| || || || || || |
448 KARA-DONG AND THE SEARCH FOR HSÜAN-TSANG'S P`I-MO [Chap. XIII
plaster had once existed. This plaster had, however, completely disappeared both in the inner divisions and in the outer walls. Hence it was most difficult to check the pouring-in of sand and loose earth from all sides. It took my men two days' hard digging—vigorously continued in spite of a fresh sandstorm which visited us on March 17, this time from the south-west—before the floor of the central passage was reached at a depth of 14 ft. from the beams of the ceiling. On examining the latter from below before the clearing had proceeded far down, I ascertained that the thick parallel beams supporting the roof of the gateway were first covered with a thin layer of reeds, evidently intended as a backing for plaster. Above this were laid closely-packed branches of Toghrak, which again were covered with a layer of stamped mud about 1.1. ft. in thickness. This had, no doubt, served as the flooring for another story, of which, however, there remained only a few stumps of posts.
When the men under my direction were proceeding to remove part of this mud flooring, with a view to lightening the weight of the roof and thus reducing the risk of its giving way while the interior was being cleared, we came upon a little store of remarkably well-preserved cereals embedded in the layer of rubbish covering the floor. There were a couple of pounds of ` Tarigh ' or millet (Se/aria ilalica), still largely cultivated about Keriya and elsewhere, together with small quantities of rice, barley, lentils, and a capful of large black currants dried perfectly hard. For the determination of the above I am indebted to the Director of the Royal Gardens, Kew. The exact nature of some roots also found here, in which Jasvant Singh, the surveyor's cook, as well as the labourers, thought to recognize some sort of turnip, has not yet been determined. A small ruined structure outside the gate (seen on the left in Fig. 53) was cleared without any finds.
The survey of the surroundings of the ruined quadrangle yielded little of direct antiquarian interest. To the west and north there extended, as far as I could see and the men sent out ascertain, dunes of great height piled up in weird desolation amidst sand-cones quite as big as those encircling the ruin. Such masses of sand would have effectively smothered the remains even of large buildings, if such were likely to have ever existed here. I may note that tracks of wild camels were reported to me on the north. To south and east the ground for about half a mile was relatively open, the dunes stretching across it being only from 5 to 15 ft. in height, But the traces of ancient occupation which the depressions between them revealed were very scanty. Small fragments of coarse pottery appeared indeed over rare patches, but for indications of structural remains such as timber posts or walls, which the low dunes could not have effectively hidden, I looked in vain, except near the southern limit of this débris area, about half a mile to the south-east of the quadrangle, where my guides knew of an ` old house'.
From the west slope of a dune about io ft. high there rose here rough wooden posts marking the position of a compact block of rooms extending about 63 ft. from north to south. On excavation the walls, which seem to have been of mud only, apart from the posts meant to carry the roof, were found to have decayed almost to the ground ; the whole structure had evidently been for a prolonged period without the protection of sand. No objects of any kind were found, except three round pots of very coarse pottery set in the mud flooring. Their clay was distinctly inferior to that of the pottery found at the Niya Site, both in point of hardness and uniformity of grain. The largest measured 18 by 18 in., being 7 in. wide at the mouth. The shoulder was decorated with a rough zigzag line, and had three small handles attached to it. Another was 7 in. high, 8 in. broad at its greatest circumference, and was provided with two elongated handles. The mouth, 3 in. wide, had a neck raised 3 in. The third, which was broken at the bottom, measured 14 in. across, with a mouth 5 in. wide. The general