628 



/ HARLEN BRETZ 



The narrowness and elongation of channels and the continuous 

 gradients of tracts as a whole suggest river valleys but these fea- 

 tures are all inherited from pre-existing valleys. Furthermore, 

 some tracts are nearly as broad as they are long. The pattern of 

 channels on a tract, like the pattern of some tracts and their isolated 

 loessial hills, is much like that of great braided streams (Fig. 8). 



Fig. 6. — ^Longitudinal profile of the deepest of the Drumheller Channels across the 

 nose of the anticline. Four rock basins are indicated, the largest of which is 75 feet 

 deep. 



J /:ij^iMi;!ii;,:n!i;i!iiMLi':j,'Ji;.;{nt^ 



1^00 

 1300 

 /200 



-)I00 

 -1000 

 ■900 

 800 

 700 

 1-600 



One mi le 



Fig. 7. — "The Potholes," longitudinal profile of the northern half, showing (i) 

 cliff along northern side of the canj^on, (2) amount of recession of the falls, (3) elon- 

 gated rock basin below the falls, (4) great gravel bar along edge of the rock basin, and 

 (5) approximate level of Columbia River when the cataract was formed. 



But the scablands are erosional in origin, while the braided stream 

 pattern is depositional. 



The evidence for the origin of channeled scabland by stream 

 erosion is overwhelming. The evidence of contemporaneity of 

 action of all channels of a given tract, at least in its early history, 

 is equally convincing. The only sequence indicated is that of 

 development of the greater channels later in the epoch and conse- 

 quent draining off of the shallower channels.^ The scablands of the 

 plateau in Washington are the beds of huge river courses in which 



' Especially well shown in Lower Grand Coulee, in Moses Coulee between Pali- 

 sades and Spencer, in The Potholes and Frenchman Springs south of Trinidad, and in 

 Palouse Canyon south of Hooper, in all of which cataract recession in main channels 

 cut off smaller channels alongside in the same scabland tract. 



