation of fields in the main valley. The comparatively high level at which this conduit runs above the rock- strewn bottom of the Nullah is easily accounted for by its purpose of carrying water to slopes which could not be reached by irrigation cuts taken from the stream at its actual mouth.
That the construction of this watercourse belongs to an early period, certainly pre- Muhammadan, is proved beyond doubt by the excellent cutting of the sockets, which shows a command of stone-craft long lost among the population of these valleys. We may well suppose that in a period of greater material culture Gilgit was more thickly inhabited than it is now, and that in consequence portions of ground on the hill sides and on alluvial plateaus more elevated than any now under cultivation were utilized for fields and required to be irrigated. Throughout the Gilgit valley it is only the water obtained from the side-streams that renders agriculture and the growing of fruit trees possible 4.
In connexion with these few notes on ancient remains in Gilgit, I may mention that the ruined mounds which Major J. Manners Smith and other officers on duty in Gilgit have noted at Hanzil and Jutial in all probability represent the remains of Stipas. I was not able to visit them, but the photograph of the mound near Hanzil (a village about nine miles above Gilgit Fort) which is reproduced in the report of the Pamir Boundary Commission, distinctly suggests this origin’. The fact that in either place the ruin shows only rough masonry of unhewn stone would render it difficult to arrive at any conclusion as to the date of construction without systematic excavation.
The valley of the Hunza river, through which I passed from Gilgit to the Taghdumbash Pamir, is, alike by the stern grandeur of its peaks and glaciers, the natural difficulty of its communications, and the strange mixture of races and languages among its population, a mountain region of exceptional interest. But: the same ice-crowned ranges and almost equally formidable gorges which have rendered Hunza until our days so secure against foreign invasion, have also effectively barred the valley from ever serving as a real line of communication or otherwise acquiring historical importance.
We have striking evidence of this isolation in the survival of Burisheski, the tongue spoken in Hunza proper, which has no relation whatever to any of the great language families (Indian, Iranian, Turki, Tibetan) that meet close to this easternmost point of the Hindukush watershed. Nor can it be doubted that these secluded communities, in their customs, traditions, and economic conditions, must have preserved much that would prove highly instructive to antiquarian students.
* See Drew, Jummoo, p. 407; compare also the remarks, that Commission. Such towers are regularly built square p. 404, illustrating the reduction of arable land in Astor throughout the whole of the Dard region. On the other which has followed long-continued decay of the artificial hand, I fail to see how ‘ the construction of the masonry’ can watercourses, : be adduced as an argument against ‘the theory of Buddhist
* See Report of the Pamir Boundary Commission, p. 93 construction ’, considering that neither the masonry nor any The decidedly circular shape of the mound makes it very other details of construction of pre-Muhammadan ruins in improbable that the remains can be those of a ‘frontier tower the valleys between the Hindukush and Kashmir have as yet of ordinary fashion’, as assumed by one of the members of been examined by a trained archaeologist.
D2
Ancient water-con- duit.
Ancient mounds in Gilgit.
Valley of Hunza.
Ruined | Stiipa near Thol.
Architecture of Thol Stiipa.
Mountain tracks in Hunza.
20 FROM KASHMIR TO THE PAMIRS [Chap. I
For such passing observations as my hurried marches along the difficult tracks of Hunza allowed me to gather, I must refer to the pages of my Personal Narrative®. But in respect of the only monument of antiquity which I was able to trace along the route, where it winds round the foot of the great Mount Rakiposhi (Fig. 3), the brief description there given requires to be supplemented by some more exact details. Close to the hamlet of Thdl, on the first of the highly cultivated little plateaus which are passed in succession on the left river bank above the historic gorge of Nilth, there stands a ruined Stiipa, shown with its west face in the accompanying photograph (Fig. 4). Considering its exposed position, the relatively good preservation of this monument of Buddhist worship is remarkable. The only serious damage it has sustained is of very recent date, being a characteristic result of the ‘advance of modern civilization’: the masonry of the base on its southern face has been removed to save a détour of a few feet to the hill road constructed since the occupation of Hunza territory at the close of the campaign of 1891.
The Stipa is constructed of unhewn slabs, more or less flat, set in rough horizontal hea with a fairly hard plaster. The surface of the rough masonry originally bore a coating of similar plaster, patches of this material being found still adhering to the north face of the upper courses of the base and to the more sheltered mouldings. The lowest story of the base forms a square of 10 ft.; and has a height of 3 ft. 9 in. from the present ground level. On the top of this base, but receding by about 1 ft. from its edge, rests a second story, which is also square, and with its projecting cornice, about 14 in. high, rises to the same height of 3 ft. 9 in. The next story is poetapenel, and, including its cornice, 3 ft. 10 in. high. It supports a circular drum, about 14 ft. in height, also surmounted by a cornice. From the latter finally springs the dome “s the Stipa proper. This, in its original shape, appears to have been hemispherical, but its top has been broken, and the extant masonry reaches only a height of about 33 ft. from the sloping plinth of the drum. Judging from what subsequent experience showed me in the case of Stipa ruins examined in the Turkestan plains, it may be supposed that treasure-seeking operations for the relic deposit likely to be contained in the centre of the little dome were the main cause of this injury. The total height of the structure, allowing for this loss, cannot have been less than 18 ft. and probably approached 20 ft.
This relatively great elevation, in proportion to the dimensions of the square base, strikingly distinguishes the Thol Stipa from the Stipas examined by me in Chinese Turkestan. Those among the latter which were found sufficiently well-preserved to permit exact measurements, like the Stipas of Mauri-Tim, the Niya Site, Endere and Rawak, in their original state showed a total elevation approximately equal to the side of their square base where it rested on the ground, i.e. to the greatest dimensions of the ground-plan. Striking differences are to be found also in the introduction of an octagonal above the square stories of the base, and in the boldly projecting and massive cornices by which these several stories and the circular drum beneath the dome are surmounted. For neither of these distinguishing features am I able to adduce exact parallels from Stipa ruins with which I am acquainted in the North-West of India and the adjoining frontier regions. But there appears in them something curiously recalling the style and general character of the Chortens of Sikkim and Ladak, and suggestive of Tibetan influence.
The small alluvial plateaus fringing at intervals the deep-cut bed of the Hunza river between Nilth and Baltit are the only portion of the valley where a track fit for regular use by laden animals could be maintained without recourse to modern methods of engineering. Above the village of Baltit, probably from early times the seat of Hunza chiefs’, the route towards the
® See Ruins of Khotan, pp. 32 sqq. 7 For a photograph of their picturesque old castle see Fig. 2.
Ee tS x Tet crete
MOUNT RAKIPOSHI, SEEN FROM ALIABAD, HUNZA.
RUINED STUPA AT THOL, HUNZA.
Sec, iii] ANCIENT REMAINS IN GILGIT AND HUNZA 21
Taghdumbash Pamir enters the long and narrow gorge in which the river has cut its way through the axial range of the Hindukush. The tracks that lead through it present, for six trying marches, all the difficulties of true Alpine climbs, including passages over large glaciers. During a considerable portion of the year they are altogether impassable except for men on foot. The transport of loads over these mountain tracks becomes possible only by the employment of the hardy hillmen, who manage to subsist on the few patches of cultivable ground to be found amidst the barren masses of rock and ice which bound the course of the main stream and the still more confined gorges of its affluents. The extreme sterility of these mountains restricts the population of ‘Little Guhyal’, as this region is now called on account of its Wakhi settlers from Wakhan or Guhyal, to a few hundred families*, It seriously taxes the resources of these scattered settlements to supply even porters for the occasional European traveller who is permitted to pass through the Hunza Valley. In the face of such natural obstacles, it is impossible to conceive how the latter could ever have served as a route for trade and general traffic,
After leaving behind Misgar, the northernmost hamlet of Hunza, the natural difficulties of the route decrease. The valley widens as we approach the watershed which separates the headwaters of the Hunza river from those of the Oxus on the one side and from the Taghdumbash Pamir on the other. Lord Curzon, in his exhaustive Memoir on the Pamirs, has duly emphasized the important geographical fact that the water-parting in this part of the Hindukush lies considerably to the north of the axial range and is also far lower’. This helps to account for the relative ease with which the Kilik and Mintaka passes, giving final access to the Taghdumbash Pamir, can be crossed, even with laden animals, during the greater part of the year.
