Sec. iii] WATCH-STATIONS ALONG THE ANCIENT ROAD TO KORLA 773
they would be certain to strike either the Konche-daryâ or some lagoon fed by it. After an anxious night, with barely enough water for us men in camp, a long wait followed in the morning before the men with the animals rejoined us. They had found great difficulty in crossing the old bed, filled in parts with salt marsh ; then, going on to the south, they had passed a wide belt of soft salt-encrusted ground, and after crossing a ridge of drift-sand suddenly found themselves by the side of a small lagoon of fresh water, evidently formed by a recent inundation. There it was possible to water the animals and to refill our two tanks. Only dead trees were to be found near the lagoon, but at dawn a line of live Toghraks was seen in the distance beyond. Abdulmalik estimated the distance they had covered on their return march to camp at about four and a half ' Pao-t`ais ', say eight miles, in a direct line, and this seems to agree well enough with the distance shown by Dr. Hedin's mapping of 1900, when he travelled from Dilpar on the Konche-darya in a northeasterly direction to a point which could not have been very far away from our Camp ccil.
When fixing my plane-table in the morning on a tamarisk-cone near the northern edge of the Remains of
jungle belt, I had sighted a large tower to the north-west. Much time had already been lost, and watch-
there was no hope of reaching water on that day at the bend of the Konche-daryâ which Dr. Hedin's Y. in. description indicated at Sai-cheke. So I decided to forgo the search for the springs of Kalta which we had left behind us, and for the tower reported near them, and to march on to the conspicuous ruin, Y. III, in sight. It was reached after close on three miles' march and proved to be a massive and in its way quite imposing structure (Fig. 347). It has now the shape of a truncated pyramid, with a base originally about 55 feet square and a top about zo feet square at its present height of about 3o feet. Large posts projecting above the top show that the original height must have been at least io feet more. Up to 7 feet from the ground the tower was built with layers of reeds interposed at intervals of 4 inches ; whether these separated brick masonry or stamped clay could not be clearly made out, owing to the masses of debris covering the lower slopes. The base of the tower was strengthened by a revetment of stout timber, which along the foot of the northern face still survived to a height of close on 4 feet. Above this base the tower shows a remarkably solid construction. The masonry courses of bricks are overlaid at intervals of 1-- feet by reed fascines, on which rest great rafters of Toghrak wood. These in turn secure massive posts carefully fitted in chules, the whole forming a timber framework of remarkable strength and cohesion. The bricks show the same size as at the previously described towers, measuring 15 inches by 8 with a thickness of 3 inches. The tower, even with its present reduced elevation, commands a very extensive view over the level plain of soft clay. It had obviously been made so lofty in order to facilitate signalling operations over ground which to the south-east was probably, then as now, covered with close jungle and tamarisk-cones. No trace of surface drainage or subsoil moisture was to be found near Y. III. Yet prolonged occupation of the post was proved by refuse layers embedded under debris at the northern foot of the tower. Amidst the quantities of reed-straw and dung found there we recovered part of a Chinese document, which by the style of its writing and the quality of its paper recalled documents found at the Lou-lan station ; also two wooden implements, apparently eating-sticks, Y. III. 01-2 (Pl. CX).
From the top of Y. iii another tower had been visible to the north-west in spite of the haze, Ruined and we reached it after proceeding about five miles over a level clay steppe with low scrub and light Ÿ tv. salt efflorescence. Before reaching it we crossed a wide but very shallow flood-bed coming from the hills to the north. The ruined watch-tower Y. iv (Fig. 348) proved in size, shape, and construction a close pendant to Y. 111 just described. It had, however, suffered a good deal more from decay. Much of the brickwork and reinforcing timber had fallen, the southern side being the least injured. A portion of the timber framework exposed above the extant top showed that the tower must have
II 5 G