576 J. H. BRETZ GLACIAL DRAINAGE ON COLUMBLi PLATEAU 



and the master stream finally enters the basin after it has skirted the 

 northern and western sides in part. Only after it enters does it receive 

 the drainage of the plateau. 



The northern rim of the plateau is the highest part. It is a broad 

 divide extending from Spokane west to the southward bend of Columbia 

 Eiver at the foot of the Cascade Eange, parallel to Spokane and Colum- 

 bia rivers and on the average only 10 miles south of them. Down its 

 southern slope flow numerous tributaries to Palouse Eiver and Crab 

 Creek. These two stream systems and Moses Coulee, farther west, are 

 intimately concerned in the glacial drainage history to be outlined. 



Palouse Eiver is the largest stream on the plateau north of Snake 

 Eiver. Its drainage area lies in the southeastern quarter of the plateau. 

 Its S3^stem consists of a large number of minor centripetal streams which 

 flow westward toward the central part of the basin. It enters Snake 

 Eiver in the middle of the plateau's southern margin. Crab Creek 

 gathers the small centripetal streams from the north-central part and 

 carries their discharge westward toward the Columbia, which it enters 

 not far south of the middle of the western margin of the plateau. 



Crab Creek has the largest drainage area of any stream on the basalt 

 plateau of Washington, but there is so little rainfall and so much evap- 

 oration in this area that no water today flows to the Columbia. It has 

 been otherwise in the past. The area possesses a complete system of 

 streamways. Some of them are normal valleys in maturity, only the 

 rainfall in some more humid times of the past having formed the 

 streams. Others are canyons of noteworthy, even spectacular, depths,, 

 intimately related to the history of the glacial drainage of the plateau. 



Stratigraphy 



The dominant formation of the plateau is the Columbia lava. It has 

 been completely covered by sedimentary deposits of silt, ash, and loess. 

 These have sufl'ered much from glacial and fluvial erosion, but still per- 

 sist and give character to the topography and soils of more than half of 

 the plateau. In the western part of the plateau this sediment is presumably 

 the Ellensburg formation. It is intimately related to the basalt, its 

 deposition beginning in some places before the great basalt flows had 

 ceased, and in other places later and minor basalt flows occurring after 

 its accumulation had well begun. If the sedimentary formation, or its 

 lower part, is Ellensburg, its presence indicates clearly that the plateau 

 is determined by the original surface of the Miocene lavas, not by any 

 later erosional plans developed on the basalt. 



