The Innermost Asia By Sir Aurel Stein, K.C.I.E

OXFORD, AT CLARENDON PRESS 1928
Learning Contributed by Quwat Khan Sunny SECTION 1.—YASIN IN HISTORY AND GEOGRAPHY (From Yasin to Kashghar)

By crossing the Sheobat pass I had gained access to the mountain territories of Gupis^ and Yasen, through which my route was to lead me straight north close to the watershed between Indus and Oxus. Ever since the Gilgit Agency was first established in 1877 more and more detailed information has become available about the geographical and kindred aspects of these tracts on the head-waters of the Gilgit river, and though some of the most useful books and surveys are not as yet within reach of the public, no general account of the ground over which this portion of my journey took me seems here called for. I may accordingly restrict my account to such observations as have a direct antiquarian or historical bearing, and to brief notes on the route actually followed by me and its successive stages. I may add that I propose to adhere to the same course in those further stages of my journey which took me across ground already fully surveyed or previously visited by me.

The region I traversed on my way from Tangir to the main Hindukush watershed presents a distinct historical interest, because the route which leads down from the Darkut pass through the open and comparatively fertile valley of Yasin must have always claimed importance as the shortest means of communication between Oxus and Indus. But the only notices shedding light on its early history are those found in Chinese records of the T'ang period, and as I have already had occasion to discuss them fully in the detailed reports on my two preceding Central-Asian journeys,. a brief summary of the main results there arrived at will here suffice. From the notices concerning ' Little P`o-lti ' ft contained in the rang Annals, which M. Chavannes was the first to render fully accessible and to elucidate,. it is certain that this territory must have contained Yasin and the valley of the Gilgit river also. It acquired considerable political and strategic importance for the Chinese when early in the eighth century the Tibetans operating from the direction of Great P`o-In or Baltistan endeavored to secure access through Little Po-hi to the Oxus valley and thus to join hands with the Arabs, the other great opponents of Chinese supremacy in Central Asia.. The necessity of keeping open the most direct route by which communication could be maintained from the Chinese side with Kashmir and other Indian kingdoms threatened by Arab conquest, made the protection of Little P`o-lff an equally imperative measure of imperial policy.

SEC.   YASIN IN HISTORY AND GEOGRAPHY   37

Kuku-nor helped to save Little P'o-lu from fresh Tibetan aggression. But it is the famous exploit expedition by which the great Chinese general Kao Hsien-chih in A. D. 747 brought an army across the Pamirs, defeated the Tibetan invaders on the Oxus, and led his troops across the Darkot pass into Little P`o-lu in the face of formidable natural obstacles that has furnished us with particularly interesting details. The topographical evidence elucidated by me when previously discussing the full record preserved by Kao Hsien-chih's biography in the Tang Shu leaves no doubt about the remarkable accuracy of that record. The description of the bold move across the ice-covered heights of Mount T'an-chti exactly fits the Dark& pass. The distances indicated conclusively prove that the town of A-nu-yiih, where the king of Little P'o-lu then resided, must be located at the present Yasin. In the same way it is certain that the bridge across the river So-i, the prompt destruction of which under Kao Hsien-chih's orders prevented the timely arrival of Tibetan reinforcements and thus ensured the immediate submission of the king and the people, corresponds to the bridge across the Gilgit river near the present Gupis, by which alone Yasen can be reached from the route leading up the main Gilgit valleys ⁶

Remote as these Hindukush valleys may seem, we can yet, thanks to the Chinese record of Chinese Kao Hsien-chih's expedition, realize the importance they assumed at a momentous juncture of garrison Asiatic history. The deep impression created by the occupation of Little   is significantly Little reflected by the closing remark of the Tang Annals on that success : Then the Fu-lin (Syria), the Ta-shih (Arabs) and seventy-two kingdoms of divers barbarian people were all seized with fear and made their submission,' But Chinese control over this region was not destined to last long. I have already had occasion above to refer to the Chinese garrison which Kao lisien-chip left behind in Little P`o-Iu, and to the difficulties of supply that its maintenance entailed. Very interesting light is thrown upon the conditions thus created by the representation which the ruler of Tokhäristan addressed in A. D. 749 to the Chinese Emperor and which has been fully analyzed by me elsewhere.'

