the gleaming expanse of the broad, shallow river, with its sand-bars, snags, and driftwood. Beyond it lay a golden plain of reeds, and a low belt of green poplar woods. The scene might almost have been on some slow river in Indiana. But the white and brown flock of fat-tailed sheep, the shepherd boy wading across the stream in a white gown and a fur cap, and above all the great two-humped camels stripping the leaves from the poplars, disturbed the illusion. The line of low yellow hills a mile away beyond the band of green was the border of the great red desert, whose slowly drifting sands have marched persistently forward for two thousand years or more in the wake of dying vegetation.
In returning to the region of villages, I followed a route in the sand several miles west of the Keriya River. There was nothing which could by any possibility be an old course of the river leading to Dandan-Uilik. There were, however, a number of isolated areas of dead vegetation lying in large hollows surrounded by sand. They became more numerous as we approached the zone of vegetation. Along the border of the zone, half or more of the plants were dead. Such phenomena furnish strong evidence that not only among the ruins, but in other parts of the now lifeless desert, vegetation flourished abundantly at no very remote period.
IIIF Color HighRes
