Sec. i] THE SITE OF FARHAD-BEG-YAILAKI 1249
and the pose of the right hand holding a patera or cup leave no doubt that the same sacred figure is represented which was first met with on two of the painted panels discovered by me at Dandânoilik." The legend in which this saintly personage figured still awaits identification.
The position of the temple close to the living-quarters of F. ii suggests that the latter were Disposition those of a small monastery, and the same was probably the case also at the ruins of F. iii. They m I quarters were situated on a small erosion terrace, about a quarter of a mile to the north-west of F. 1. The disposition of the rooms as seen in the plan, Plate 57, showed a close resemblance to that at F. Ii. Here, too, the quarters were placed in two wings, adjoining at right angles, with the temple standing apart to the north-east. In both places the large room, at the south-west corner of the court, but with no entrance from it, is likely to have been intended as a guest-room for visitors. An exactly corresponding arrangement is regularly adopted for the mihmân-kheinas of modern houses in Chinese Turkestân so as to allow guests to enter their quarters -without passing through the rest of the house.
The temple, i, stood here also on ground some feet higher than the rest of the ruin, and had in Fresco and
consequence but a few feet of sand to protect it. As seen in Fig. 309, the walls of both cella and other re-
enclosing passage were broken from a height of a few feet above the floor. Those of the cella F. icy.
retained in the south-east corner portions of a well-painted fresco dado, showing on dark red ground a valance-like design with pendent triangles and tassels between them. A square base in the centre of the cella and four hexagonal bases in the corners had once served for the accommodation of statues. In the porch of the cella the lower portions of four stucco images survived (Fig. 309). The figure to the right of the cella entrance, evidently a Lokapâla, stood on a crouching demon of which the grotesque head, F. III. i. 003, was recovered, though damaged on the surface. The fragment of a paper manuscript in Brâhmi was found close to the inner wall of the cella, and on its floor a Wu-chu coin. The chronological value of this coin find and of the evidence furnished by the other Chinese coins from the site, all pre-Tang pieces, will be discussed below. The well-built quarters were found completely empty. But in the room ii, which may have served as the place of assembly of the little ` Samgha', a large wooden tablet inscribed in Brâhmi was recovered. A detached room to the west yielded the fragment, F. III. iii. ooi, of a small statuette delicately carved in wood.
To the north-west of the groups of ruins so far described small ` Tati ' areas strewn with 'Tati' areas.
ancient pottery debris could be traced for a distance of close on 3 miles, on patches of wind-eroded ground interspersed between sand ridges and belts of close-set tamarisk-cones. The considerable extent of the area once occupied by dwellings, probably built with mere mud walls and hence completely decayed, is thus clearly marked. The winding course of an ancient canal, now raised dyke-like 8 feet and more above the eroded ground on either side, could also be followed quite clearly for about a quarter of a mile from near F. n. In one or two places I came upon indications of fields cultivated at no very distant period. They were found on patches of open ground lowered by wind-erosion, to which water may have been temporarily brought while ` Old Domoko ' was inhabited or even later. Structural remains of antiquity were found only near three of the small ` Tatis ', but it is probable that others may lie hidden under tamarisk-cones or accumulations of dunes. At F. iv (Plate 57), about 12 miles from F. i, the much-eroded remains of some quarters yielded no finds ; but from a large refuse-heap near by small objects of domestic use were recovered, including the ` female' portion of a fire-stick, F. iv. 004, and an appliqué grotesque ornament in terra-cotta of a type familiar from Ybtkan, F. iv. 002. Two detached rooms cleared at F. v yielded nothing.
More interesting was the result which attended the clearing of a small mass of masonry seen to
" See Ancient Kholan, ii. PI. LIX, n. vii. 5; Pl. LXII, D. X. 5; also i. pp. 261, 278.