wall portions rose only about two feet above the original ground level. But the small fragments of decorative stucco relievos picked up near them, and the presence of an outer passage some three feet broad running along the south-east wall, left no doubt about their having once enclosed a temple cella. But what was once the interior of this shrine now showed as a hollow lying more than six feet below the bottom course of bricks in the extant wall remains (Fig. 38). The soil' washed out under the latter, as if by a current, showed plainly how wind erosion was proceeding in its work of destruction.
Complete as the havoc had been among all structural features of the ruin, it had not sufficed to efface all traces of its original decoration. Thus the small fragments of hard white stucco which were picked up on the eroded ground once occupied by the shrine, make it quite clear that the relievos adorning its walls must have closely resembled in character and style those brought to light in the Rawak Vihâra. Among the pieces described in the list at the end of this section are details from the drapery of large images (Ki. or. b, 06, 07, 001 r) ; and from the lotus wreaths which formed the border of vesicas (Ki. or. a, or 2). Flame-tongues once belonging to large aureoles are particularly frequent (Ki. o3,08-14, etc.). From life-size relievo images also a few small fragments have survived (Ki. 04.015). In character all these tiny relics of a Buddhist shrine curiously recalled those recovered in much greater abundance from the Kighillik site near Ak-sipil.t° It is possible also that the great hardness of the Kine-tokmak fragments, all in a white stucco resembling plaster of Paris, may be due, just as there, to accidental burning when the shrine was abandoned. Yet bleached pieces of carved wood were also found among the small débris, one of them, Ki. oi. c, showing clearly a Buddhist rail pattern.
From here I proceeded north across low dunes to where Ahmad, a ` treasure-seeker ' from Suya, had to show me other ` old houses ' about a mile distant. They proved to be the remains of some modest dwellings built with timber and wattle walls, first destroyed by erosion and finally by the burrowings of those who had searched them for treasure whenever the march of the dunes left them exposed during the course of centuries. In the least injured among them, two rooms about fifteen feet long and some thirteen feet wide could still be distinguished with their walls showing horizontal reed bundles between the plastered faces. In one of the rooms a row of circular holes in the plaster flooring, about two feet six inches across and six inches deep, evidently marked the position once occupied by large jars. No finds apart from pottery fragments rewarded the clearing made at these scanty ruins. Some more of the same type were said to be scattered among the dunes further north. These dunes rose to twenty-five feet or more.
Turning thence to the south-east, we reached ground where the dunes were rapidly getting lower and erosion had run its full course. Here the marks of ancient occupation were abundant, but showed distinct approach to the familiar ` Tati ' type. Pottery débris reddened large stretches of eroded ground. Yet here and there more interesting traces still survived in the light drift sand. There were the low stumps of fruit trees and poplars which once surrounded the cultivators' humble dwellings, now completely vanished. At one place a row of Jigda or Eleagnus-trees clearly marked an orchard. Among the small objects picked up here was a fine piece of cut-glass, Ki. 0017, the rim of a vessel, of pale yellow-green colour and decorated outside with a well-modelled festoon ornament. It is noteworthy that the only coins among the day's ' finds' at Kine-tokmak were both uninscribed Chinese pieces of a type ascribed to about the fifth century A. D.
For a distance of nearly three miles from the northernmost of the Kine-tokmak ruins, these traces of ancient occupation continued unbroken until the line of the old river bed already