* For photographs of Wakhis and Kanjitis settled in ® See Curzon, Zhe Pamirs and the Source of the Oxus, ‘Little Guhyal’, see Figs. 5, 6. p. 40.
Meeting of ethnic areas near Kilik Pass.
Position of Taghdum- bash Pamir.
Importance of Tagh- dumbash route,
CHAPTER I] SARIKOL AND THE ROUTE TO KASHGAR
Section I—THE GEOGRAPHICAL POSITION AND ETHNOGRAPHY OF SARIKOL )
TuE great alpine valley of the Taghdumbash Pamir, to which I crossed on the 29th of June, 1900, over the Kilik Pass (15,800 feet above the sea, Fig. 8), was certain to prove interesting ground from more than one point of view. At its head I was close to the point where the drainage areas of the three great river systems of the Indus, the Oxus, and the Tarim meet, the representatives as it were of the still greater ethnic areas of India, Iran, and Turkestan. That the view from the height of the Khushbél Peak, my first survey station, simultaneously comprised the confines of British India, Afghanistan, Russia, and China, was the best illustration of the abiding influence which geographical facts must exercise over political developments even in this desolate region. The wide expanse of level or easily undulating grazing land which the same view disclosed at the bottom of the Taghdumbash Valley, formed a striking contrast to the rocky and almost impassable gorges of Hunza, and could not fail to impress me at the outset with the advantages for communication which the Pamirs offer, notwithstanding their elevation and severe climate. But it was only in the light of subsequent observations, gathered on my marches along the whole length of the Taghdumbash Pamir from the Wakhjir Pass and the Oxus source down to Tash-kurghan, that I fully realized the historical interest of the route which leads through this valley. |
In order to explain the importance which may be claimed for the Taghdumbash Pamir as an early line of communication between Chinese Turkestan and the Oxus Valley, it is not necessary to consider the historical topography of the whole Pamir region. Lord Curzon, in his justly famous memoir on ‘The Pamirs and the source of the Oxus’, has furnished a lucid and critical summary of that much-discussed subject, and to it the reader may be referred with confidence for information on all general questions’. An important geographical fact, aptly recognized in Lord Curzon’s analysis*, helps to limit the scope of our inquiry. The Taghdumbash Pamir, though it shares the designation and also, in its upper portion, the main physical characteristics of the other Pamirs, lies in a different watershed from the rest, and is thus plainly marked as part of a different system. We find this physical division emphasized at the present day by the political boundaries, which leave the Taghdumbash the only Pamir under Chinese jurisdiction, and we may safely attribute to it a determining influence upon the earlier historical conditions of this mountain tract.
While all the other Pamirs are situated within the drainage area of the Oxus, the waters of the Taghdumbash Pamir discharge themselves eastwards into the great Turkestan Basin.
1 Reprinted from the Geographical Journal, 1896; for * See Zhe Pamirs, p. 19. an analysis of early travellers’ routes see in particular pp. 63 sqq.
Sec.i} GEOGRAPHICAL POSITION AND ETHNOGRAPHY OF SARIKOL 23
The river of which they are the main feeders, and which takes its best known name from Tash-kurghan, the chief place it passes, breaks through the great meridional range flanking the Pamirs on the east, and ultimately joins the Yarkand river or Zarafshan. The collection of valleys which the river of Tash-kurghan drains, together with some minor alpine tracts adjoining them towards the Upper Yarkand river, constitutes the well-defined mountain district now known as Sarikol®. Topographical facts and historical evidence alike prove that the position occupied by the present fort of Tash-kurghan and its neighbouring villages has from very early times been the political centre of this whole territory. At Tash-kurghan ends the open valley of the Taghdumbash Pamir, and the importance of the latter as a route is mainly due to the exceptionally easy access it provides to the central tract of Sarikol, over more than a hundred miles’ distance. In view of this close connexion, it will be well to begin with a review of the main data which throw light on the ancient topography of Sarikol as a whole.
Small in extent, and devoid of natural resources, the territory of Sarikol derives its importance solely from the advantages of its position with regard to the routes which from early times have connected the Upper Oxus Valley with the oases to the south of the Turkestan Desert, and hence with China. All the routes leading from the Oxus in that direction, whether they ascend through Roshan, Shighnan, or Wakhan, have to cross the water-parting on the east of the Pamirs, and subsequently to surmount the still more elevated meridional range culminating in the Muztagh- Ata Peaks, which forms the link between the T‘ien-shan system in the north, and the extreme points of the Hindukush and Kun-lun in the south*. The chief valleys of Sarikol extend between this meridional range and the watershed which fences in the true Pamirs on the eastern side. To this position they owe their relatively great width and also their prevailing direction, the valley of the Taghdumbash descending mainly from south to north and that of Tagharma from north to south. Every route that crosses the Pamir watershed to the south of the Muztagh-Ata masszf, is compelled to debouch at one point or the other into either of these great valleys; in the opposite direction the same holds good of all the routes that lead from the drainage area of the Yarkand river westwards into the Oxus Basin. The Taghdumbash and ‘T'agharma Valleys meet at the point where the Tash-kurghan river takes its sudden turn to the east; and Tash-kurghan, only some eight miles further south, is the place towards which all the above routes, whether from east or west, naturally converge.
It is easy to show that this centre of the Sarikol District must at all times have been
* The above spelling, Sarzkol, reproduces the pronuncia- tion of the name as generally heard by me during my passage through the district, from both its Tajik (Iranian) and its Kirghiz inhabitants. I am unable to decide how far this modern pronunciation conforms to the etymology of the name; for the derivation is not quite certain. If Mirza Haidar, whose mention is the oldest I can at present trace (artkh-i-Rashidi, ed. Elias, pp. 297, 312), is justified in writing Sarigh-kul, the name would be Turki, meaning ‘Yellow-Lake.’ As ol is a common variation of ku/ through all dialects of Eastern Turki, and as 7 for 7h would easily be accounted for by the well-known phonetic processes of assimilation and subsequent ‘ supplementary lengthening’, the modern form of the name could be readily explained on the basis of this etymology. But the form S¢rdkol (Sirzku/) is also met with in Oriental records, and has found acceptance among numerous European geographers, perhaps on account of its
supposed semi-Persian etymology (Svr-z-kul, ‘head of the lake’). It is curious to note the same uncertainty of spelling in the case of one of the several alternative native names for Wood’s Lake on the Great Pamir; see Curzon, Zhe Pamirs, p. 43.
* Though accurate surveys of this meridional range have been accomplished in recent years, the admirably lucid account which Baron Richthofen, in his C/zma, vol. i. pp. 195 sqq-, has given of its salient features and its importance for the general orography of Central Asia, may still be consulted with advantage. The merit of having first systematically explored those portions of the range which lie between the Tash-kurghan river in the north and the junction of Hindu- kush and Kun-lun in the south, belongs to Captain H. H. P, Deasy, whose book, Three Years in Tibet and Eastern Turkistan (1901), furnishes also a description of the smaller Sarikol valleys comprised in those portions ; see chapters viii, xiii, xiv,
Importance of Sarikol.