From the Chinese records we know that in A. D. 75o effective Chinese intervention, once again Later under Kao Hsien-chih's leadership, relieved Po-hi and the mountain territories to the west from Chinese Tibetan pressure. But with that general's complete defeat in A. D. 751 by the Arabs, Chinese non. power in Central Asia was destined to decline rapidly, and the withdrawal of its distant outpost isolated in the midst of the Hindukush cannot have been delayed for many years. Yet as late as A. D. 753 we are told of an expedition led by Kao Hsien-chih's successor against Great P'o-1u or Baltistän, which can scarcely have been undertaken from any other base than that furnished by the Gilgit valley ; 9 and the arrivals of embassies and tribute from Little P`o-lit is recorded right down to A. D. 755.9

38   FROM YASIN TO KASHGAR   [Chap. II

Modern Local History: 

From the time when the Hindukush region passed out of China's sphere of interest in the eighth century reliable historical records concerning Yasin and the adjoining valleys fail us for nearly a thousand years. The account of local history which Colonel Biddulph and others have gathered from oral tradition assumes a definite shape only with theadvent to power, towards the end of the seventeenth or early in the eighteenth century, of the family, apparently of Badakhshan origin, from which are descended both the Katur rulers of Chitral and the Khushwaqts originally established in Mastuj." The Khushwaqt branch appears very soon to have asserted its power over Yasin also ; and owing to the superior capacity for war or for intrigue possessed by many of its members, the whole of the Gilgit valley likewise passed at different times under its temporary domination. There is no need to examine here the tangled web of a story in which struggles, marked uniformly by treachery and murder and waged between close relations or with members of the rival house of Chitral, prevailed right down to the close of the nineteenth century?' A few points, however, deserve mention. It is interesting to note that, as I have had occasion to point out elsewhere, we owe the earliest exact record connected with Khushwaqt rule in Mastuj and Yasin to Chinese intervention in 1749.. Traditions of Chinese or ' Kalmak ' invasion still survive in Yasin, but are too vague to be fixed chronologically.

Advantages of geographical position and Fertility of Soil:

Significance attaches to the fact that though the Ydrichun valley below and above Mastuj is the original seat of the Khushwaqt branch, yet Yasin was always preferred by them as a residence.. This preference is fully accounted for by the advantages which Yasin offers by its geographical position and natural features. The fact that in its main valley open ground of comparatively great width extends for a distance of about forty miles would alone suffice to give it importance. There are here none of those narrow defiles, formed by precipitous spurs of rocky or vast debris shoots, which in other great valleys to the south of the main Hindukush greatly reduce the area of arable ground and render communication between them difficult. The glacier-fed waters of the Yasin river and its side streams make irrigation easy, and if considerable portions of the available ground are now left uncultivated, the cause is certainly not want of water but an inadequate population. The same high flanking ranges, showing peaks over 20,00o feet in height, which assure this abundant supply of water, also protect Yasin against attack on all sides except the south. There, too, as the account of Kao Hsien-chih's expedition shows, the Gilgit river, unfordable for the greater part of the year, serves as a very effective obstacle to invasion, especially as the extremely precipitous spurs on either side of the outlet of the Yasin river form flanking defences of exceptional strength.

That Yasin could, and once did, support far more than the present population, estimated at about five hundred families or about 4,700 souls, is proved by the extent of the ground capable of irrigation and by the fertility of the soil. The fact that the whole of the main valley from Dark& village down to the point where it debouches opposite G anis lies at the moderate elevation of between 7,000 and a little over 9,000 feet would alone account for this fertility. But the north

Sec.!]   YASIN IN HISTORY AND GEOGRAPHY   39

to south direction of the valley, whereby all the land obtains sufficient sunshine and a shorter season of severe cold, adds greatly to the favourable character of its climate. The scantiness of the present population is ascribed by Colonel Biddulph, no doubt quite correctly, to oppression and misgovernment," and that this has been long continued is sufficiently clear from what we know of the history of Khushwaqt rule during the last two centuries."

