by three families of in all 15 persons. The Katik-arik is in reality an old bed of the Tarim, and at the time of our visit was pretty full, although the current was so slow that it was hardly noticeable, and consequently the water was perfectly clear. It was evident that this bed had only quite recently been discarded by the river. The characteristic alluvial deposits and erosion banks were as distinctly marked and as sharp-cut as any part of the river in which its activity is most pronounced. Then we came to a point where the old bed of the Tarim wheels away to the north and is quite dry, while the Katik-arik, which has hitherto flowed along the old bed, turns off to the south in numerous winding loops. At its very first loop it divides into two arms, of which the one on the left goes to Toktaghono-mandscharlighi on the old Tarim, and then enters the Ojman-köl, whilst the one on the right continues on to the Begelik-köl. Where the latter contracts, its current was pretty lively, but in the open reaches there was scarce any current at all. In one of these the depth was 5.50 m.; the other soundings that we obtained were 3.50, 2.50, 3.95 4.I0, 3.30, and 4.o m. Thus the canal is tolerably deep throughout; in one place only did we find a very shallow ledge with a lively current. Just above a district called Usunkara-tschilan the Katik-arik (the Hard Canal) is joined by another branch-canal known as the Tschong-aghis, or the Big Mouth. It was at that time dry, except for a little stagnant water at its end. The expansion of the canal bearing the name of Gölmekadaghan is encircled by reed thickets, tall tamarisks, and poplars, growing on sandy soil. These, although young, are tolerably vigorous, but neither so dense nor so decayed as those alongside the old river-bed. Close to the right bank is a ridge of sand. Small bays or coves, surrounded by reeds and suitable for fishing, are called here also köuruk or köjuk. Another expansion is known as Kosch-arukdeboldschemasi. Then come the district of Istam Begi-kadasi and a second Gölmekadaghan. Immediately below this last there are yet other huts on the left side of the stream; these are said to have been inhabited by the people of Tikenlik before they migrated to their existing quarters. My guide, Kirghuj, had been one of them; hence his minute acquaintance with the geography of that region. Most of the people removed from this site 20 or 3o years ago, though some lingered on until 16 years back. Crossing yet one more expansion, and traversing a narrow aghis, we emerged upon the first basin of the lake, Aghis-köl. Beside this section of the canal and its last expansion there is forest of young poplars, luxuriant though dead. Most of the trees stand in the water, and more or less deeply; they are grey and withered, and nothing but the stems and the larger branches are left. Probably this forest, which is thin and of no account, flourished at a time when the supply of water was not so plentiful. Then from some cause or other the inflow into the lake was stopped, and the locality dried up completely and the poplars withered; but the water having returned, they are now drowned. Amongst the branches that remain were several nests of the karaune, a species of very large black waterfowl.
The lake, as we paddled out upon it, lay as still and bright as a mirror; there was not a breath to crumple its surface, and the outlines of the dunes were etched upon it with remarkable vividness. Then, threading a contraction of the lake, Beglarningboltasi, we issued upon the middle basin known as Ghol-köl, which we crossed to-
IIIF Color HighRes
