776 ON THE ANCIENT ROUTE ALONG THE KONCHE-DARYA [Chap. XX
The masonry was reinforced by timber, but the use of this was much more restricted than in Y. III and Y. iv, owing to the far greater solidity of the masonry.
In view of the constructive features just noticed, it appears to me probable that this watch-tower, too, like those previously described, goes back to Former Han times. The enlargement may have been undertaken soon after the original tower was built, perhaps with a view to greater height and thereby wider outlook and visibility. Among the debris on the southern and eastern faces refuse, consisting mainly of reed-straw, was found up to a height of about 8 feet from the ground. In this, and only a.few inches from the surface, were found two small leather pieces from lacquered scale armour. Their ornamentation with red designs on black ground closely resembled that of the leather armour scales excavated from the rubbish layers of the Tibetan fort at Miran.n Leather thongs like those used as fastenings on the Miran scales still adhered to these pieces. The resemblance to the Miran finds and the place where the scales were picked up suggest that they found their way into the refuse in Tang times, when the towers, though already decayed, may still have served as convenient halting-places. On a narrow tongue of the Sai terrace and about a hundred yards to the north of the tower there were graves scattered over an area about 5o yards across. They had all been dug up, evidently a long time ago. Ibrahim, on the strength of a statement by his father, said that this had been done by Korla people about forty years ago, i. e. during Yagab Beg's régime. No traces of coffins survived and but little trace of bones. The graves were probably those of soldiers or wayfarers who had died at this dreary roadside station.
From Y. vii the ruined watch-station known as Yàr-karaul could be seen, and we reached it after a march of seven miles over ground where tamarisk-cones, and farther on reeds also, again made their appearance. The terrace-like edge of the gravel glacis, to which the route keeps near, is broken up near Yar-karaul into a number of bold Mesas. One of them is occupied by the ruined post and accounts for its name. They have been carved out of the clay underlying the gravel surface of the Sai by wind-erosion, which finds here a very effective instrument in the coarse sand washed down from the foot-hills. The Mesas stretch from north to south, clearly showing the prevailing wind direction. About a quarter of a mile before Yar-karaul we passed a smaller terrace of this kind, whose walls of clay cut through or hollowed out by erosion gave it the appearance of a ruin.
The Mesa bearing the ruined post, Y. viii (Fig. 346 ; Pl. 38), rises to a height of about 5o feet and on its top extends over a length of about II 2 yards, with a maximum width of less than 6o yards. The remains of the small tower or guard-room occupy approximately the centre of the flat wind-eroded summit. Of the walls, 4 feet thick, only that facing south and containing the entrance still stands to a height of about io feet. The other walls, forming with it an enclosure 19 feet square on the outside, have been demolished to within a foot or two from the ground by treasure-seekers. These have also burrowed into the foundations, which are made of rough blocks of clay so as to enlarge a small natural terrace into a base. The bricks used in the masonry of the walls are of the usual size of i5" x 7-8" x 3". The clearing of a rubbish-heap below the southern wall of the ruin, composed mostly of reed-straw and remains of fuel, yielded only the fragment of a Chinese paper document and a few small pieces of plain silk. That the top of the Mesa was gained, in ancient times as now, by a steep and narrow couloir from the south-west, was proved by a small layer of refuse like the above found under the sheltering cliff about half-way up. About a dozen graves, all opened long ago, were traceable in two groups to the south and south-east of the ruin. The hollows marking them showed a north to south direction corresponding to the slope of the plateau.
11 See Serindia, i. p. 464 ; iv. PI. L. I much regret that the two leather scales from Y. vii cannot be traced at present in the collection.