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

Sec. iv]   SEARCH OF RUINED LIMES STATIONS T. XXVIII–xxx   603

sanctuary of the ` Gate ' by the ruined station T. xxix. Local worship dies hard—in China quite as much as elsewhere.

T. xxx, the westernmost tower I could trace on this part of the Limes, proved to be a square mass of stamped clay, rising to about fifteen feet in height, but too badly decayed through erosion to permit of exact measurement at the base, which is likely, however, to have been, as usual, a square of approximately twenty feet. Raised on a small clay terrace, about twelve feet high, it made a conspicuous landmark on the dismal shôr-covered flat. Apart from potsherds of the hard dark-grey kind described above, no ancient remains of any sort could be found. Neither of the wall nor of other ancient buildings could traces be expected to survive on such ground, where wind-erosion above, and salt moisture below, the soil had full scope for destruction. Though the view from the top of the terrace was quite open, no other ruin could be sighted to the west, except the tower T. xxiv already examined on my way north of Shih-pan-tung. It just showed its top above a maze of erosion terraces. I was unable to spare time to search the ground westwards for remains of the Limes. But I am inclined to believe that its line may well have run in the direction of Shih-pan-tung and then, after crossing the Tang Ho delta, have joined on to the section of the wall which I traced in 1914 for some distance to the south-east of the Khara-nbr.

SECTION V.—SURVEY OF LIMES LINE TOWARDS AN-HSI, T. xxxi–xxxv

Before turning to the remains of the wall explored north-eastwards, a few remarks on the general topography of the ground along this section of the Limes may conveniently find a place here. Looking from T. xxx to the south and east, I could see a belt of absolutely bare salt-encrusted soil extending far away. Such scattered old tamarisk-cones as rose above it to heights of eight to ten fcet had long ago been completely cleared of their dead wood. I crossed this belt on my return march to Tun-huang, and found, as I expected, that its abundant salt-crust was probably the result of the overflow, or ` spill ', from the eastern canals of the oasis which is allowed to empty itself over this area. Compared with the ground marked by clay ridges, and from T. xxvi eastwards by gravel ' Sai ', which the line of the Limes follows, this shôr-covered belt seemed to form a shallow but distinct depression. This observation has been confirmed by the experience of the ground further east, gained in April, 1914, when, starting from Ko-ta-ching (Map No. 81. A. 4),' I set out to strike the. Limes to the north-north-east and on my way to it had to cross a wide depression of salt marsh, which at that season proved almost impassable.

I have thus been led to conclude that there extends from east to west a long stretch of low ground, water-logged for a great part of the year and salt-covered bog for the rest, which occupies a large portion of the area shown in Map No. 81. A—C. 3. It runs parallel to the relatively narrow belt of higher ground over which the Limes wall was carried between Tun-huang and An-hsi, and lies to the south of it. This raised belt stretches itself parallel to the Su-lo Ho bed with its riverine marshes and divides it from the southern depression. To the west of T. xxvt it has, as we have seen, a surface of alluvial clay which retains steppe vegetation but is, all the same, undergoing wind-erosion, as shown by the low clay ridges on which the Limes stations were invariably built here. East of T. xxvi the surface changes to that of a low gravel plateau, flanked on the south by a zone of drift-sand, which again forms the edge of the marshy depression already mentioned. With the geographical explanation of the latter we are not concerned here. But I may mention in passing that its marshes appear to be fed on the east and west by the ` spillage' of the canals of An-hsi and Tun-huang respectively, and in the centre by the floods which the torrent-beds crossed on the high road between those oases occasionally carry down from the outermost ranges of the Nan-shan.'a

' The Lo-to-ching of the map is a misreading.   la See also below, chap. xxv. sec. i.

4 H 2

Western-Most watchtower, T. xxx.

Topography of N.E. Limes section.

Marshy depression south of Limes line.