THE MECHANICS OF GLACIERS 9^5 



^dacier bcintr uniform, wherever there is melting the cross-section 

 must change so as to produce a smaller flow as we descend the 

 glacier. For glaciers which fill narrow valleys this simply means 

 a diminution in both the breadth and thickness of the glacier; 

 but for those which spread out, like the Davidson glacier in 

 Alaska, the flow is diminished by the smaller thickness, in spite 

 of the increased breadth. 



We may infer a similar thinning of the glacier above the 

 neve-line ; indeed, with many glaciers, the great breadth of the 

 reservoir, notwithstanding the smaller density of the ice, requires 

 a very small depth in order that the flow should not be too 

 great. 



If we look upon the ice-sheet, which covered North America, 

 diagrammatically as radiating from a center, and suppose that the 

 annual accumulation and melting per unit surface were propor- 

 tional respectively to the distance inside and outside of the neve- 

 line, we find, in order that total melting and accumulation should 

 be equal, that the neve-line would have been at about two-thirds 

 of the distance from the center to the circumference of the circle. 

 It is probable, however, that the melting was larger at the edge, 

 and the accumulation less at the center than this, and, therefore, 

 that the neve-line was nearer the circumference. The greatest 

 flow was of course through a section through the neve-line, and 

 the greatest velocity must have been somewhere in the same 

 neighborhood. 



Pressure. — Let us see if the different parts of a glacier, grad- 

 ually thinning out to its ends, can have the normal velocities due 

 to their cross-sections. We know next to nothing about the 

 sliding of the ice over its bed, but we are safe in assuming that the 

 normal velocity of sliding would be no greater in a thin than in 

 a thick section.^ 



Consider then that the bottom velocity of the region between 

 the two sections y^^ and CD of the dissipator is uniform (Fig.i). 

 For the sake of simplicity we will consider the thickness of the 



' The only experiments bearing on this point are those of Hopkins {Phil. Mag., 

 London, 1845, Vol. XXVI, pp. 3-6). By means of a wooden frame he held together a 



