VOLCANOES. GLACIEES. 43 



centre of a glacier, like that of a river, moves more 

 rapidly than its sides. But how and why do glaciers 

 move at all ? Rendu, afterwards Bishop of Annec,y, in 

 1841 endeavoured to explain the facts by supposing 

 that glacier ice enjoys a kind of ductility. The " viscous 

 theory ' of glaciers was also adopted, and most ably advo- 

 cated by Forbes, who compared the condition of a glacier 

 to that of the contents of a tar barrel poured into a slop- 

 ing channel. We have all, however, seen long narrow 

 fissures, a mere fraction of an inch in width, stretching 

 far across glaciers a condition incompatible with the 

 ordinary idea of viscosity. The phenomenon of regela- 

 tion was afterwards applied to the explanation of glacier- 

 motion. An observation of Faraday's supplied the clue. 

 He noticed in 1850 that when two pieces of thawing 

 ice are placed together they unite by freezing at the 

 place of contact. Following up this suggestion Tyndall 

 found that if he compressed a block of ice in a mould it 

 could be made to assume any shape he pleased. A 

 straight prism, for instance, placed in a groove and sub 

 mitted to hydraulic pressure, was bent into a trans- 

 parent semicircle of ice. These experiments seem to 

 have proved that a glacial valley is a mould through 

 which the ice is forced, and to which it will accommo- 

 date itself, while, as Tyndall and Huxley also pointed 

 out, the ' veined structure of ice ' is produced by 

 pressure, in the same manner as the cleavage of slate 

 rocks. 



It was in the year 1842 that Darwin published his 

 great work on i Coral Islands.' The fringing reefs of 

 coral presented no special difficulty. They could be 

 obviously accounted for by an elevation of the land, 



