the level of the lake itself; but it was firm enough to bear, so that our Tibetans were able to make a short cut that way with their yaks. The water in these lagoons, as well as in the Oman-tso itself, was perfectly fresh, and all were then frozen, except for some open spaces which showed near the west side of the principal lake; probably there were springs there. On the southern shore of the Omantso there is a very distinct terrace, apparently at the same level as the terrace which we saw near the obo. It is very strange that there should be fresh water in this little lake at a time when it is shrinking; one would naturally expect to find it salt. It may be that it lies a shade higher than the Dadap-tso and communicates with it by an underground channel. The water of the latter lake is, my Tibetans de-
Glared, salt, but all the same it can be drunk in case of necessity. The Oman-tso means the Milk Lake, the meaning being that its water is as sweet as milk. In the lower part of the narrow latitudinal valley that runs towards the west-southwest, there exist several springs just above the lake, and again during the rainy season its water is freshened by big floods from the southern mountains. Moreover, as the isthmus between the Oman-tso and the Dadap-tso is so extremely low, it only needs a very inconsiderable rise in the former to set up a broad, shallow current out of the first-named into the latter. True, we did not observe direct traces of any such temporary overflow, but on the soft gypsum surface they would very soon be blotted out. If however my supposition is correct, the Omantso does not fill a self-contained basin, but discharges into at any rate the Dadaptso. This would explain the lake's immunity from saltness.
From the northern shore of the big pool we inclined towards the south-west, and continued for a space along the north-western shore of the Oman-tso, having
IIIF Color HighRes
