GLACIAL EROSION AND TRANSPORTATION 277 



abrasion during the settling and consolidation of the neve in 

 the amphitheatres in which it accumulates. At the bottom of 

 the depressions, however, the conditions are different. Intense 

 glaciation there takes place, as is attested by the rounded 

 and striated surfaces, and by the occurrence of rock-basins. 

 The ice filling a cirque impinges with great weight upon its 

 bottom and in its motion outward tends to deepen the exca- 

 vation. At the same time the blocks loosened from the walls 

 of the cirque are carried away by the outward flow of the ice. 

 There are thus at least two processes which unite in enlarging 

 and deepening these peculiar features of glaciated mountain- 

 tops. 



When a glacier leaves a cirque and flows down a canon 

 the grade of which is uneven, the erosion of the ice-stream will 

 also be uneven. The reason is that the ice in descending a 

 steep slope exerts its greatest force at the base of the incline in 

 the same manner as in the excavation of cirques. The tend- 

 ency of a moving ice-stream in descending a steep slope is to 

 increase the inequalities of its bed ; this tendency, it seems 

 probable, will lead to the formation of both scarps and cirques 

 when the drainage of a high-grade valley is changed from a 

 liquid to a solid form. To illustrate : The grade of mountain 

 streams increases toward their sources, and when their gorges 

 become occupied by ice, the irregularities of their channels 

 — caused principally by the meandering of streams, thus leav- 

 ing projecting bosses on either side — may cause ice-cascades 

 in the glaciers. An ice-cascade exerts the greatest erosive 

 power at the base of the scarp which it descends, thus aug- 

 menting the inequality. At the same time the canon is broad- 

 ened and the minor features resulting from stream-erosion are 

 erased. The steeper the grade the more pronounced would be 

 the action of the ice in remodeling and strengthening the ma- 

 jor inequalities of its bed. The resulting scarps and terraces 

 should therefore be most numerous and best defined near the 

 heads of the channels in which they occur.* 



The prevalence of cirques is also graphically described 

 by Dr. G. M. Dawson in his "Report upon the Forty-ninth 



* "The Quaternary History of Mono Valley, California," pp. 352-355, 



