along the eastern edge of the old jungle. The men promptly called it the ` Ustang ' (canal) of the " old town." But I was unable to find any proof of its artificial origin.
Further, down we had to pass through a belt of steep, conical sand-knolls from 15 to 30 feet high, rising close together and all covered on their tops with tangled masses of living and dead tamarisks. On the northern slopes • the snow that had fallen a week before still lay plentifully to the depth of an inch. In the midst of thfs belt, extending for a distance of about three miles from south to north, I came upon broken pottery remains and any enclosure made of thickly-packed rushes. Inside it the men recognised trunks of fruit-trees and planted poplars, or ` Terek.' We had evidently passed the site of some ancient farm. Beyond, the sand-hills were lower, but also bare. Living tamarisk bushes could be seen only on isolated sand-cones rising here and there over the low dunes. Pottery fragments strewn over the sand, with bits of slag and similar hard refuse, assured my guides that we were near the goal.
Soon I sighted the first two " houses," standing on what looked at first sight like small elevated plateaus, but which subsequent observation proved to be merely portions of the original loess soil that had escaped the erosion proceeding all round. The wooden posts of these buildings rose far higher above the sand than in the case of the dwellings at Dandan-Uiliq. A rapid inspection sufficed to show that their mode of construction was materially the same ; but the dimensions here were larger and the timber framework was far more elaborate and solid. That these remains were of far greater antiquity became evident almost immediately when, in a room of one of the houses, I came upon some finely carved pieces of wood lying practically on the surface, which displayed ornaments of a type common to early Gandhara sculptures. Marching about two miles further north across fairly high dunes, I arrived at a ruined structure of sun-dried bricks, which Abdullah had already mentioned to me at Keriya as a ` Potai.' It proved, as I had expected, to be the remains of a small Stupa, buried for the most
IIIF Color HighRes
