Sec. iv] REMAINS OF A WALLED ENCLOSURE 387
tamarisk bundles embedded in the wall of stamped clay at intervals of about two feet. Beyond this section, the line of the wall could still be made out in places by tamarisk bundles which had once served as its foundation and now survived on the top of small Yardangs. The original thickness of the wall could not be determined with any certainty, as it was evident that its sides had everywhere suffered much from the paring effect of the sand driven along it. The subsequent plane-table survey of the site showed the distance between the north and south walls I had thus traced to be approximately 1,020 feet.
It was a more difficult task to ascertain the position of the west and east walls, which obviously Difficulty of
must have once completed the defences of the small fortified station. Among the close-set Yârdangs and
which furrowed the ground everywhere immediately beyond the area occupied by ruins I looked in walls.
vain for any sign of a continuous line suggesting a wall. The time I could then spare from other pressing tasks was too limited to permit of a prolonged examination of the problem, and as the Surveyor was again hors de combat with rheumatic pains, which exposure to the bitter cold and the constant cutting wind had increased—on December 22 the minimum thermometer showed 46 degrees Fahr. below freezing-point—I was deprived of such guidance as a large-scale plan made with the plane-table might have furnished.2 It • was owing to these difficulties that I failed altogether at that time to discover• the true position of the east, or to be more precise east-northeast, wall, and that the two small segments of the west (recce west-south-west) wall which I actually did notice I first erroneously interpreted as remains of two projecting towers which might have flanked the western gate of the little stations
In reality these two clay mounds close to L.A. iv, as my fresh survey in 1914 clearly showed, Position of
fall exactly into the line of the west wall, running N. 330° W. to S. 150°E. The southern one is west ate
about twenty-four .feet long, and shows at its base a thickness of about fifteen feet. It rises steeply to about sixteen feet above the eroded ground level ; but it was impossible to determine how much of this height belongs to the superstructure and how much is merely the result of the lowering of the ground through wind-erosion. To the north, beyond a gap of some thirty yards, rises a second and smaller segment of the wall, about fifteen feet long and nine feet thick. Its height is about nine feet, and on its top two layers of tamarisk bundles could still be made out, separated by about three feet of stamped clay. The survival here of these small remnants of the wall is easily accounted for by the protection which the ruined dwellings, L.A. IV–v1, closely adjoining must have given. The ground between the two segments is strewn with heavy timber débris, and as our surveys of both 1906 and 1914 show this gap to be exactly in the middle of the west wall, we may, I think, safely assume that the western gate of the fortified station stood here. I found closely corresponding examples of large gateways, built with heavy timber, in the ancient fort of 'Kara-dong on the lower course of the Keriya River s and in the fort L.K. explored in 1914 on my way to the Lou-lan Site.b
Of that portion of the circumvallation which had faced east-north-east I had been unable in East face of
1906 to discover any trace. So I was forced to the conclusion that the constant scouring of that circumvalla-
terrible east-north-east wind, which, as our experience showed, does not cease completely even in the winter, and of the sand it drives before it, must have first breached this wall face directly obstructing its way at every point marked by a Yârdang trench, and ultimately broken down and
2 The site plan of the L.A. station, reproduced in 3 See Desert Cathay, i. p. 388.
Pl. 23, has been prepared from a fresh survey of the site ' Cf. Ancient Kholan, i. pp. 447 sq., Fig. 53; ii. Pl.
carefully executed in 1914 by Min Afrazgul under my super- XXXVIII.
vision. The details of particular structures shown in it have ° Cf. my Third Central-Asian Expedition, Geogr.Journal,
been inserted from the plans made in 1906. - 1916, xlviii. p. 1 2 1.