xxvi Appendix. 
a depth of somewhere about thirty-five miles has a temperature below the 
melting point of ordinary rocks at the surface, while all below this depth has a 
temperature above its melting point at the surface. Consequently, we have an 
outer crust in which the attraction of cohesion among the molecules is greater 
than the repulsion caused by heat, surrounding a nucleus in which the 
repulsion caused by heat among the molecules is greater than the attraction of 
cohesion. 
The outer crust must therefore be more or less rigid, while the superheated 
interior must be in such a state that if the pressure that keeps it in its place 
is decreased at one point it will expand, and this expansion will permeate 
through the whole mass until the pressure is again equally distributed 
throughout. Conversely, if the pressure is increased on any point, this 
pressure will affect the whole mass and distribute itself evenly through it. Of 
course I need hardly say that the rigid state of the crust is not separated from 
the superheated state of the interior by a marked division, but the one passes 
imperceptibly into the other. Now each portion of this rigid crust must be 
maintained in its place by three forces, viz.—its weight, the lateral thrust of 
the arch, and the outward pressure of the superheated interior. While these 
three forces remain constant equilibrium will be maintained, and no move- 
ments will occur on the surface. But if one or more of these forces change in 
amount, the equilibrium will be subverted and movements of the surface will 
take place. If also the equilibrium be disturbed at one place, it follows, from 
what I have said about the distribution of pressure in the superheated interior, 
that the equilibrium will also be disturbed in all surrounding areas. If, for 
instance, an upheaval of the crust should take place at any point, the under- 
lying superheated rocks, being thus relieved from the pressure above them, 
would expand and rise up, and fill the hollow; but this expansion would 
spread through the mass, and would therefore lessen the outward pressure of 
the interior in all the surrounding areas, which would consequently subside, 
and equilibrium would only be once more restored when the mass of the 
subsided areas equalled the mass of the elevated area. Consequently elevation 
implies subsidence, and vice vers. Where now must we look for the causes 
that are in operation to disturb this equilibrium? The most obvious is the 
radiation of heat into space by the earth, and the consequent cooling and 
contraction of the superheated interior. This is at present almost universally 
accepted by geologists as the cause of the movements of the surface and the 
upheaval of mountain chains, but many arguments have been urged against it, 
and although I am willing to allow that it must have some effect in producing 
movements, these effects are, I think, completely absorbed by the much 
larger ones that flow from causes that I shall presently describe ; and it is 
quite impossible that it can be the only cause of movement, partly because 
