The Theory of Glacier motion. 105 



the variety and size of the several segments into which it 

 is divided by the crevices, until at length a sufficient change 

 of position is effected to permit the escape of the imprisoned 

 waters, when these, rushing forth, empty the icy caverns 

 which they had filled, and the mass of the glacier, whole or 

 in parts, descends by a certain distance into the valley. The 

 motion cannot hence be uniform, but per saltum, which is 

 found to be the general fact, though often difficult of ob- 

 servation" {Trans. Dub. Geol. Soc. I. 319 — 320). This 

 operation, Mr. Mallet argued, was facilitated by the ice 

 being of less specific gravity than water, and he urges that, 

 when the glacier was raised by hydrostatic pressure, it 

 was so in a direction perpendicular to its bed, but, when by 

 the withdrawal of the water it was again deposited, it would 

 rush in a direction perpendicular to the horizon ; hence the 

 ■cause of its motion {id., 321). By means of this theory 

 Mallet professes to account in an ingenious way for many 

 of the phenomena of glaciers. The theory, I need not say, 

 is no longer held by anyone, although it hardly deserves 

 so severe a criticism as was passed upon it by Charpentier, 

 when he says of it : — " Voila une explication qui a besoin 

 de commentaire, mais non pas de refutation" {Essai siir les 

 Glaciers., 38 note.) This sharp phrase is due to the fact that 

 lie altogether misunderstood the argument of the ingenious 

 Irish writer. It is answered completely, however, by an 

 -appeal to the fact that glaciers have been clearly shewn not 

 to move by jerks and jumps, but continuously in winter 

 and in summer, and also that they have a continuously 

 differential motion. While his main theory no longer lives, 

 we must not forget that it was Mallet who first, in the paper 

 just cited, noticed that the transverse crevasses in glaciers 

 have a curved form with their convex side presented 

 downwards, and in the direction of the glacier motion 

 (pp. cit.^ 321), and he urged that the central parts of a 

 glacier must descend much more rapidly than its lateral 