The fact that the constant wars of the Yasin rulers since the beginning of the century have ', Natural as Colonel Biddulph has rightly observed," ' been the most powerful agent in depopulating the dorMin. country ', supplies an additional reason for drawing attention here to a geographical factor of interest. I believe that this warlike activity of the Yasin rulers is itself largely explained by geographical conditions. A look at the map shows that the peculiar position of the Yasin valley makes it a favorable base for aggression in the direction both of Chitral and of the main valley of the Gilgit river, access to the upper portion of which it completely commands. On the other hand Yasin itself is easily defended both on the north and south, as already pointed out, and its remoteness alone tends to make it secure from local attack by tribal communities or petty chief-ships. It is only when command of the direct route connecting Indus and Oxus by way of Yasin and the passes of Darkot and Baroghil becomes an important object for distant but powerful' neighbours, that the seclusion of Yasin fails to protect it. It was thus when Tibetans and Chinese in turn were striving for a hold upon Little   and again in our own days when political
developments between two big Asiatic powers affected remote Yasin in a curiously similar manner."

This relative remoteness and seclusion of Yasin are reflected in an interesting ethnological Burushaski and linguistic fact. The language spoken in Yasin by the bulk of the population is Burushaski, trnr"" locally known as Wurishki, a tongue wholly distinct from the Dardic languages of the Hindukush region and without any known relationship. The name of the tongue is derived from the designation Wurish which the Yasin people apply to themselves, and this appears again in the name Wurshighm or Warshigum, by which Yasin proper is still known locally."' Outside Yasin Burushaski is now spoken only in Hunza and Nagar, both of them territories that in position exactly correspond to Yasin and are even better protected by natural difficulties of access. But there is plentiful linguistic evidence that in earlier times the area where this strange language was spoken extended much farther to the south. Traces of its influence have been found in almost all Dardic languages, however far removed from the valleys where Burushaski (Wurishki) still survives.

It has been long recognized that the present restriction of Burushaski to the most remote Recession valleys south of the main Hindukush range and to a very small portion of the total population of the whole mountain region points to a gradual withdrawal and absorption of the race that spend, originally spoke it, due to a wave of `Aryan' invasion represented by the tribes speaking Dardic languages.

This conclusion is supported by the fact that the Burish or Wurish are considered to be of the same caste as the Yashkun who form the entire population of Punial, below Gupis, and a very strong element numerically lower down also in the Gilgit valley, as well as in Astor and Darel. There is apparently little at present to distinguish the Burish or Wurish people (Fig. 42) in physical features from the type prevailing among the Dard-speaking population, whether of the Yashkun or of other castes in those valleys. But in view of the absence of adequate and systematically collected anthropological materials the relation between linguistic and racial divisions in this area can scarcely as yet be examined with profit.

In my preceding observations on the geographical features of Yasin I have specifically had in view the main valley, i.e. Wurshigarn. But in a political sense Yasin all through its modern history included also the small mountain tracts of Kho (Kuh) and Ghizar on the uppermost Gilgit river, and having regard to their geographical position it is clear that this must have been the case also during earlier periods. A brief reference to them is, therefore, needed here. The tract, designated as Kho by Colonel Biddulph but spelt Kula by more recent authorities, comprises the very narrow strip of valley through which the Gilgit river passes immediately above and below its junction with the Yasin river at Gupis. The fact that with a total length of about thirty miles this portion of the valley counted in 1900 a population of only some 1,200 souls, sufficiently illustrates its very confined nature. Of the small side valleys which join it from the south and are included with it, only the Batres-gah Nullah supports a few hamlets. A very narrow defile, defended by an old tower some miles above the mouth of Batres-gah, divides Kho from Ghizar, while eastwards, in a still more difficult gorge extending for about nine miles, lies the boundary towards Punial and Gilgit... It is this easily defended defile which forms the true natural barrier between Yasin and Gilgit and explains the inclusion of Kho in the former territory.

The Ghizar tract extends right up to the watershed towards Mastuj and Chitral, formed by an easy saddle (12,250 feet above the sea) immediately west of the Shandur lake. The upper part of the valley from above Ghizar proper is fairly open, and this accounts for the presence of a population about twice as large as that of Kho. The importance of the district, however, is due solely to the easy route which it affords for communication between Laspur and Mastuj on the one side and Yasin and Gilgit on the other. The facilities afforded by this route explain why Mastuj and Yasin, though on opposite sides of a mighty mountain range rising to peaks over 21,000 feet in height, were yet for a long period united under one rule. They help us also to understand why the Chinese mission which Wu-kung accompanied about A. D. 751, coming from the Pamirs via the Broghol chose the detour through Mastuj, Laspur and Ghizar, as I have elsewhere shown that they probably did, in order ultimately to reach Udyana from the Yasin side.

