THE KARAKORUM PLATEAU 75
snow. During the preceding cloudy day, so little snow had melted that the river had completely dried up. This morning, an unusually hot sun had melted the fresh snow of the preceding night so rapidly that before eleven o'clock the river had revived.
A month later, on the northern slope of the Kwen Lun mountains, I proposed to take a guide and a servant, and go down the gorge of the Sanju River, where the fords are very bad. The guide said that it would be possible if the cool, cloudy weather continued, but not if it were clear and sunny. It rained the evening before we started, but that neither caused the river to rise, nor disturbed the guide. The next day, however, it was warm and clear. The guide became nervous, urging us to gallop down the valley whenever possible, in order to cross the last ford before the daily flood from the high snowy mountains overtook us. We outstripped the river, but had not been long in camp when a boy called out : -
" The flood has come."
Sure enough, the river had suddenly become muddy, and was visibly rising and broadening into an impassable torrent. Similarly in August, on the upper Chira River, a week of rain, mostly drizzle, did not cause the stream to rise so much as did a single day of bright sun. Farther east, near Lop-Nor on the same northern slope of the Kwen Lun range, I was told that in summer the erratic Vash Sheri River becomes a mere brook during periods of three or four days of cloudy, rainy weather among the mountains, but expands to a violent flood when, for a few days, sunny weather melts the snow. Various writers have noticed simi-
IIIFカラー高解像度
