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

:Sec. ii.]   THE TAKLAMAKAN DESERT

..( Sheet No.. 5 ). In the east a zone of high and utterly barren sand-ridges spreads beyond the -Tarim, filling '.a big triangular space between the foot of the Kuruk-tagh and the belt of wind-eroded desert along the western shores of the ancient salt-encrusted Lop sea bed ( Sheets No. 25, 29, 30 ).

In the south, the border of the Taklamakan lies along the northern ends of the oases,

mostly small, which line at intervals the foot of the K'un-lun glacis from Karghalik to Niya (Sheets Nos. 6, 9,14,19). Further east this line finds its continuation in patches of sandy jungle intermittently watered

by rivers of small size and extending below that gravel glacis ( Sheets No. 19,23 ) as far as the small oasis of Charchan. From there onwards to the vicinity of Lop-nor the narrow belt of vegetation which accompanies the Charchan river right down to its junction with the Tàrim forms the well-marked border of the Taklamakan.

With the exception of the Khotan-daryà, not one of the numerous rivers descending

from the snowy K'un-lun succeeds in making its way through the Taklamakan. All the rest are lost in this sea of sand ' at a greater or lesser distance from the line occupied by the existing oases or by

areas of coarse desert vegetation. But within historical times, as proved by remains dating from the third century A. D. onwards, a number of these terminal river courses carried a greater volume of water and permitted ground to be cultivated lying considerably further north than the present ' terminal oases '. 1 My explorations of such ancient sites abandoned to the desert led to extensive surveys in this southern portion of the Taklamakan, and these were supplemented by others along the routes followed right across it from the Tarim to the Keriya river delta and along the bed of the Khotan-daryà. 2

The topographical record of these travels in the Taklamakan illustrates in a striking manner the uniformity prevailing in the character of this huge desert.

Physical features of   Whether the traveller enters it from the edge of cultivated around in

desert.   t!   b

oases, or from the jungle belts along river courses and in tracts where subsoil drainage from the mountains approaches the surface again at the foot of the gravel glacis, he first passes through a zone with desert vegetation, mostly in the shape of tamarisks, wild poplars and reeds, surviving amidst low drift-sand. A very peculiar and topographically interesting feature of this zone consists of ' tamarisk-cones', hillocks of conical form and 'often closely packed, which the slow accumulation of drift-sand around tamarisk growth has in the course of centuries built up to heights reaching fifty feet and more. 3 Further out in the Taklamakan only shrivelled and bleached trunks of trees or sand-cones with dead tamarisk growth emerge from the dunes, until finally these too disappear among high and utterly bare ridges of sand. 4

SECTION III.—THE OASES OF THE TARIM BASIN.

The areas left between the Taklamakan and the encircling mountain ranges to the north, west and south have geographically so much in common that they might well be treated as one region. If I prefer here to separate the western and northern margins of the

43

Southern border of
Taklamakân.

' Terminal oases.'

See, e. g., Sheet No. 14. C. 1 for the Dandânoilik site; 14. B, C. 2 for the ancient sites north of Chira—Domoko line; 18. B. 4 and 19. B.1 for the site beyond the Niya river end; 19. D. 1 for the Endere site. Cf. also above pp. 8 sq., 14 sq., 19 sq.

2 Cf. above pp. 19 sa., 27.

3 Regarding the representation in the maps of these tamarisk-cones' and of desert vegetation, living or dead, see below Chap. III. sec. ii.

• 4 Abont the formations shown by the sands of this innermost and largest desert area it will suffice

here to mention that while the shape of individual dunes conforms to the prevailing wind-direction, the big hill-like ridges (dawdns), into which they are heaped at intervals, according to my observations, seem generally to stretch parallel to those river-beds which lie nearest, whether still receiving water or dry since long ages. See for such high Dawiins parallel to river courses, e. g., Sheet 14. B. 2, C, 1, 2, D.1-3; 17.A. 4; 18. A. 1-3; 19. B-D. 1; 29.1;.4; 30.A. 1. Cf. also Sesindia, pp. 241, 451 sq., 1239.