Ghizar, now separately administered by a' Governor ' of Khushwaqt descent, lay at a distance Head oi from my route. But of the Kho tract I was able to see something on my way from the Sheobat pass to Yasin proper. The descent from the former took us on August 23rd over ground clearly `alter. marked as the bed of a former glacier and past steep slopes of rock debris to the high grazing ground known as Kuterao-ferao. Next day, a couple of miles lower down, we struck at Mayurai the Batres-gah valley. Here an area of carefully levelled ground, not less than half a mile square, attested former cultivation at an elevation of over to,000 feet. The Batres-gah valley looked comparatively open. A track practicable for laden animals ascends the valley south-eastwards to the Suj-gall pass leading to Nyachut in Darel, and this is the route by which Fa-hsien and his Chinese fellow pilgrims probably proceeded to Ta-li-lo and the Indus. For our own journey to Yasin two marches down the Batres-gah Nullah to its mouth and then another along the Gilgit river to Gupis would have offered the easiest route. But in order to save a day I chose the short cut that was reported to lead above Gafar-bado to a pass giving direct access to the head of the Gupis Nullah due north.

The mountains to the south of the Gilgit river have in these parts not yet been adequately Pass above surveyed. So it was scarcely surprising that this unexplored pass proved almost impossible for' our porters. For fully eight hours we scrambled over huge masses of rock debris and boulders (Fig. 32), the worst I ever encountered in this region, relics of an ancient glacier, before we reached the narrow rock gully forming the pass at an elevation of just under r6,000 feet. The entire absence of glacier mud or other soft soil over all this ground seemed a clear indication of the dryness of the climate on this side of the Indus watershed, far advanced denudation resulting from it. On the steep north slope of the pass nevi-beds were still found, the last remnant of what must have been, within a relatively recent period, a glacier of some size. Nightfall obliged us to camp amidst old moraines at an elevation well over 13,000 feet.

The following day's march led down to Gupis through a steep and gradually narrowing Descent to valley. In its upper portion it was of interest to observe a succession of ancient terminal moraines Gapis. marking at intervals the points to which the glacier had advanced at different geological periods. Stretches of sloping meadow land used for grazing separated these steep falls of rock debris, the lowest of them being met with some nine miles below our camping-place. Below the junction with the Bashkar-gah branch, which descends from the south-west and still holds an active glacier, the valley contracts into an extremely narrow gorge flanked on either side by rocky precipices. These cliffs, which at their base showed in places clear marks of glacier ' grinding ', appeared to rise to a height of 3,000 feet or more above the canon-like bed of the stream. Their name Upaiyat, interpreted to mean in Shina ' higher than birds can fly ', suggests some local legend similar to that which accounts for the origin of the ancient name Paropanisus (Avestic Upairi-cadna) borne by the Hindukush main range north of Kabul. A very steep spur jutting out above the main valley near Gupis and known as I shkerbal was pointed out as a natural place of refuge resorted to in old times.

August 24th, spent at the village of Gupis, was our first day of halt since leaving Kashmir. Halt at We took advantage of it for work of many kinds, in which we were assisted by the presence of the Gepis Fort. small garrison of Imperial Service troops holding Gupis Fort. This effectively guards both the mouth of the Yasin valley opening on the opposite side of the river and the route that leads to Mastuj and Chitral. It was of direct historical interest to find this testimony to the strategic importance of the point under present, as under past, conditions ; for, as mentioned above, it is certain that the bridge across the river So-i, e. the Gilgit river, the timely destruction of which played so decisive a part in Kao Hsien-chih's successful operations against the Tibetans, must have stood in the close vicinity of the present Gupis. The modern wire suspension bridge giving access to Yasin crosses the bed of the united Gilgit river almost opposite to Gupis Fort. But in view of the considerable width of this beds of the big volume of water carried by the river during the greater part of the year, and of the materials available in this region, it is unlikely that any bridge other than a mere rope-bridge could have been constructed here before the days of modern engineering.

