TJie Theory of Glacier motion. 99 



from the fire, or does the outside particle refuse to melt ? 

 A candle, too, ought to melt in the socket instead of near 

 the lighted wick, and to bulge out into abnormal obesity. 

 When heat is applied to ice the heated side is first raised to 

 the melting point, and as we recede from that side the 

 temperature gradually diminishes. If more heat is applied 

 it is spent in keeping up this state of the ice as to tem- 

 perature against all possible losses, and in melting that part 

 of the ice nearest to it. No amount of this additional heat 

 will alter the state of the interior, except so far as it may 



alter the other conditions on which it depends No 



amount of internal heat could possibly bring about internal 

 melting in a uniform mass of ice. It is simply a myth " 

 {Id. 497). 



Again, as Heim says, Croll neglects the fact that ice is 

 transparent to heat rays to some depths ; but when we have 

 considerable thickness, as on a glacier at 0° of temperature, 

 the heat rays are not transmitted, but go to melt the surface 

 layers. Croll's theory does not account for the quicker pro- 

 gress of the centre than the sides of a glacier, nor for 

 crevasses, etc., etc. (Heim, 308). 



Again Croll's theory requires that his glacier shall 

 be at the temperature of 32°, whereas we have every 

 reason to believe that it is much below that temperature 

 during a large part of the year, not only on account of 

 the rarefied atmosphere in which it lives but the continual 

 radiation from its surface at night and evaporation at all 

 times, and except in its lower layers, when the glacier is 

 constantly passing into a liquid form, it is most improbable 

 that the temperature in any part of a glacier is so high as 32. 

 It cannot be urged that when a minute crystal of ice 

 melts, the liquid thus formed drains away between the 

 interstices of its neighbours, or we should have every slab 

 of ice subject to solar rays sweating away its substance 

 from its nether surface, which is not the case. So far as 