Routes through Tash- kurghan.
Resources of Sarikol.
Earlier settlements in Tash- kurghan Valley.
24 SARIKOL AND THE ROUTE TO KASHGAR [Chap. II
an important goal for travellers. From whichever side we may approach Sarikol, there is an inhospitable belt of high mountain land to be crossed first, practically devoid of permanent habitations, and throughout incapable of furnishing supplies and places of shelter to caravans. The elevated Pamir region stretching westwards can never, during historical times, have permitted of cultivation. The routes which, starting from Tagharma, connect Sarikol with Kashgar to the north-east and Yarkand to the east, lead by a succession of high passes over barren spurs of the great meridional range with narrow uninhabited gorges between them. It is true that the difficult and rarely frequented tracts which cross the mountains between the Tash-kurghan and Yarkand rivers in the direction of Karghalik and Koék-yar, pass through some of the minor Sarikoli settlements. But the produce raised on their isolated plots of cultivable land does not suffice even for the maintenance of the small pastoral population scattered over this region. Finally, if the route be followed which leads northward past Muztagh-Ata and then descends along the Yaman-yar river into the Kashgar plain, as described in my Personal Narrative®, an even greater distance has to be traversed before permanent habitations are reached.
This situation of Sarikol, in the midst of desolate mountain tracts and yet at the junction of important routes, necessarily invests with exceptional value whatever natural resources the district possesses. Given an adequate population, and an administration capable of protecting it, these resources would, undoubtedly, be far larger than they are at present. The lower part of the Taghdumbash Valley, for a distance of forty miles from below Tash-kurghan to the hamlet of Dafdar, shows a remarkably broad and uniform expanse of fertile ground at its bottom. Cultivation is now restricted to a small continuous stretch of the valley above and below Tash-kurghan, and to certain isolated settlements, such as Dafdar and Pisling, recently started at points higher up the valley where it is easy to obtain water for irrigation from side streams.
But the uniform tradition of the inhabitants, as reported to me during my stay at Tash- kurghan, asserts that the compact cultivated area once extended much higher up the valley. In support of this belief, reference was made to the remains of extensive irrigation canals traceable along the foot of the mountains, especially on the east side of the valley, as well as to deserted village sites, such as Bazar-dasht, found at a considerable distance beyond the limits of the present belt of village land. I believe that these statements may be accepted as based on substantial facts. On the one hand, the almost unbroken stretch of alluvial land which I passed on my route from Dafdar to Tughlan-shahr, seemed only to wait for systematic
‘irrigation in order to yield the crops for which the climate is adapted. On the other hand,
it is impossible to doubt that the periodical raids from slave-hunting Kanjitis and Shighnis, to which we know Sarikol to have been subjected before and after Yaqtib Bég’s rebellion, and which did not cease completely until the British occupation of Hunza, must have resulted in a partial depopulation of the country °.
From the devastating effects of these raids, the upper portions of the Taghdumbash Valley, lying nearest, necessarily suffered most. But they extended also to the Tagharma Valley, which, in its well-watered central flat, about twelve miles long by seven broad, offers ample ground for agricultural settlements’.
® See Ruins of Khotan, chaps. v-vii. are illustrated by the information recorded by Capt. H. Trotter
® For the modern history of Sarikol, compare Col. T. E. in 1874 about a fort village in the south-east of the plain, the Gordon, Roof of the World, pp. 109 sqq., where an instructive Tajik population of which had been reduced from fifty to four account is given of the general conditions of the main valleys families within less than a generation; see Yarkand Mission at the time of his visit (1874). Report, p. 269.
7 The destructive results of Kanjiti raiding in Tagharma
Sec. i] GEOGRAPHICAL POSITION AND ETHNOGRAPHY OF SARIKOL 25
It is difficult to judge now, when the whole tract is slowly recovering from the results of long-continued aggression, what its population and resources may have been during periods when the support of a strong paramount power, established in Eastern Turkestan, assured to Sarikol safety from its predatory neighbours to the west and south. The political value of this hill tract as an advanced post guarding the frontier towards Badakhshan, Gilgit, and the Pamirs was, as we shall see, realized early by the Chinese. But far more than the spasmodic assertion of sovereignty, which, as far as the Chinese administration was concerned, is likely to have been often purely nominal *, the racial tenacity of the inhabitants themselves must have helped to preserve to the little alpine chiefship its continued existence and historical individuality.
The population of Sarikol, apart from the nomadic Kirghiz herdsmen who visit its grazing grounds, consists of hill Tajiks, who by physical appearance and language alike are unmistakably proved to belong to the so-called Galcha stock. The Sarikoli tongue, first recorded by the late Mr. R. Shaw with his usual accuracy and care, is very closely allied to Wakhi, the language of Wakhan, of which it may be said to represent a mere dialect. The Sarikolis whom I saw showed all the racial characteristics of the Iranian ‘Tajiks’ who form the bulk of the population in the Oxus region ®, and the knowledge of Persian common among all classes indicates the influence exercised by the culture of Iran even in these distant valleys. To the prevalence of Iranian traditional lore in the local legends I shall have occasion to refer later. The antiquity of these legends, as attested by Hsiian-tsang, the Iranian elements in the local nomenclature, as well as the absence of any historical tradition to the contrary, render it highly probable that the population of Sarikol has from early times borne the same ethnic character as at present.
The geographical position of Sarikol is such that it could never have lain on the route of one of those great waves of invasion which from time to time have swept across the more accessible parts of Central Asia, and in their progress have radically changed the ethnology of those regions. Sarikol, with its severe climate!°—the elevation of the main valleys sinks nowhere below 10,000 ft.—and its limited area of productive land, could not possibly attract occupation and settlement by a powerful invader. The predatory inroads to which the main valleys were exposed, during recurring periods, from the petty hill-states to the south and west, no doubt subjected the more peaceable population of Sarikol to great vicissitudes. But considering the passing character of these raids, and the ready shelter offered by the many secluded side- valleys, they were not likely to threaten it with extermination. Experience in any case shows that the Sarikoli settlements during recent times have recovered from severe trials of this kind, including even a forcible transportation of the greater part of the population to Kashgar, which occurred during Yaqib Bég’s rule™. It is manifest that this small Iranian community, placed in such close proximity to a great Turki-speaking population, and sharing its political fortunes, could not have preserved its language and racial characteristics intact without the isolating effect of its mountains and that tenacious attachment to inherited lore which is peculiar to hill people.
° The relations which existed during the period preceding YAqiib Bég’s rebellion between the hereditary chief of Sarikol and the Chinese authorities were characteristically indicated in the nominal tribute paid by the former and the valuable presents received in return. The payments in silver and gold made to the Sarikol chief were represented ‘as a subsidy for the military protection of the frontier and the road towards Badakhshan’; see Gordon, Roof of the World, p. t10.
® For a photograph see /. Anthr. Inst., 1903, Pl. XXVII,
where the individuals in the back row and the central figure STEIN
in the front row are Wakhis settled in Sarikol, the rest Kirghiz. Wakhis and Sarikolis proper are not distinguishable in outward appearance.
© The character of the climate of Sarikol is well indicated by the saying which Col. T. E. Gordon recorded from the mouth of its governor at the time of his visit ‘that there are only two seasons, summer and winter, the former lasting three months, the latter nine’; Roof of the World, p. 114.
4 See Gordon, Roof of the World, p. 111.
Ethnic affinity of Sarikolis.
Ethnic in- dividuality of Sarikol.
Early exten- sion of
Galchastock =
eastwards.
Connexion with ancient population of Khotan.
The Pakhpo tribe.