 A rope-bridge of the old type such as is usual between Kashmir and the Hindukush, con-
structed with ropes of twisted twigs, actually existed at Gupis before 1895, and it is probably to a bridge of this kind that the Chinese record contained in Kao Hsien-chih's biography refers where it speaks of a pont de rosin'? It is true that a rope-bridge would not have beenpracticable for the horses, or rather ponies, of a mounted force such as the Chinese account mentions as forming part of the Tibetan troops. But animals might be swum across the river, as they are elsewhere at the present time. Nor should the possibility be excluded of a bridge of a somewhat more substantial kind having been available at a point above the junction of the rivers of Yasin and Ghizar (Fig. 37). There only the latter would have to be crossed in order to gain access to the Yasin valley. In such a position, about four miles above Gupis Fort, a rickety bridge constructed of poplars was maintained until recent years across the Ghizar branch of the river, though liable to he carried away by summer floods..

An easy march of some thirteen miles on August 25th brought me fromGupisto Yasin, the chief place of the valley. The openness of the ground at the bottom of the valleywas the more impressive for the extreme steepness and height ofthe bare rock walls which confine it on either side. Much abandoned cultivation below the hamlet of Gindal bore witness to the vicissitudes through which Yasin had passed. I found some of these abandoned lands in course of being reoccupied by recent emigrants from Badakhshan. Their presence, together with many imported articles of dress and the like, were a sign of the vicinity of the Oxus region and of the influence that its civilization has always exercised beyond the Hindukush watershed.

Yasin proper, the largest village in the district, stands in the stretch of fertile ground, covered visit to with rich fields and orchards, that extends for about four miles along the right bank of the river Lasitin, below the issue of the Nasbar valley (Fig. 41). A day's halt there enabled me to visit the late Raja Shahid-u1-`Ajam, then Governor of Yasin, in the old and now half-decayed castle that had sheltered his Khushwaqt ancestors during generations of strife and bloodshed. In the plentiful wood-carving of its tumble-down halls the predominance of Persian architectural ornament was unmistakable, clearly pointing to models from distant Badakhshan. It was interesting also to observe signs of the time-honoured feudal devotion linking Vasa/ people with the race that for two centuries and down to quite recent times had almost constantly misruled them.

From the rest-house situated not far from the outlet of the Nasbar-gol, on ground now under- Old remains going reclamation after prolonged abandonment, I proceeded to examine the spot, a quarter at Yash' of a mile to the south-west, where some old remains were reported to have been brought to light by shepherds, about three years before my visit. I found there, at the rock-strewn foot of the hillside, the' remnants of a walled platform measuring about 18 by 3o feet, and on it a small circular mound formed of rough stones and rubble and probably marking the last relic of a completely decayed Stripa. The mound had been levelled almost to the ground by digging ; but among the debris thrown down on its north side a careful search brought to light eight small clay seals,. of which the best preserved showed the relievo representation of a Stapa with five ' Chhattras' and around it traces of the Buddhist formula ye dharmaprabhavalt, &c., in Nagari characters of the late type common in Tibet. These clay seals, similar to those found at the sites of numerous Buddhist shrines in India and Central Asia, had evidently formed part of a votive deposit disturbed when the mound was dug into.

Just to the north of Yasin and the mouth of the Nazbar valley there extends on the right bank an almost level plateau, about three miles long and one mile wide, known as Dasht-i-taus.It is known to local tradition to have once been cultivated, and its position is such as to permit of its easily being brought under irrigation again by a canal from the Nasbar stream. At the south-western extremity of this plateau, where it falls off with precipitous cliffs of conglomerate towards the bed of the Nasbar-gol, there rise the much-decayed walls of a ruined fort known as Chumarkhan. As seen in the plan (PI. 1), it forms an irregular quadrilateral, with a face about 17o feet long crowning the cliffs above the stream and a keep-like structure in the centre measuring 18 by 20 feet within (Fig. 40). The walls, built of flat pieces of slaty stone and large rubble pieces inserted between them, are as much as three feet in thickness, but now rise nowhere to more than five feet above the ground. That it was meant to guard the approach to the Dasht-i-taus plateau from the Nasbar valley, across a narrow saddle immediately below the north face of the fort, there can be no doubt. The advanced state of decay of the walls points to their considerable antiquity. Popular tradition ascribes the fort to the same age as the cultivation of the Dasht-i-taus.

