Sec. v] EXPLORATION OF SOUTHERNMOST RUINS 241
undercut by wind action. But the wind itself is likely to have been guided or at least influenced by the direction of the old river bank, running here, like the terminal bed still traceable below Tulküch-köl, a little to the west of north. To the north-west of the bridge rows of garden-trees with remains of fences could be made out for over a hundred yards. Immediately beyond there appeared a well-marked depression which, as seen from the panoramic station, had the appearance of a large rectangular tank or reservoir. On examination it proved about a hundred yards across from north to south. Accompanied by Surveyor Rani Singh I then proceeded to follow the dry river course downwards (see site-plan in Plate 7). After continuing for less than half a mile in the direction above indicated, the ancient bed was found to make a sharp bend to the south-west. With this bearing it could be traced for about a mile across an area of curiously open ground, almost clear of vegetation and showing, besides great patches of bare loess soil, only low dunes eight to ten feet high. The contrast with the belt of closely set tamarisk-cones passed further south and seen in the panorama was striking. In places the ancient bed was completely covered up by drift sand, but at short intervals it emerged again, recognizable by its steep-cut banks.
Then the bed resumed again its north-westerly direction, and when after crossing it we had Terminal ascended a big sandy ridge westwards to a height of fifty feet, we could see it joining a broad valley- ant se of like depression stretching far away to the north-west, with living tamarisks and wild poplars. river. Immediately to the north there lay before us scattered groups of large Toghraks, evidently of great age but still flourishing. From where we stood the view extended over miles of this wide silent valley, flanked by big ` Dawans' of dunes rising up to two hundred feet or more and appearing over the flat sandy waste like chains of true hills. One of them was the same which had limited our view westwards while at work on the northern and central group of ruins. It was impossible to mistake here those great riverine ridges of sand which I knew so well to accompany the terminal courses of all rivers losing themselves'in the desert, and which I was able subsequently to study with particular clearness, when crossing in February, 1908, the Taklamakan from north to south towards the dead delta of the Keriya River.2 Nowhere could the Surveyor's sharp eyes or my own, aided by strong binoculars, see any trace of ruins or ancient cultivation. Ibrahim, too, who stood by my side, declared that he had vainly searched this great Nullah and others west of it for several marches in the hope of more ruins. Here was clearly the depression into which the waters of the Niya River at the time of summer floods had once emptied themselves, below the head of the canals irrigating the ancient oasis. The vigorous growth of wild poplars showed that even now a course of subsoil water deep down must find its way to this desolate valley. Moreover, at a point to the north-east where we again struck the old river-bed, I found that the light sand covering its banks revealed in places shrivelled trunks of dead Jigda trees and low stubble of withered reeds. But certainly this bed had seen no water for long ages, and over all this strange ground desiccation was written quite plainly.
With the survey of the short portion of the ancient river course just described my work at the Departure site was concluded, and nightfall of October 3o saw my camp shifted back once more to the shrine of Sfrom Niya Imam Ja'far Sadiq and thus to the purlieus of the living. The great and almost uninterrupted toil i:e. of the twelve days spent among the ruins of the deserted oasis had been rewarded by finds so abundant that I felt less keenly the strict limitation of time which regard for the manifold tasks awaiting me far away eastwards imposed upon me. It would have been of geographical and antiquarian interest to determine the exact course followed by the river about the third century A.D., as far as possible both above and below the southernmost group of ruins, and also to trace with
also Ruins of Kholan, p. 330 ; Ancient Kholan, i. pp. 383 sq. ; and Deserl Cathay, i. p. 422; ii. 385, 401, etc.