56 miles of my cyclometer, reckoning from Toksun to Kumush. By Lü-kuang on the
frontier of Yen-chti or Kara-shahr, 40 li distant, must be meant a halting-place at or near the present Kara-kizil (Map No. 51. D. 4), where water is obtained from a deep well, and the natural frontier of Kara-shahr, corresponding with the watershed towards the Bagrash Lake, is entered. Thence, beyond the Fan-shin ;;; 4 tract, the military station of Chang-san-mieh A= ,r•' is reached after Too li more. There can be no doubt that the cultivated area of Ushak-tal (Map No. 51. A. 4) is meant, though the actual road distance from Kara-kizil is over 3o miles. ` Going towards the south-west for 145 li one passes the road station called Hsin-ch`êng Jïtt&(" the new city ") and crossing the
Tan i l River (the Kara-shahr River) arrives at the garrison town of Yen-ch`i .' That by
the station of Hsin-ch`êng a place approximately in the position of the present town of Kara-shahr must be intended is made clear by the bearing and distance. The situation of the ancient capital of Kara-shahr will be discussed presently.
We must regret that Hsüan-tsang begins his Hsi-pi-chi from Yen-chi, or A-ch`i-ni(p~ t f , as he calls Kara-shahr, but does not describe the route by which he arrived there from Kao-ch`ang or Turfan ; for the account we receive of this journey in his Life4 is not as clear as it might have been in the great traveller's own record. M. Chavannes has assumed that he followed the route which the Tang itinerary describes,6 and on general grounds this appears probable. But obscurities of detail remain, and it must be remembered that the present high road is not the only route by which to reach Kara-shahr from Turfan.° If Hsüan-tsang followed the main route, which certainly is the easiest, I think that we may identify the miraculous ` spring of the Master A-fu ', about the origin of which the Life tells a lengthy legend, with the remarkable spring which issues from a sheer wall of rock in the deep gorge passed by the route about a mile below the station of Arghai-bulak (Map No. 54. B. 3), and from which this derives its name. There is no water to be found elsewhere on the route above this point until Kumush. But it must be mentioned that the Life describes the spring as issuing from a ` monticule de sable, au sud de la route ', whereas the Arglaai-ôulak issues from a cliff of what seemed to me granite or gneiss, and flanking the route on the west. On the other hand, the statement that the Master, after passing the night with his companions by this spring, started by daybreak and traversed, evidently the same day, ` the Yin-shan or " mountain of silver ", which is very high and large', would well accord with the long march by which the elevated plateau above mentioned is crossed from Arghai-bulak to Kumush. That the pious pilgrim was attacked by robbers when proceeding west of this mountain would well agree with the topography of the route beyond Kumush ; for the broken ground crossed there would specially facilitate such exploits from the higher valleys north which afford fair grazing for nomads.?
My stay in the Kara-shahr region was too short and the extent of the ground that I actually visited, away from the line of the main route and certain ruined sites, too limited to justify my attempting here either a systematic survey of its' geography or a review of the data we possess regarding its early history. But among the geographical features distinguishing the Kara-shahr territory there are some so striking, and of such obviously great importance as determining its history, that a brief account of them seems called for here.
Kara-shahr in some respects occupies a unique position among the districts comprised within the Tarim Basin. Immediately adjoining from the north-east the great flat trough which extends in
b Cf. Chavannes, Turcs occid., pp. 6 sq.
° Another route through the mountains further west was followed by Roborovsky (see his Map II), and there may be more in the hills south-west of Toksun.