Sec. ii] RUINED SITES WEST OF THE MUZ-ART RIVER 813.
shows, at the mouth of one of the small barren ravines that descend here from the crest of the range. I was surprised to find, within a hundred yards of the principal ruin, a well holding tolerably fresh
water, by the side of a dry flood-bed. Close to the well was a well-preserved small karaul or
guard post, and a planted poplar and willow tree. The presence of this subterranean drainage from the range explained both the ancient and modern occupation of this otherwise barren spot. The
post was said to have been built and occupied for a number of years after the Chinese reconquest
in 1877, in order to keep watch on the direct desert route from Ak-su to Kuchâ. mentioned above. The ruins of Tajik and those of the Toghra-akin site farther to the west were surveyed between April 24th and 27th. They comprise a much-dilapidated quadrangle, and the remains of a Buddhist
sanctuary, situated on a natural clay terrace at the mouth of the valley, and a group of small shrines
scattered over low ridges higher up. The arrival of an adequate number of labourers from Yulduz-bagh by the evening of the second day of my stay enabled me to clear completely the
remains of the ruined quadrangle, Taj. 1 (see plan in P1. 41, 43). They had evidently been dug into
more than once. The structures once occupying its south-eastern and north-western sides had been almost entirely destroyed. But enough remained elsewhere to reveal a certain peculiarity of con-
struction. A low natural plateau of clay had been cut down to a uniform level in the centre, while its rims were left standing to form raised platforms at the sides, or to serve, after excavation, for the lower portions of walls. The latter method had been adopted in the western corner, where the walls of three apartments, partly cut from the natural clay and partly built with flat slabs of clay, still stood to some height (Fig. 353).
Complete clearing of the debris proved one of these apartments, i, to have been a small Buddhist shrine (see plan, Pl. 43). There was a central niche for an image in its back wall ; and traces of
wall-painting survived in the two arched passages leading, on either side of it, into the room behind.
In i were found the stucco relief head of a Bodhisattva, Taj. 1. i. 01, and the wooden right hand, Taj. I. 01 (Pl. LXVI), which had probably belonged to the Buddha figure shown in the Abhaya-
mudra. A Chinese coin found in i could not be identified. The floors in i and in the room adjoining on the west were covered with hard plaster of Paris. No finds were made on clearing. The very ruinous remains some 14 yards off, in the line of the north-western side of the quadrangle, seemed to be those of small shrines, one quadrangular, the other circular. In the area iii, adjoining the shrine i, we discovered three small paper fragments with Kuchean writing.
The clearing of the rest of the quadrangle brought to light further structural remains in the northern corner. There in room ii we found a Chinese coin, apparently uninscribed ; a fragment of thick green glass, Taj. 1. ii. 01, from the wall of a vessel, and three large jars of coarse pottery, set in the floor but broken. Numerous fragments of hard plaster from the flooring showed that the other sides of the quadrangle had also once been occupied, probably by monastic quarters ; but none of these remained except two small rooms on the north-east, which still showed their floors cut into the hard clay.
The group of small shrines (Taj. 11), previously mentioned, occupies two low but steep ridges of clayey rock which rise about 200 yards to the north-east of the ruined quadrangle (Pl. 41) on either side of a little ravine. The shrine 11. i (Pl. 43), excavated from the soft rock, consists of a domed cella, 10 feet square, with a passage behind ; the cella was approached by openings on either side of the back wall, which once, no doubt, supported an image in relief. Slight traces of painting could be seen where plaster strengthened with reeds still covered the foot of the passage walls. A small fragment of paper with Brâhmi characters was recovered from the sand covering the floor. A more interesting find was a well-preserved wooden tablet of the ` Takhti ' shape (Taj. 02 ; PI. CXXIII), 13 inches long and 2 inches wide, which was picked up under a slight layer of