D. Burns — 0)i the Mechanics of Glaciers. 299 



8° than on one of 4° ? What proof is there that these lines of mole- 

 cular melting (if there be such at all) are parallel to the bed of the 

 glacier? And even if they do vary to a degree with the bed of 

 the glacier, it seems to me very confident theorizing to calculate 

 on such nice action of a vv^eak force like gravity on such a small 

 body and through such a minute distance. Will heat travel down 

 a slab of ice at 32° faster if held at an angle of 8° than if held at 

 one of 4P ? It follows /rom Mr. Croll's reasoning that such is the 

 case, and it may be so, but he is a bold man who would assert it 

 without experimental proof. But granting that the molecules move 

 faster down the 8° slope, I still fail to see that a crevasse would be 

 formed. What causes the pulling that produces the tension ? A mole- 

 cule that moves a little on and becomes ice again cannot pull more 

 than it did before. Certainly each molecule on the 8° slope pulls 

 more than on the 4? slope, but that has no connexion with Mr. 

 Croll's theory, nor is it to that he refers. It is the faster motion of 

 the molecules that somehow does the pulling, but in what way lies 

 beyond my comprehension. To see what would actually happen, 

 let us think of the glacier as constituted of molecules stretching in 

 lines along its length. Let us suppose that some of these rows melt, 

 and some do not, so that the melting molecules move among the 

 others as through tubes. Suppose the end of the glacier to be at 

 the foot of the 8° slope, and that the 4° slope is the next above. 

 Suppose further that a molecule on the 8° slope can melt, move and 

 freeze again in half the time that one can on the 4° slope. Concen- 

 trate attention now on one of these tubes, and let heat be applied to 

 the end of the glacier, so that molecule A melts and moves down a 

 little. B, C and the whole row now do likewise, and when a mole- 

 cule an inch distant from A is melting, suppose A to melt again, and 

 to continue to do so at regular intervals. This would produce in 

 time a melting phase at intervals of an inch all the way up the %° 

 slope. When these phases get on to the 4° slope, they will move at 

 half the speed, and the result will be that they will crowd on each 

 other, and follow at intervals of half an inch. Still as many mole- 

 cules move on the 4° slope as on the other, and through as great a 

 distance. From these considerations we see that Mr. Croll does 

 not explain the origin of crevasses, but explains how a glacier can 

 move down a low gradient as fast as a high one, if such be the case. 



If Mr. Croll, however, will adopt my amplification of his theory, 

 he need not despair of crevasses of a kind. Given one part of a 

 glacier exposed to the sun's rays for a sufficient time, while a higher 

 part is protected by snow or cloud, and you will have a crevasse 

 formed between. The lower part will flow away while the higher 

 remains rigid and motionless. There would, however, be no frac- 

 ture, but a rigid motionless glacier above and a continually de- 

 creasing one below. Altogether crevasses are difficulty number three. 



I cordially agree with Mr. Croll when he claims that his theorj^ 

 explains why glaciers move at a greater rate in summer than in 

 winter. I only wish it were as satisfactory in other respects. 

 Though I cannot accept it as an explanation of the motion of glaciers 



