National Institute of Informatics - Digital Silk Road Project
Digital Archive of Toyo Bunko Rare Books

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

terminal basin. In some of these I clearly observed the marks of quite recent inundation, both in 1907 and in 1914. At the same time the soil, still wet over extensive areas, showed so little salt efflorescence as to make it obvious that the water which reached them had been relatively fresh and been drained away by percolation before it could become completely evaporated.'oa

The ground over which the route leads, where it skirts or traverses this ancient terminal basin, is very deceptive. Neither in the soft, sandy soil of the depressions, nor among the dunes of the broad sand ridges which divide them, can any traces of the track survive from one season of caravan traffic to another. The difficulty which travellers experience about discovering and following the right track is greatly increased by the hundreds of high clay terraces which, scattered in clusters or rows, rise like islands or towers over great portions of the wide basin (Fig. 148),Iob It would be easy for wayfarers, if unguided, to lose the right bearing where these mazes of fantastic clay terraces are encountered, and to stray away into the hopelessly barren desert north or south of the basin, which, with its wilderness of Mesas and sand-dunes, acts like a great curtain.

That first march from Bash-toghrak had brought us close to the eastern end of the strange basin just described. Our guide had failed to strike the brackish well which hereabout serves as a halting-place. But on the following morning, March 7, we had only gone about three miles when the track, now quite clear in coarse sand and gravel, brought me to a deeply-cut and well-defined flood-bed descending from the east. It was easy to realize that we had reached here the debouchure of an old terminal branch of the Su-lo Ho, and on following the track up the cliffs of its steep right bank I was struck by the sudden and complete change of the ground. Eastward, there spread out a flat gravel-covered expanse, broken only here and there by shallow depressions. To the north, the view was bounded by the low Kuruk-tagh hills in the distance. The atmosphere was not clear enough then to permit the big snowy range about Anambar-ula to be sighted as it could be on later occasions. Yet there was no doubt possible here that the route had entered the great open valley trough of the Su-lo Ho and approached the marshy expanse forming its present terminal basin. The edge of this lay within only six miles or so of the point on the gravel plateau to which the route had now brought me.

It is true that the ground previously traversed was also an old terminal basin, and that, through it and the valley of Bash-toghrak beyond, the waters of the Su-lo Ho may at an earlier, but geologically, perhaps, not very distant, period have made their way down to the ancient Lop sea, since dried up. But acceptance of this theory can in no way weaken the impression that the route at this point enters its third and last section. It brings us in five convenient stages, making up a total marching distance of about 97 miles, to the centre of the large oasis of Tun-huang, the westernmost outpost of China towards Eastern Turkestan and the base of its earliest Central-Asian operations. Along the whole length of this section the caravan track, following the line of the ancient Chinese route, leads close to the bed of the Su-lo Ho or else past a string of freshwater lagoons fed by the Tun-huang drainage. With plentiful good water and abundance of grazing at convenient intervals, movement along this line is easy at all seasons. Though the ground still continues incapable of cultivation, it is fit for grazing over considerable areas of the riverine belt, and the traveller soon begins to feel that the true desert has been left behind.

This marked geographical change finds its striking reflection in the fact that the westernmost

Deceptive ground of terminal basin.

Terminal bed of Su-lo Ho.

Last section of route to Tun-huang.

'oa The fact that flood water from the Su-lo Ho could find its way even now towards this earlier terminal basin, whether by percolation or otherwise, is clearly proved by the dry riverbeds traced north of the present terminal course of the Su-lo Ho and forming part of its delta (Map No. 74. A, B. 3).

iob There is no doubt that they: are ` witnesses' due to

erosion, and the fact that they are found also round Lake Khara-nbr further east, and near the actual terminal basin of the Su-lo Ho on the south, bears strong testimony to the lacustrine character of the depressions crossed by the route east of Besh-toghrak. Cf. below, pp. 575 sq.,589, 642 sq.,

717.