This is so contrary to our ordinary experience and ideas, in which loss of heat tends to change from gas to fluid and solid, that we must look into it a little to make it sound reasonable. The recent brilliant work of P. W. Bridgman (contrary to the earlier speculations of Tammann) indicates that the effect of increased pressure, at high temperature, makes a substance solid and crystalline. Crowd any atoms close enough together, and no matter how fast they expand or contract under the influence of heat the crystalline atomic forces will get to work when they are crowded within their range, and the closest packing, hence that which will yield most to the pressure, hence that which is likely to take place, is when they are all regularly arranged facing the same way. Such an arrangement we call crystalline. Just so when they want to pack the most people into the car of an elevator they ask them to all face to the front. Keep this metaphor a moment. Any one who should try to penetrate such a crowd would find it a hard job. They would offer a very effective rigidity. Now suppose them to sweat in those confined quarters their fat away, their phlogiston, their caloric. If the walls of the car remained rigid while the individuals therein shrunk they might after a while be able to turn around or even move around in a car. Such is then the supposed condition of the atoms in the FOURTH, the central, layer of the earth's crust. This assumes that the middle layer is rigid and sustains itself, like the shell of a nut, as in the figure, while within the atoms are in a less rigid condition. That such a shell might be self-sustaining is suggested by an experiment of Bridgman, who put a marble with a gas bubble in it under a pressure of something like 150,000 pounds to the square inch without producing any perceptible change.
As loss of energy from the earth's interior went on this central core of gas would enlarge until the middle shell was hardly self-supporting. Then, probably at some time of astronomic strain when the earth's, orbit was extra elliptical, it would collapse, in collapsing generate heat, and so stop the process. The collapse would be transmitted to the viscous layer which might be increased, motions set up in it, and so a wrinkling of the outer thin crust on which we live.
Then there would be four layers to the earth like the butternut of the figure. First, the inner kernel of gas; second, the hard shell or endocarp; third, a viscous layer like the sarcocarp or pulp, and outside of all the wrinkled crust of exocarp. If such is the structure of the earth we may have in the very structure of the earth itself a reason why from time to time there are collapses of the middle layer leading to elevations of portions of the outer rind, and marking off the chapters in geological history, the lines between geological systems.
There are reasons in facts of observation for believing that such is the structure of the earth, of which I have as yet said nothing. We see the interior of a glass marble, I saw the bubble in the interior of Bridgman's glass marble, how? By waves, vibrations, which start from the sun or some other source, and going through it reach my eye. Though the earth is not penetrated by sunlight it is penetrated by the waves and vibrations that start from that jar produced by a crack which we call an earthquake. These vibrations can be received by that eye of the geologist called a seismograph. The seismologist tells us there are three kinds of waves sent out in an earthquake. If you notice the explosion of a blast at a little too close distance you will notice that you see it first, then hear it, and then perhaps a little later a few chips of rock may come flying past your ears. These three things correspond somewhat to the three kinds of waves which spread forth from an earthquake. But in the case of the explosion we see the blast first, then hear later. The waves which produce the sensation of sight are, we know, lateral disturbances, the waves which produce the sensation of sound are waves of condensation, whose motion is in the direction of their propagation and they come later. In the case of the jars of earth, the reverse is true. The first set of waves to arrive are the waves which are due to compression—vibrations in the direction in which the waves are produced—and correspond to sound waves. Later come waves which are transverse sidewise disturbances of the solid mass of the earth. As we can easily see, in an earthquake jar traveling from the opposite end of the earth, there should be no insurmountable difficulty in recognizing the jar, which is a direct upthrow from one which would tilt it to the right or left. Now there is a law of Laplace by which the velocity of spread of sound waves through gas may be calculated. That this law should hold at temperatures and pressures so high as those that must exist in the middle of the earth is, of course, a question, but it will be interesting to see how nearly the actual velocity of about 10 kilometers a second compares with the velocity which such waves should have in gas of a density and under a pressure such as a gas near the center of the earth must have. Using Oldham's figures (and they seem to be confirmed by the recent investigations of E. Rudolph and S. Szirtes[18]), we find that the time of transmission of these first and fastest preliminary compression tremors is about twice the velocity of such a jar according to Laplace's law in as dense a mass of gas, provided the ratio of the specific heat of a gas at constant pressure to that of a gas at constant volume remains 1.4, which is for many substances. But as it is 1.6 for mercury the discrepancy is not more than I had expected.
[5] Gerlands, "Beitrage zur Geophysik," XI., Band, 1 Heft, 1911, p. 132. "Das kolumbianische Erdbeben am 31 January, 1906."
The second preliminary tremors arriving later are due to the lateral disturbance. Their propagation is much less rapid when the point of origin is nearly opposite the point of receival. In other words there is a core within the earth about 0.4 of the radius in radius, in which according to Oldham, these lateral waves have much less velocity. Now in a gas there is less resistance to lateral displacement than in a solid, and the less the resistance the less the velocity, so that this fact fits in with the idea of a gaseous core perfectly. If there is such n core, moreover, of less rigidity it would have less refraction. Consequently waves not striking the border above the angle of total reflection would be totally reflected, and just as around a bubble there is a dark border where the light does not get through so at a certain distance from the source of an earthquake there would be a circle (it is really about 140 degrees of arc away), where no second tremors would be felt. Here again, though seismograph stations are as yet few, fact and theory are apparently going to correspond.
The last type of earthquake waves follow around the outer layer of the crust.
There is one farther line of verification to which I had addressed myself. Is it likely that the loss of heat and energy from the central nucleus, at the rate which we know at the surface from a central nucleus of anything like 0.4 the radius of the earth, would give a shrinkage of anything like the amount indicated by the mountain ranges, in anything like the time which we are led to assign on other grounds to the geologic periods?
Rudski has also attempted to connect the shrinkage and age of the earth. Both these methods depend on how fast the earth is losing heat, that is on the geothermal gradient. Since at present, owing to the apparently large but unknown contribution of radioactivity to that gradient we know very little about what the other portion is, it seems unwise to give any figures, especially as almost all the numerical data are largely guess work. It will, however, be fair to say that very long times for the age of the earth seem to be indicated, nearer millions of millions than millions unless the radius of the gaseous core was mainly small or its rate of contraction with loss of temperature high.