Sec. V] ALONG THE WESTERNMOST KUN-LUN 89
case, it deserves to be noted that the annual taxes payable into the Karghalik treasury on account of sales of animals alone were assessed at twelve ` Yambus ' or horseshoes of silver, corresponding to about Rs. 1,500,—not counting the Beg's additional octroi. In the days before Hunza came under effective British control (1891), the Pakhpu valleys were a favourite hunting-ground for Kanjutis raiding across the Shimshal Pass for cattle and slaves. Not less than 170 Pakhpuluks had thus been carried off within living memory and never heard of again.
On two points the information obtained from my Pakhpu visitors presented some antiquarian interest. Hsüan-tsang tells us, evidently from stories heard on his passage,16 that there was on the southern border of the kingdom of Chê-chii-chia, or Karghalik, a high mountain with very elevated passes and peaks piled up one above the other. Plants and trees are stunted by the cold. From spring to autumn the torrents of the valleys and the mountain sources spread on all sides. There one sees niches in the flanks of the mountain, and cells among the rocks. They are disposed in a regular fashion among the grottos and woods. Many Indians, having obtained Arhat-ship, display their supernatural faculties, rise into the air to travel afar, and come to settle in these places. A multitude of Arhats have entered Nirvana there. On that account a great number of Stûpas have been constructed. Even now there are three Arhats residing in these rock caverns. They are plunged in the state of ecstasy producing " extinction of mind ". Their bodies are shrivelled ; their beards and hair continue to grow, so that monks come from time to time to shave them.' There can be no doubt that the story here recorded by Hsüan-tsang rested on old local tradition, for the same account of the miraculous Arhats was heard already by the Indian Buddhist traveller J inagupta when he passed through Chê-chii-chia about A.D. 556 on his way from Gandhara to China.'7 It is equally clear that the legend must have been localized at some natural rock caves.
It was, therefore, of special interest to me to learn from my Pakhpu visitors that there are at least four well-known caves in their mountains, and that two, if not all of them, are looked upon as Mazars, and thus sites of local worship. One, said to be ` large enough to hold over 200 sheep', is situated near the point where the Kara-kâsh J ilga debouches into the main valley of Pakhpu, and is held sacred as the resting-place of a saint, Sultan Köputwali'. Another cave exists at Kulânarghu, a place apparently to be located at the head of the high valley between the Takhta-koram and Kukalyang Passes. A third, in the Kilda Valley, is passed by the route from Kök-yar to the Yangi-dawan. It was particularly curious to hear of a small cave in Chukshü, near the head of the valley leading to the Yangat (*Yangi-art ?) Pass, which is looked upon with much awe by the hill-men as the resting-place of a miracle-working Faqir who had died there ` in old times '. By their position and sacred character these caves seem to correspond closely to the sites which figured in the legends heard by Jinagupta and Hsüan-tsang, and thus to furnish a striking fresh instance of the survival of Buddhist local worship in these parts. Nor need Hsüan-tsang's reference to woods in this region appear as strange as the now generally barren appearance of these mountains might suggest. For, rare as tree growth must be under their present climatic conditions, Rai Ram Singh on the surveys for which I had dispatched him towards the Karlik-dawân, actually discovered considerable fir forest still surviving in the Akchik Valley between Tatligh and Tarishilagh-öghil, at an elevation close on io,000 feet.'R Nowhere else in the Kun-lun do I know of firs or similar tree growth. The discovery of this forest accords remarkably well with the peculiar luxuriance of trees in the Karghalik oasis. This might, perhaps, be due to a local climate rendered less dry by
'a See Julien, Mémoires, ii. pp. 221 sq. ; Beal, Si yu-ki, ii. P. 308.
" Cf. M. Chavannes' paper on Jinagupta in T'oung pao, 1905, pp. 332 sq. His narrative, apparently more detailed