THE EARTH'S THERMAL HISTORY The primitive earth was presumably very thoroughly stirred up. But the heavy metallic constituents refused to mix with the rocky ones and quickly settled to the centre to form the core. At some stage during the solidification the granitic and basaltic layers sep arated from the dunite one, and, by some process so far very im perfectly understood, the granitic matter became collected into large patches, which we know now as the continents. In a few thousand years a thick solid crust formed, and subsequent cooling took place by conduction through it.
The later history is substantially modified by radioactivity. The amount of heat being conducted out of the earth per unit sur face per unit time is the product of the thermal conductivity of the surface rocks, say o•01 c.g.s., and the increase of temperature per unit depth, which is about 0.00032° C. per centimetre. Thus calories per sq.cm. per sec. are being conducted out of the earth. But the experimental work of the present Lord Rayleigh, J. Joly, J. H. J. Poole and others has shown that average gran ite contains enough radioactive matter for 1 cu.cm. to produce calories per second. Hence about or 23 km., of average granite would supply all the heat leaking out at the surface. The actual thickness of the granitic layer, as we have seen, is about okm.; but other rocks are also radioactive. We are forced to suppose that unless the radioactivity is practically confined to a surface layer with a thickness comparable with 2okm., the heat coming out of the earth would exceed its actual amount.
The temperature at any depth in the crust may be regarded as made up of two parts, one due to the original heat, the other to that developed by radioactivity since solidification. The age of the earth and the other relevant physical data being roughly known, the former part can be evaluated. Subtracting from the rate of conduction out of the earth the portion due to original heat, we are left with the part due to radioactivity. This is found to be equivalent to the rate of generation of heat by about r 5km. of granite, or by iokm. of granite with 2okm. of basalt under it;— in excellent agreement, so far as we can tell, with the results of seismology. If we adopt the latter view, the present temperature at the base of the intermediate layer should be about 56o° Centi grade. At a depth of about 3ookm. the rocks should have cooled by about 28o° since solidification; below 600km. the present tem perature in the rocky shell is nearly the melting point. We can now easily see why the central core is liquid. The melting point of dunite at ordinary pressures is 1,400°-1,6o0° ; that of pure iron is about 1,5oo°, but that of the material of the central core might be lower by some hundreds of degrees on account of im purities. In addition, the melting point of dunite is raised by pressure; that of iron, at any rate at low pressures, is lowered by pressure. Accordingly we should expect that the central core would be more fusible than the surrounding rock; and if the latter is near its melting point the iron must be liquid.
The foregoing discussion concerns mainly continental condi tions. Radioactivity beneath the oceans must be less, and the cooling at great depths there may be up to 3o% more. The upward concentration of radioactive matter requires an explanation. It is perfectly genuine, for, without it, it would be impossible to recon cile the radioactivity of the surface rocks with the rate of increase of temperature inwards. Also it can be shown easily that if the radioactive layers were twice as deep, or even if the total amount of radioactive matter were the same, but were uniformly distrib uted to a depth of iookm., the steady temperature at the base of the radioactive layer would be so high that the interior could never have become solid. A difficulty sometimes expressed about the concentration near the surface is that the elements concerned are heavy, and might have been expected to collect near the centre of the earth. But when the total amount of a substance is too small to affect the density of its solvent appreciably, the density is often of much less importance than solubility or volatility in determining the distribution.
Actually it is found that the radioactive generation of heat in basalt is under a third of what it is in granite, and in denser rocks such as dunite the amount is much lower again. This points direct ly to a strong upward concentration, especially as the dunites ana lysed are, of course, surface samples, and have probably been en riched by contact with the more radioactive rocks that surround them. Further, Holmes has called attention to the fact that in successive igneous outpours, even of material of similar composi tion, the material from the higher levels seems to have been en riched in radioactive materials at the expense of that lower down. The evidence that these substances actually do tend to collect up wards is very strong. The explanation is probably connected with the facts that they form volatile compounds, and that an impor tant fraction of the water on the earth's surface has probably been extruded during geological time. The ascent of the water from rock magmas would carry the volatile constituents upwards. But it is to be noticed that the effect of radioactivity on deep seated temperatures is proportional not only to the total amount of radioactive matter present, but also to its mean depth. The upward concentration of this matter would continue until it be came too near the surface to prevent any farther solidification down below. Later denudation would tend to transfer the radio active substances to the ocean bottom, as Joly has pointed out.
A further complication is the influence of change of state. If a liquid gradually solidifies by cooling from the top a solid crust forms and steadily thickens. The rate of thickening is largely controlled by the latent heat of fusion, which has to be conducted away before new matter can solidify. With a latent heat of ioo calories per gm., which is about correct for the few silicate rocks yet tested, it can be shown that solidification could not have reached a greater depth than 3ookm. during geological time. Yet the seismological evidence is clear that the earth is solid to nearly ten times this depth. It seems to follow that no change of state involving so great a latent heat has taken place at great depths; the change has possibly been from a liquid state to a glassy one, and not to a crystalline one.
