Sec. ii] THE ANCIENT COURSE OF THE KONCHE-DARYA 767
move together westwards until we reached the Konche-daryâ.. Then Afraz-gul was to follow the river south to Tikenlik, where he would connect his traverse with that brought by Lal Singh in January, 1913, from the south, and continue the survey of the Lop ` high road ' along the Yarkand-daryâ. to Kara-kum and along the Konche-daryâ, to Korla. I myself would turn to the north and, moving across the desert, strike the ancient Ying-p`an–Korla route at the ruined watch-station of Kurghan described by Dr. Hedin as the nearest to Ying-p`an.
We started on March 21st from our Ying-p`an camp and following the wide well-marked bed of the Kuruk-daryâ, head in a westerly direction, found it filled with marsh and luxuriant reed-beds up to a point about one and half miles higher up.1ß There several deep-cut ravines, which represent the end of the Shindi river's drainage, join the bed from the north-north-west. It was quite evident that the periodical floods which descend in the former account for the moisture to be found in the Kuruk-daryâ, bed at this point.19 Similarly it appears very probable that the southwestern direction which, as already stated, the bed followed for some distance from here is a deflexion due to the alluvial deposits formed by the Shindi river's flood channels at their extremities. We were following what looked like an old trail leading westwards, and noticed that where it crossed the mouth of certain channels, it was marked by tamarisk bundles embedded in these channels as if intended to form a causeway over wèt ground for laden camels. We made out traces of this track, which had a decidedly old appearance, over a total distance of more than ten miles from camp, in places where it had worn into the bare clay or Sher. No certain explanation of its origin suggested itself.
After we had left the bank of the ancient river-bed, living Toghraks, so far plentiful, soon disappeared and dead trunks also became scarce. About three miles from camp we entered on a bare clay steppe, showing slight marks of wind-erosion and very scanty tamarisk scrub, mostly dead. After travelling seven miles we struck a shallow but unmistakable river-bed some 5o yards wide and, like the rows of dead trees near by, running approximately west to east. Dead reed-beds covered the banks. The traces of the track reappeared beyond, and in one place, significantly enough, lay along the top of a small clay terrace cut off by wind-erosion from neighbouring ground of the same level. Either the track was distinctly old or else wind-erosion was proceeding rapidly in this area. We had covered about ten miles when we reached a wide and well-defined dry riverbed, over 30o yards across for the most part and lined with rows of dead branchless Toghraks showing signs of great age. Its average depth seemed nowhere less than 10 to 12 feet ; but much drift-sand lay within it.
This large bed came from the west and manifestly joined up in its farther course with that which we had traced on the preceding day's reconnaissance to a point south-west of Ying-p`an camp. That it formed a connexion between the Kuruk-daryâ, and the present Konche-daryâ, bed—still, as it proved, about five miles away to the west—was quite clear ; but, of course, no close estimate could be formed of the time that had elapsed since it ceased to carry water. We followed the right bank of the bed westwards for about a mile until it became more and more smothered under high dunes ; we then turned off WSW. in order to make sure of reaching water before nightfall. We crossed a succession of short but high ridges of sand, all aligned from north to south, before again coming upon a dry river-bed, of much smaller size but showing signs of great age. All the dead