PROGRESSION OF GLACIERS. 
269 
of clay and marl, or even of compact limestone, 
especially in large mountain masses, I have fre- 
quently observed that the rock presents a net- 
work of minute fissures pervading the whole, 
without producing a distinct solution of continu- 
ity, though generally determining the lines ac- 
cording to which it breaks under sudden shocks. 
The network of capillary fissures pervading the 
glacier may fairly be compared to these rents in 
hard rocks ; with this difference, however, that in 
ice they are more permeable to water than in stone 
How this network of capillary fissures is formed 
has not been ascertained by direct observation. 
Following, however, the transformation of the 
snow and neve into compact ice, it is easily con- 
ceived that the porous mass of snow, as it falls 
in the upper regions of the Alps, and in the 
broad caldrons in which the glaciers properly 
originate, cannot pass into solid ice, by the pro- 
cess described in a former article, without retain- 
ing within itself larger or smaller quantities of 
air. This air is finally surrounded from all 
sides by the cementation of the granules of neve, 
through the freezing of the water that penetrates 
it. So enclosed, the bubbles of air are subject to 
the same compression as the ice itself, and be- 
come more flattened in proportion as the snow 
has been more fully transformed into compact 
ice. As long as the transformation of snow into 
