701 
going on between them, they would ultimately reduce each other to 
their mean quantity and (not to take the extreme case of infinity) 
at some very moderate depth, the fluctuations above and below the 
mean temperature of the air, as the successive nocturnal and diur- 
nal waves pass through a particle of the stone there situated, will be 
rendered very tr ifing, and may for our present purpose be regarded 
as evanescent. Beyond this depth, whatever mass of stone may 
exist, may be regarded as a slow conducting mass, interposed be- 
tween a surface of ice constantly maintained at 32°, and a surface 
of stone constantly maintained at the mean temperature of the air, 
which by hypothesis is very little above it. Through this then the 
heat will percolate uniformly but feebly, and the ice below will be 
very slowly melted, and the more so in proportion to the thickness 
of the interposed stratum. Let us now consider what happens to 
the ice on the parts undefended by the stone. In the day time these 
experience the direct radiation of the sun, and therefore melt and 
run off in water. At night, it is true, the remaining surface cools 
by radiation; but this cold is propagated downwards, and on the 
return of day the superficial lamina is necessarily put in equilibrium 
with the air and melted by the sun, and however cold the interior of 
the mass may be, the surface will still be kept all day in a state of 
fusion. Thus the degradation of the general surface of the ice will 
be in proportion to the direct intensity of the sun’s rays and the time 
they shine, while that of the surface beneath the stone will only be 
in pr pportion to the excess of the mean temperature of day and night 
above 32°, diminished by the effect of the thickness of the stone. 
This of course will produce a difference of level, and a relative ele- 
vation of the stone sunk as really observed. One curious and, at 
at first sight, paradoxical consequence seems to follow from this 
reasoning, viz. that the ice of a glacier, or other great accumulation 
of the kind, may, at some depth beneath the surface, have a per- 
manent temperature very much below freezing, though in a situa- 
tion whose mean annual temperature is sensibly above that point. 
In fact (continuing to use the metaphorical expression already em- 
ployed), there is no reason why waves of cold, vf any intensity be- 
low 52°, may not be propagated downwards into the interior,of the 
ice ; but waves of heat above that pomt, of course, never can. Thus, 
the cold of winter and the frost produced by radiation in the clear 
nights of summer, will enter the mass and lower its internal tempe- 
rature, while the heat of the summer air and that imparted by solar 
radiation will mainly be employed in melting the surface, and will 
run off with the water produced. 
I am not aware of any observations on the internal temperature 
of glaciers—they are of course difficult from their usual rifty state ; 
but the point may not be unworthy the attention of the scientific 
traveller. May not this be the cause of those natural formations of 
ice which have been observed in caverns, in Teneriffe, and on some 
elevated points of the Jura chain, below the level of perpetual snow ? 
It is obviously no matter whether the interior mass in the above 
