378 THE LOU-LAN SITE [Chap. XI
window, six feet wide and flanked by posts, above the centre of this platform. As I had found an exactly corresponding arrangement in the ruins N. xxiv and N. xxvi of the Niya Site, described above,6 we have here fresh evidence how closely the constructive features of dwellings agreed in widely distant parts of the Tarim Basin during the third century A. D. In room vi was found the wooden measure L.A. ii. vi. col (Plate xxxv) already mentioned.' In room vii the posts and wattle showed signs of having been partially destroyed by fire, and charred fragments of wood were found under the drift-sand covering the floor.°
The structural remains marked L.A. in (see Fig. 102, Plate 25) seem, as already explained, to have belonged to the south-western wing of the ` Ya-mên'. Of the large room, L.A. ni. iii, measuring about J5 by 28 feet, the walls of timber and horizontal reed wattle were clearly traceable on two sides only, the rest being badly eroded. Some of the posts, however, still stood upright ; two among them, which must have carried one of the beams supporting the ceiling, still rose to a height of 13 feet. This room had been searched by Dr. Hedin, who had found in or near it a large and well-preserved earthenware jar, a decorative wood-carving, and what he took to be the solid wooden wheel of a cart.° Among the pieces of architectural wood-carving left behind, Fig. 99 shows the wooden circular base with socket on which, one of the central pillars had once stood, a badly splintered double cantilever on which one of the roof beams must have rested, two volutes probably also from cantilevers, and two turned balusters with elaborate mouldings similar to those subsequently recovered from L.B. iv (Plate XXXIII). All these pieces were of Toghrak wood.
On carefully clearing the floor of the small and almost completely eroded apartment ii, we came upon two fragmentary Chinese slips, and when this search was extended to the ground immediately adjoining on the south-west, which I first assumed to have been occupied by another portion of the structure, L.A. iii. i, Chinese records on wood and small miscellaneous relics rapidly cropped up in numbers. Thus some thirty-seven slips bearing Chinese writing were recovered here, besides two small fragments of Kharosthi documents on wood. Among miscellaneous finds may be specially mentioned a bronze arrowhead, L.A. iii. ooi (Plate xxIx) ; the portion of a wooden bowl, lacquered red and black, L.A. III. 004 ; a lacquered wooden style, perhaps intended for writing, L.A. iii. i. 002 ; and a small fragment of an earthenware dish, L.A. iii. oo2, decorated with a faintly lustrous slip in green, resembling Chinese work of Han times. In reality, we had here struck the easternmost portion of that huge refuse-heap L.A. vi. ii, the subsequent discovery of, and abundant yield from, which we shall have to describe presently.
The ruin to which I turned on December 19, after clearing the last-named apartment of what we called the Ya-mên, was that of a relatively large dwelling, L.A. iv, close on a hundred yards to the south-west. It comprised, as the plan in Plate 24 shows, a number of rooms of varying sizes up to 21 feet square, roughly built with walls of timber and vertical tamarisk rushes, and in addition to these, in the centre, a group of apartments more solidly constructed of stamped clay or of timber and plaster.10 Sand filled it to a height of three to four feet. The room i, about 13 by 12 feet inside and built with walls of stamped clay about three feet thick, was provided with a low sitting platform of clay along three of its walls and with a fire-place, also of clay, on the fourth. The internal arrangement thus corresponded exactly to that observed in many of the
g Cf. Hedin, Central Asia, ii. pp. 633 sq. with the excellent photographs in Pl. i,XX, LXXI.
10 For a good view of part of this ruin see Hedin, Central Asia, ii. PI. LXXII.