26 SARIKOL AND THE ROUTE TO KASHGAR (Chap. II
The hillmen of Sarikol at the present day form the extreme outpost of Iranian nationality towards the east. But if we may judge from several important indications, settlements of an Iranian-speaking race must in ancient days have extended much further eastwards, especially in the direction of Khotan. The documents in Brahmi script which my excavations brought to light from the ruined temples of the Dandan-Uiliq site, and which, with other records of the same character previously obtained from the same locality, have been partly deciphered by Dr. Hoernle, make it appear highly probable that the language spoken in the eighth century by the indigenous population of Khotan was of Iranian origin’ On the other hand, the anthropometric data collected by me in the Khotan region, according to the careful analysis made of them by Mr. A. T. Joyce, in independence of any historical or linguistic arguments, plainly mark in the Khotanese population of to-day the prevalence of a racial element closely related to the hill-tribes generally designated as Gathas, who are settled in the highlands of the Oxus and Zarafshan %,
The ethnographical importance of these observations is confirmed by the statement of the Chinese historical record to be discussed below, dating back to the period of the T’ang dynasty’s rule over Eastern Turkestan (7th to 8th century a.p.), which describes the external appearance and language of the people of Sarikol as identical with that of the Khotanese**, In view of this convergence of linguistic, anthropological, and historical proofs, we can scarcely avoid the conclusion that the modern Sarikolis represent but the remnant of a larger Galcha population which once spread as far as the territory of Khotan, but further east has since undergone considerable racial amalgamation and abandoned its language for Turki. The great difference in geographical position, and consequently in accessibility between the valleys of Sarikol and the oases of the plain of the northern foot of the Kun-lun range would amply account for the thoroughness with which this transformation has proceeded in the latter region.
Finally, it may be pointed out in passing that an ethnic link between the Iranian Sarikolis and the present population of those oases is, perhaps, to be found in the small and little known hill-tribe of the Pakhpos, who partly as herdsmen, partly as cultivators, dwell in the narrow valleys near the headwaters of the Tiznaf and Yarkand rivers. Dr. Bellew, to whom we owe what scanty information has so far been recorded about this curious people, describes them as of ‘pronounced Caucasian features’ and very fair. He was much struck by the difference of their physical type from that of other races he had come into contact with in Eastern Turkestan. He notes that the few Pakhpos whom he met and was able to examine, denied having any language of their own apart from Turki; but he remarks also upon the extreme shyness of these hillmen, which ‘led them to conceal all information regarding themselves’. I myself was unable to proceed near enough to their hills, south of Karghalik, to obtain an opportunity of meeting Pakhpos. But during my short stay at Karghalik, Daud Bég, a local official, who had for some years held charge of the tract they mainly inhabit, distinctly told me that, though all Pakhpos know Turki, and though its use is extending owing to frequent intermarriages with people of the Karghalik oasis, another language is talked by them among themselves which
® Compare Ruins of Khotan, pp. 309 sq.; Hoernle, Central Asian Antiquities, ii. pp. 32 sqq.; below, chap. 1x. 8063 -Vare:
8 See below, chap. vi. sec. iv., on the racial origin of the people of Khotan, and Mr. Joyce’s paper in /. Anthr. Inst, 1903, pp. 322 sqq.
“4 See Chavannes, Zurcs occid., p. 124.
® Compare Varkand Mission Report, pp. 61 sq. It must
be remembered that Dr. Bellew, with the main portion of Sir D. Forsyth’s mission, did not visit the Sarikol valleys, and consequently could scarcely be expected to notice any similarity. in physical appearance that might exist between Sarikolis and Pakhpos. The photographs of Pakhpos re- produced on p. 46 of the Varkand Mission Report show types which to me do not appear to differ materially from those ordinarily met among Sarikolis and Wakhis.
NVOSIN LV SILAINVS
‘VZNOH ‘UVEIVHY LV GATLLAS SIHNVM
Sec.i] GEOGRAPHICAL POSITION AND ETHNOGRAPHY OF SARIKOL 27
is understood by Sarikolis, and is apparently a dialect of the latter’s tongue. If this information should prove to be correct, our presumption in favour of the Pakhpos being another remnant of a Galcha population, formerly more widely spread, would become an established fact 1.
Section II—EARLY CHINESE ACCOUNTS OF SARIKOL
The important position occupied by Sarikol in respect of the routes leading across the Pamir region accounts for the ample information to be gathered from early Chinese records regarding its historical topography. The narrative of the pilgrim Hsiian-tsang, who on his return journey from India to China traversed Sarikol, supplies most of the details. It was in his itineraries that the old name of this mountain chiefship, which the Chinese transcriptions variously render as Chieh-p'an-t‘o, Han-p‘an-t‘o, &c., was first correctly identified by General A. Cunningham}. But for a systematic review of all available data we shall find it convenient to take as a basis the official description of the territory which the T‘ang Annals furnish, and which has now become accessible in M. Chavannes’ translation.
This account records for the Sarikol tract the several names Ho-f‘an-t‘o be Se te or flan-to Yi ip, or Ko-hoan-t'an YB fit fH, or Ko-lo-c'o Yh Be pe, and indicates its position with unmistakable clearness. The kingdom was reached by proceeding from Sw-/ or Kashgar to the south-west through the gorge of Chzen-mo, which must correspond to the present Gez defile, and lay at a distance of 600 li, or approximately six days’ march. This estimate agrees remarkably well with the route which leads from Tashmalik, at the south-western edge of the Kashgar oasis, along the Gez or Yaman-yar river, and past Muztagh-Ata to the head of the Tagharma Valley, and which I myself followed in the opposite direction on my journey from Tash-kurghan to Kashgar. o-p‘an-t‘o is correctly described as being situated directly to the west of Chu-chii-o, which is represented by the modern Karghalik; in the north it touched the territory of Sw-/é or Kashgar, in the west Au-md or Wakhan, while to the north-west there adjoined the territory of P‘ax-zan in which, with M. Chavannes, we may recognize Farghana. The administrative centre of the territory lay in the middle of the Ts‘ung-ling or ‘Onion mountains’, which are said to encircle the whole of the kingdom* This designation has been generally applied by the Chinese to the meridional range or ranges which buttress the Pamir region on the east*, and divide it from the Tarim Basin. The position of the Ho-p‘an-t‘o capital, as here marked with reference to them, fits exactly the present Tash-kurghan. The river Sz-¢o, on which the capital is stated to have stood, can, in view of Hsiian-tsang’s account to be noticed below, be no other than the river of Tash-kurghan.
Of the people of Ho-p‘an-t‘o the Annals record, as already stated, that their appearance and language were identical with those of the people of Vii-t‘ien or Khotan. They are further described as strong and given to violence. Murder and brigandage were alone liable to be
* It deserves to be noted that Sung Yiin distinctly mentions the language and customs of the people of Chu-chii-po as closely resembling those of the Khotan people; see Chavannes, Voyage de Song Vun, p. 20. Chu-chii-po corresponds to the present district of Karghalik, including the valleys towards the headwaters of the Yarkand river, in which the Pakhpo settlements are found; see Chavannes, Turcs occid., pp. 123 note, 311.
’ See his paper in /.A.S.B., 1848, vol. xvii, referred ‘to by Yule, 7.R.A.S., 1872, ‘Notes on Hwen Thsang’s Account of Tokharistan,’ p. 117.
? See Chavannes, Zurcs occid., p. 124.
* It will be seen below that the term 7s‘ung-ling was used during the T‘ang period as a special designation of the Sarikol territory itself; as to the meaning and origin of the name, comp. Richthofen, Chzna, i. p. 221.
E 2
Sarikol in T‘ang Annals.
The people of Sarikol in the T‘ang period.
