________________
200
THE INDIAN ANTIQUARY
SEPTEMBER, 1917
their local knowledge, even when combined, proved very inadequate, and after less than half of the journey it gave out altogether. We were thus obliged to trust largely to the guidance of the faint caravan tracks traceable and to what information we opportunely obtained at the single small Mongol camp encountered. The scarcity of wells and of grazing implied serious riske in this mode of progress and made it an anxious time for me, especially as I had found the strain of riding too painful and was obliged to direct our moves from an improvised pony litter.
It was reassuring when, after passing the well of Ming-shui, the great snowy mass of the Karlik-tagh came into view, far away to the north-west, and served to direct us in the rough. But great difficulties still awaited us in the last barren hill range through which we had to make our way, owing to want of water and the very confused and, in places, rugged configuration of its valleys. It proved an easternmost extension of the T'ien-shan system. When we had safely emerged from it through narrow tortuous gorges, which ever threatened to stop our camels far away from water or grazing, it was a real relief to look down on the open Dzungarian slopes and sight some 15 miles away a tiny spot of dark trees. It was the little village of Bai, for which I had wished to make all the time, and after nearly four weeks of continuous travel it was no small satisfaction to have safely reached it without the loss of a single animal. There was reward for our troubles in the extensive plane-table surveys, supported here as all through our journeys by astronomically observed latitudes and by many careful height observations with mercurial barometer and clinometer. They will throw fresh light, I hope, on the morphology of the Pei-shan ranges.
A rapid journey subsequently carried me during October along the north foot of the eastern portion of the T'ien-shan range, already bearing its first winter snow, to Barkul and Guchen (Ku-ch'êng-tzu). The ground crossed here, topographically better kaown, had a special interest for me, as it helped to acquaint me with the peculiar physical conditions of a region, through which many of the great historical migrations westwards, like those of the Yüeh-chih or Indo-Scythians, Huns, and Turks, must have passed. These valleys and plateaus of Dzungaria, favoured by a climate less dry and possessed of abundant grazing. grounds, have often played an important part in the history of Eastern Turkestan. They have again and again afforded a temporary home to nomadic tribes. They could never have maintained their flocks and heards in the arid planes of the Tarim Basin, but they were always able from across the T'ien-shan to carry out their raids into it and exact tribute from its flourishing oases. I could observe a curious if faint reflex of those great tribal movements in the numerous camps of Muhammadan Kazaks, fine men of Turkish speech and descent, whom the Mongols had driven south under Chinese protection, since they secured the "independence" of Outer Mongolia.
After leaving Guchen I survoyed, near Jimasa, the remains, extensive but badly decayed, marking the site of an ancient capital of this region, which under the names of Chinman and Pei-ting often figures in the Chinese Annals from Han to T'ang times. Its connection with the Turfan oases to the south had been a very close one from an early historical period, and as Turfan was to be my base for the winter's labours I was very glad to march there by the most direct route, hitherto unsurveyed. It led me across the Bogdo-ula range, & rugged portion of the T'ien-shan rising to numerous snowy peaks, by a pass close on 12,000 feet and once again confirmed the accuracy of the early Chinese itineraries in which this route is described.