576 WILLARD D. JOHNSON 



a smooth curve, lessening in declivity in the direction of flow. The 

 glacier, however, by ablation, is diminished in volume as it lengthens ; 

 it is normally deepest close to its head ; and possibly it is most effective 

 in scour-erosion in proportion as it is deep. It must, in that event, 

 tend to produce a valley "down at the heel." 



The reverse grade, on amphitheater floors especially, occurs with 

 sufficient frequency to be regarded as a type form. Rock-basin 

 lakes, beginning at the amphitheater head, sometimes have notable 

 length, several times the canyon width. The upper surface of the 

 glacier here, on the other hand, invariably declines forward. Thus, 

 in specific instances, it is not merely inference, but fact, that the glacier 

 is deepest at the rear, and excavates there to a forward-rising grade. 



It is, furthermore, implied that forward inclination of bed is not 

 essential to glacier movement. It is not necessary, merely to deter- 

 mine that question, to inquire intimately into the nature of glacial 

 motion. Fundamental in that motion, apparently, is the weight of 

 the ice; and if the glacier at bottom, under its own weight, is not 

 strictly viscous, it is apparently at least viscoid, responding in effect 

 to the law of Hquid pressures. 



A viscous substance, heaped upon a level surface, spreads in 

 mounded disk form, deepest at the center. Its flow-curve, in any 

 radial vertical plane, advances from the bottom. The tendency to 

 flow movement is proportioned to depth — to load; it diminishes 

 toward the outer margin. The outer portions, therefore, move too 

 slowly, and are affected by horizontal, forward thrust. They are 

 retarded at the same time by basal friction, and in consequence present 

 a bulged and swelling front, implying, over a broad marginal tract, 

 rising Hnes of flow. But the glacier is terminated forward, and is 

 thinned toward its termination, by combined melting and evapora- 

 tion— i. e., by ablation; and, by ablation, it may be inferred, the 

 constantly bulging front is planed away. The glacier may be regarded 

 as made up of two layers — a superficial, relatively rigid layer, and a 

 basal layer, mobile under the weight of the other; or of a zone of 

 fracture and a zone of flow. In the thinning frontal region, the upper 

 layer, or cover, is brought into contact with the bed. Rearward, it 

 is lifted; though at the same time there it is planed away. Hence, 

 rising hnes of flow in effect extend to the surface ; for the cover is to 



