国立情報学研究所 - ディジタル・シルクロード・プロジェクト
『東洋文庫所蔵』貴重書デジタルアーカイブ

174   ON THE WAY TO LOP-NOR   [Chap. V

top of a tamarisk-cone close by afforded a distant view of an extensive belt of tamarisk-cones to the north and east and of a glittering line of shôr far away on the horizon where the terminal marshes of the dying Tarim spread their salt wastes. It was an inexpressibly sad landscape, befitting this Ultima Thule of inhabited ground in the south of the Lop basin.

More interesting than this old watch-tower, but also far more puzzling, was the ruin M. xv, to which I was guided by Niâz, one of the Lopliks of MIrân, on January 18th. It lay about a mile to the north-east of M. v, amidst low tamarisk-cones and not far from the eastern edge of the riverine belt of vegetation,20 and presented itself as a shapeless mound of sun-dried bricks mixed with hard lumps of clay (Fig. 119). Nowhere on the exposed slopes did first inspection show brickwork in proper horizontal alignment, nor was any found in the subsequent clearing, which took the best part of two days. The height of the mound was about 15 feet, and the masses of fallen masonry composing it proved closely compacted throughout.

The first indication of a structural feature was secured when, at the eastern foot of the mound, we came upon a short segment of what obviously was the inner face of a circular wall. The surviving portion rose nowhere to more than two feet, and of the colours with which its surface had been painted there remained but faint traces. The original thickness of this wall, which I first took to have enclosed a circular passage, could not be determined either here or on the north-west, where I subsequently traced some scanty remains of its inner surface to a height of only 6—io inches above the floor. But from the relative position of these two wall segments the diameter of the enclosed area could be roughly calculated as 19 feet 8 inches. That this circular wall had not formed the enclosure of a Stûpa, like those of M. III and M. v, but belonged to a round chapel or Vihâra sheltering large Buddhist images in stucco, became clear as the removal of the hard masses of debris proceeded. No trace was found of a Stûpa base, such as would certainly have survived the collapse of the enclosing wall and of the dome that once spanned the rotunda. But instead there came to light a series of stucco heads embedded in the debris and attesting the former existence of images which the masonry from wall and vaulting had completely crushed in its fall.

Thus from the south-west portion of the interior there were recovered the two heads, M. xv. 014, 017 (Pl. XX), probably of Buddhisattvas, both well modelled and retaining in part the colouring of the faces, notwithstanding the very friable condition of their plaster. On that side, too, was found the life-size head, M. xv. 015 (Pl. XIX), interesting on account of the unusual treatment of the eyes and the peculiar arrangement of the hair in heavy tresses festooned over the forehead. The two Buddha heads in relief, M. xv. 018, 019, somewhat over life-size, were found in the western portion of the mound, about three feet above the floor level, and near them the fragmentary grotesque head of a demon, M. xv. 013 (Pl. XX). The right half of the leg and body of a colossal seated Buddha was found close by, but broke when the heavy brickwork in which it was embedded was being removed. The treatment of the drapery of this torso closely resembled that seen in the colossal seated Buddhas, of which portions were discovered at M. 11.21 On the east the head of a colossal Buddha, badly broken, lay face downwards, so firmly wedged amidst the debris that it could not be removed except after complete demolition. The same was the case with the knees of a seated Buddha discovered in the very centre of the mound and at a height of some four feet above the floor.

We searched in vain amidst the amorphous masses of brick and clay for indications of the original position of the images to which these fragments had belonged. That any remains at all of stucco sculptures should have escaped complete destruction must, in fact, seem surprising. From the condition in which the fallen masonry was found there could be little doubt that the enclosing

20 See Serindia, iii. Pl. 29.   21 See ibid., i. p. 488, Figs. 123, 124.

Ruined mound M. xv.

Clearing of collapsed rotunda M. xv.

Sculptural remains recovered.

Features of collapsed dome.