Fa-hsien’s notice of Sarikol,
Sung Yiin’s passage through Sarikol.
28 SARIKOL AND THE ROUTE TO KASHGAR [Chap. II
punished with death; for other offences compensation by fine was allowed. Taxes were paid in clothing materials, an observation illustrated by what a native informant, in the Report of the Yarkand Mission, notes of the barter carried on in Sarikol with cotton goods imported from the plains*. The military force of the territory was reckoned at a thousand men. Its ruling family is stated to have originally come from Kashgar, and to have transmitted its power from generation to generation.
During the period corresponding to 435-439 A.D., under the later Wei dynasty, Ho-p‘an-t’o is said to have first entered into relations with China, ‘In the ninth year Cheng-kuan (635) it sent an envoy to do homage at the Imperial court. During the period K‘ai-yiian (713-741 a. p.) China conquered and pacified this kingdom ; it established there the military post of Ts‘ung-ling, which is the extreme point under military occupation on the frontier of An-hsi’, i.e. of the Chinese protectorate then comprising Eastern Turkestan. M. Chavannes, in his comments on this notice, points out that another passage of the T‘ang Annals distinctly identifies the ‘ military post of Ts‘ung-ling’ with the ancient kingdom of Chieh-f'an-t‘o; and it is under this appellation or simply ‘ Ts‘ung-ling’ that we find Sarikol repeatedly mentioned in the Chinese records translated by him ®.
The earliest Chinese travellers of whose visit to Sarikol we have any record are the pilgrims Fa-hsien and Sung Yiin. But in the case of the former our knowledge is not only exceedingly brief, but dependent on a conjectural identification®, Fa-hsien and his fellow-pilgrims, when proceeding, about 400 A. D., from Khotan towards India, reached first the kingdom of Z2z-ho F- £. A notice of the T‘ang Annals translated by M. Chavannes plainly shows Tzt-ho to be identical with the territory known under the T‘ang as Chu-chii-po, i.e. the present district of Karghalik’. From there the pilgrims ‘went south for four days, when they found themselves among the Ts‘ung-ling mountains, and reached the country of Yu-hwuy, where they halted and kept their retreat’*. The name Yz-huc J PF, otherwise wholly unknown, presented a puzzle until M. Chavannes, by a slight emendation, restored it to Yd-mo fh B, an abbreviated form of the name Ch'tian-yii-mo }if Fy PH, under which Tash-kurghan is mentioned in the Pe shzh. We shall have occasion to follow elsewhere the ingenious and convincing arguments by which M. Chavannes further traces Fa-hsien’s route from Tash-kurghan to Czeh-ch'a or Kashgar, where he appears to have gone in order to rejoin some companions before attempting the passage of the Pamirs. But it may be noted that the four days’ march south of Tzii-ho or Karghalik to where the Tsung-ling mountains were entered, could well be explained on the assumption that Fa-hsien’s party for the journey to Tash-kurghan chose a route which first took them to Kok-yar, south of Karghalik, and from there westwards into Sarikol through the mountains adjoining the course of the Upper Yarkand river ®.
The account which the next Chinese traveller, the pilgrim Sung Yiin, has left us of his passage through Sarikol (519 a.p.), is less laconic. But a want of proper sequence in the
in a letter dated Sept. 4, 1903, the main points as discussed
“ See Varkand Mission Report, p. 56 (‘the rate is one by him in a note of his forthcoming translation of Sung Ytin’s
sheep for thirty yards of cloth. ..... No coin is current
in Sarfgh Kul, everything is by barter’). ® See Chavannes, Zurcs occid., p. 125 note; also the passages quoted, s. v. Ts‘ung-ling, p. 373, especially the one bearing on Kao Hsien-chih’s expedition to Gilgit, p. 152 note. 6 I am indebted for the first information concerning Fa-hsien’s probable route from Khotan to the confines of India to M. Chavannes, who was kind enough to communicate to me,
itinerary. [See now Voyage de Song Fun, p. 54, note 3.|
7 See Chavannes, Zurcs occid., p. 123.
8 Compare Fa-hien’s Zravels, transl. Legge, p, 21.
® The route here suggested would partly coincide with the one discussed below in connexion with Sung Yiin’s account of Sarikol; see note 13, p. 30.
Sec. ii]. EARLY CHINESE ACCOUNTS OF SARIKOL 29
details—a defect to be observed also in other portions of Sung Yiin’s narrative—formerly made it difficult to ascertain which of his data refer to Sarikol itself and which to the traveller’s subsequent route across the Pamirs. The annotated translation of Sung Yiin’s itinerary recently published by M. Chavannes, enables us to recognize clearly some characteristic observations bearing on Sarikol, and to trace the probable line of route}.
Sung Yiin, like Fa-hsien, travelled westwards through Khotan and entered the limits of the kingdom of Han-p'an-t'o Ye HR IR from the side of Chu-chii-po, which M. Chavannes has correctly identified with the modern district of Karghalik™. Going west for six days the traveller ascended the Ts‘ung-ling mountains, in which we must recognize the ranges separating the valley of the Raskam Darya from the plains about Karghalik. After three more days to the west he arrived at the ‘town of Po-yii’ or ‘Po-méng.’ Thence he ascended, in three days, the mountains called Pu-k'0-yz (meaning literally ‘the mountains on which it is impossible to find a rest’), where the cold was extreme and the snow was lying summer and winter. On these mountains a lake or the site of a lake was pointed out to the traveller as the former habitation of a poisonous Naga, who for his misdeeds had been banished through an early king’s magic. By difficult mountain routes he thence reached in four days the capital of the kingdom of Han-p‘an-t‘o. This territory was supposed to occupy the very summit of the Ts‘ung- ling mountains and to be the centre of heaven and of earth.
Sung Yiin notes that its inhabitants trained the watercourses to irrigate their crops, and when told that the cultivators of China depended on rain alone for their produce, they would not credit such a story. East of the city of Han-p‘an-t‘o was the river Méng-chin, which flowed to the north-east towards Su-lé or Kashgar. On the high Ts‘ung-ling mountains neither herb nor tree would grow. On their west side all rivers were flowing westwards. At the time of Sung Yiin’s passage, in the eighth month of the Chinese year, the temperature had already turned cold; the north wind was driving away the wild geese, and snow was falling over the whole region.
The description of the territory which Sung Yiin here gives is not as detailed as we might wish, but there can be no doubt that it fits accurately the central valleys of Sarikol, as recognized by M. Chavannes?*. The mention of the river flowing to the north-east, the bed of which lay east of the city of Han-p‘an-t’o, clearly indicates for the latter the site of the present Tash- kurghan, described also by Hsiian-tsang. A reference to the brief description of the Tagh- dumbash River Valley given above will suffice to show that Sung Yiin’s observations on the climate and character of this mountain region, and on the system of irrigation by which parts of the valley are made to yield crops, are quite accurate.
The most direct and frequented route from Karghalik to Tash-kurghan leads through the mountains towards the Raskam Darya. After crossing the latter near the group of Sarikoli villages known as Tong, it continues to the Kandahar (or Khandar) Pass, from which Tdash- kurghan is reached in four marches. It appears very probable that this was the route which Sung Yiin followed. Not having visited the route myself, and knowing only its western portion _ from the descriptions of Dr. Hedin and Captain Deasy, I must restrict myself to pointing out that the position of the village group of Tong would well correspond to Sung Yiin’s ‘town of Po-yii (or Po-méng)’, and that the Kandahar Pass, being about 16,600 ft. above the sea and
See Chavannes, Voyage de Song Yun, pp. 20 sqq. also below chap. tv. sec. i. “ See Chavannes, Zurcs occid., pp. 123 note, 311; *% See Chavannes, Voyage de Song Fun, p. 21, note 4.
