76 FROM KASHGAR TO THE KHOTAN RIVER [Chap. III
and to point out its bearing upon the question what broke the continuity of the ancient hill range that once probably extended across the Taklamakân from the vicinity of Marâl-bâshi to the Mazdr-tâgh on the Khotan river.
March to Some two miles beyond the gap we came, to my surprise, upon a large patch of living vegeta-
Lal-kan. tion including reeds and tamarisk scrub, for which it seemed difficult to account. Stranger still
appeared a broad belt of dead Toghrak forest, through which we made our way in darkness under Barat's guidance towards a point of the Lâl-tâgh where he remembered an old mining pit and two rock cisterns near by. We struck the point correctly after a total march of some 28 miles—but the ` Kaks ' were found no longer to hold water. It was as well, therefore, that I had not carried out my original intention of making for Lal-tagh direct from our Camp xvi on ` the old route '. The now abandoned pit known as Ldl-kan was said to have been worked at one time for semiprecious stones, and from it the low hill chain of the Lâl-tâgh derives its name.
Old river- The range, which we followed next morning along its foot towards the south-east, was cut
bed at foot by erosion into fantastic forms and, like most of these island-like hills near Marâl-bâshi, proved of Làl-tàgh.
to be composed mainly of feldspar-carrying sandstone, stratified almost horizontally. At first only dead tamarisk-cones were to be seen on our right. But after rounding a little promontory of strikingly bold cliffs, four and a half miles from camp, there appeared lines of large dead Toghraks, and amongst them, a mile or so beyond, an unmistakable ancient river-bed skirting the foot of the Lâl-tâgh. It seemed to come from the south, and could manifestly originate only from some ancient branch of the Kâshgar river. This at the present day dies away in the marshes to the south and south-east of the hill islands of Tumshuk (Map 8. B. I). The belt of Toghraks, among which I noticed one big old tree still green at its head, continued to skirt the foot of the L5.1-tâgh to near its end, and after covering a total distance of seven and a half miles from Lâlkan we reached the ruined site that we were seeking, in a picturesque little combe opening from the south-east.
Ruined The remains of the site, as I had been led to expect by the fragments of stucco relievos brought
shrines of from it and received in 1908 through Ayûb Mirâb,b soon proved to be those of a Buddhist sanctuary. I,âl-tizglz
site. As shown by the rough plan (Pl. 3) and Fig. 8o, they comprise in the centre a shrine (iii),
completely destroyed, covering a rectangular area, approximately 76 by 5o feet, and raised on a solid brickwork platform about 12 feet high ; on the slope close above it the foundations of a smaller structure measuring about 48 by 26 feet (ii) ; and adjoining the north-east side of the former a solid masonry tower (i) measuring at its foot about 36 by 20 feet. Fragments of stucco relievos, mostly small and all completely calcined, which a superficial cleaning brought to light both on the top of the platform that once carried shrine iii and below its south-west side, clearly belonged to the images and decoration of a Buddhist sanctuary, dating approximately from the Tang period.6 Among them I specially noted a piece from the drapery of a colossal image and the well-preserved head of an elephant, which probably formed part of a relievo frieze of the type represented at the temples of the ` Ming-oi ' site near Shôrchuk, Kara-shahr.7 From the appearance of the exposed masonry and from what the plentiful burrowings of local ` treasure-seekers ' had disclosed of the debris, it was certain that the shrines in the centre had been destroyed by fire. Of a smaller structure to the east of the central pile and of two other little cellas nearer to it on the north and south, also burnt, only the bare outlines could be traced. On a low terrace about 3o yards to the south of shrine iii there survive the walls of an oblong building (v),
b Cf. Serindia, iii. p. 1312. objects recovered, see the Descriptive List at end of the
6 For the specimens brought away and for other small section.
7 Cf. ibid., iii. pp. 1191 sqq.