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

Ruined towers by Khara•nbr.

576   THROUGH THE LOP DESERT TO TUN-HUANG   [Chap. XIV

and scrub, spreading also into a succession of big bays on the south. At two points to the north of, and not far from, the track, I passed remains of small watch-stations, T. xxi, T. xxit, occupying the tops of isolated Mesas. Near the second of these we came upon a couple of Muhammadan Chinese or Tungans grazing cattle and horses, the first human beings met with since the start from Abdal.

Beyond this, the route still leading due east crossed a series of long and narrow gravel-strewn ridges which from the plateau on our right jutted out northward like the fingers of a hand. The depressions between them contained spring-fed marshes, and were connected with a wide salt-covered basin on the north. But this was too far off for close survey, and it was not till later that it could be recognized as containing the lake usually designated on European maps by its Mongol name of Khara-nôr, ` the Black Lake ', and erroneously believed to be the termination of the Su-lo Ho until our surveys proved that it was not. It was the presence of this large sheet of water,4 forming a safe line of defence by itself, which explained why no towers could be sighted here to the north.

Then the narrow continuous ridges gave way to a wide bay, bare of vegetation in its northern portion and covered with long rows of those characteristic clay terraces that were familiar to me in the vicinity of lake basins dried up or undergoing desiccation (Fig. 146). The terraces here had their long side invariably stretching from south to north. It was easy to recognize that they represented the remnants of earlier continuous ridges, such as the route had just crossed at right angles, which the erosive force of the violent east winds prevailing in this region, and of the coarse sand driven before them, had slowly sawn through and cut up. The origin of the ridges themselves could be accounted for with equal ease. They owed their existence manifestly to depressions scooped out between them by the drainage which once, during moister periods, came down the gravel glacis from the foot of the mountains on the south and cut deep into the clay sediments of a far more ancient lacustrine basin.4a I have thought it useful to record here this quasi-geological observation ; for the surface features thus produced have largely determined the line chosen for the ancient Limes in this region.

At last we emerged from between these terraces to an open stretch of flat ground extending northward, and there first came in sight of the Khara-nor, a large sheet of dark blue water at a distance of some four miles. The wide salt-encrusted edges showed that its level at a later season would rise higher and the area covered by its water considerably expand. A number of isolated clay terraces, regular Mesas, rose scattered over the flat shore to the north-east, where it was covered with abundant vegetation. They were manifestly the last survivals from terrace clusters and ridges which the slow but relentless force of erosion had long ago ground clown and carried off. On two of them, not far from what looked like the eastern end of the lake, I sighted ruined watchtowers.6 A third, T. xxiii, perched at the end of a long ridge projecting into the flat basin from the south, rose immediately above the route just where it took a final turn to the south-east. The wide depression containing the bed and marsh basins of the Su-lo Ho was now left behind, and moving over bare, gently rising ground, evidently part of the alluvial fan once formed by the Tang Ho, or

Lake

Khara-nör.

Eroded clay terraces.

' Our surveys made at the close of March, 1914, have proved that the area covered by the lake at that time extended some four miles further east than shown on Map No. 78. B. 3. On the other hand, a series of additional towers then discovered along this eastern extension of the lake showed that its bed on this side could not have formed a permanent barrier in ancient times. Probably it lies dry even nowadays for a considerable portion of the year.

" Cf. below, p. 589.

These towers, T. xxiii. b, c, have not been correctly marked on Dlap No. 78. B. 3. They could not be explored until 1914, and the survey then made proved that T. xxxIi. c is situated about four miles to the north-west of C. 157 and T. xxxI1. b about a mile further. It is at the latter tower that the line of wall abandons its general east—west direction to turn towards the south-east and thus approach the Tun-huang oasis.