Remains of other remains of the earlier settlement. About eighty yards to the south of E. vi I found the badly small Stupa. decayed ruin of a small Stûpa, rising only eleven feet or so above the original ground-level. Of the
lowest base, only the south face could be traced for about eight feet. Of an upper story, also square, the outlines survived for about fifteen feet on the west and ten feet on the north face. The whole was built solidly in sun-dried bricks, measuring about twenty by thirteen inches with a thickness of three and a half. On the south face of the lower base and about four feet above the ground, a row of tamarisk sticks fixed in the masonry had evidently served to support stucco mouldings. A cutting had been carried right through the little mound, no doubt, in search of treasure.
Clearing of More puzzling at first sight were the remains of a tower-like mound, E. vII (Fig. 79), adjoined
ruin E. viz. by scanty remains of a dwelling, which were traced about a quarter of a mile to the west of the Tang
fort, and of which the plan is shown in Plate 21. The timber and plaster walls of the house survived only to a height of about eight inches from the floor, but still showed clearly horizontal reed wattle. The fragments of two rectangular Kharosthi tablets found in room i definitely proved this ruin to date from the earlier settlement. Some small finds of pottery, silk fabrics, and painted wood were also made here. The mound to the east of the dwelling proved on clearing to be the ruin of a tower measuring about twenty-five feet square outside, of which the interior was completely filled with fallen masonry débris. The north wall, still rising about eighteen feet high, and about three and a half feet thick, alone retained its facing. This showed that the walls were constructed of courses of sun-dried bricks measuring twenty by thirteen inches and about four inches thick, with intervening layers of stamped clay one foot in thickness. Bricks of identical measurement seem to have been used in the other traceable ruins belonging to the earlier period of occupation. A badly decayed wall continuing the south face of this tower westwards in slightly reduced thickness for about fifty
feet, might have formed part of an enclosure. But the true character of this and its relation to the ruined house could not be definitely determined.
Numismatic About forty yards to the west of E. NTH I traced the floor of a large but completely eroded
evidence. structure, but was rewarded by no finds apart from a clipped copper coin of the Wu-shu type and
a piece of pottery showing a rich blue-green glaze. The havoc here wrought by far-advanced wind Î
erosion obviously accounted for the scarcity of structural remains of the earlier period ; for pottery débris of distinctly ancient appearance was visible in abundance on all patches of bare soil for nearly
a mile south of the fort. Moreover, the copper coins which were picked up near the fort showed the
Wu-shu type. Their fragmentary state also attested the force of erosion. I shall have occasion further on to refer to the numismatic finds as a whole.
Discovery While the clearing of these remains was still proceeding a chance find led me to discover that
of Kharosthi the clay rampart of the fort, built within a generation or two of Hsüan-tsang's passage, was in document
on leather. one place actually raised over a bank of consolidated refuse undoubtedly dating from the earlier Tu-huo-lo settlement of which he had seen the completely abandoned ruins. At a point about a hundred feet to the west of the fort's single gate (see plan, Plate 20, and Fig. 77), wind erosion had badly breached the circumvallation of stamped clay. Searching on the surface thus laid bare one of the men who passed by on his way to camp had his attention arrested by a small piece of folded leather sticking out from a hard layer of rubbish. On being called to the spot and extracting it with my own hands, I found it to be the well-preserved fragment of a Kharosthi document on leather, measuring when unfolded about 4 by 31i inches (E. Fort. oo i . a ; Plate xxxviiI). The writing of the nine lines seen on the inner surface and of the single line outside agreed, just as did the general arrangement and manner of folding, with the numerous similar documents on leather which N. xv, the richest refuse heap of the Niya Site, had yielded up in i9oi.10
10 Cf. Ancient Khotan, i. pp. 344 sqq. ; ii. Plates XCI—XCIII.