Sung Yiin’s description of Han-p'an- fo.
Identifica- tion of Sung Yiin’s route.
Hsiian- tsang’s jour- ney over the Pamirs.
Hsiian- tsang’s route into Sarikol,
30 SARIKOL AND THE ROUTE TO KASHGAR [Chap. II
very steep, seems to represent the Pu-k'o-yi mountains as regards height, difficulty, and relative distance 1°.
Hsiian-tsang, whose itinerary has already been referred to as our main authority concerning ancient Sarikol, reached the district from the opposite direction, i.e. from the west. Returning after his long travels in India towards his native land about the summer of 642 a.p., Hsiian- tsang passed through Badakhshan into the kingdom of Ya-mo-hst-t'ce-t, which undoubtedly corresponds to the present Wakhan’*. In the valley of Pa-mz-/o, reached by the pilgrim after seven marches to the north-east, his earliest European interpreters could not fail to recognize the Pamir region. The salient features of his description, as Lord Curzon duly notes, ‘ stand out as an unmistakable picture of the Pamir country , and leave a doubt only as to the particular valley or Pamir by which the traveller crossed it’’ With Lord Curzon and the majority of Hsiian-tsang’s commentators, | believe that there is preponderant evidence in favour of the route which leads through the Great Pamir and past Lake Victoria; for only in the latter can we find a real approach to the position and size of ‘the great Dragon Lake’ which the pilgrim passed ‘in the middle of the Pa-mi-lo Valley’ '%
Lord Curzon has observed that this identification is distinctly supported by what we are told of Hsiian-tsang’s immediately succeeding marches. ‘On leaving the midst of this (Pa-mi-lo) valley and going south-east, along the route, there is no inhabited place. Ascending the mountains, traversing the side of precipices, encountering nothing but ice and snow, and thus going 500 li, we arrive at the kingdom of Chieh-p‘an-t'o 3% #% [1 The south-eastern direction, here indicated, of the further journey to Sarikol could not be accounted for if Hsiian- tsang was supposed to have travelled by the Little Pamir and past its lake, the Chakmaktin- Kul. A look at the map shows that a traveller proceeding from the latter towards Sarikol would have to follow the valley of the Ak-su for some distance to the north-east before he could reach a practicable route across the watershed range to the Tash-kurghan river.
No such difficulty arises if we assume that Hsiian-tsang’s journey lay over the Great Pamir. From the latter, two main routes are open to the traveller whose goal is the inhabited centre of Sarikol. He can either make his way in a generally eastern direction to the Naiza- tash Pass, the descent from which, in the Shindi valley running north-east, would bring him straight to Tash-kurghan. Or he may direct his route first into the valley of the Ak-su river, where it leaves the Little Pamir, and thence reach the upper portion of the Taghdumbash
13 The route here indicated may be conveniently traced in the map accompanying Dr. Hedin’s Rezsen in Zentral-Asten (drawn by Dr. Hassenstein) and in the Map of Portions of Western China and Tibet explored by Captain H. H. P. Deasy, published by the Survey of India Department, 1900. For a description of the route from Tash-kurghan eastwards as far as Unkurluk, see Hedin, Through Asia, ii. pp. 702 sqq. Dr. Hedin, who took six days from Tash-kurghan to Langar, one of the Tong villages, describes the summit or ridge of the Kandahar Pass (16,610 feet) as being ‘as sharp as a knife’ and experienced near it heavy snowfall on Sept. 19-20, 1905. I am unable to refer at present to Captain Deasy’s Three years in Tibet, which contains an account of the same mountain tract.
I have not been able to trace any information as to the existence in this region of a lake corresponding to that mentioned by Sung Yiin. But this could scarcely surprise
us if, as I suspect, Sung Yiin’s reference is really to a spot which no longer contained a lake but only the legendary site of one. The story of the Naga Susravas, whose earlier lake habitation is placed by Kashmir legend near the ancient city of Cakradhara, and who is believed to have subsequently banished himself to a distant mountain lake (Susram Nag, near the Amarnath Peak), seems to offer an exact parallel. The story is told by Kalhana, Raaz. i. 203-70, and has been discussed by me in my notes, Rajat. i. 201-3, 267.
* For a critical review of all available data bearing on the early topography and designations of Wakhan, Marquart, Eransahr, pp. 223 sqq., should be consulted.
® See The Pamirs, p. 69.
16 S7-yu-kz, transl. Beal, ii. p. 297.
% See Sz-yu-kz, transl, Beal, ii. p. 298; transl. Julien, ii. p. 209.
Sec. ii] — EARLY CHINESE ACCOUNTS OF SARIKOL 31 Pamir by going southwards, the watershed range being crossed here by a series of passes of which the Payik or Bayik Pass is the most frequented and easiest }8,
The initial portion of the route here sketched may be accomplished either by traversing the so-called Nicolas Range, which divides the Great and Little Pamirs, through the well-known depressions of the Benderski Pass (14,705 ft.) or the Urta Bél Pass (14,090 ft.) or by turning the eastern end of that range over the probably still lower Kizil-Rabat Pass1!®% Whichever the passes crossed, the general direction of the second route is south-east. This exact accord with the definite statement of Hsiian-tsang may, in view of his remarkable and often proved accuracy in the matter of bearings, be considered by itself a sufficiently strong argument for the belief that the pilgrim actually followed this route. But there are other indications, too, to sup- port it.
Several allusions in Hsiian-tsang’s itinerary from Badakhshan to Khotan, as well as statements contained in his Lzfe, show that he effected this portion of his return journey during the spring and summer*. At this period of the year the Great Pamir route is much preferred to that over the Little Pamir or by the Wakhjir Pass, inasmuch as the ascent along the Pamir river, coming from Victoria Lake, offers none of the difficulties to be overcome in the Ab-i- Panja Valley, when the swelling of the stream renders the tracks in the narrow gorge between Sarhad and Langar almost impracticable for laden transport ”1.
Now it is noteworthy that at the same season and through the same cause, viz. the melting snows of the mountains, similar difficulties arise beyond the Naiza-tash Pass which offers the nearest approach to Tash-kurghan from the side of the Great Pamir. The descent from the pass to the open Tash-kurghan Valley leads, for a distance of upwards of twenty miles, through the narrow rocky defile of the Shindi stream. This carries a very considerable quantity of water even early in the season, and having to be crossed and re-crossed in numerous places, according to the testimony of a very competent observer, renders ‘the road exceedingly difficult for laden horses’*., On the other hand, the Payik Pass, though its elevation (15,078 ft.) is slightly greater than that of the Naiza-tash (14,920 ft.), is entirely free from the risk of such
8 Compare for these passes, Lord Curzon’s Zhe Pamirs, pp. 57 sq.; Report of Pamir Boundary Commission, pp. 41 sqq.
1 Compare for these lines of communication, Lord Curzon’s The Pamirs, pp. 558q.; Report of Pamir Boundary Commission, pp. 17, 22.
* In the Zzfe (transl. Beal, p. 196) we are told that Hstian-tsang and his fellow-travellers, after traversing other parts of the old Tukhara country (Tukharistan), stopped in Po-to-chang-na or Badakhshan ‘on account of the frost and snow, for a month and some days’. Only after this long halt, which no religious or other obligations called for, did they march on to the Kokcha Valley or Yamgan (Yin-po-chien) and Kuran or Zebak (Chiti-lang-na) ez roufe for Wakhan. It is thus clear that they awaited the spring in the sheltered central valley of Badakhshan (probably near the present Faizabad, see Yule, /.R.A.S., 1872, pp. 109 sq.) before setting out for the passage of the Pamirs. Subsequently, when the pilgrim, after a twenty days’ stay in Chieh-p‘an-t‘o or Sarikol, was crossing the mountains to the north-east, probably by the Chichiklik route, we read of an attack by robbers, in the course of which ‘the elephants being driven about in
the pursuit, were engulphed in the water and perished’. This incident, related in the Zz, p. 200, clearly points to the time of the summer floods. The gorges passed by that route only then contain enough water to account for such a loss (comp. Gordon, Roof of the World, p. 108).
