i'oundly modify tlieir channel profiles, such a form would be 

 considered as direct evidence of the ineffiiciency of ice as 

 a corrader. The rounded back of the roche moutonnee, 

 according to them, evidences the abrasive action of ice, 

 while the rough and jagged down stream aspect marks the 

 original preglacial contour of the rock, and hence the failure 

 of the glacier to strongly erode the channel forms. 



But if this were really so, then glacial motion would be 

 strong at some points where the planes of the thalweg and 

 the stream were opposed and negligible, at others on the 

 down stream aspect, where a similar opposition of the 

 general planes of thalweg and stream exists a mechanical 

 impossibility. Consider for example the accompanying 

 figure (Fig. 10) illustrating the structure of a roche 

 moutonnee. 



Illustrating formation of roche moutonnee. 

 Any pressure of the glacier at A is tending to crowd the 

 structures together and compress them and corrasion here 

 is accomplished mostly by sand-papering action with minor 

 fracture of rock blocks. In this case abrasive action vastly 

 predominates. On the other hand the passage of the glacier 

 over the lip B of the roche moutonnee is accompanied by 

 heavy quarrying because there is no support to the rock 

 mass down stream of the profile B C, and the glacier can 

 therefore easily pluck or quarry the blocks from the far.- 

 IS C. Moreover, in the eases of those roche moutonmV 

 forms which form the whole floor of the channel— the point 



