Sec. iv] EXPLORATION OF N. XXVI AND S.E. GROUP OF RUINS 237
observed in this vicinity I may mention that the small ruin N. x, situated to the south-east of the ` Mazar', and excavated in 19oI,7 was now adjoined by an eroded depression which lay in one place fully twenty-eight feet below the original ground-level.
While still encamped near N. xiv I had sent out a select band under Ibrahim, my old guide, to search eastwards for the ruins which Islam Akhûn had alleged he had seen, but which he had proved unable to find again. After three days' search the party succeeded in tracing a number of ruined dwellings south-eastwards which, hidden away amidst high and closely packed sand-cones, had escaped discovery five years before. A reconnaissance upon which I dispatched the Surveyor had fixed their position in a long-stretched line to the east and south-east of the southernmost ruins, N. III, iv, explored in 1901. So on October 27 camp was shifted to the centre of this newly-discovered group, the distance proving close on five miles. En route I was able to examine the ruined Stûpa again, and noted that, though there was little or no change apparent since 190I in the conditions of the drift sand immediately around the small structure, yet the movement of the sand had now left bare the line of an enclosing fence to the south and southeast, thus showing the ancient ground-level (Fig. 65). At the ruin N. III neither the cover of sand nor the state of erosion showed any perceptible change. From here we marched for over a mile to the south-east across broad dunes, fifteen to twenty feet high and often broken by chains of big tamarisk-cones, until a convenient camping-place was found on deeply eroded ground near three small ruins, which rose on ` witnesses ' of striking appearance (Fig. 66).
The northernmost of these, N. xxvii, was situated on steeply eroded slopes by the west side of a tamarisk-cone. It consisted of the exposed and badly eroded remains of two rows of small rooms built in timber and plaster, forming a block of about fifty by forty feet. The only objects recovered by clearing them consisted of a rectangular covering-tablet on which all writing had been bleached, and a wooden comb (Plate xxviii). N. xxviii was another small ruined dwelling occupying the top of an island-like ` witness ', seen in Fig. 66, which rose with almost vertical slopes some twenty feet above the eroded ground. Abundant débris of timber lay around, but only three small rooms and a passage to the west of them were still traceable. The floor had been swept clear by the winds, except in a corner of the passage, where there lay sand but only to a height of one foot. It was here that Ibrahim alleged he had found the fifteen tablets which he brought to me on October 24 on returning from his reconnaissance, and which are shown at the end of the Descriptive List marked N. Ibr. 1, 001-00I1, 0013, 0014, 0016. I have no special reason to doubt the correctness of his statement, and the generally bleached or warped condition of these tablets, as seen in the specimens, N. Ibr. oo5, oo16, reproduced in Plates XXVi, XXVII, distinctly supports it. A smaller ` witness ', some fifteen yards to the south, which had been completely cut off from the main plateau by the progress of erosion, retained on its top a stratum of horse- and sheep-dung. About 15o yards to the west-south-west there emerged from the protection of a sand covering the shrivelled trunks of an arbour, with a fence traceable for about forty yards, seen in the foreground of the photograph, Fig. 66. At the south-west end the remains of a single-roomed structure, N. xxxiv, could just be traced.
Proceeding about three-quarters of a mile to the south, past scattered remains of more arbours and gardens, I reached the ruin of a relatively large house, N. xxix, previously reported by the Surveyor. The plan in Plate 16 shows the disposition, etc., of the rooms, which were built partly in timber and plaster and partly with plastered rush walls. Room i, to the north-east, had evidently served as an office and yielded sixteen well-preserved Kharosthi tablets, mostly wedge-shaped or rectangular. From the adjoining room ii came, besides some small oblong tablets, a large double-
? See Anciint Khotan, i. p. 379.