552 TO GUCHEN AND ACROSS THE T`IEN-SHAN [Chap. XVI
Ch`i-ku-ching, crosses the watershed on this depressed portion of the range by a saddle about 5,600 feet above sea-level. In the valley leading down north-westwards to the station of Ta-shih-t`o (Fig. 290) a small stream gathers from springs, which permits cultivation to be maintained for a couple of miles. Here the second section of the ground may be considered to end ; for beyond, a waterless stony waste extends along the foot of the range, practically without vegetation, for a distance of over thirty miles.
When discussing above the topographical notices furnished by the Later Han Annals and the Wei h o of the territories along the northern slopes of the Tien-shan, I called attention to the definite
indication furnished by the latter text as regards Eastern Chii-mi It is mentioned
as the first and easternmost of the territories reached north of the range by the ` new route of the north ' after emerging from the desert to the south-east.4 In view of the plain and unalterable geographical facts, there can be no possible doubt that the ` new route of the north ' leading to those territories from the Jade Gate must have crossed the Tien-shan, just as the present high road does, by the saddle above Ta-shih-t`o. The fact, I believe, justifies us in assuming that the ground described by me as the second section belonged to Eastern Chü-mi. It is possible that the first section crossed by our route was also included in this territory.
From the circumstance that the description in the Later Han Annals mentions only Eastern Chü-mi and not Western Chü-mi, which the Wei lio's list names as lying next to the west, Hsi" Sung, the commentator of the Hou Han shu, concludes that the latter territory was at the time of the Later Han dynasty absorbed by Eastern Chü-mi.5 In support of this assumption it may be pointed out that the notice in the Later Han Annals, which describes the people of Eastern Chü-mi as nomads, living in huts and tents and having but little cultivation, ascribes to them a total of three thousand households, while attributing a total of only a thousand to the ` kingdom of I-chih' g t which, we have seen, must be identified with the valley of Barkul.6 For this comparatively large population of Eastern Chü-mi we should find it easy to account, if the territory at the time included not only the valleys and plateaus west of Barkul which we have briefly described, but also the much better watered slopes of the rising portion of the range above the road between Mu-li-ho and Guchen. There we find abundant forest clothing the spurs and higher valleys, while cultivation can, to a fair extent, be practised lower down.
On October i3th we crossed the barren stony and utterly waterless plateaus that lie between Ta-shih-t`o and the wretched roadside station of San-ko-ch`üan (Map No. 31 B. I). The march of twenty-seven miles was covered in a trying blizzard. These plateaus offered a characteristic sample of the region that stretches along the northern foot of the depression in the range. This third section of the ground may be said to extend westwards for about another fourteen miles, taking the form of a clayey steppe with very scanty scrub. Throughout this section there is practically no grazing to be found, except perhaps in some glens where the range again rises to heights of over io,000 feet.
A marked change in the aspect of the country occurred when we reached the first cultivation at the village of Mu-li-ho. It was quite Chinese in its appearance and population ; but among its inhabitants was found a well-to-do Yârkandi trader who claimed to be a British Indian subject and offered hospitable shelter. He had come to this place by reason of the recent migration of the Kazaks, with whom he had previously traded for years in their old seats on the Altai, while he himself resided in Kobdo, Uliassutai, &c. The information he was able to give about the new grazing grounds occupied by his Kazak clientèle on the Tien-shan slopes to the south was the more
4 See above, pp. 542 sq. 6 See Chavannes, T`oung-pao, 1907, p. 210 ; above, p. 542.
5 Cf. Chavannes, Toung-pao, 1905, p. 557, note I.