double-bracket, nearly seven feet long, which I had noticed on my first return to the site. It is seen photographed in the foreground of Figs. 63, 64 after transference to the central room, iii. Its ornamental wood-carving showed closest resemblance to the floral motifs displayed by the central panels of the double-bracket found in room iii and already described.
The excavation of a small dwelling close to my camping-ground and without any special interest, was still proceeding under the Surveyor's supervision when, on the afternoon of October 26, I revisited the ruin N. v, where in 1901 I had cleared the precious rubbish heap of N. xv with such abundant results. What drew me back there, apart from the attraction of a scene of successful work, was the wish to recover some items from the refuse which I remembered with regret that I had thrown aside then as valueless. These were small oblong pieces of hard ` green ' leather, rounded at one end and peculiarly punched with holes, in which I had since learned to recognize pieces of leather scale armour.3 To my satisfaction I found the refuse we had cleared out still undisturbed under a light cover of drift sand, and my conscience was relieved when, after some careful scraping, we recovered the missing scales, N. v. xv. 004-006, 0011. As their description in the List below shows, they closely resemble in appearance the one already illustrated, though there are slight differences in size and in the position of the holes which served for threading the scales together. A find even more gratifying was a small Stûpa model in wood, N. v. xvi. 001 (see the drawing in List, p. 247), which I picked up on the surface of the ground close to the adjoining structure N. xvi. It had evidently been laid bare by a slight move of the drift sand, and now afforded striking confirmation of the belief expressed in Ancient Khotan that this small and badly decayed ruin was that of a Buddhist shrine.4 The model which, no doubt, had served as a votive offering, shows quite clearly the threefold base, the cylindrical drum, and the dome characteristic of the Stûpa type prevailing among the ruins of the Tarim Basin. It deserves, however, to be noted that the proportions of the successive bases in this little model differ materially from those observed in the only Stûpa of the Niya Site, the top base being shown as the highest, while the bottom one is so in the real Stûpa.5
On the same occasion I also examined a spot situated a little under a mile to the north-northwest of N. xxvi (see site-plan, Plate 7), where Ibrahim and a companion had, while ` treasure-seeking' two years before, come upon remains of skeletons. They had promptly christened it the ` Mazar', an unconscious preliminary, perhaps, to resumed local worship. There, by the side of an isolated tamarisk-cone some thirteen feet high, I found, indeed, unmistakable indications of an ancient cemetery. Not only were there plentiful human bones scattered over the eroded slopes of a small plateau, especially to the east and north, but also bleached and splintered boards which, to judge from their sizes, must have belonged to coffins. The larger pieces measured on the average six to seven feet in length ; the shorter ones clearly showed their use as head and foot pieces of coffins by the tenons and mortises intended for dovetailing. The fragments of dark brown and red fabrics, marked N. x. 001, were picked up near a heap of bones here. The discovery of this cemetery is of interest as it furnishes conclusive evidence, not otherwise available at this site,° that interment was practised by at least a portion of the inhabitants, the majority of whom we may safely assume to have been Buddhists. Unfortunately it is impossible to base much archaeological hope upon this fact seeing how little chance any objects buried with the dead would have of escaping destruction on ground exposed to the full force of wind erosion. Here it had left no intact skeleton, not even a single complete skull for measurement. As an instance of the burrowing power of the winds