304 THE ANDES OF SOUTHERN PERU 



the snowline and the snow itself, though wet and compact, is not 

 underlain by ice. Yet the process of hollowing goes forward 

 visibly and in time will produce serrate forms. In neither case 

 is there the faintest sign of a bergschrund; the gradients seem 

 so well adjusted to the thickness and rate of movement of the 

 snow from point to point that the marginal crack found in many 

 snowfields is absent. 



The absence of bergschrunds is also noteworthy in many locali- 

 ties where formerly glaciation took place. This is notoriously the 

 case in the summit zone of the Cordillera Vilcapampa, where the 

 accumulating snows of the steep cirque walls tumble down hun- 

 dreds of feet to gather into prodigious snowbanks or to form 

 neve fields or glaciers. From the converging walls the snowfalls 

 keep up an intermittent bombardment of the lower central snow 

 masses. It is safe to say that if by magic a bergschrund could 

 be opened on the instant, it would be closed almost immediately 

 by the impetus supplied by the falling snow masses. The explana- 

 tion appears to be that the thicker snow and neve concentrated at 

 the bottom of the cirque results in a corresponding concentration 

 of action and effect; and cirque development goes on without 

 reference to a bergschrund. The chief attraction of the berg- 

 schrund hypothesis lies in the concentration of action at the foot 

 of the cirque wall. But in the thickening of the snow far beyond 

 the minimum thickness required for motion at the base of the 

 cirque wall and its change of function with transformation into 

 neve, we need invoke no other agent. If a bergschrund forms, its 

 action may take place at the foot of the cirque Avail or high up on 

 the wall, and yet sapping at the foot of the wall continue. 



From which we conclude (1) that where frost action occurs at 

 the bottom of a bergschrund opening to the foot of the cirque wall 

 it aids in the retreat of the wall; (2) that a sapping action takes 

 place at this point whether or not a bergschrund exists and that 

 bergschrund action is not a necessary part of cirque formation; 

 (3) that when a more or less persistent bergschrund opens on the 

 cirque wall above its foot it tends to develop a schrundline with 

 a marked terrace below it; (4) that schrundlines are best devel- 



