国立情報学研究所 - ディジタル・シルクロード・プロジェクト
『東洋文庫所蔵』貴重書デジタルアーカイブ

~

tt

Sec. 1]   THE TERMINAL STATION ON THE WALL   637

those who chose it had not rested content to let the wall make its bend here. It was then that my attention was first drawn to what looked like two straight lines of mounds and a third linking them at right angles, all rising above the scrub-covered, salty ground in the direction of the last tower, T. iv. a, westwards. Seen from above and at some distance, they looked distinctly like the remains of much-decayed earth ramparts, and their position, within the bend of the wall and about half-way between its last two watch-towers, at once suggested an entrenched camp. Naik Ram. Singh, who had previously taken the labourers to the ruin T. iv. a to clear its remains while I was reconnoitring elsewhere, had noticed these lines independently, and in fact, when I subsequently visited that tower, they showed up quite as clearly.

Descending from T. iv. b in the direction of the last tower to the west-north-west I found survey re-

scanty, yet unmistakable, remains of the Limes wall on the slope of hard clay, in the shape of the neÏed in

1.

characteristic layers of reed fascines. But beyond I lost the wall amidst the abundant scrub and 9 4 reeds covering the low ground in spite of much salt efflorescence, and as I rode across it, pursued by clouds of mosquitoes and other insects, I found it difficult to follow up the lines of mounds so clearly noticed from above. Hence the account given in Desert Cathay of my observations here remained necessarily imperfect. Fortunately I was able to revisit the ground on March i7, 1914, under less trying conditions, and the following supplementary details are taken from the survey then made.

From the foot of the isolated clay terrace occupied by T. iv. a the line of the Limes wall Camp ram-

could be traced quite clearly, running as a straight, narrow mound, about four feet high, across parts S. of

T. n•. a-b

the flat ground covered with reeds, first for about 48o yards to S. Jos° E. and then for about line. 390 yards further in the direction of S. 94° E. Here, as elsewhere on soil subject to moisture and shôr, the stamped clay and fascines of the agger had uniformly decayed into soft earth. At the point thus reached, which is almost exactly at half the distance between T. iv. a and T. iv. b, the line of the wall strikes a rampart-like bank of earth, rising close on fifteen feet above the level of the ground and about 25o yards long. It faces due west and, though not quite straight throughout nor of uniform height, distinctly suggests artificial origin. Close to the southern end of this mound is another, rising almost as high in places and equally distinct, which turns off at right angles to the east and could be traced for close on 400 yards. From the northern end a similar line of raised ground strikes off, also turning due east and hence parallel to the mound last mentioned ; it could be traced for about 280 yards.

On the east face of the rectangle which these lines might have been meant to enclose the Decay of

ground was uniformly flat. But the very absence of mounds here seemed to favour the conjecture camp ram-

parts ex-

that these banks of earth marked completely decomposed ramparts of clay. It is just on the east plained.

side that the force of wind erosion would have full play, and what the effect of this can be on massive ramparts of the same type, even when situated on ground still supporting a good deal of vegetation, is strikingly illustrated by the remnants of the walls of the ruined town south of Ch`iao-tzû, seen in Fig. 185. There, too, the eastern face of the outer walls has almost completely been destroyed, just as at the Lou-lan Site.10 It remains to be mentioned that I found these mounds thickly covered with tamarisk growth and other scrub, and strewn in places with dead Toghrak trunks and branches, while in the shôr-covered interior vegetation was scant. It was impossible to expect structural remains of any sort to survive on ground like this, subject to the decomposing effects of subsoil water close to the surface and salt efflorescence. Even the small, hard débris of the ` Tati ' type, which alone could withstand them, would be bound to be completely hidden in such soil. In fact, eight months later I could study exactly corresponding results which permeation with water had produced at ruined sites by the side of the salt-impregnated shores of

•   10 Cf. above, pp. 387 sq. ; also below, chap. xxvi. sec. iii; and for the ruined towns near An-hsi, chap. xxvr. sec. ii.