264 THE RUINS OF DANDAN-UILIQ [Chap. IX
SECTION V.—DISCOVERY OF DATED DOCUMENTS
The scattered group of ruined structures situated about half a mile east-north-east of my camp was the next to which I turned attention. The first to be explored were the closely adjoining ruins, marked D. iv and D. y in the plan, in which I could without difficulty recognize the remains of an oblong temple-cella and of a large dwelling-place. These ruins had suffered badly from erosion which, in the unprotected soil immediately to the north and east of them, had produced broad depressions to the depth of about 20 feet below the original ground-level. The ruins, owing to this lowering of the adjoining ground, as marked in the detail plan (Plate XXVI), seemed now to occupy a raised tongue of land quite clear of dunes, and nowhere retained more than two or three feet of covering sand. Above this there rose the splintered short stumps of posts which once had held the framework of wattle and plaster walls, as seen in the photograph of D. y (Fig. 35). In places the rows of these posts still clearly marked the position of dividing walls, though the walls themselves had practically disappeared down to the ground-level. Elsewhere there remained under the thin layer of sand only the massive beams, usually 6 to 8 inches in thickness, which had formed the foundation of the walls, while near the edges of the ground actually undergoing wind-erosion even these remains had disappeared, or were strewing the slopes in shapeless fragments of bleached brittle timber. The plaster work had survived only in a few portions of walls, which in the detail plan can readily be distinguished by being marked in black. Nowhere did these remnants of walls exceed a height of 2 feet.
The exposed condition of these ruins had, of course, attracted the visits of ` treasure-seeking' parties, including some that Turdi had personally conducted in former years ; and their burrowings had left visible marks in the débris of plaster, timber, potsherds, &c., which lay scattered about on the surface of the sand, as seen in the photographs. Notwithstanding the damage thus caused, my careful clearing of the ruins was rewarded by some interesting relics. Within the cella D. iv (Fig. 34), which measured 13 ft. 8 in. inside from east to west with a width of II ft., only some much decayed fragments in friable stucco were found, which seemed to belong to the drapery of a colossal figure. This had no doubt occupied the base, 3 feet square, which faced the entrance on the east and still rose to a height of 8 inches. At the foot of the west wall, behind the base, there remained traces of a painted frieze, probably resembling that illustrated from D. xi'. Of the walls enclosing a passage around the cella, 5 feet wide, very little remained but the foundation beams and stumps of posts. Yet even this scanty shelter had sufficed to preserve on the west side the two painted panels, D. Iv. 4 and 5, but their colours have suffered much.
The obverse of the panel D. iv. 5, measuring 17â in. in length with a width of 44 in. and shown in Plate LX III, is of special interest : for in the half-length human figure with the head of a rat and wearing a diadem, which occupies the central semi-ellipse, we have manifestly a representation of the chief of those sacred rats which were worshipped as having saved Khotan from an attack of the Hsiung-nu, and whose legend, as heard by Hsüan-tsang at the site corresponding to the present ` Kaptar-Mazar' at the western edge of the oasis, we have already discussed in detail 1. Though the characteristic features of the rat's head are unmistakably indicated, the painter has cleverly imparted to it a human expression. The colour visible on the face and on the part of the neck left bare by the robe is distinctly yellow, thus agreeing