Page:Quarterly Journal of the Geological Society of London, vol. 28.djvu/498

404 'High Tartary, Yarkand, and Kashgar,' p. 97, thus alludes to some he noticed in the Karakash valley:—

"Nov. 2 [1868]: Marched down the Karakash stream, which now flows freely between ice-borders. It is fed by numerous warm springs; hence its freedom from ice. But these springs give the whole water a brackish taste.

"A couple of miles from last night's camp we crossed a little plain dotted over with small craters, four or five yards across and two or three feet deep in the middle. The bottom of these craters is occupied by a deposit of common salt or saltpetre. The servants took a supply for common use."

Again, at p. 98:—

"Nov. 5: A succession of fine meadow-plains full of salt-craters, larger than the former ones (some six or seven yards across). Some were full of concentrated brine (unfrozen in most), which on evapo- rating will give the usual salt deposit, I suppose. In this valley, wherever there is grass there is also saline efflorescence in the soil. I fancy both depend on the presence of moisture, and hence occur together."

Mr. Shaw makes no mention of the pits near Tarl Dat.

My theory of the formation of these pits is as follows. I suppose that under the surface there is a layer of sand, and under the sand a stratum of clay; and that the water which sinks into the ground at the head of the valley flows in sand or gravel under the latter. In the Karakash valley, quicksands and quagmires are very common; and whilst exploring the pits there, my horse repeatedly sank up to the girths of the saddle. In the upper part of its course the Karakash river sometimes disappears for miles and flows under the surface of the ground. Where the pits are formed I assume the existence of a layer of clay which keeps the water down until it issues in a series of springs at Tarl Dat, where the ground slopes more rapidly. I suppose that the water, flowing in very varying quantity at different times, gradually eats away the clay in certain places, and allows the sand to escape, and circular patches of the surface subside and form the pits. Depressions formed in this way are very common in the Punjab, particularly about Umballa and between Jhelum and Rawil Pindi. Sometimes at Umballa many square yards of ground thus subside and leave an enormous hole twenty or more feet in depth, with vertical sides. This process has gone on to such an extent in some localities between Jhelum and Pindi that more than half the surface-area has been lowered thirty or forty feet, and the whole country has been cut up into ravines with nearly perpendicular sides. I account for the mud-discharges at Tarl Dat by supposing that after a fall of rain or snow the air contained in the water-bearing stratum would get churned up with the water and mud, and be ejected as frothy mud at Tarl Dat. I believe the brine- and salt-pits in the Karakash valley are formed in the same way. This river rises and falls several feet every day; at some seasons it is almost dry, at others it overflows its banks. It is fed entirely by