420 THE ENDERE RUINS [Chap. X1I
with low sand-cones. The dead tamarisk scrub which strewed their slopes sufficed for our camp fires, while the ice brought along from the Yartunguz-Tarim saved us the trouble of digging a well. A fairly strong wind had been blowing for the greater part of the day from the north or north-east, and the dust-haze it raised showed an ominous persistence.
Our march so far had been almost due east. But on leaving Yantak-chaval on February 19 our guides struck to the north-east. The reason for the change of direction became clear when, after crossing a great ` Dawan ' some i 20 ft. high, we found ourselves traversing a broad stretch of low undulating dunes where progress was relatively easy. Plenty of living tamarisk and occasional little groups of Toghraks showed that subsoil water was near, and the general impression I gained was that we were crossing here what may once have been the terminal area of .a river. The latter could only have been an earlier course of the present Endere river. Towards the east this area was bordered by a long-stretched chain of high dunes, in which it was easy to recognize from afar the ridge flanking the left bank of the actual Endere river-bed. From its darker colour it was known to our guides by the appropriate name of Kizillik, the ` Red (sands)' 1. For fully two miles our march lay over this great accumulation of dunes, and then we emerged on the western edge of the belt watered by the Endere stream.
Deserted The channel which we now ascended towards the south-east was called by our guides
o 1 terminal of the old Darya' of Endere. Here, too, the river has in recent years shown a tendency to
Endere shift westwards, so that we found quite a respectable sheet of ice, from io to 20 yards in
river. width, covering what previously had been a deserted dry bed. On the other hand, this return
of the summer floods to the earlier channel was causing the ` New Darya', said to have been formed further westwards some twenty years before, now to receive annually less and less water. It was with this change of course, and not with any diminution in the quantity of the water brought down by the river, that my guides connected the abandonment of the little colony of ` Endere-Tarim ', which had been formed on the ` New Darya ' at a point apparently four to five miles to the north-north-west of where we first struck the river. It was impossible to spare time for a visit to this modern representative of the terminal oasis of Endere, or for a survey of the river-courses down to their actual termination. Judging from the guides' statement, the summer grazing-grounds of the Endere shepherds may extend northward for a day's march beyond the now deserted ` Tarim '.
March On February 20 we marched along the ` 1(51e-Darya' up to the point where, near the
along of Kokul-toghrak, the New River branching off from it, and thence fol-
grazing-ground gra ew ver was ranc g
river. lowed the main river bed upwards. Not far from a rustic Ziarat we crossed to the right
bank. The river was here about 20 yards broad, holding under its ice-sheet about 2 feet of water, but the well-defined, steeply-cut banks rising 6 to 7 feet above the level of the ice indicated a considerable volume of water at other seasons. At a deserted hut of rushes marking the shepherd station (le chik) of Kara-öchke-öltürgan (` where the black goat sat ') we left the ponies behind, and struck into the desert south-eastwards. The belt of vegetation, the width of which, owing to the persistent haze, it was difficult correctly to estimate as long as we kept by the river bank, proved here very narrow. Hence, where we camped for the night, though scarcely more than two miles in direct distance from the river, there was only bare eroded ground with here and there low sand-cones covered by scanty dead tamarisk scrub.