Sec. ii] ACROSS THE EASTERNMOST TIEN-SHAN 533
The descent was gentle at first, with low cliffs on either side and the scrub-covered ground Descent in
between them fairly open. Some mountain sheep were sighted and also big coveys of partridges. But deep-cut
at a distance of about a mile the valley bottom turned into a deep-cut gorge, with steep rock walls gorge.
rising hundreds of feet on either side. The cliffs consisted of granite striking WNW.–ESE. with an almost vertical dip. The rock walls and the narrow bottom were quite bare of vegetation, except at one point about four miles from the watershed where a bed of reeds was found at the mouth of a side gully (Fig. 295). The gorge wound more and more as it descended and the rock coulisses seemed often to close in completely. Accordingly when dusk obliged us to halt, in the narrow bottom of the gorge, I was not free from anxiety lest it might, perhaps, end in a couloir impracticable for laden camels. No cairns of the Chinese sort to mark a route had been noticed anywhere along the day's march. But the men drew hope from the sight here and there of little stone heaps blocking turns into wrong gullies after the fashion prevailing in Turkestan. It meant for them a return from the region of the heathen ` Khitai '. The gorge all the way down recalled scenery such as I had seen near Ara-tam at the southern foot of the Karlik-tagh and between the depression of Turfan and the Kara-shahr valley. There could be no doubt that we were in mountains belonging to the Tien-shan system.
The north-eastern trend that the gorge had been taking had caused some uneasiness both to Escape
Lal Singh and myself ; for it seemed likely to take us farther away from the assumed position of from
Bai, where alone we could now hope to find water. The disappearance, moreover, of all traces of gorge. a track for some distance back raised doubts whether the exit from this gloomy gorge, as far as our camels were concerned, might not lie through some gully on either side. However, when we had continued our march on the morning of September 26th for about two and a half miles, relief came at last. Ismail, one of my Yarkandi followers, stimulated to unwonted initiative by the common anxiety, had climbed the cliffs to the west of the gorge and was heard thence shouting excitedly. On rejoining he reported that he had seen a wide unbroken Sai beyond the rugged spur above us and far away in the distance a dark spot which he took for trees and houses. As we moved down, the flanking spur on the left soon grew less rugged, and at one point a gully was found which allowed the ponies carrying my litter to gain the crest.
There a vast view unfolded itself before us. To the west and west-north-west the snowy peaks of View along
the Karlik-tagh far away rose in perfect clearness, the northern slope of the range down to about N slopes of
9,000 feet powdered with what obviously was fresh snow from a recent storm. A confused mass of black spurs was seen descending from them towards a bare plateau ; this was cut through by a valley evidently containing the stream that carries water to the outlying settlements of Adak and Nam. Beyond this plateau to the north-west rose another and still more distant snowy range—the mountains overlooking Barkul. Behind us to the south-west and south barren foot-hills (Fig. 286) masked the height of the eastern continuation of the Karlik-tagh. Almost due north our binoculars showed a dark patch of cultivation which was evidently Bai, the goal towards which we had been steering. Deceptive as I knew the bare Sai to be that stretched unbroken towards it, it was clear that Bai lay much lower than where we stood (4,450 feet above sea-level by aneroid), though the latest map in my hands placed it at 6,000 feet. Impressive in the uniformity of its vast expanse was the view to the north-east. There one huge desert valley of gravel seemed to stretch away to the foot of misty hill chains, fading away into Mongolia and clearly forming part of the Altai. I could scarcely have gained elsewhere an impression more comprehensive of the varied character of the ground comprised in the great region north of the Tien-shan for which the name of Dzungaria offers itself as a convenient designation.
An easy descent of eleven miler over the gentle Sai of bare gravel brought us to the first vegeta- Arrival at tion on a clayey steppe dotted with tiny tamarisk-cones. The sand lay all heaped up behind them Bai.