Sec. iii] AN ANCIENT TERMINAL BASIN 333
at work here could be judged from the fact that the width of the Wadi a little before the caravan track left it for the plateau adjoining on the south was still about a mile, and that to reach the edge of this plateau an ascent of over a hundred feet had to be made between the steep cliffs of a small side gully.
The configuration of the ground here corresponds exactly with that observed along the actual course of the Su-lo-ho both above and below Toghrak-bulak, where the river has cut its way between gravel-covered plateaus of the same description. Our surveys of 1907, as supplemented by Läl Singh's work on his return from the present terminal basin of the Su-to-ho, have shown that this Wadi debouching into the south-eastern extremity of the old lacustrine basin lies exactly in the continuation of two old beds, no longer filled by the Su-lo-ho but undoubtedly forming part of its earlier delta. The distance between the point where Ram Singh crossed the more southerly of the two beds and the place where the caravan track leaves the Wadi is only five miles, and the connexion between the two may therefore be considered as certain."
SECTION IV.—THE DELTA OF THE SU-LO-HO
Before I proceed to sum up the conclusion which may be drawn from our survey of the dried-up basin just described, as regards its relation to the lowermost course of the Su-lo-ho on the one hand and the Besh-toghrak valley on the other, it will be convenient to indicate briefly the facts which the surveys of 1914 have established as to the actual termination of the river. As recorded in Map No. 35. B, c. 4, they show that the Su-lo-ho bed which passes Toghrak-bulak, and alone at present carries a regular supply of water during spring and summer, finds its end in a lake holding, at the time of the Surveyor's visit, a sheet of salty water about six miles in length and two miles across at its widest.
Considering that the volume of water carried by the Su-lo-ho bed at Toghrak-bulak, as measured by me on March 17, 1914, amounted only to about 18o cubic feet per second, as against 1,800 cubic feet measured on May 2, 1907,1 it is obvious that the area covered by the lake during the time of the spring and summer floods must be much greater. Probably most of the salt-encrusted ground shown to the east and south of it is then under water. To the west and north the seasonal expansion of the lake is limited by a great curving ridge of dunes which rises above it. The relative height of this was estimated by Lai Singh at about three hundred feet. This ridge resembles in character the high ` Dawàns' of sands which are found along the terminal courses of the rivers that lose themselves in the Taklamakâ n.2 It obviously owes its origin mainly to the silt which is brought down by the Su-lo-ho when in flood and after deposition is heaped up by the prevailing easterly winds.
Before reaching this terminal lake the actual course of the Su-to-ho, for a distance of over twenty-four miles above and below Toghrak-bulak, occupies a narrow trench-like bed sunk deep between the gravel plateaus to the north of the western extremity of the Limes line. About eight miles below Toghrak-bulak the plateau on the right bank of the river gives way to a scrub-covered sandy depression ; that flanking the left bank continues some nine miles farther down. For the
Old bed of Su-lo-ho delta.
Survey of Su-lo-ho termination.
Terminal lake reached by Su-lo-ho.
15 Notwithstanding this, I should not have forgone the chance of following up the Wadi farther in the direction of the old bed, had I not felt obliged to retrace the route to Toghrak-bulak in person for the sake of solving an archaeological question. It was whether the line of ancient watchtowers thrown out along the route beyond the western end of the Limes had a continuation or not beyond T. ti. The
reconnaissance made on May 2, 1907, in this direction could not be held conclusive owing to dust-haze, and consequent ` low visibility ' ; cf. Serindia, ii. p. 640. However, the renewed search on March 16, 1914, was attended by the same negative result.
1 See Serindia, ii. p. 639.
z Cf. ibid., i. pp. 241, 451 sq. ; iii. pp. 1239 sq.