Resuming on August 27th my journey up the main valley, I rode along this now utterly barren plain and was shown in several places traces of an old canal coming from the side of the Nasbar stream. Its line was quite distinct from that of a smaller canal derived from the Thui stream farther north and much better preserved ; by the latter canal Sulaiman Shah, a Khushwaqt ruler of Yasin and Gilgit at the beginning of the nineteenth century,. had endeavoured to bring water once more to Dasht-i-taus. This enterprise of reclamation is supposed to have been abandoned when the rule of that energetic but unscrupulous prince came to a violent end. That the occupation of the Dasht-i-taus belongs to an earlier period is proved by the remains of a large walled enclosure called Bahri-khan (Pl. t), situated about two miles above the northern end of Yasin and opposite to Ghujalti village on the left bank (Fig. 38). Tradition connects it with a Chinese or Ka.lmak ' invasion, possibly the same that I have had occasion to discuss before in connexion with the Chinese record of Khush-amad's reign about the middle of the eighteenth century.• The enclosing walls, built of large water-worn stones from the river-bed below and fully five feet in thickness, form an irregular pentagon of which the three best-preserved faces measure about 264, 273 and 153 feet respectively. Their present height nowhere exceeds four or five feet, and the remains of large quarters traceable near the centre of the enclosure are even more decayed. Apart from debris of hard dark-grey pottery within the ruined fort, I could trace no signs of prolonged occupation. But there can be no doubt that if the old canal were restored, or even if that of Sulaiman Shah were completed, the amount of arable land, and with it the population, in the Yasin valley could be greatly increased.

Continuing our march up the valley we followed the line of Sulaiman Shah's canal for nearly four miles before crossing the mouth of a large river that descends the Thui valley. Through this leads an important summer route, by which the upper Yarkhun valley can be reached on the Masuj side across the high Thui pass. Beyond the junction the main valley, known from this point upwards as Warshigam, affords room for a succession of picturesque villages with rich fields and orchards, extending almost without interruption up to Hondur. They are comprised in the subdivision of Salgam that forms at the present day the best-populated tract of Yasin. Passing the large fort of Mir Wall, called after Pakhtan Wali's father, Hayward's murderer, I visited at Barkulti the fine but much-neglected house of the local ' Hakim' (Fig. 45). The style and abundance of the excellent wood-carving in its large halls recalled what I had seen in Hakim Obaidullah's house at Miragram on the Mastuj side.• It was unfortunately too late in the day to take any photographs of the interesting interior. The house was said to have been built some five generations ago. A photograph (Fig. 42) taken at Hondur, where we halted for the night, may show how closely the fine physical appearance of the people of Warshigum resembles that of their neighbours in Chitral and Mastuj. Notwithstanding their Burishaski speech, wholly distinct from either the Hardie or the Iranian language group, the ' Burish ' whom I saw seemed to exhibit all the physical features characteristic of the pure Homo Alpimes type.

From above Hondur cultivation becomes intermittent, as the valley northward narrows. But in spite of the stupendous rock walls rising on either side progress along its bottom remains quite easy right up to the village of Darkat, a distance of nearly twelve miles. Here, at an elevation of over 9,000 feet, the head of the valley opens out into a huge amphitheatre, forming a wide grass and jungle covered flat flanked on its sides by mighty ice-clad spurs. The streams issuing from the glaciers which fill the side valleys between these spurs unite close to Darkat to form the river of Yasin. The plain extending around their junction abounds in grazing and fuel, and seems as if created by nature as a resting-place for a force of invaders from the north, such as Kao Hsienchih had successfully led across the Darkot pass. It struck me as a significant indication of the vicinity of the uppermost Oxus valley that one of the headmen of Darkat was an immigrant from Wakhan, and that the plentiful ornamental wood-carving in his panelled ' Aiwan ' was of a distinctly Persian style."

