the latter the poplars are both young and few in number. The bed of the former is deeply and energetically excavated, and its high steep banks keep the water within reasonable bounds, except when it overflows at the season of high-flood, and this is extremely seldom; whereas the Tarim is a broad and shallow stream, with shifting banks. In a word, the Kok-ala is older than the Tarim, and without doubt served as the channel for the whole of its flood at a time when the present main channel did not exist, and when the lofty masses of sand in the south-west perhaps reached as far as the right bank of the Kok-ala, in the same way as they now overhang the right bank of the Tarim. When one sees, as in this region, several more or less parallel rivers possessing serpentine beds excavated to irregular depths, it is impossible to dissociate from one's mind the idea that they really do represent successive stages of the main bed of the Tarim, or, in other words, that each of these in turn did for a certain period serve as the principal artery of the entire hydrographic system, and as the outflow for the entire mass of its water, or at all events for the greater part of it. During that period they excavated their beds so deeply that they still continue to carry water even though the main stream has flitted farther to the west. As in the case of the Ugen-darja, it is only the deep winding trench of the actual current that now survives; those parts of the general bed in which alluvium was deposited are now overgrown with vegetation, and held fast by its power of detaining drift-sand and dust. Very often this is evident at a glance.
Approximately south-west of Kurm-uj lies a series of lakes — Arpa-kumghundung, Talaschtik-mughu, Soghujtschagh-jatghan-kotan, Pavane-kotanu, Ak-köl, Bolto, Matijasi-togh, Matijasi-kotanu, Toj-bolghan, Sejt Asarni-ilegi, Janijas jatghan, Dungutuva, Ottok-köl, Ullugh-köl, Chodschem Värdi-köl, Üssupti jajiri, Jaman jajir, Örlesch, Scharkurun, Konkolluk, and Schama-tschapghan. These are traversed by the Kudschek arm, which at the same time supplies them with water. The actual series terminates with the lake of Scharkurun, from which the river issues in the form of a narrow ilek, while the two lakes last-named lie at the side of the series.
The natives of the district of Tikenlik told me, that the aghis, which is spanned by a bridge (vide pp. 159-16o) not far above the Jangi-köl, was once the main stream of the Tarim. Although I • have never visited the locality myself, and consequently am unable to answer for the correctness of this information, it nevertheless has the appearance of being correct, especially as I obtained incidental corroboration from various other sources. I will therefore record here the results of my inquiries. The river in question is said, then, to have divided immediately below the existing bridge, each branch diverging to its own series of lakes, of which all are now dry, except the first three. All the same, traces of their former basins are still easily discernible, and the former existence of stationary water is proved by the character of the soil, which is either bare or sparsely dotted with steppe plants. Add to this, that the names still remain. The left or northern branch went to the following lakes — Jaghatsch-alghutsch, Tongus-jardi, Kärälik, Tajir-jajiri, Tenger Kullunu-uji, Döschötu-arghaji, Köldölön-köl*, and Tughu-baschi, where the branch entered the Kontsche-darja.
IIIF Color HighRes
