772 ON THE ANCIENT ROUTE ALONG THE KONCHE-DARYA [Chap. XXI
masonry courses. This reinforced masonry forms a facing to the tower of a thickness of about 2 feet. The core inside is formed of alternating layers of clay and reeds, the latter inserted at intervals of 12 inches. On the northern side the masonry facing has completely fallen owing to the foundation having given way, and the reed layers of the core here show signs of burning.
Relics from Towards the west a slightly higher portion of the terrace bears the remains of what evidently
ancient had been living quarters, consisting partly of low walls of stamped clay, but mainly of wooden
posts standing in rows, which evidently formed part of wattle and plaster walls otherwise completely destroyed. The rectangular area enclosed by them measures about 27 feet by 19. In one corner a large round hole with fragments of pottery in it shows where a large jar has been fixed in the floor ; on the inside of the north-east wall traces of a fire-place built of plaster could just be made out. The interior of this rectangular area was covered with thick refuse, made up mainly of reed-straw and horse-dung. But near the southern corner a number of miscellaneous articles were recovered. Among these were the fragment of a large wooden comb, Y. 11, of (Pl. CX) ; a rudely made ` plasterer's float ' in wood, Y. II. 09 (Pl. CX) ; a pair of well-made string sandals, Y. ff. O10—I I, of the same type as found at L.A. and along the Tun-huang Limes ; a wooden winder for string, Y. If. oz (Pl. CX), &c. Special mention must be made of the small bone die Y. ft. 03 (PI. CXI), with numbers on the sides of the cube marked in the same way as on a die from the Tibetan fort at Mirân ; and of a roughly executed painting on paper, found torn into several pieces, Y. II. 014 (Pl. CVII). It shows a Chinese building in front elevation, with a grotesque figure approaching ; a grotesque beast also appears on other fragments. Many of the outlines are punched with small holes as if intended to be used as a stencil.
Revetment A very curious feature of the ruined watch-station is the carefully constructed revetment of
of platform. alternating layers of clay and reed fascines, each about 4 inches thick, which has been used to enlarge the platform on the south. I t closely recalled a similar arrangement applied at T. XLIII. h, far away on the Limes of Hua-hai-tzti, to secure wider space on the top of a tamarisk-cone chosen for a watch-station.' In order to strengthen this extension of the platform and at the same time to facilitate access to the top, two solidly built ramps had been constructed leading up from below at right angles to each other. These ramps, 3 feet wide, were made of short rafters set into the clay and layers of reeds, and kept in position by double rows of stout vertical posts. The outside of the ramps was revetted with hurdles of brushwood. The whole gives an impression of great solidity and equals the constructive skill and neatness that I had observed in the wall line of the ancient Han Limes west of Tun-huang. It seemed clear to me that we have here the work of Chinese hands, trained in the same school of engineering skill as the builders of that other rampart, which, after two thousand years, still survives all the vicissitudes of the bare wind-swept desert.
Failure to The examination of these remains had delayed us, and when, after returning to the edge of
locate the gravel Sai, we marched on in order to gain the springs which Dr. Hedin's description indicated
uPlna in a locality known to his guides as Kalta, we failed to discover the ruined tower he mentions as
close to them. A high tamarisk-cone that looked like a watch-tower in the falling dusk induced us to pass on beyond two small stone heaps, where probably we ought to have turned south into the adjoining belt of tamarisk and scrub. When at last, after a march of some twenty-seven miles, darkness obliged us to halt amidst thin Toghrak jungle, no water could be found anywhere, only a stretch of soft shôy-covered ground to the south, evidently marking an old bed of the Konchedaryd. As the ponies had tasted no water during two marches, and it seemed very doubtful whether by turning back the springs could be found in such deceptive ground without much loss of time, I sent them, with a couple of men and some camels under Abdulmalik, to the south, where
7 Cf. above, i. p. 392.