348 Mr. Bowman on Natural Terraces 
proaches the nearest to snow, and the opposite or lower end, 
to ice, the change being gradual and the consequence of the 
alternate melting and freezing of the surface. Above the 
height of about 7000 feet, the temperature of the air is seldom 
high enough to melt the snow; and as all ice has previously 
been water, it is clear there can be but little of it in a solid 
state at still greater elevations. The blocks and fragments of 
rock that are detached from the surrounding precipices, ac- 
cordingly sink through the mass to the bottom of the snow. 
At less elevations, where during summer the melting process 
is more active on the surface of the glacier, the water perco- 
lates between the particles of the porous mass, till it reaches 
the rocks on which it rests. Here, the temperature being 
lower, it is reconverted into ice, which cements together the 
stones and gravel, and by its expansive property in passing 
into the solid state, has a tendency to detach them from the 
bottom. Thus a new layer or skin of ice is continually ac- 
cumulating between the lower surface of the glacier and the 
face of its rocky bed, which, as it thickens, acts a wedge and 
imperceptibly keeps forcing up the whole superincumbent 
mass, and with it the stones and gravel which it had en- 
tangled at the bottom. But the cooperation of another agent 
is necessary to bring them to the surface. This is the heat of 
the sun, which during summer, especially in the day time, 
melts the upper layers of the ice, and consequently diminishes 
its thickness ; till at length, by this double action, the stones 
which lay upon the bottom are lifted up till they lie exposed 
upon the surface of the glacier. So that here, a kind of cir- 
culation or interchange of particles takes place, as in a lake, 
modified by circumstances ; their ascent being retarded while 
in a state of congelation, and their descent accelerated by the 
perpendicular cracks and fissures which everywhere intersect 
the mass. 
This expansion from below is going on at the same time in 
the direction of both the longitudinal and transverse axes of 
the glacier; andas the whole mass lies in a hollow or inclined 
trough, the power of gravity prevents it from moving in any 
other direction than downwards. But the motion is not uni- 
form; the sides advance with greater rapidity than the mid- 
dle ; so that if a row of stones were placed at equal distances 
in a straight line across the glacier, they would soon arrange 
themselves into a curve or arch; those at each side being car- 
ried down more rapidly, would form the base; while those in 
the middle having comparatively little motion, would be left 
behind and form the highest part of the arch. This is owing 
to the following cause: a slight elevation of temperature in 
