17 
1 879.] Relations to that of the Glacial Epoch . 
ground heat produces scarcely any sensible influence on the 
temperature of the surface, which is determined almost 
wholly by that of the air and other external agencies. The 
temperature, in short, is determined from above , and not 
from beneath. In warm countries, where the temperature 
of the air is high, that of the surface is high also. And 
so likewise in cold climates, the low temperature of the air 
gives a comparatively low temperature to the surface. 
Suppose our globe to be enveloped for some thousands of 
years with a covering at the uniform temperature of say 
ioo° ; and suppose, further, that 5000° should represent the 
temperature of the earth’s mass; then, in such a case, there 
would be a gradual decrease of temperature from 5000° at the 
centre to ioo° at the surface. Let us suppose now the warm 
covering is removed, and replaced by one at — ioo°. In the 
course of some thousands of years there will be a gradual 
decrease of temperature from 5000° at the centre, as before, 
to — ioo° at the surface. Internal heat limits the temperature 
at the centre, but external heat limits in every case the tem- 
perature at the surface. 
To maintain, as Sir Wyville Thomson does, that 32 0 is 
the temperature of the floor on which the Antarctic, ice- 
sheet rests, is virtually to beg the whole question at issue. 
It is the temperature of the ice that determines that of the 
floor on which it rests, and not the latter that determines 
the former. 
What the temperature of the ground under the Antarctic 
ice-sheet may be is a question which at present we have no 
means of determining with certainty ; we know only that 
it must be far below the freezing-point, for the ice resting on 
it is considerably under that point. 
Although the temperature of the ice must impose a limit 
to the thickness of the sheet, underground temperature can- 
not do so, for the temperature of the ice is not determined 
by underground heat. 
But supposing we knew the temperature of the Antarctic 
ice, yet this knowledge would not enable us to determine 
with certainty the limit imposed by temperature on the 
thickness of the ice. For, excepting in cases where the 
temperature is but very little below 32 0 , we are at present, 
in the absence of experiments, unable to say what would be 
the amount of pressure necessary to lower the melting-point 
to any assigned temperature. The experiment of Mousson 
shows that Prof. Thomson’s formula, t = 0*0137°^, does not 
hold true when the pressure is excessively great. A pressure 
of 73 atmospheres will lower the melting-point from 32 0 to 
VOL. ix. (n.s.) c 
