between them is that, from the superior elevation of the ferry-boat I was able to note various sedimentary islands and channels running through the alluvium which had escaped my observation when I sat in the lower-lying canoe. Thus the map gained from being laid down from the platform of the ferry-boat, for this big craft, which was entirely dependent upon the current, was carried in and out to the far extremity of every bend, whereas the canoe-men generally took short cuts by paddling across the shallower places.
On the left we passed first a canal, leading to Kara-tschatsch and Katschkenedschajiri, two uninhabited homesteads; below them there are no inhabitants except shepherds. On the 8th of May this canal contained a little water, but by the 19th it was all gone and the canal was dry. Immediately opposite, on the other side of the river, there is a tiny lake, Motel-uktusi, scooped out of the level ground, with reeds all round it, but high sand at its south-west angle only. Beside it stood formerly the village of Jangi-köl, and the site is still called Jangi-köldake-kona-uj. Farther down on the left is Jamandake-uj. At the eastern foot of the spur of sand, which touches the eastern shore of Gölme-käti, stands the deserted homestead or village of Nias Supa Bajne-uji. On the left of the river are two other homesteads, Dovlet Bajini-uji and Talei Kullunu-kötschetu, with a canal leading from the river. Two other canals, still on the same side of the Tarim, connect with the little lake of Bajini-kölu, and two homesteads that stand beside it. Immediately below this comes the lake of Kaghune-kölu. On the north-west shore of the Karaunelik-köl are several huts, which have been unoccupied for fifteen years.
On the loth 1\'Iay we did not get any farther than the entrance to the lake of Ullugh-köl. On the left bank stands the mälaallä, or »village», of Ullugh-köl, also known as Ilias Murabne-mähallä, with 20 to 3o families. One half of this population, who are estimated to number 30o persons in all, dwell in Ak-tarma; the other half in this village, which we saw a few hundred meters back from the river. At a place called Ak-jolu-basighi the flood-bank of the river is a couple of meters high. On the right we noted a canal which goes to Tajiri-kakmasi and Ullugh-köl; both times when I passed it was empty. On the other hand the Tägirmän-arighi, on the opposite side of the river, carried water on the 8th 1VIay; this watercourse proceeds first to the mill of Ullugh-su, and then empties into the old river-bed of Kudschek. This last is now perfectly dry, although twenty-five years ago (1879) the Tarim poured its flood through it and then at Tunghan-tüschken joined the Ara-tarim (see below). Thus from this point the existing Tarim is a quite young stream; hence too we can understand the absence of poplar woods along its banks and the quite recent growth of such trees as do appear at intervals. This circumstance is not of course incompatible with the former existence of lakes amongst the sand; only then they must have been supplied by longer canals. And that there really were indeed lakes of that description is evidenced by the abandonment, it is said thirty years ago, of the village of Kona Ullugh-köldake-uj, situated beside the upper canal (basclai) of Ullugh-köl. Here we noticed, in the face of the perpendicular scarped bank, two distinct water-marks, at 84 and 23 cm. respectively above the then existing level. The upper probably indicates the maximum level of the spring flood; the lower a level at which the river remains some time
IIIF Color HighRes
