106 THE ROUTE FROM KARGHALIK TO KHOTAN [Chap. V
subsequently receded again northward, ought not to be lost sight of. Such an assumption
would at least furnish a ready explanation of the fact that the surface of the Stûpa is best
preserved on the north-western side. It is well known that the semi-lunar dunes in the
western portion of the Taklamakan, in conformity with the prevailing winds, mostly show
a line of progression from north-west to south-east. It is evident that, whenever these dunes
encounter obstacles in the form of buildings, they will pile up sand highest against their north-
western face and consequently assure to that better protection than to the rest b.
Like all other Stûpas, Topa-Tim must at one time or other have attracted the attention 1
of ` treasure-seekers '. But the digging made on the top of the mound was apparently not carried deeper than about seven feet. The square well and chamber, which other Stûpa ruins
examined by me show in the centre line of the dome, was not traceable here ; but this might I
possibly be due to the much-decayed condition of the dome and the excavation just referred to. On account of the ` weathering of the surface it was impossible to find any trace of the plaster coating which once probably covered the whole structure. Nor could I ascertain the size of the bricks used in it, as this would have required a cutting to be made into the masonry, and would in all probability have led to further destructive digging by others bent on more practical objects.
On my first attempt to reach Topa-Tim, I had noticed on the right bank of the Sughaz- yar a débris-strewn area far more extensive than the one seen to the west of Mokuila. Stretching to the north of the caravan route it seemed to cover fully three square miles, perhaps more. To the south of the road and again eastwards at a distance of about two miles from the ravine the débris disappeared under low dunes of drift-sand, invaders from the desert. The lateness of the hour at which I reached this area prevented me from determining its precise limits northward. My guides called it the ` Tad ' of Kakshal, applying to it the name of the nearest village in the Mokuila oasis.
The relics of ancient habitations—and as such they could not fail to be recognized even by the most casual wayfarer—lay here scattered in patches of varying extent and thickness. They comprised, besides pottery fragments of all sorts, small pieces of stone and burned brick, slag, broken bones, much corroded bits of metal and similar hard refuse. The potsherds were without exception of remarkable hardness, but generally coarse in texture. The overwhelming majority of them showed varying shades of red and, where lying thick, imparted a reddish glow to the whole ground. But pieces of black pottery were not infrequent. Fragments showing any ornamentation, or giving clear evidence as to the shape of the vessels to which they had belonged, were very rare, and the specimens of this type brought away and described in the list below were the most distinct among those picked up during a search of over an hour. The surface of all the potsherds was peculiarly rough, looking as if it had been subjected to ` grounding '—a manifest mark of the force of erosion which had been at work here. Fragments of bones, apparently of animals, turned up in many places, just as they would be found now in most of the rubbish-heaps of Turkestan villages. In one much-eroded piece my men recognized the knuckle-bone of a sheep, used probably, as nowadays, as a substitute for dice.
The occurrence of slag was restricted to particular patches. The most striking of these was a loess bank rising some fifteen feet above the general level of the ground, not far from