the relievo fragments found in clearing the cella floor, flame and lotus-wreath ornaments, once undoubtedly forming part of large vesica borders, were most numerous. These made it quite certain that the style of decoration must have been the same here as in Kha. i and Kha. ii. Of the larger sculptures in stucco only the fragment of a head, Kha. vii. 2, slightly over life-size and gilt, had survived as well as a finely modelled hand, Kha. vii. 009. A mass of miniature Stûpas in clay, between two and three inches in height and all uniform in shape, of which Kha. vii. ooro (Plate Ix) is a specimen, lay closely packed in one corner. They alone had escaped thorough destruction. It was of interest to note that they showed a cylindrical topmost base, supported by eight round buttresses, and a dome surmounted by a square crown just as is commonly seen in small stone-carved Stûpas of Gandhâra.'5
Immediately to the north of Kha. i there were laid bare under a cover of four to five feet of sand the small structures where those quarrying the large shrines for timber had established their workshop. The walls were uniformly built of timber and wattle, and the good preservation of the lower portion of the north wall in Kha. viii has made it possible to illustrate the system of construction by the section reproduced in the detailed plan (Plate 6). In Kha. viii, a room seventeen by twelve feet, the clay-built fire-place, and a plastered sitting platform in the corner beside it, still survived. In front of the seat there was found a trough of unbaked clay, two feet by one, of uncertain use. On the top of the heap of chippings which filled part of the room to a height of about two feet from the floor, was found the large but partly rotted packet of paper leaves from two Pothis in Brâhmi script which I have already mentioned. In addition to various fragments of other manuscript leaves the finds here consisted of some turned pieces of wood, belonging no doubt to Kha. i, including a baluster and two knob-shaped finials. The small structure, Kha. x, unearthed further west, had suffered more damage ; but the identical arrangement of fire-place, etc., showed that this, too, had served for quarters. Apart from plentiful wood chippings the only find made here was the portion of a panel painted on either side with the seated figure of some divinity. Though its colours were badly effaced the painting still showed some interesting details of dress (see Kha. x. r in List) which recall similar representations on the Dandân-oilik panels.
Far more abundant remains and of a varied kind came to light in the small ruin Kha. ix (see Fig. 42), situated to the north of Kha. x and within five yards of it. Here were discovered at first, scattered among wood chippings on a level about two feet above the floor, very numerous single leaves and fragments from a Sanskrit Pbthi, written in imposingly large Brâhmi characters on paper which had an original size of about twenty-two by nine inches. There were numerous packets also, all broken, containing closely packed layers of leaves from the same manuscript, and curiously enough these seemed to have suffered more from damp than any of the detached pieces. It was evident that they must all have been scattered about after the ruin had become partially filled with sand ; but no definite indication could be found as to where this big Pôthi or portions of it had originally been deposited. That its dispersion had begun earlier was, however, proved by a curious little convolute which was found in the small structure just north of the cella. In it there were rolled up fragmentary leaves of the large Pôthi, a narrow folio of another Sanskrit manuscript, and part of a document in cursive Brâhmi script written on both sides of a thin sheet of paper.
The place richest in these finds proved to be a small cella, about eight feet square, marked by a plastered floor which rose about six inches above the rest of the ground here. No traces of its