1879*] Relations to that of the Glacial Epoch. ix 
In short, the question to be determined is — What is the 
temperature of the Antarctic ice ? Now, if nearly all the 
heat possessed by the ice has been received from the 
upper surface of the sheet, the temperature of the mass 
must be mainly determined by that of the surface, and 
cannot be far above the mean temperature of the surface. 
If so, the temperature of the ice must evidently be very 
considerably below the freezing-point. 
(1.) If we suppose the heat to be transmitted from the 
surface downwards by Conduction , we must necessarily 
conclude that the surface is at a higher temperature than 
the ice below ; for conduction can only take place from a hot 
to a colder body, and this process could not possibly 
maintain the mass of the ice below at a temperature equal 
to the mean temperature of the surface. The general ten- 
dency of conduction would, therefore, be to keep the ice be- 
neath at a lower temperature than that at the surface. 
(2.) The work of Radiation, however, would probably 
have the opposite tendency. The heat received by direct 
radiation from the sun could not possibly raise the temper- 
ature of the ice above 32 0 , but the heat lost by radiation 
might lower the temperature to far more than 32 0 below 
zero. If the heat received from the sun’s rays should keep 
the surface of the ice at, say, 32 0 during the summer, and 
the heat lost by radiation should keep the surface at — 32 0 
during winter, which is not an extravagant supposition, the 
mean temperature of the surface would then be o°. But 
the mean temperature of the underlying ice would not 
be so low ; for the low mean temperature of the surface 
is almost wholly due to loss by radiation into stellar 
space during winter, and this loss would be chiefly confined 
to the surface. Had the surface been rock instead of ice 
the rise of temperature during summer would have been 
about as great as the decrease during winter, and conse- 
quently the mean temperature would have been 32 0 instead 
of o° as in the case of ice. Hence the difference between 
the mean temperature of a rock surface and that of the 
rock below would not be so great as in the case of ice. 
The tendency of diredt radiation, therefore, is to maintain 
the surface of the ice-sheet at a lower temperature than 
that of the underlying mass. 
(3.) This tendency is strengthened by another circum- 
stance which comes into operation. During summer a 
large portion of the diredt heat from the sun is spent in 
melting the surface ice. The melted ice passes down 
through crevasses and openings in the sheet, thus carrying 
the temperature along with it. The heat of summer is by 
