Sec. iii] AN ANCIENT TERMINAL BASIN 329
in the area liable to inundation from the Tarim to the north-east of Abda1,6 were quite clear of reeds. Here the foot sank deep, and the sand underneath was saturated with water, apparently fresh. Slightly higher portions of the ground bore thin growth of young reeds, while the sand showed a crusty surface as the result of moist soil contracted under a hot sun. Here, too, the sand immediately below the surface was moist, and throughout there was a striking absence of shôr. For this it seemed impossible to account except on the assumption that the water which periodically spreads over this basin is carried off again by subterranean drainage before it becomes evaporated. Our line of march allowed the western limit of this basin to be seen quite clearly. It was edged everywhere by the belt of dunes already referred to. The fact that these seemed lowest towards the north-west near the foot of the gravel Sai suggested the possibility of an outflow towards the area of soft shôr that Muhammad Yaqùb had noticed north of his line of levelling from Béshtoghrak.
After having thus covered a distance of five and a half miles across the open flat of the basin we arrived at a big Mesa, the northern outlier of a line of clay terraces rising within the basin on our right (Fig. 187). It rose steeply to a height of about 120 feet and showed seven horizontally stratified layers of reddish clay separated by thinner strata of yellowish sand. From its top, an extensive view was obtained over the dry lake basin to the south and west, as well as over the absolutely bare Sai of dark gravel sloping up gently northward. The edge of the basin in the latter direction was well defined by a narrow belt of small sandy hillocks, only four to six feet high, bearing thorny scrub and fringing the shore-line.
To the east the eye was met everywhere by serried rows of those high wind-eroded clay terraces which had already presented themselves as a most striking feature of this area on my passage in 1907 farther south.' Their lines, as was subsequently noted on many occasions, stretched everywhere from north-north-east to south-south-west, showing the same regularity as the familiar Yardangs of the Lop desert, but a different bearing. That these big terraces rising to heights from eighty to a hundred and twenty feet or thereabout were, in fact, not different in character and origin from those Yardangs was testified by their invariably long and narrow shapes and the comparative rarity of the gaps separating those which belonged to the same row.
The successive rows themselves were separated by intervals which in this portion of the basin did not exceed a quarter of a mile or so. These rows of big terraces afforded clear evidence of the powerful action of wind-erosion on this ground during an earlier epoch. The total absence of Yardang ridges and trenches on the open ground of the basin and between the rows of terraces was all the more curious. I could attribute this contrast only to the protection afforded to the soil in the present epoch by periodical inundation, coupled with the nearness of moisture below the surface, and the consequent growth of some vegetation above it.
A well dug at the foot of the Mesa which had served for our plane-table fixing yielded perfectly fresh water at a depth of only three feet, conclusive evidence that a constant flow of moisture finds its way also to this northern edge of the basin. No doubt, this accounted likewise for the plentiful scrub on the low hillocks of sand which, as already mentioned, fringe the basin below the foot of the gravel glacis. We had these well within view while continuing our march for about two miles to north-by-east. After threading our way between the fantastic forms of Mesas, upon which wind-erosion, using the gravel close at hand as its weapon, was obviously able to exert its full strength, we finally emerged on the open slope of the Sai.8
e See above, p. 181 ; Map No. 30. c. i ; Serindia, i.
pp. 351 sqq•
7 Cf. Desert Cathay, i. p. 533.