Sec. iii] THE SITES OF UZUN-TATI AND ULUGH-ZIARAT 461
Whatever of these scanty structural remains had escaped erosion displayed plain evidence Character of the destruction dealt here by treasure-seekers' diggings. This was not to be wondered at, of remains. seeing how relatively near this ruined site was to the several oases southwards. This easy accessibility and consequent thorough exploitation left no hope of ruins capable of excavation having survived here. In their absence the archaeological indications I needed as to the date when the site was last occupied could be forthcoming only from among the small débris to be picked up on the ground. I accordingly utilized all the available men for a careful search under my own supervision, and this soon yielded interesting chronological evidence.
By the side of small objects in bronze, including a well-shaped clasp, a buckle-like object, Chrono-
some finger rings, and little fragments in the same material (see D. K. ooz, 003, 005, shown in logical
Plate LI) ; small broken pieces of glass and true Chinese porcelain (see for specimens D. K. of debris.
002, 003. k, in Plate LII) were common. As porcelain did not make its appearance among the keramic products of China until the beginning of the Sung period (963 A. D.) s, its discovery here was a clear indication that the site had been inhabited down to a far later time than, e. g. Dandan-Uiliq, Endere, or Kara-dong, where no trace of this material could be found. But even more definite chronological proof was obtained in the shape of the two copper coins which were picked up under my own eyes at the Tati. One of them (see Plate CX, 35) has been recognized as belonging to the Pao-yüan period (1o38-39 A. D.) of the Emperor Jên tsung, while the other (see Plate XC, 43) is a piece bearing the name of Muhammad Arslan Khan, whom we know to have reigned over Khotan and other parts of Eastern Turkestan in the eleventh century. Thus the occupation of the site well into Muhammadan times was established beyond doubt.
As Uzun-Tati did not offer further scope for systematic work I decided to proceed to the Search for
second ' kone-shahr ' of our guides, which the people of Domoko knew of by the name of Ulagh-
Ulûgh-Ziârat or Ulûgh-Mazâr ` the holy (or high) shrine.' Though it was known to be Zirat.
situated somewhere between Lachin-ata and Uzun-Tati, and in the end proved only about three miles distant in a direct line to the south-south-east of the latter, it took us nearly two days and very tiring marches and countermarches in the sand before I was able to complete my survey of its remains. Our soi-disant guide first led us over fairly high and closely-packed dunes to the south-east. At two points within a mile of Uzun-Tati old pottery débris cropped up again over small patches of bare ground, plainly marking an extension of that site. When we had covered about five miles in a direct line, pottery fragments appeared again in detached spots over large stretches of eroded ground. Our guide' now declared that we had got too far towards the Lachin-ata Mazar, and asked to be allowed to turn back once more in quest of the site. Halting the caravan, I accompanied him for some hours with a few of the men in his devious wanderings, but was finally obliged to retrace my steps to the caravan without any clue having been obtained for the location of Ulûgh-Ziarat. As water was getting short I resolved to make for Lachin-ata, where there was said to be a well with brackish water. But after marching for an hour and a half eastwards between tamarisk-covered cones and over dunes 20 to 3o ft. high darkness obliged us to halt. During the night the labourers sent out in search managed to discover Lachin-ata about two miles to the south of our dreary camp.
Next morning I dispatched our Darogha with an urgent requisition for efficient guides to Remains of
the Bég of Gulakhma, to the north of whose tract we now were. While I was utilizing my Ulûgh-
enforced halt for anthropometrical measurements on the labourers with me, the luckless ` guide' Ziârat.