156 RUINED SITES NEAR DOMOKO [Chap. V
thorough this destruction had been, there was comforting evidence that the relics still remaining had yet escaped damage from either fire or moisture.
With a large additional contingent of labourers which the vicinity of Domoko allowed me to summon, the work of clearing was vigorously continued on September 25. Soon the first indication of the structural disposition of the building was obtained in a line of low broken posts marking the position of a timber and plaster wall. Faint traces of plastered steps leading up to this showed that it was the outer wall of the main shrine facing south. Similar remains of the east wall were soon laid bare, and with the remnants of cella walls of the same material subsequently discovered I was able in due course to reconstruct the ground-plan of the temple as reproduced in Plate 6.
This shows it to have formed a quadrangle of which the external measurements were seventy-five feet on the east and west and seventy-three on the north and south. The middle of this quadrangle contained a cella, measuring on the inside a little over twenty-eight feet. The centre of the cella was occupied by a badly decayed base or platform, ten and a half by nine and a half feet. It rose in its completely broken state to a height of about two and a half feet above the plastered floor, without a trace of its original stucco facing. The space left between the outer walls and the cella, twenty-one feet on the north and south and twenty on the other sides, seemed a priori far too wide for the enclosing passage for which the example supplied by the majority of the shrines excavated at Dandân-oilik and elsewhere in the Khotan region led me to look. So I was not surprised when the complete clearing of the temple area north of the cella revealed manifest remains of intermediate walls which seem to have divided this outer space into no less than three concentric passages.
One of these intermediate walls could be traced with practical certainty by a large and well-carved foundation beam (marked a), inserted in the flooring parallel to the outer north wall and at a distance of about six and a half feet from it. Numerous detached pieces of frescoed wall stucco turned up in a corresponding position parallel to the west wall. To a second intermediate wall must be attributed the large stucco wall surface, measuring nine by five feet and covered with stencilled rows of small Buddha figures, seen in Fig. 41. It was discovered lying with its frescoed face downwards close to the floor at about the same distance from the north cella wall and approximately parallel to it (see b in Plan). The identical fresco pattern which is known from Dandân-oilik to have been frequently used for the decoration of lower wall surfaces as a kind of dado 4 was found in corresponding positions further east on smaller fragments of wall stucco. Finally, it is possible that certain cuttings in the floor which I noticed chiefly in the temple area south of the cella, running there parallel to the outer wall, may be the result of the quarrying done to remove other foundation beams.
In spite of all efforts, and of the large number of men kept at work, the excavation of this shrine (marked Kha. i) was not completed until the evening of September 26. So great was the mass of sand and débris which had to be shifted, and so exacting the care needed in securing the rich yield of manuscript leaves, fragments of stucco relievos and painted wooden panels, and pieces of frescoed wall plaster which lay mixed up with it.6 There was plentiful evidence that the walls had been constructed of timber and wattle after the fashion of those found in the shrines of Dandân-oilik,6 and had been almost completely destroyed at an early period for the sake of
4 Cf. Ancient Khotan, ii. Pl. II, IV.
The marks Kha. i. C, E, N, W, used in the Descriptive List indicate objects found approximately in the centre, east, north, and west of the area of the shrine.
6 For a specimen of this construction which has survived in a wall of the detached room Kha. viii, see the inset in Plan, Pl. 6.