of to-day. The abandonment of the ruined towns cannot have been due to the encroachment of sand, for in some places, such as Miran, Vash Sheri and Ying-pen, there is no sand ; in others, such as Endereh, the amount is small ; and even in places such as Dandan-Uilik, where there is much sand, it appears to have encroached upon the inhabited area after, not before, the vegetation had begun to die for lack of water. It is possible that in one or two minor cases the diminution of the water supply may have been due to the diversion of part of the headwaters of a stream, but in most cases this is impossible. Manifestly, such diversion in one place would cause a larger stream and more abundant vegetation elsewhere; but as a matter of fact, the amount of vegetation has diminished everywhere. Neither the theory of the diversion of tributaries, that of the encroachment of sand, that of the former existence of a better irrigation system, nor that of the destruction of forests offers any explanation of the fact that many streams have grown much more saline. This can be accounted for only on the theory of a change of climate, a theory which not only is in harmony with all the known geographic facts, but which also explains many of the events of history.
We have seen that in the Lop basin, as in Kashmir, desiccation has not been a steady process. Fluctuations or pulsations have been the rule, some large and long, some slight and short. From the third to the sixth century of our era, there seems to have been a time of intense aridity succeeding a period of relatively abundant moisture, during which the rivers and hence the population were much larger than at present. During the time of aridity, many old towns
IIIF Color HighRes
