572 WILLARD D. JOHNSON 



complex channeling seemingly had passed into the phase of conflu- 

 ent glaciation, presented a much less intelligible ground-plan pattern. 

 In every case, however, there was still an approximately central 

 draining canyon. It was the sprawling high-walled masses of the 

 residual uplands that told a clear story. 



The legitimate inference was that the suspension of glaciation had 

 suspended as well a process which had threatened truncation of the 

 range. It was obvious, postulating recession, that the canyons of 

 this summit region were independent in their courses, and had 

 developed independently, of the initial upland drainage plan. It 

 was clear, from their grade profiles, that they were not stream-cut. 

 The inference was not only legitimate, but necessitated, that, pro- 

 foundly deep as they were, they were essentially of glacial excavation. 



Here, then, were facts of observation in support of one of the two 

 extreme views referred to at the outset. That, however, ice-streams 

 had descended on the long western slope to the range foot, or even 

 close to it, I subsequently found to be untrue. 



The canyons of that flank I now regard as stream work below, as 

 ice work above, and as the joint product of streams and of glaciers, 

 alternately, in between. But wherever they are abnormally deep, I 

 infer from the evidence of the floor profiles that they have been thus 

 deepened by glaciers. 



The range had not been domed over by a continuous ice- sheet; 

 it had been glaciated rather against its upper slopes. The summit 

 tracts, narrowed by flank attack, had remained bare, perhaps because 

 wind-swept ; the ridges and peaks of degradation continued emergent, 

 because sharp. Obviously, though its period had been short, the 

 action of the process had been relatively rapid; for in the shallow 

 preglacial canyons of the broad foothill zone, in which lay the moraines 

 of the outer glacial boundary (the relatively insignificant, coarse 

 products of degradation), the normal processes of erosion had accom- 

 pHshed so httle that, seemingly, their action there had been suspended. 



The summation of the hypothesis was that retrogressive cutting 

 in large part had carried away the uplands, along an approximately 

 definite, and an approximately level, plane of attack. Deep canyons 

 had resulted, indirectly, because recession, directed horizontally, had 

 been directed into a rising grade. This action seemed distinct from 