In Hsiian-tsang’s own narrative we find various references to climatic conditions of the mountains around Sarikol which distinctly suggest personal observation during the season above indicated. Thus of the valley of Pa-mi-lo we are told that ‘the snow falls both in summer and spring-time’, while again in describing the passage already referred to through ‘the eastern chain of the Ts‘ung-ling mountains’, north-east of Sarikol, the traveller points out that ‘even at the time of the great heat the wind and the snow continue’ (see Sz-yu-hz, transl. Beal, ii. pp. 297, 303; Julien, ii. pp. 207, 215).
*1 Speaking of this gorge Col. Gordon observes (Roof of the World, p. 129): ‘In summer the swelling of the stream makes this road extremely difficult, and sit is then that the Great Pamir route is followed in preference.’ Compare also Farkand Mission Report, p. 270.
*” See Captain (now Colonel) H. Trotter’s remarks in Farkand Mission Report, p. 270.
Season of Hsiian- tsang’s Pamir journey.
Route over the Wakhjir Pass.
Supply diffi- culties onthe Pamirs.
32 SARIKOL AND THE ROUTE TO KASHGAR [Chap. II
obstruction and, as noted by a good authority, ‘must on the whole be classed as exceptionally easy 78,’ |
The distance of 500 li, or approximately five marches, which the /7s¢-y#-ch¢ records for the journey from the midst of the Pa-mi-lo Valley to the kingdom of Chieh-p‘an-t‘o, agrees well with the route sketched above. According to the latest map of this part of the Pamir region prepared by the Survey of India, the journey from the eastern end of the Great Pamir by the Kizil-Rabat and Payik Passes to where the Payik Valley debouches into the Taghdumbash Pamir would cover about eighty-four miles, while a march of only some sixteen miles further down would bring the traveller to the village of Dafdar and the commencement of the cultivated part of the main Sarikol Valley.
Past Dafdar and the approach to the Payik Pass there leads also the third alternative route connecting Sarikol and Wakhan, to which we had occasion to refer in the opening of this chapter. It ascends the whole length of the Taghdumbash Valley, and then crosses the Wakhjir Pass at its head to the source of the Ab-i-Panja branch of the Oxus (Fig. 7). That part of the valley which properly bears the designation of Taghdumbash Pamir may be said to extend from the Wakhjir Pass to a short distance above Dafdar, where the river makes its sharp bend to the north. Forming an unbroken continuation of the central valley of Sarikol, the Taghdumbash Pamir, with the broad grassy flat along its bottom for a distance of upwards of fifty miles, looks as if created by nature for a convenient thoroughfare from Sarikol to Wakhan.
The Wakhjir Pass, close on 16,200 ft. above the sea, is, it is true, higher than either the Naiza-tash or the Payik Pass, and is certainly deep in snow in mid-winter. Its approaches, both from the east and the west, are, however, remarkably easy. Similarly, the descent in the Ab-i-Panja Valley to Bozai-Gumbaz, where the Little Pamir route joins in, and further down to Langar offers no difficulty of any kind. In accord with these observations we find it attested by Col. H. Trotter’s inquiries, made in 1874, that the Wakhjir Pass was in former times ‘much used by the Bajaori merchants who used to go from Badakhshan to Yarkand by the Taghdum- bash and Tung Valley roads’. Considering that for centuries past the commerce of the mountain region north and south of the Hindukush has to a very large extent been in the hands of these enterprising traders from Bajaor, the evidence recorded by Col. Trotter as to their usual route over the Taghdumbash Pamir possesses some historical interest ™.
At any period of regular trade intercourse between Sarikol and Wakhan, the Taghdumbash Valley route was bound to receive attention on account of a topographical fact which deserves brief notice. There is one feature of the Pamirs which seems to have impressed travellers of all ages with equal thoroughness: it is the total absence of permanent habitations and the want of all local resources. We can trace this feeling of utter desolation and sterility through the accounts of the Chinese pilgrims no less than in the records of European travellers since the days of Marco Polo and Goéz. From the elaborate arrangements for supplies and_ shelter which were deemed necessary on every occasion when large parties of western visitors such as Col. Gordon’s expedition or the Pamir Boundary Commission had to cross, or to camp on, the Pamirs 2°, we may judge of the difficulties which must always have attended the movements of trading caravans or military bodies across that region.
In view of this serious obstacle in the matter of supplies, it must always have been an important consideration to travel, if other physical conditions permitted, by a route on which
% See Sir Th. Holdich in Report of Pamir Boundary % Compare, e.g., Gordon, Roof of the World, pp. 124, Commission, p. 41. 163 sq. * See Varkand Mission Report, p. 270.
hein
GLACIERS AT SOURCE OF OXUS, SEEN FROM APPROACH TO WAKHJIR PASS.
(PHOTO-THEODOLITE VIEW.)
KILIK PASS, LOOKING SOUTH FROM KHUSH-BEL SPUR.
“ai % as
}
Sec. ii] EARLY CHINESE ACCOUNTS OF SARIKOL 33
the distance over wholly uninhabited ground was reduced as much as possible. Now in this respect the route over the Wakhjir Pass offers an undoubted advantage. If starting from Langar, the highest point in the Ab-i-Panja Valley where village remains and traces of old cultivation have been observed ?*, a journey of scarcely more than 106 miles would suffice to bring the traveller to Dafdar, and thus to the commencement of cultivated ground in the main Sarikol Valley. The shortest route from the same starting-point over the Little Pamir and the Naiza-tash Pass to Tash-kurghan measures fully 120 miles, while on the road that leads from the latter place by the Great Pamir to the central part of Wakhan, some 180 miles have to be traversed before a permanently inhabited locality (Langar-Kisht) is reached 2”,
Section III—HISTORICAL SITES OF SARIKOL
From this review of the Taghdumbash Valley route, which was needed to complete our survey of the old lines of communication between Sarikol and Wakhan, we may turn once more to Hsiian-tsang’s itinerary, and to the account it furnishes of Sarikol itself or ‘the kingdom of Chieh-pan-t'o’. Before we examine the details which possess a definite antiquarian bearing, it will be well to note how closely the pilgrim’s general description of the territory agrees with the natural features observed to this day.
In Chieh-p‘an-t‘o, to which is ascribed an approximate circuit of 2,000 li, or about twenty days’ marches, ‘the mountain chains run in continuous succession, the valleys and plains are very contracted. There is very little rice cultivated, but beans and corn grow in abundance. Trees grow thinly, there are only few fruits and flowers. The plateaux are soppy, the hills are waste, the towns are deserted.’! The account we receive of the inhabitants is in keeping with these stern surroundings. . Their manners are described as ‘without any rules of propriety’, and very few of them gave themselves to study. ‘They are naturally uncouth and impetuous, but yet they are bold and courageous. Their appearance is common and revolting; their clothes are made of woollen-stuffs.’ Notwithstanding the coarseness which Hsiian-tsang’s description indicates, and which the material conditions of life among a hill community thus placed adequately account for, the people of Chieh-p‘an-t‘o had received their share in the benefits of Buddhist religion and culture. ‘They know how to express themselves sincerely, and they greatly reverence the law of Buddha.’ There were some ten convents, in which about five hundred monks studied ‘the Little Vehicle according to the school of the Sarvastivadas’; and in regard to the letters in use we are told that they much resembled those of the CZ‘ca-sha country or Kashgar.
