MUSEUM OF COMPARATIVE ZOOLOGY. 203 
To the above mentioned sources of heat we have clearly to add a 
third element of action; that which arises from the constant escape of 
temperature from the earth’s interior. The amount of this is not well 
ascertained. It is probably not sufficient to melt more than a few 
millimeters of ice each year, but it has to be reckoned on in estimating 
the influences which are at work to bring about melting at the base of 
the glacier. From these three direct heat-giving actions, we are entitled 
to assume that a certain amount of the ice at the base of a moving mass 
of the material will inevitably be melted. In addition to these direct 
influences we must also take into account the probability, indeed, we 
may say the certainty, that the pressure of the superincumbent glacier 
indirectly favors the liquefaction of the ice at the lower parts of the 
section. As long ago as 1849, James Thomson theoretically estab- 
lished the conclusion that the effect of pressure was to lower the freezing 
point of water.* Shortly after, Wm. Thomson experimentally proved 
the verity of these theoretical considerations, and it now appears that 
where a glacial sheet has a thickness of three thousand feet, the freezing 
point of water is lowered to about 30°, or approximately two degrees 
below that at which it congeals in ordinary atmospheric conditions. 
It may be regarded as fairly well established that the ice in several 
points in Eastern New England attained a thickness of more than three 
thousand feet ; it is possible, indeed, that when the glacial envelope was 
most completely developed its depth much exceeded this amount. As 
we do not yet know the temperature at which ice enters a glacial mass 
such as recently covered the northern portion of this continent, we can- 
not make sure that this pressure alone would be sufficient to bring 
about melting at the base of the mass, even if it had the thickness of 
amile. There can be no doubt, however, that the effect arising from 
the lowering of the melting point would be greatly to favor the liquefac- 
tion arising from the sources of heat before mentioned. It seems to me, 
therefore, that we are justified in assuming, at least as a working hy- 
pothesis, the existence of large amounts of molten water next the bed 
rock on which a deep glacier lies. This hypothesis seems to be ade- 
quately supported by the physical conditions of our drift deposits. 
Wherever these have been observed, we note the existence of many 
- accumulations of washed sand and gravel, commonly known as kames or 
osars, which distinctly map out the position of subglacial streams of 
fluid water, which often coursed for scores if not hundreds of miles 
under the ice arches before they escaped at the margin of the glacier. 
* See Transactions of the Royal Society of Edinburgh, Vol. XVI., Part 5. 
