Sec. iii] ACROSS THE SALT-ENCRUSTED LOP SEA-BED 299
much more open and allowed us once again to take the direction N. 55° E. To the northward, Mesas
only small Yârdang-like ridges could be seen, while to the north-east rows of reddish tower-like
Mesas showed in striking contrast above them. They seemed far away and probably only their
tops were visible ; but the distance could not be determined as farther on we lost sight of them.
It has occurred to me since that these far-off Mesas were probably identical with the belt of eroded
terraces on somewhat higher ground which Lal Singh's survey shows to the west of his camp
C. 89 (Map. No. 32. B. 2).' As we proceeded the Yârdangs sank away rapidly, and soon only light
swellings of the ground, in which abundant gypsum flakes were mingled with decomposed clay,
remained to suggest the former existence of such formations.
After covering about ten and a half miles from the Mesa where the coins were found, we reached Progress
perfectly level ground of the same kind, slightly salt-encrusted and showing a thin layer of coarse over easier
drift-sand. From a small isolated hillock rising two miles farther on above this plain, I sighted to
the east a continuous line of the familiar white salt-coated ridges. It could be seen stretching far
away to the north-east, and is, I think, probably connected with the Yârdang belt which Lai
Singh's plane-table records about five miles to the south-east of his camp C. 89. The view obtained
here decided me to abandon our north-easterly course and to steer due eastwards for the point
where the fence of salt-encrusted ridges seemed thinnest. For another mile or so we still had the
benefit of an easy surface of decomposed clay. Then followed what seemed a shallow depression
covered with rows of low salt-encrusted hillocks, with flat beds of hard Mar between them. Fortu-
nately the winds had driven a thin coating of coarse dark gravel across these, thereby making the
ground less trying.
The direction of the rows of salt-covered Yârdangs was here, too, from N.20° E. to S.20° W. Edge of
Their lines after about a mile thinned out, and when we arrived in the evening near the eastern drieup
edge of this Yârdang belt I had at last the perfectly open view before me of a vast salt-encrusted reached.
plain. I t was the true bed of the ancient sea, which I knew that the old Chinese route to Lou-lan
must have crossed, and which we had to face on the morrow. With its trials before us, I was glad
to find a spot near the edge of the Yârdang belt where the salty ground, hard indeed but fairly
smooth, allowed men and camels to lie down without discomfort, after the day's total march of
close on twenty-one miles. But it was only with great difficulty that the iron tent-pegs could be
driven into the surface of salt.
Much of the night had to be spent by the men in ` re-soling ' those camels which had again View of
become footsore on the hard shôr between the Yârdangs. After that dismal ground, it was almost sea-like
with a sense of relief that in the light of the early morning of March ist we beheld from the last salt basin.
Yârdang ridge the bed of the ancient sea stretching away to the east and south-east as a perfectly
level plain. Only a few isolated knolls rose island-like slightly above it far away in the distance.
Through powerful glasses it was just possible to make sure that these were the tops of hills, mani-
festly those with which the low desert ranges of the westernmost Pei-shan abutted on the huge
dried-up basin. It was encouraging to point to them as affording assurance that ` land ' was in
sight, however long the difficult crossing before us might prove. To the north and north-east
a continuous fringe of low hills showed above the horizon, very distant also, yet confirming my
belief that the ancient route—which from evidence discussed below I knew to have crossed the
sea-bed--could not be sought for much farther to the north than the point where we stood. On
ground like this it seemed useless to search for direct evidence of the line actually followed by the
route, even if considerations of safety had allowed time for the purpose.
7 The Yârdang symbol was used by Làl Singh without been retained in the map in the absence of further evi-
distinction for all formations due to wind-erosion, and has dente.