A move of some three miles northward across a boulder-strewn plain, no doubt once the bottom Approach of an ancient glacier basin, brought us to the entrance of the narrow gorge by which the stream from the glaciers adjoining the Darkot pass has cut its way through a transverse ridge. This ridge, which the route crosses, is known as ' Darband ', from old walls intended to form a chiusa. It was from the crest of this ridge, at an elevation of about 10,000 feet, that the ice-covered depression in the range forming the Darkot pass, 15,38o feet above the sea (Fig. 44), first came into view from the south. Both to the west and east of it glaciers, of far greater size than that below the pass, descend from peaks towering to heights well above 2o,000 feet. The abundant moisture assured by the vicinity of the streams that drain all these ice-clad slopes favours vegetation. I was therefore not surprised to find that besides pasture and numerous birch-trees, terraced fields cultivated by the Darkot people extend above those streams to an elevation of about t i,000 feet. Our march that day was continued past these fields of Gakushi and up the broad but steep spur which gives access to the pass, until at an elevation of about 12,305 feet we encamped on the last level bit of ground, known as Khamba. The smiling alpine landscape spreading to the south struck me by its contrast with the great wastes of ice and rock that I remembered so well on the north side of the pass from my ascent to it seven years before.

It was the desire to see the scene of Kao Hsien-chih's great exploit that had then induced me to visit the Darkot pass in spite of the early season, bad weather, and an exceptional winter snow- fall. It was the thought of it, too, that now invested the actual crossing of the pass with special interest for me. It was effected on August 29th, with all the advantages of the late summer season and ample arrangements for transport. All the same it served to impress me again with the seriousness of the natural obstacles presented by the glaciers of the DarkOt. I realized more than ever that the Chinese general's passage in A. D. 747 with a relatively large force, already severely tried by their march across the whole width of the Pamirs, deserves to rank, as a great military achievement, side by side with the most famous alpine feats of leaders such as Hannibal or Napoleon, if it does not surpass them. I have discussed elsewhere this remarkable exploit in full detail,'' and have shown how closely the topographical features of the Darkot pass agree with the exact account of this expedition which Kao Hsien-chih's biography in the Tang Annals has preserved for us. I need therefore only record here such of the observations made on my renewed visit to the Darkot as will help to supplement my previous description of the pass.

The track above our camping-place, Khamba, ascended very steeply along bare rocky slopes, but when free of snow it was practicable for laden animals. After about a mile it passed at an elevation of about 13,10o feet the large inscribed boulder of which I had first heard at Yasin (Fig. 46). It lies a few yards above the track, and its top, sloping at an angle of about 45°, presents an almost flat surface measuring approximately five feet by four. The stone appeared to be a dark-grained granite ; the surface on the exposed top has weathered to an almost black appearance. This surface shows in its middle portion and engraved to a depth of about one-fifth of an inch the outlines of what manifestly is meant to represent a Stripa, and by its side to the right five rows of Tibetan characters, two in each for the most part manifestly coeval with the Stupa. In

all these the engraved lines show a uniform brownish colour, while over the middle of the stone and to the left of the Stupa, there appear much shallower graffiti of Arabic writing, easily distinguished by their light colour. This difference of colouring and the fact of the graffito in the middle running across the central part of the Stupa leave no doubt that these graffiti are of relatively far more recent date. In addition to these graffiti there appears between the top of the Stupa and the Tibetan inscription the rough sketch of a horse and rider, showing a colouring similar to that of the former. The surface immediately below the Stupa base has suffered by peeling, but retains traces of a flower or fruit-like design, also old. The roughly scratched figures of four-horned animals, evidently meant for mountain sheep, which are to be seen on the left side of the stone between the graffiti in Arabic characters, seemed to have been exposed to weathering somewhat longer than the latter.

The Stupa design (Pl. 1) is very peculiar, but the inverted bowl on its top, surmounted by what is manifestly meant for a chhattra, suffices to establish its character. The two lowest steps of the base and the much higher member above them certainly recall the threefold base of the Stupa engraved on the boulder of Charrun in Mastuj which Fig. 6 and Pl. 2 of Serindia reproduce." But the cross-like design intervening between the shaft and the inverted bowl bears but a very distant resemblance to the projecting plinth and the drum that the rock-engraved representations of Stupas at Charrun and Pakhtornihni both display in a corresponding position below the dome. Peculiar, too, is the substitution of an inverted bowl for the hemispherical dome of the Stupa, though Buddhist tradition has from an early date sought to recognize in this dome a symbolic representation of Buddha's pcitret or begging bowl.' With the curious presentation of the pedestal or supports meant to carry the Chhattra ' may be fitly compared the equally coarse design which the PaPhtaridffii rock-carving shows in the same place. Finally, poor as the drawing of the umbrella at the top is, there can be no doubt what it is intended to signify.

