the whole channel floor irregularities. That is, there could 

 have been no bergschrund or sympathetically curving 

 crevasses but rather a slight depression in the central 

 portion of the cirque glacier and a slight heaping up at the 

 sides near the base of the cirque. For a moment's reflec- 

 tion will suffice to show that a glacier which possesses a 

 bergschrund and is much crevassed has not formed its 

 channel irregularities, but has rather been moulded to them 

 so as to have been put into a state of tension with conse- 

 quent destruction of stream characteristics. 



Fifthly, the least percentage of ice action which is 

 expended ascorrasion is that accomplished while descending 

 the heavy slope of the cirque, while the greatest occurs at 

 the foot of the cirque or that spot where the plane of the 

 channel floor is most opposed to the motion of the descend- 

 ing ice. 



We have then a complex action in a cirque. There is 

 relatively heavy corrasion at the base of the cirque. The 

 action there however is mostly abrasive, because of the 

 tendency to compress the structures instead of quarrying 

 them at this point. The base of the cirque will be lowered 

 until the vertical measure of strength has been expended. 

 This results in heavy sapping which tends to induce the 

 atmospheric slope of repose in the cirque walls. This 

 sapping action is however only a partial factor in the form- 

 ation of the cirque walls, because of the corrasive action 

 of the ice in descending the cirque slopes. The reasoning 

 here should be followed closely, because it is necessary to 

 clearly grasp i he <-un<litimis obtaining in t his portion of the 

 cirque. Firstly, then because; tin- cirque is situated on the 

 glacial divide, the ice or snow action at the lip of the cirque 

 is negligible, but increases in intensity with progress down 

 the cirque wall. With increasing volume the factor of flow 

 increases, and because the cirque structures are practically 



