418 Professor Forbes's Eleventh Letter on Glaciers. 



tlie bottom, seems a simple corollary, from the fact that the 

 centre of a glacier flows past its sides. Even admitting the 

 irregularities of the bottom to be less than those of the late- 

 ral expansions and contractions of the valley, the enormous 

 pressure on the bed must generate a friction proportionally 

 great. Some persons have, however, found so much diffi- 

 culty in conceiving the fact of varying velocity in a vertical 

 plane (notwithstanding the evident analogy of a river), that 

 I was glad to take an unexceptionable opportunity of demon- 

 strating it. 



I have already shewn, at the close of my Sixth Letter, that 

 the effect of friction in retarding the rate of motion, must be 

 most sensible nearly in contact with the soil ; and that when 

 the glacier is of great thickness, the upper part of it (to which 

 alone we have access) may be expected to move uniformly, 

 or sensibly so, which accounts for the approximate verticality 

 of most crevasses, during the limited period of their exist- 

 ence. It is, therefore, near the contact of a glacier with the 

 subjacent soil, that the most sensible effect may be looked 

 for. The circumstances which first suggested themselves to 

 me as the most favourable for such an observation, were 

 where a glacier emerging from a gorge, or from between a 

 double mound of its own formation, falls into a valley, and 

 presents for some space lateral faces of ice, not, indeed, quite 

 vertical, but still very highly inclined, and which repose im- 

 mediately on a bed of rubbish, which, if not flat, but sloping 

 somewhat towards the centre of the glacier, might be consi- 

 dered, beyond cavil, as the floor or bed on which it rests. But 

 a careful examination of several glaciers with a view to such 

 an experiment convinced me that, even if successful, it would 

 not be conclusive. For it almost invariably happens, under 

 these circumstances, that the glacier, being no longer con- 

 fined laterally, tends to scale off by means of fissures parallel 

 to its length (as in fig. 2.) ; and even if these fissures do not 

 give rise to a sensible sliding of the surfaces, they indicate 

 the direction of the twist to which the ice is exposed by the 

 more rapid motion of the centre. To avoid misapprehension, 

 I here repeat that such a tendency to scale by means of lon- 

 gitudinal fissures, occurs only where lateral compression is 



