FROM KURUK-ASTI TO THE CONFLUENCE OF THE KODAJ-DARJA. 55
to say, it was a quite inconsiderable and imperceptible quantity. The reason there was a current at all in its lower course, seeing that all accession of water had been so long cut off from above, was due to the fact that the deeper parts of its bed, as well as the occasional marginal lagoons which depend upon it, empty themselves only gradually. At the angle of the confluence there was a miniature pool. To judge from the shape of the Kara-jilgha-darja, and the sedimentary deposits in its mouth, it must, at the season of high flood, be a stream of some magnitude. On the other hand, relatively to the existing and comparatively new bed of the Jarkentdarja, it is itself growing year by year smaller. It was described as being narrow
A shepherds' track leads across the river from the confluence to Avat; while another traverses the forest at a distance of two or three kilometers from the left bank. On the edge of the sandy desert, about II/2 kilometers east of the river, stands the little masar of Ghesi-gumbes. Thence the stream flows for the most part to the north. The shepherds were of opinion, that at the next flood season the river would cut off the first big loop by forcing its way through the narrow isthmus at its base. The banks in the loop were as much as 4 m. high. Upon reaching the forest region of Tongus-burun we found ourselves opposite to Dungsarat, Ottus-kemi, and Jaka-kuduk — all places lying to the north-west. Then came a loop abandoned two or three years ago, but still containing a deep, narrow pool. Both extremities of the loop had been swept so thoroughly by the last full flood that we were able to see to its head. A little bit lower down we passed the beginning of an old bed deserted about 3o years ago. It is now completely overgrown, though its left bank, 3 m. high, still remains unimpaired. In the locality of Tapsi, again, the river left behind it, two or three years ago, an abandoned loop, which is now called Tapsi-köl. There we found shepherds living. On the right bank there is a shepherds' path leading to the saint's tomb of Masar Chodscham; and just below that yet another Kona-Jarkent-darja struck off from the channel we were navigating. The place where this begins is called Kum-ägis (also Ägis-kum), or the »High Sand», from some sand-dunes which have been formed in the bed of the Kona-darja itself. At Jar-kotan we again passed a loop on the left, which, strange to say, was a double loop. Its inner spiral was three years old; its outer one was only abandoned a year ago.
During the course of the day the river again altered its character to some extent. Mature forest was practically non-existent; there was nothing but young wood or bushes, a circumstance due in part to the changeableness of the river in that region. Steppe predominated; sandy ground, dotted with scrub, was corn-mon; and often the banks were marked by sand-dunes, upon which the erosive force of the current was playing with swift and disastrous effect. The large number of abandoned loops shows that the river hereabouts frequently changes its direction, though generally only for a short distance at a time. The chief cause of the river's variability is the small coherency and loose character of the soil, which owing to the absence of forest is not sufficiently bound together by strong, interlacing roots. Time after time we had unmistakable evidence of the quite recent desertion of this or the other loop. Time after time, too, we perceived the poplars standing
IIIF Color HighRes