Dr. A. H. Francke, to whom I submitted photographs of the rock-engraving together with carefully drawn copies of the Tibetan characters, was kind enough to furnish me, in a letter dated September 15, 1921, with a note on them reproduced in Appendix L. From this it is seen that the inscription names a certain Lirnidor together with his family or clan designation 2-Me-'or, probably taken from a locality, as the donor of a Stripa. The fact of the personal name being put in the genitive is taken by Dr. Francke as an indication of the early date of the inscription, and this is in harmony with the palaeographic character of the letters, which, in that scholar's opinion, ' show the characteristic marks of the Tibetan script of the eighth and ninth centuries '.

From the relative position of the two, it is obvious that the representation of a Stupa is contemporary with, if not older than, the Tibetan inscription. This chronological indication has its special archaeological interest with regard to the peculiar cruciform type of Stapas which Dr. Francke has noted before among Ladakh rock-carvings.

Local tradition, so far as it goes back, knows nothing of Tibetans having ever established their Buddhist worship in Yasin. Hence all the more interest attaches to the statement in Dr. Francke's note that a Tibetan text mentions the conquest of the Gilgit region as having taken place under a Tibetan ruler of the eighth century. In view of this collateral evidence one is tempted to connect the Buddhist rock-carving on the Darkat with that short-lived Tibetan advance on the uppermost Oxus which the Tang Annals record towards the close of the second quarter of the eighth century and which Kao Hsien-chih's adventurous expedition successfully arrested.

The difference in the weathering shows that the graffiti in Arabic characters must be considerably later than the Tibetan inscription. They consist of the Shiah invocation Ya Allah, Ya ali Madad, scribbled across the cruciform portion of the Stupa, with another Ya Muhammad ya Muhammad ya 'Ali below it. To the left of the Stupa, written vertically, we read the names Khawar Shah Shah Khushwaqt and in the left top corner Ya ali Madad Badshahe Mardan. Finally across the lowest base of the Stupa is scrawled the record, ba kalam Murad Beg. None of the persons here named seems capable of definite identification at present ; for the name of Khushwaqt appears to have been borne not solely by the founder of the family ruling Yasin and Masuj since the beginning of the eighteenth century, but also by at least two of his descendants."

Close above this spot the track took us to the moraine on the east side of the much-crevassed Ascent of glacier that descends from the pass. After scrambling up this to an elevation of nearly 14,000 feet, we crossed the glacier in a zigzag line to avoid crevasses. Further progress lay along its pats. snow-covered western moraine, past precipitous cliffs and avalanche slopes, until easier ground was gained at about 14,600 feet on snow-beds flanking the head of the glacier. At last, after four hours' steady climbing, we reached the broad and flat expanse of snow that forms the crest of the pass (Fig. 43). The reading there taken with the mercurial mountain barometer indicated a height of 15,050 feet, closely approximating to that recorded in the Trans-frontier map sheet (15,38o feet).

This great fire bed was the same to which my ascent of May 17, 1906, had brought me, and full description then recorded in my Personal Narrative" makes it unnecessary to note further topographical details about this remarkable pass. It will suffice to mention that though the gathering snow clouds soon deprived me of a full view to the south, I was once more greatly struck by the contrast which the steep glacier bed on that side, confined between precipitous spurs, presented to the wide snowy expanse of the summit and to the gentle slope of the mighty ice stream flowing down from it towards the Yarkhun gorges (Fig. 48). It was easy for me to realize again on the spot the feelings of dismay which filled Kao Hsien-chih's Chinese ' braves' when they had been brought to this height and found what a descent lay before them, and to appreciate the great qualities of their leader, who by combined boldness and stratagem assured final success in his adventurous enterprise.

Please share it with: