1878.] 333 [Niles. 



If this is granted, it must then be admitted that the abrading 

 power of glaciers is much less than if the fragments of rock were 

 usually firmly set in the ice. This is one of the many reasons which 

 I have for believing that the erosive power of glaciers is not suffi- 

 cient, in itself alone, to account for the excavation of those valleys 

 in which they are found. 



Among the phenomena which attract the attention and obstruct 

 the progress of the explorer under a glacier, is the abundance of 

 streams. A short distance below the edge of the glacier the ice is 

 constantly melting, and in every place accessible to the observer the 

 water falls, usually in large drops but sometimes in streamlets. Thus 

 the surfaces not covered by the ice are exposed to a constant fall of 

 water, which, first forming numerous rivulets, soon collects in small 

 and rapid streams. The dropping of the water and the rushing of 

 the torrents, the frequent slipping of smaller fragments of stone which 

 have been started by the rivulets and the occasional tumbling or 

 plunging of a larger mass, the incidental cracking of the glacier and 

 the frequent crash of pieces of falling ice, all unite in impressing upon 

 the listener that this is a busy place. Where the glaciers rest upon 

 the upper portions of the roches moutonnees, the streams are formed 

 in the hollows between them which the ice does not fill; therefore, 

 under such conditions their erosive power is exercised upon those 

 lower portions of the rock-surface which are not effected by the move- 

 ments of the glacier. 



In estimating the erosive power of a stream we must take into 

 consideration, not only its volume and velocity, but also the more 

 important factor of the materials with which it is charged. The 

 importance of this is well illustrated by the modern appliance called 

 the sand-blast, in which it is not the violence of the current of air or 

 steam but the .sand which it carries with it, which cuts away the 

 surfaces of stones, metals and glass with such astonishing rapidity. 

 Sometimes this element has been overlooked, as, for example, when it 

 has been argued that because pure water may rush violently over a 

 rock for a long period without producing any perceptible change, 

 therefore, iha valleys which now are or formerly were occupied by 

 glaciers must have been excavated by the ice rather than by the 

 streams below it. A sub-glacial stream, considered as an agent of 

 erosion, should never be compared with a stream of pure water. All 

 of the streams beneath a glacier are charged with small, angular 

 fragments of stone, such as glaciers transport in immense quantities. 

 Any one who has walked over the middle and lower portions of a 



