84 Mr. H. H. Howorth on 



were so, then we ought to find that in the early winter the 

 surface layers move faster than the bottom ones, whereas 

 later on, when the surface water has been some time sealed 

 up and the frost operates deeper and deeper, the motion of 

 the deeper layers ought to be greater than that of the 

 surface ones, which seems contrary to experience. 



Again, if the glacier motion is due to a general swelling 

 of its bulk, due to the enlargement of its component grains, 

 how is it that it does not swell upwards in the direction of 

 least tension, as well as downwards in the direction of the 

 most compact ice. Its behaviour ought assuredly to be that 

 of quicklime, etc., when charged with water, which swells and 

 pushes out in the direction of least resistance. 



Again, if the motion of the glacier is due to growth of 

 its grains, how are we to account for its moving more 

 quickly in the centre than at the sides ? The walls of the 

 glacier no doubt act as a drag on the movement of the glacier 

 by means of friction, but that they exercise a crushing 

 influence on the ice near the sides so as to pr^event the 

 creation of fissures or the infiltration of water there more 

 than in the centre, seems an assumption at issue with the 

 evidence. 



Again, as Heim says, if the size of the glacier grains is a 

 function of the number and rate of the cooling and melting, 

 and also of the time during which the process has continued, 

 the grains near the sides, which move much more slowly 

 than those in the middle, ought to be ten, or twenty, or 

 thirty times larger, which they are not. 



Heim again makes an elaborate calculation to shew that 

 if the movement of the glacier is merely the sum of the 

 movements caused by the growth of the crystals, then the 

 observed rate of melting of a glacier would not compensate 

 for the growth, but would be very much too small, so that 

 we ought to find glaciers continuously and rapidly growing 

 in length and size. He concludes, in fact, that the swelling 