Further, the rocks beneath the oceans are more basic, and there fore are probably less radioactive and have cooled more than sub continental rocks. Even at the same temperature basalt seems to be stronger than granite and when it is cooler it will be stronger still. Hence when an ocean floor is compressed against a contin ent, the latter will be the first to yield. This is the probable explanation of the Pacific type of mountains—the long series of the Rockies and Andes with the smaller ranges running nearly parallel to them and to the neighbouring coast.
Mountain formation, on almost any mechanical theory, would be expected to be intermittent in time and localized in position. The upper rocks are elastic solids, which have to be under a con siderable stress before they give way, and when they do so the yield is at a definite place and becomes complete almost instan taneously. If then the stress is one that grows very slowly and gradually, there will be long quiescent periods when the stresses are accumulating, separated by short intervals when they are relieved by folding and fracture, with mountain formation as a result. This corresponds with the facts as known to geology.
The theory has, however, still to meet several difficulties in points of detail; while the phenomena of igneous activity, shown in volcanoes and intrusions, seem at present to stand right outside it. The temperatures that seem most probable within the inter mediate (basaltic) layer are enough to make a hard glass readier to flow than the basaltic layer actually is; nevertheless, basalt driven up to the surface seems to be much more completely fluid and, to judge by its effects on neighbouring rocks, at a much high er temperature. A curious fact noticed by Aston provides what looks like a promising clue. The "inert gases" (helium, neon, argon, krypton and xenon) are far scarcer on the earth than one might expect from comparison with the abundance of other ele ments of neighbouring atomic weights in the crust : the ratio is in all cases somewhere about a millionth. A very natural explana tion is that when the primitive earth was highly heated and prob ably distended its gravitation was insufficient to control the freely moving molecules of a gaseous atmosphere ; consequently the truly gaseous constituents were mostly lost from the earth and transferred to the resisting medium. The present earth consists of the constituents, such as iron, that liquefied readily, and of the materials that formed liquid or solid compounds at the high temperature concerned. From this point of view it is interesting to note that the amount of oxygen in the earth's crust is within I % of that needed to combine with all the other elements ; and the rarity of the inert gases is to be attributed precisely to the fact that they are inert gases, and could not find even a temporary place of safety in the earth's interior. But if we adopt this expla nation we must suppose that most of the water and carbon com pounds on the earth's surface at present have been expelled from the interior since the earth shrank to its present size. Now there is ample independent evidence that chemically active gases and water vapour are continually being expelled from volcanoes, and T. A. Jaggar and A. L. Day have given reason to believe that their reactions with one another and with the oxygen of the air are the principal factors in maintaining volcanic temperatures.
Continuous heating in the deeper crust would produce expan sion and hence tension in the upper crust. This would lead to vertical cracks and violent igneous outpourings. On the views just given concerning the way radioactive materials become concen trated near the surface, the earth's crust must have had such a history in its earliest days, but denudation and other factors have obliterated all traces of it. On the moon, however, there has been no denudation, and an extensive system of fissures and craters is the moon's most prominent feature. Lunar vulcanism is then to be referred to the moon's earliest days and the upward move ment of radioactive matter produced by it provides the reason why the internal heating ceased and there is no vulcanism on the moon to-day.
This theory was actually suggested as an explanation by Sir G. B. Airy when attention was called to the facts by Archdeacon J. H. Pratt. Extra mass on top would force the crust down, the fluid below would flow out, and the process would continue until just enough had flowed out to restore the balance; in the "iso static" state every vertical column of the same section would have the same mass inside it. Pratt did not accept this explanation, and proposed as an alternative the idea of a "depth of compensation"; according to him the extra mass of a mountain is compensated by a uniform reduction in density of all the matter below it, down to a fixed depth. This hypothesis requires that the matter should diminish in density when an extra pressure is placed on top of it, and therefore is physically unplausible ; and there are also weighty geological arguments against it. But most geodesists, especially J. F. Hayford and W. Bowie in the United States, have adopted Pratt's view and found it to correspond well with the facts. It is only recently that W. Heiskanen has shown that the Airy theory fits the facts at least as well, and that H. Jeffreys has shown that this result was inevitable from the nature of the law of gravita tion. The only modification needed in the Airy theory is that the substratum is not a fluid, but a very stiff solid, capable of trans mitting distortional waves, but not of enduring great stresses for a long time. The weakness is, of course, to be attributed to its high temperature.
The variation of gravity over the outer surface depends some what on the thickness of the elastic outer layer, for this affects the depth where the outflow takes place. A horizontal displace ment of a given mass at the surface would clearly affect gravity in a different way from a similar displacement at the centre of the earth. Heiskanen, on the hypothesis that the outflow was al ways vertically below the added matter, found that the thickness of the upper layer should be 4o-8okm. in different great mountain ranges. Actually the elastic bending of the upper layer will spread the compensation horizontally a little, and, when allowance is made for this, the most probable thickness is only about 3o kilometres. Thus the strong region extends at most to the bottom of the inter mediate layer; all matter below that has a strength less than about a fiftieth of that of granite or basalt under surface conditions.
(H. J.)