48 THE REGIONS SURVEYED [Chap. -H
of the Âltin-tàgh glacis. 1° It is apparently in this bay that the depression reaches its deepest point, and here it may be supposed in earlier geological times to have received also the drainage from the terminal basin of the Su-lo-ho which adjoins eastward. Down the southern side of this valley and beyond along the clearly marked southern shore-line of the ancient sea, leads the difficult desert track from Tun-huang to Lop, graphically described by Marco Polo and still used by. rare caravans during the few winter months when it is practicable. 11
Before leaving the Târim basin for regions further east a brief account may couveniently be.given here of the small but geographically very interesting Turfiln depression. • basin of Turfàn north of the Lop depression, to which a good deal'of our survey .work was devoted in 1914-15. Quite detached from the Tàrim basin it shares so many of its physical characteristics that it appears like a small scale reproduction of it. As Sheet No. 28 shows, it is enclosed in the north by à rugged snowy portion of the T'ien-shan, rising to peaks over 14,000 feet in height, by an outlying range of the same in the west, and by utterly barren hills and plateaus of the Kuruk-tagh in the south and east. Within these limits it contains a succession of well-defined zones exactly corresponding to the gravel glacis, the belts of cultivation and desert vegetation, and the dune-covered areas of the Tàrim basin.
The terminal sea-bed of the latter has also its pendant in the narrow salt lake, for the
most part dry, 12 stretching along the south-eastern edge of the basin. Into its lowest portion at the time of the summer floods gathers whatever drainage from the mountains escapes evaporation. To the
east of it we have a miniature edition as it were of the Taklamakàn in the plateau-like area covered by high ridges of dunes known as Kitt-tàg4, the `Sand Hills'. 13 Its position seems to be determined by the direction of the prevailing winds which as a result of 'aspiration' sweep down from the cooler region in the north-west when the great heat of the spring and summer causes the air to rise from the lower parts of the basin. A very remarkable feature of the Turfàn basin is the depth of its terminal depression. Along the lake-bed above mentioned it descends to a level which according to our mercurial barometer observations lies in places close on 1,000 feet below the sea, while most of the principal oases lie also about or below sea-level. 14
To the very high summer temperatures resulting from this low position may be attributed, at least partly, the peculiar conditions affecting the water
Irrigation in Turfan supply. of the basin and in consequence the cultivation in its oases. 15
oases. pp y q
The streams which carry down the melting snows of the T'ien-shan in the spring and summer lose most of their water on the descent over the bare glacis of gravel. A portion of the water absorbed in the ground, it is true, comes to light again, like the kara-sa of the Khotan region, in marshy springs at the northern foot of the low and utterly arid hill range stretching across the middle of the basin from east to west and dividing its cultivable area into two unequal belts. 16
But this water supply, too, would permit of irrigation only over very limited ground were it not at the present time supplemented on a big scale by means
Kilrèz irrigation, of sub-terraneous channels or `Ktirêzes' which catch the subsoil water
beneath the gravel slopes and carry it, protected from evaporation, over considerable distances to ground otherwise hopelessly sterile but under irrigation extremely fertile. The use of Klutzes is unknown elsewhere throughout Chinese Turkistàn, and in the rl'urfân district, too, it can apparently not be traced further back than the 18th century. Yet
Character and depth of Pullin depression.
1Ô See Sheets Nos. 32.0, D. 4 ; 35. A. 4.
" See Serindia, ii. pp. 549 sq., 560 sqq.
12 See Sheet No. 28. 0, D. 3.
'3 See Sheets Nos. 28. D. 3; 81. A. 3.
'' See Sheet No. 28. C, D. 8.
16 These and other physiographical aspects of the
Turfin oases have been lucidly discussed in Professor
7y. Huntington's Pulse of Asia, pp. 806 sqq.
The physical features of the Turfiin basin as â whole will be fully treated in the paper which I hope to prepare for publication along with a detailed map based on the one-inch survey of the central area of the territory.
16 See Sheet No. 28: C. 2, 3, D. 8, for springs near Murtuk, Singim, Su-bishi, Lanijin; Sheet 31. A. 3
for those in the bed above Pichan. ~'`