96 W J MCGEE — SHEETFLOOt) EROSION. 



by a meteorologic change, chiefly diminution of precipitation to such 

 extent that the rivers shrunk and ceased to carry detritus or even to flow 

 into the sea, and that the storm freshets flowing northeastward were 

 paralyzed while those flowing southwestward were relatively stimulated 

 (both in high and cumulative degree because of the delicate adjustment 

 between pecipitation and absorption), so that every divide in the district 

 began migrating northeastward. It is deemed probable that the tilting 

 began somewhat gradually late in the Neocene, and thus that the trans- 

 formation in the face of the district took place rather slowly. It is also 

 deemed probable that essentially the present attitude has now been 

 maintained for a longtime, though the tilting is doubtless yet in progress 

 at a slow rate. It may be noted that the records of neighboring districts, 

 so far as known, are consistent with that interpreted for this district. 



STREAM EROSION IN THE DISTRICT. 



Character of streams. — The current maps of the Sonoran district com- 

 monly represent a considerable number of rivers gathering from many 

 tributaries in the mountains and flowing northward into the Gila or 

 southward and westward into the Altar and Sonora or directly into the 

 gulf. Better maps based on actual surveys, such as those of the General 

 Land Office in the United States, show multitudes of small streams flow- 

 ing down the mountain slopes and either ending on the plain or unit- 

 ing in rivers which wander toward some principal waterway and end 

 blindly ; but the general maps express hypotheses based on the be- 

 havior of streams in humid lands, and even the best maps represent the 

 sand washes produced by great storms in lieu of permanent water. In 

 the more elevated eastern part of the district, and during the rainy 

 seasons, especially that of winter, when the storm water in the moun- 

 tains is supplemented by melting snows, most of the sand washes are, 

 indeed, converted into streams, which, although shallow, are rapid and 

 even torrential, and carry vast volumes of sediment-charged water down 

 the slopes; but throughout most of the territory the sand washes are 

 hardly wetted during the normal rainy season, and are transformed 

 into torrents only during great storms or cloud-bursts occurring at in- 

 tervals of years or decades. So the map representing sand washes as 

 rivers is misleading unless its purely conventional character is clearly 

 understood. Although the aggregate stream length represented in such 

 a map is considerable, perhaps several times greater than the aggregate 

 length of the streams actually flowing on a given date during the wettest 

 season, yet during the dry season the aggregate length of actual water 

 lines is reduced to a minute fraction of the aggregate length of sand 

 wash, and nearly all of the channels are dry. So the stream of living 



