THE COASTAL DESERT 115 



a dry valley to its head lie will find there a set of broad and shal- 

 low tributaries. Sand drifts may clog them and appear to indi- 

 cate that water no longer flows through them. They are often re- 

 ferred to by unscientific travelers as evidences of a recent change 

 of climate. I had once the unusual opportunity (in the mountains 

 of Chile) of seeing freshly fallen snow melted rapidly and thus 

 turned suddenly into the streams. In 1911 this happened also at 

 San Pedro de Atacama, northern Chile, right in the desert at 

 8,000 feet (2,440 m.) elevation, and in both places the dry, sand- 

 choked valleys were cleaned out and definite channels reestab- 

 lished. From a large number of facts like these we know that the 

 dry valleys represent the work of the infrequent rains. No desert 

 is absolutely rainless, although until recently it was the fashion 

 to say so. Naturally the wind, which works incessantly, partly 

 offsets the work of the water. Yet the wind can make but little 

 impression upon the general outlines of the dry valleys. They re- 

 main under the dominance of the irregular rains. These come 

 sometimes at intervals of three or four years, again at intervals 

 of ten to fifteen years, and some parts of the desert have probably 

 been rainless for a hundred years. Some specific cases are dis- 

 cussed in the chapter on Climate. 



The large valleys of the desert zone have been cut by snow- 

 fed streams and then partly filled again so that deep waste lies on 

 their floors and abuts with remarkable sharpness against the bor- 

 dering cliffs (Fig. 155). Extensive flats are thus available for 

 easy cultivation, and the through-flowing streams furnish abundant 

 water to the irrigating canals. The alluvial floor begins almost 

 at the foot of the steep western slope of the lava plateau, but it is 

 there stony and coarse — hence Chuquibamba, or plain of stones 

 (chuqui=stone ; bamba=plain) . Farther down and about half-way 

 between Chuquibamba and Aplao (Camana Quadrangle) it is partly 

 covered with fresh mud and sand flows from the bordering valley 

 walls and the stream is intrenched two hundred feet. A few miles 

 above Aplao the stream emerges from its narrow gorge and thence- 

 forth flows on the surface of the alluvium right to the sea. Nar- 

 row places occur between Cantas and Aplao, where there is a pro- 



