774 ON THE ANCIENT ROUTE ALONG THE KONCHE-DARYA [Chap. XXI
risen well above the present height of about 3o feet. The Toghrak beams were all fitted with chules. Those lying on the slopes and along the foot of the northern and western faces were all remarkably massive, having lengths up to 20 feet or more and measuring about 8 inches square in section. Up to a height of io feet from the ground the masonry consists of single brick courses with a layer of reeds on the top, the two together having a thickness of 4 inches. Higher up, each five or six brick courses carry a layer of reed fascines and embedded in these horizontal beams secured by big uprights. At the foot of the eastern side we discovered some refuse, which yielded fragments of coarse woollen canvas, Y. iv. of ; vegetable fibre rope, Y. iv. 02 ; tangles of string, &c. About 18 feet from the southern face traces of a clay or brick wall probably marked the position of quarters.
No ruin could be sighted from the top of the tower ; but by continuing to steer on the same north-westerly course as close as the numerous tamarisk-cones, i 2 to 15 feet high, now studding the plain, would allow us, we came upon another ruined watch-station at a distance of less than four miles. Its much-decayed remains are far less conspicuous than those of Y. iii and iv, which explains their having escaped Dr. Hedin's guides. The tower Y. AT was built entirely of alternating layers of reed fascines and earth, 3 and 2 inches thick respectively. I t appears to have originally measured about 24 feet square at the base, and now stands to a height of only 12 feet or so. But to this is added the height, about 8 to io feet, of the mound on which it is built. This could only have been an old tamarisk-cone, and the fact that its soft earth could not carry a massive timber structure probably explains why the method of construction with lighter materials was here resorted to. In order to secure more cohesion between the fascine layers short posts, 3 to 4 inches in diameter, were driven through them. But this had not prevented the whole mass from sliding down towards the south-east owing to subsidence of the loose earth in the old tamarisk-cone. On the west face I noticed the same curious curving of the fascine layers that I had observed in places along the frontier wall west of Tun-huang, of similar construction and dating from Han times, where, owing to sloping ground, no firm foundation could be secured for it.8 It was evidently, in both cases, the necessity of rapid construction that made it difficult to guard against a defect which would come to light only after the lapse of considerable time.
No other tower could be sighted beyond, though the ground now became more open. A continuous line of big Toghraks was to be seen far away to the west and south-west, clearly marking the present course of the Konche-darya. We halted for the night near the bank of a well-defined curving bed, probably a lagoon which, as appeared from the abundance of reeds and scrub, had received water from the river until recent years. Here the terraced edge of the Sai, which had passed out of sight beyond Kalta, again drew close in from the north. But it was here much lower, and beyond Sai-cheke, about five miles farther on, became quite indistinct. At that point a fairly well marked track we had picked up on the morning of March 25th brought us to the bank of the Konche-darya which, by the big bend (cheke) it here makes, has given the place its name. A short distance before reaching it, we luckily fell in with Ibrahim, a young hunter from Shinega, towards Korla, and thus secured the local guidance we so sorely needed. The sight of the river flowing in a deep-cut reed-lined bed, with a fine volume of perfectly clear water, was very grateful to eyes which, for many long months, had seen nothing that could be called a river. The width of the Konche-daryà was here fully 40 to 5o yards and its current about two feet per second.9 We had no means of ascertaining its depth ; but it could scarcely be less than 6 to 7 feet at its shallowest, and Ibrahim talked of ten gulach in the middle, which, if obviously an exaggeration, was yet significant.
8 See e. g. Serindia, ii. Fig. 189. conjectured course above and below has been shown too
9 The width of the Konche-daryâ here and along its wide in Map No. 25. B. z, through a draughtsman's mistake.