Hsiian-tsang, according to the statement of his biographer Hui-li, remained in Sarikol for about twenty days’, and to this halt we probably owe the detailed information he gives of the traditions and legends attaching to the royal family and the capital of the country. The account recorded by him as to the origin of the former is of particular interest, for it proves the high antiquity which popular belief in Sarikol then ascribed to the race of its rulers, while we can still trace a characteristic feature of the story in a legend actually surviving at the identical locality.
* See Gordon, Roof of the World, p. 129. It also ' See Sz-yu-kz, transl. Beal, ii. pp. 298 sq.; Julien, ii, p. deserves notice that firewood, another requirement of the 209. The latter version has: ‘Les plaines hautes et basses sont traveller, is plentifully found up to Langar, see ibid. and désertes, les villes et les villages sont [presque] inhabités.’ Report of Pamir Boundary Commission, p. 16. ? See Life of H. T., transl. Beal, p. 200; Vie de Hiouen-
7 Compare Gordon, Roof of the World, pp. 153, 163. Thsang, transl. Julien, p. 274.
STEIN F
Hsiian- tsang’s de- scription of Sarikol.
Traditions of Chieh- p‘an-t‘o.
Legend of foundation
of dynasty.
Legend of Kiz- kurghan.
34 SARIKOL AND THE ROUTE TO KASHGAR [Chap. II
After describing the reigning king of Chieh-p‘an-t‘o as a man of upright character, a pious Buddhist, and a lover of learning, Hsiian-tsang proceeds to tell us that ‘since the establishment of the kingdom many successive ages have passed. The king gives himself the title Chzh-na-t'r- po-chii-tan-lo (Cina-deva-gotra), meaning descendant of China and the sun-god*. Formerly this country was a desert valley in the midst of the Ts‘ung-ling mountains. At this time a king of Po-la-ssi (Persia)* took a wife from the Han country (China). She had been met by an escort on her progress so far as this, when the roads east and west were stopped by military operations. On this they placed the king’s daughter on a solitary mountain peak, very high and dangerous, which could only be approached by ladders, up and down; moreover, they surrounded it with guards both day and night for protection. After three months the disturbances were quelled. Quiet being restored, they were about to resume their homeward journey, when the lady was found to be enceinte’.® Thereupon the king’s envoy held council with his companions how to meet the consequences of this disgrace. From an attendant he learned that a spirit, coming from the sun’s disk and mounted on horseback, every day at noon visited the princess. Afraid of the punishment awaiting him on return to his own country, the envoy decided to seek safety by remaining and gaining time.
‘On this he built on the top of a rocky peak a palace with its surrounding apartments ’°; then having erected an enclosure round the palace of some 300 paces, he located the princess there as chief. She established rules of government and enacted laws.’ In due time she bore a son possessed of extraordinary beauty and miraculous powers, who made his strength felt in the countries around and everywhere gained recognition for his laws. When this king died of old age he was buried in a stone chamber constructed in the caverns of a great mountain, about a hundred li to the south-east of his capital. There his body was believed to remain undecayed, though shrivelled up and thin, looking as if he were asleep, and worshipped with regular oblations. |
‘From that time till now his descendants have ever recollected their origin: that on their mother’s side they were descended from the king of Han, and on their father’s side from the race of the sun-god; therefore they style themselves “descendants of the Han and sun-god”’’. Hsiian-tsang adds the observation that the members of the royal family resembled in their bodily appearance the ‘people of the Middle Country’, i.e. China, but notes that the dress worn by them was that of barbarians. For this he evidently means to account by stating that ‘in after ages these people fell under the power of barbarians’ *.
We are unable even approximately to gauge what shreds of historical fact may be interwoven with the legend which Hsiian-tsang heard about the family ruling Sarikol at the time of his visit. But we can scarcely doubt that this legend was widespread and firmly rooted in popular belief; for we find an unmistakable trace of it preserved in local tradition to the present day. To
3 I take the translation of this sentence from Julien, whose version is evidently more accurate than that of Beal (see Mémoires, ii. p. 210). The simple deva, ‘god,’ in the transcribed Sanskrit title is amplified by the Chinese text into ‘sun-god’.
* Thus Beal, Si-yd-h, ii. p. 300. Julien, in translation and index (ii. pp. 210, 524, resp.), spells the name Po-d-sse We Fi] ple but the graphic variation of the second character, i Fil, from the character /a HI in Po-la-sse, the usual Chinese transcription of the name *Parsa or Persia,
is so slight that there can be little doubt as to the latter country being intended.
5 Julien, Mémoires, ii, pp. 210 $q., translates: ‘ Au bout de trois mois les brigands restérent tranquilles. L’ambassadeur voulut alors emmener la princesse de Chine dans les états de son mattre; mais la jeune fille se trouvait déja enceinte.’
® Julien, Mémoires, ii. p. 212: ‘Alors il batit, sur le sommet de la montagne, un palais et un hétel.’
7 See above note concerning this title Cina-deva-gotra.
® See Mémovres, transl. Julien, ii. p. 213; Styu-hz, transl. Beal, ii. p. 302.
Sec. iii] HISTORICAL SITES OF SARIKOL 35
the remains of ancient stone walls perched on precipitous cliffs, which rise above the left bank of the Taghdumbash river some seven miles above Dafdar, there clings the story commonly known to Sarikolis and Kirghiz alike, that king Naushirwan, an ancient Persian ruler, had once placed there his daughter for safety. On account of this story the ruins bear the popular designation K7zz-kurghan, meaning in Turki ‘the tower of the daughter (or princess) ’.
Owing to an unfortunate chance I did not hear of the existence of these ruins until after I had reached Tash-kurghan (July 7, 1900), having failed to notice them two days earlier when I passed the site on my march down the valley. Want of time did not permit me to retrace my steps a distance of some fifty miles, however anxious I was to examine them personally. ~But I was at least able by repeated inquiries to make sure of their approximate position, and of the widely spread knowledge of the legend attaching to them.
There can be no doubt that we have in the Kiz-kurghan legend a genuine relic of the fuller tradition current in Hsiian-tsang’s days, and consequently we are justified in attributing historical significance to the place in which we find it localized. In the light of the story as recorded by Hsiian-tsang, popular tradition could not have supposed the Kiz-kurghan ridge to have been the temporary place of safety selected for the Chinese princess while the road westwards was blocked, unless at the time when the story was current a main route in that direction passed up the Taghdumbash Pamir at the entrance of which the ridge rises. Kiz-kurghan, like the modern fortified post of Ghujak-bai (Ujad-bai of the maps), which nearly faces it on the opposite bank of the river, must, in fact, be passed by all travellers who wish to reach the Upper Oxus, whether by the Wakhjir or the Payik Pass. Thus the legend localized at Kiz-kurghan affords direct evidence that the Taghdumbash Pamir was used as a general line of communication in ancient times, and further helps to support the assumption explained above that Hsiian-tsang himself travelled by it on his way to Tash-kurghan.
From the account of the /sz-yi-chz, already quoted, we learn that the palace in which the Han princess and her miraculously conceived son, the founder of the dynasty, first established themselves and ruled Sarikol, was built ‘on the top of a rocky peak’. In the absence of any notice to the contrary, we may assume that the site of this royal palace was within the capital of Chieh-p‘an-t‘o which Hsiian-tsang visited, and of which he tells us that it ‘rests on a great rocky crag of the mountain, and is backed by the river Sita’®, The position here indicated agrees so closely with that of the present Tash-kurghan that the identification