Mr. Hopkins 07i ihe Mechanism of Glacial Motion. 167 



piece of ice should form a portion of an actual glacier, where 

 the extension may be produced so slowly as to be continued 

 for a number of years without producing fracture, it is very 

 conceivable that the above ratio might become very much 

 greater than in the former case. 1 have explained how the 

 tension of the mass will be relieved in the direction of maxi- 

 mum tension by the formation of fissures. If the elasticity q{ 

 the ice were 'perfect^ a physical line in that direction would 

 return to its original unstretched length, but will otherwise 

 remain permanently elongated in a degree depending on its 

 want of perfect elasticity. In a similar manner another line 

 may become permanently compressed, and thus, by a repeti- 

 tion of the same processes, the linear dimensions of an element 

 of the mass may become much altered in the course of a long 

 period of time. In like manner a rectangular element may, 

 in a long period, very slowly change its form so as to deviate 

 very considerably from rectangularity without fracture. Thus 

 it may happen that the flanks of a glacier may become dislo- 

 cated with a much less degree of extension than the central 

 portion, because the extension takes place much more rapidly 

 near the sides than near the centre. As an example we may 

 take the results of observations made by Prof. Forbes on the 

 Mer de Glace. He ascertained the velocities at four stations 

 in a transverse line, the first station being near the side of the 

 glacier, and the second about 100 yards, and the third about 

 230 yards from the first; the fourth was near the centre of the 

 glacier. The velocities were in proportion to the following 

 numbers: — 



A B C D 



1-000 1-332 1-356 1-367. 



Consequently the mean relative velocity between A and B may 

 be taken as =-332, and that between B and C 



= (1-356-1-332)^, 



(since the distance B C = 130 and A B = 100) 



= 0-2 nearly. 

 The distance between C and D is not given, but assuming it 

 to be nearly the same as between B and C, the mean relative 

 velocity between C and D would be less than -01, i. e. only 

 about one-fortieth of that between A and B. Now, in the in- 



& 

 vestigation of articles 4< and 5, — is proportional to the rela- 

 tive velocity, and therefore the maximum extension, which 

 = ^, is proportional to the relative velocity. Consequently, 



