of the Tarim basin.3 I was struck, too, by the show of good clothing worn by the men who were in charge of the hired ponies, including in many instances materials from far-away Europe or China.
Considering that these men were not chiefs or even headmen, it seemed a good illustration of that nomadic wealth and comfort for which historical students in the West are apt to give inadequate credit to the Asiatic invaders of Europe during the period of the great migrations. A large party of burly Kazaks whom we met on the way were driving some sixty camels and as many mares as the wedding price for a girl to be married from a Kazak family in the hills near Urumchi. It looked like a picture out of distant times and scenes. That my Chinese Mandarin friends of Barkul had reason to apprehend trouble from these tribal visitors to the district could be gathered from what some of the Kazaks told me of their scant satisfaction with the pastures allotted to them in these tracts, and of their eager wish to regain their old grazing grounds in the Altai. No doubt, it might mean fighting, and the possession of the latter would be less secure. But then they were so much richer in water, grass and game. There was no mistaking the spirit alive in these men and the ancient inheritance that it betokened.
As our journey from Barkul to Guchen lay along a regular route already followed by European travellers, there is no need to give a detailed description of the ground we traversed. But a brief indication of the chief physical features which characterize its several sections may be useful ; for those features help to throw light on points of ancient topography to which reference has been made above in the discussion of the historical notices concerning Barkul. The first section comprises the western end of the Barkul basin. It extends from the wide grassy plain that surrounds the lake and is watered by numerous springs, up to the ill-defined watershed dividing the basin from the valley of the stream that drains towards Chi-chi-t`ai-tzû (Map No. 31. D. 1). This area affords plentiful grazing not merely around the lake, but also on the open steppe rising above it westwards and at the bottom of the numerous small valleys that descend from, and intersect, the plateau-like watershed. This comparative plenty of vegetation is due to the fact that the range to the south, though not reaching the permanent snow-line, is yet high enough to attract adequate moisture, especially during the winter months. In consequence of this the northern slopes support a more or less continuous forest belt down to a level of about 8,000 feet ; in places tree growth extends even lower down on the eastern faces of side spurs. Springs are to be found at intervals along the route and probably are numerous on the slopes descending towards it from the south.
After the watershed is passed between low broken ridges at an elevation of about 7,40o feet, between the roadside stations of Lo-t`o-ch`üan-tzti and Wu-tun-shui, vegetation becomes distinctly sparser. No trees were observed on the northern slopes of the main range. Yet this still attains a height of close on 9,000 feet at a point south of the halting-place of Tê-shui-ai-tzû and receives sufficient moisture at all seasons to feed the small stream that we passed there (Map No. 31. D. 1). The main valley farther north, into which this stream drains and which the cart-road (not followed by us) descends to Chi-chi-t`ai-tzi , probably contains springs ; for at this pleasant little station, guarded by a few soldiers, we found a lively stream and some cultivation. The same stream some four miles farther west turns due south and, in a gorge, breaks through what apparently is the lowest portion of the whole Tien-shan range between the Karlik-tagh and the Bogdo-ula massif southeast of Urumchi. At the small oasis of Ka-hsün-kou which is irrigated by this stream and occupies the fertile bed of the valley, the elevation indicated by the aneroid was less than 5,000 feet above sea-level. From here an open valley appeared to descend gently towards the depression of Ch`i-kuching, which lies on the H ami—Turfan high road. Thence the route ascends south-westwards, as Map No. 31. c, D, I shows, to plateau-like ground, and after striking the high road coming from
3 See Mr. Joyce's Appendix C, Serindia, iii. pp. 1368 sq. ; also his App. C below.