Sec. THE SITES OF L.L. AND L.M. 193
was completely eroded for a distance of about 74 feet. To the south of this portion the east wall Enclosure in face projected about 42 feet outwards to afford space for a small inner enclosure adjoining it, about S.E. corner. 68 feet long, which extended to the south-east corner. This inner enclosure was marked off from the rest of the interior by walls on its north and west sides, about 8 feet thick and consisting mainly of
rough tamarisk fascines set in clay. The ground of this enclosure was raised about 8 to io feet
above the rest and was thickly covered with masses of reed-straw and refuse, mainly the dung of
sheep and other animals. In clearing these layers of refuse there were found numerous fragments of fabrics, mainly of felts and woollen materials (L.L. oi, 03, 06-7, 013-15).
A fragment of printed silk, L.L. 02 (Pl. LXXXVI), found here shows on blue ground a diagonal Find of trellis pattern of white dots, and may claim particular interest because Mr. Andrews' examination
indicates that its weave is the same fine warp-rib which his study of the figured silks from the cemetery silk and
printed site L.C., discussed below,' has proved to be peculiar to our earliest specimens of Chinese silk
fabrics from Han times .2 The chronological indication furnished by this printed silk was not recognized by me at the time, nor was any to be derived from the small fragments of bronze or iron (L.L. 04-5, oio) found in this place. But fortunately a small piece of torn paper was discovered in the same refuse deposit, with a few characters distinctly resembling Early Sogdian writing, and
these alone sufficed to make it highly probable that the occupation of this little fort dated back to the same period as that of the Lou-lan station L.A.
The absence of any structural remains in the interior made it impossible to determine whether Erosion of the refuse layers just referred to were due to occupation during the time when the fortified post interior.
served its original purpose or to such later use of its shelter by herdsmen as was proved to have occurred at various ruined structures at Miran and the Lou-Ian Site.3 Apart from the small inner enclosure which the refuse deposits described had helped to protect, the interior of the circumvalla? tion was found completely scoured by erosion. Though less striking in its effects (the maximum depth of erosion was here only about 6 feet), it had followed exactly the same course as at L.K., the sand having been driven into the interior through the breach on the east side, and out again by a gap towards the north-west corner. Of a timber and plaster structure that had once stood outside about 6o yards from the east wall, only splintered fragments of foundation beams, &c., survived on the slope of a low mound.
From L. L. I proceeded to the north-west under the guidance of one of the men who had accompanied Afraz-gul on his reconnaissance, and after a march of about three miles reached the nearest of the ruined dwellings reported. Our way led over eroded ground covered in many places with detached dunes 6 to to feet high, and along it worked stones, as well as potsherds and small bronze and glass remains, were picked up in increasing numbers as the site L.M. was approached (see List). Among the small glass objects, the beads in coloured or gilt glass (L.L—L.M. i. 02 I-2, 026 ; L.K.—L.M. 01-3, o10-14) and pieces of cut or moulded glass (L.L.—L.M. 1. 02, 07 ; L.K.—L.M. 04-9, Pl. XXIV) may be specially mentioned. The ruins of the L.M. site were found to be scattered on either side of a depression, manifestly an old river-bed, lined with rows of dead Toghraks and winding in a general direction from west to east (see the plan, PI. II). The first ruined dwelling,•
L.M. I, which was reached after crossing a similar but smaller depression, rose on an erosion terrace about 14 feet high, and by its position and appearance closely recalled the ruins of the Niya Site. Low walls of timber and wattle partly covered with drift-sand indicated the position of rooms on the original ground level ; that of others badly eroded was marked only by big pieces of timber, which strewed the eroded slopes of the terrace all round.
it 1 See below, pp. 232 sqq. (Burlington Magazine, 192o, September).
2 Cf. Andrews, Ancient Chinese Figured Silks, p. r9 3 Cf. Serindia, i. pp. 245, 400, 402, 427, 490.