100 REPORTS ON THE STATE OF SCIENCE.—1912. 
The general result of these examinations indicates that heat gradients 
beneath high grounds and continental areas are markedly less than 
those beneath low grounds and the oceans. Because seismic activity 
beneath certain portions of ocean beds is, as I have already said, 
at least four times greater than it is along shore lines or well inland, 
and if the gradient beneath continental surfaces is 1° in 75 feet, we 
might expect a gradient beneath the deeper parts of oceans of about 
Pati 19. 
Another method by which an approximate estimate may be made of 
suboceanic thermometric gradients is to assume that the steepness of 
these increases as we descend from a shore line to a sea-bed at the same 
rate as they increase as we descend from a high level to a shore line. 
In the tables given by Professor Prestwich I find seventeen entries which 
refer to gradients obtained at elevations lying between 1,017 and 9,529 
feet above sea-level. Ten of these observations were made at metal 
mines, six at coal mines, and one in a borehole. The mean height of 
these stations is 2,723 feet. The mean of the gradients is 1° F. for 
68 feet of descent, which, it will be observed, is somewhat less than the 
average gradient given by Professor Everett. The mean gradient from 
low-lying stations is about 1° F. for 60 feet descent. The difference 
between these gradients is therefore 8 feet, and if this difference steadily 
increases as we descend beneath sea-level, at a depth of 12,000 feet we 
should expect to find a gradient of 1° F. for 25 feet descent. This value 
and the gradient of 1° in 19 feet already suggested, considering what 
has been observed in mines under the sea, may be rough approximations 
to thermometric gradients beneath deep oceans. With rock conductivity 
constant, the rate at which heat is lost beneath our ocean Would there- 
fore be about three and a half times that at which it escapes from con- 
tinental surfaces. If this is so, we may assume that the suboceanic 
crust of the world is either thinner or a better conductor of heat than 
that beneath the land. The plumb-line and observations made with 
pendulums show that high ground and mountain ranges have a deficiency 
in their gravitational attraction. To account for this Sir G. B. Airy 
advanced the hypothesis that materials of which they are constituted 
bulged downwards into a heated denser nucleus beneath. This, -and 
the fact that the value of gravity increases as we approach the seaboard, 
means that the superficial covering of our earth beneath mountains is 
not only thicker, but it is also less dense than it is beneath lowlands and 
near the sea. 
It may also be added that rocks which are heavy and those which 
are metamorphic or crystalline have a slightly higher conductivity for 
heat than many other stratified rocks which are comparatively light. 
The crust of our earth beneath a suboceanic depression, partly, per- 
haps, because it is continually bathed by an oceanic circulation of cold 
water, is therefore a region where we should expect to find the greatest 
flow of heat, and consequently it is one where sudden contractions 
which accompany solidification should most frequently occur. 
Diabase, which is a volcanic rock, when it passes from the fluid to 
the glassy state contracts about 14 per cent., but at the time of solidifica- 
tion, which takes place at a temperature of about 2,000° F., there is a 
