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. ii]   THE VALLEY OF BFSH-TOGHRAK   325

signs of moisture in the soil and had a well dug. At only two and a half feet from the surface water was struck in abundance and, according to Afraz-gul's statement, ` as fresh as that of any river '. This completely tallies with what I observed at my own well, more than eight miles farther up, and makes me inclined to believe that by sinking wells at suitably selected points drinkable water could even now be found probably on the northern side of the valley as far down as our Camp cvii.

For about a mile and a half before reaching the point where they came upon the embanked line above discussed suggestive of an ancient canal, the Surveyors' route led along a well-marked ancient track, and this, as shown on the map (No. 35. A. 4), they were able to follow, with breaks here and there, for another six miles or so. Considering that near the vicinity of the above point three small cairns were found on little elevations close above the track, and that another cairn was noticed about a mile and a half farther to the south-west, similarly near this track, it is difficult to resist the conclusion that it owed its origin to, or was used by, men. How long ago, it is impossible to guess. No footprints of wild camels were noticed on or along it, though numerous in the scrubby belt farther south which affords good grazing.

Of the observations recorded by the Surveyors on their farther progress towards Bésh-toghrak the following may deserve brief notice. To the north-east of Camp 97 their route crossed two broad Nullahs descending from the hill range and cutting through the gravel-covered clay plateau, much after the fashion I observed to the east of the present terminal basin of the Su-to-ho.10 The projecting plateau tongues thus produced showed a relative height of 120 to 15o feet. Beyond the mouth of the second Nullah a stretch of ground was passed which for about two miles was covered with Yârdang-like clay terraces from six to twenty-five feet in height. There can be no doubt that these terraces, all stretching from NNE. to SSW., just like the belt of Mesas I had myself passed farther to the south-east, were produced by wind-erosion. They obviously represent the last remnants of former plateau tongues which first had been carved out by drainage from the hill chain and had been subsequently cut through by the erosive force of the wind. Their origin was thus exactly the same as that of the clay ridges and Mesas observed in 1907 at a number of places south of the terminal Su-to-ho.11

Beyond this belt of Yardangs and an outlying isolated Mesa a large patch of boggy ground was met with, resembling that I had passed near my Camp cx. Its occurrence here affords further proof of the abundance of subsoil water towards the head of the Bésh-toghrak valley. At Camp 98, which stood a couple of miles to the south-west of the cliffs facing ` K`ai-mên-kuan', water was found at two feet from the surface. It proved distinctly brackish, though drinkable, a circumstance fully accounted for by the fact that the camp stood near the northern edge of an extensive area where hard shar mingled with patches of salt-impregnated soil supporting reed-beds. It was after crossing this ground, no doubt once a bog, for fully four miles that the Surveyors reached the sandy belt with abundant vegetation which makes Bésh-toghrak so convenient a place for halts on the desert journey to Tun-huang.

The account here given of the physical aspects of the ground through which the Lou-lan route must have passed after emerging from the Su-lo-ho basin may be appropriately completed by some observations on the results of Muhammad Ya.qùb's levelling operations. The object with which I had arranged these was to ascertain whether the assumption suggested to me by the observations of my previous journey to Tun-huang in 1907, regarding an earlier connexion between the terminal basin of the Su-lo-ho and the ancient Lop sea-bed,12 was supported by the configura-

10 See Map No. 35. c. 4 ; Serindia, ii. p. 642 ; Desert   n See Map No. 38. A. 4 ; Serindia, ii. pp. 576, 589.

Cathay, ii. pp. 139 sq.   12 Cf. Serindia, ii. pp. 551 sq. ; Desert Cathay, i. pp. 533-7.

A well-marked old track.

Belt of terraces carved by wind-erosion.

Subsoil water at head of valley.

Levelling operations.