Sec. i] ALONG THE FOOT OF THE TIEN-SHAN 789 .
straight line towards the large oasis of Bugur, thus avoiding the considerable detour made by the present road. This detour is necessitated by the difficulty in obtaining water. This is nowadays
to be found only at the little oases of Charchi, Eshme, Châ.dir, all nestling close to where small
streams, fed by springs among the foot-hills or on the lower edge of the Sai, can be used for irrigation before they lose themselves in the scrubby desert. The probability that in ancient times these
streams may have carried water to ground much farther south, now reached only on the occasion
of rare rain floods, and have made its permanent occupation possible, is suggested by two observations. One is that in 1908, on my quest for the imaginary ` kône-shahr ' of Kök-darwâza along the
Charchak-dary5.,2a my self-styled guides' informed me that the bare clay steppe northward towards Charchi had at times been searched by ` treasure-seekers ' for small objects of value laid bare by wind-erosion, just as on the ` Tatis ' of Khotan. It was impossible to spare time now for testing the truth of this statement from Charchi, where no local information on the subject was obtainable.
The other reason in favour of the hypothesis mentioned above is that due south of Châdir, at a distance of about fourteen miles, lies the old site of Aghrak (Map No. 21. B. 1), on a continuation of the line of patches of old cultivation to which the stream of Yangi-hissar formerly extended. The site falls on the straight line connecting Korla with Bugur, and its examination showed that it had been occupied down to Muhammadan times.3
On the way to Charchi the route leading along the foot of the glacis was fringed on the south by a continuous belt of sandy ground with scrub and tamarisk-cones and farther on also with Toghraks. None of the beds descending from the range above contained water, but at the point where one of these debouched we found the well of Yantak-kuduk-langar holding water at a depth of 15 feet. Charchi itself is watered by a small brook rising from springs at a Mazâr, some six miles higher up, and carrying about two cubic feet of water per second. This suffices for the cultivation carried on at the tiny oasis by sixteen resident families. Of traces of old cultivation farther south my informants knew nothing.4
The region traversed on April 8th on the way to the small oasis of Eshme was of a very similar character, the route leading, for most of the way, close to the foot of the Sai along the northern edge of scrub and Toghrak jungle. At the abandoned roadside station of Kuruk-eshmelangar the road strikes the eastern edge of a wide alluvial fan formed by drainage beds descending towards Eshme. A canal from the bed which now carries water to Eshme was said to have once reached this point, permitting of some cultivation, but to have been destroyed by a big flood. The configuration of the ground makes this likely enough, and a well about 8 feet deep shows that subsoil water still finds its way here across the fan. Eshme itself had not yet received its spring water from the ` Eshme-akin ', which takes off some five miles higher up from a broad flood-bed
2a See Serindia, iii. pp. 1232 sqq.
3 See below, ii. p. 791.
4 Dr. Herrmann, Seidenstrassen, pp. 121 sq., has suggested the location at Charchi of the territory of Chieh-chih~~ which a memorial addressed to the throne about 90 B. c. and reproduced by the Chien Han shu mentions in conjunction with Ch`ü-li as situated to the east of Lun-t`ai (Bugur) and suitable for the establishment of military colonies ; cf. Wylie, J. Anthrop. Inst., xi. p. 96.
Of Ch`ü-lic I believe that I have shown that it can
be safely identified, in view of the topographical indications furnished by a special notice in the Chien Han shu, with the riverine tracts on the Inchike and Yârkand rivers below
Shahyar ; cf. Serindia, iii. pp. 1236 sq.
For the location of Chieh-chih, which does not appear to be otherwise mentioned, no definite evidence is available. The description given of both Ch`ii-li and Chieh-chih (` The land is broad and fertile, and water and herbage are everywhere plentiful . .. the soil is excellent and might be improved by drains and watercourses ') seems to point rather to some riverine tract than to the ground about the present Charchi ; this from the comparatively low range above it could never within historical times have received a large supply of water. Possibly the broad belt of riverine jungle along the Charchak river-bed, receiving water in certain years from the side of Kucha, might be meant ; cf. Serindia, iii. p. 1233.