IN the Journal and Proceedings of the Asiatic Society of Bengal (vol. 8, No. 9; 1912) there is a paper by Dr. L. L. Fermor entitled “Preliminary Note on the Origin of Meteorites.”In approaching his subject Dr. Fermor deals with the effects of pressure in determining the mineral constitution of rocks. He refers especially to the case of eclogite, wherein the development of garnet, and occasionally of diamond, is remarkable. The first represents the allocation of the normal elements present in a gabbroid magma in such a manner as to give rise to the development of minerals possessing the maximum density; the second, the same effect in the relatively rare case of carbon being present. He instances the well-known occurrence of both diamond and garnet in certain eclogites of South Africa. Dr. Fermor concludes that there must be what he calls an infra-plutonic zone deep in the earth's surface materials, and probably extending far downwards, composed of rocks of eclogitic type and, probably, at such a temperature as maintains them in a plastic-solid state. He explains the occasional appearance of this infra-plutonic rock at the surface as due to special conditions whereby cooling forestalls the effects of reduction of pressure during the ascension of the rock; which in this way attains a region of negligible pressure while preserving a mineral structure proper to very great depths. The well-known kelyphite rims, sometimes surrounding the garnets of eclogites, indicate a partial break-down into mineral structures of lesser density. The recent results arrived at by Dr. Harold Jeffreys (NATURE, Sept. 25, 1926), based upon the velocity of transmission of compressional seismic waves, point to the existence of a zone of the density of basaltic glass underlying the granitic continents; the basaltic zone being succeeded downwards by one which Dr. Jeffreys suggests may be dunite. The granitic layer may possess an average thickness of from 20 km. to 30 km. The basaltic zone may have a thickness of about 20 km. The underlying layer may extend to a depth of 1500 km.Daly has contended for the existence of a general basaltic layer extending beneath the continents and oceans. Other eminent petrologists have held this view. We desire here, in the first place, to cite some recently added evidence in its favour. Washington's discovery of the chemical resemblance of the plateau basalts ejected at various times and at different points of the earth's surface, might in itself be regarded as conclusive. The resemblance, however, seemed to fail respecting the radioactivity of the rocks; the Oregonian basalt showing a much higher radioactivity than the Deccan and Hebridean (Phil. Mag., Nov. 1924). However, we pointed out at the time that the authenticity of our 'Oregonian' material was not perfectly assured. Thanks to the courtesy of Prof. Daly, Prof. Landes, and Prof. W. P. Smith, we have been able to examine specimens of Oregonian basalt of undoubted authenticity, and even a fragment from the specimen chemically investigated by Washington. The average radium and thorium contents of Oregonian rocks from twelve localities come out as closely alike with those already determined for the Deccan and Hebridean areas. Those who are inclined to limit radioactivity to local conditions will find it difficult to explain away the chemical and radioactive similarity of these enormous and widely sundered outpourings. Coming now to the third terrestrial layer-the high-density medium underlying the basaltic-it appears that there is no necessity, so far as seismic evidence is concerned, to suppose this layer to differ from the basaltic save in the matter of its mineral structure. The gabbros and the eclogites are magmatically the same. We have found that an eclogite possessing the density 3-415 yielded upon fusion a glass which when cold possessed a density of 2-746, which agrees with the results obtained by Day, Sosman, and Hostetter (Am. Jour. Sci., 37,1914) for the density of basaltic glass. The density of eclogitethe piezocrystalline form of the magma-would agree with the seismic evidence. It ranges from 3 2 to 3 5. That of dunite is 3-3. We see, then, that seismic evidence is not opposed to the simplifying assumption that the sub-continental materials as revealed in the plateau basalts may extend downwards to a depth approximating to 1500 km. In short, the assumption of the eclogitic character of the third terrestrial layer involves little more than the acceptance of Dr. Fermor's inference that the pressure conditions which convert carbon into diamond in eclogites is also responsible for the high-density mineral structure of these rocks.However, there is something more to be said. We have found recently that the eclogites possess on the average barely one half the radium and thorium contents of the plateau basalts. An explanation, we believe, can be offered for this apparently contradictory result; an explanation which, if it is correct, seems to throw light on the origin of terrestrial surface structure and surface history. Yet the explanation we would suggest is very elementary and simple. We assume that the outer material of the primeval earth was originally compounded of all those siliceous aggregates afterwards differentiated into the layers we have been discussing; and that it was not throughout uniform in chemical composition. There was heterogeneity on, probably, a very varied scale; in some places coarse, in others fine: and this heterogeneity of distribution and association affected the stable chemical elements as well as those that are radioactive. These assumptions seem to be the most general we can make as well as the most probable.Let us consider first the effects of heterogeneity in the distribution of radioactive elements. This would carry with it thermal heterogeneity. Some parts would melt before others, and when their surroundings were melted would retain a higher temperature and lower density. These parts would gravitate upwards. Again, some parts of lesser radioactivity would retain for longer periods the solid state. In the depths this condition would be especially effective, for here the pressure confers upon the medium the maximum density, as we have seen. Thus the temperature and pressure conditions conspire to preserve to, or confer upon, the medium a high density; and accordingly it gravitates downwards when the fusion of its surroundings permits. Hence it would come about that such gabbroid magma as was poor in radioactive elements-poor in uranium, in thorium, and in potassium-would retain the solid state longest and sink into the depths. It is a fact that eclogite is poorer than the plateau basalts in all three radioactive elements. Respecting potassium, the mean content of K2O in the five plateau basalts analysed by Washington is 0-89 per cent. The mean K20 content of seventeen eclogites cited by Rosenbusch is 0 70 per cent., and of eleven cited by Mile. Briere (Bull. Soc. Franeaise de Min., 43, 1920) it is 0-37 per cent. They are, in fact, at once the poorest in the heat-producing elements and the densest rocks known. There would, also, exist a lack of homogeneity respecting the distribution of the non-radioactive elements. Some parts would be richer in silica, alumina, etc.; others in metallic oxides, etc.; such parts would for ever seek to ascend or to descend. Or, in times of thermal loss, certain well-known factors concerned in magmatic differentiation would operate in the same directions.The final results should be precisely what we find; a highly siliceous and aluminous surface layer rich in radioactive elements and-what seismic evidence reveals-in the depths, rocks of maximum density and, as we now find, of minimum radioactivity. Should these inevitable final conditions be disturbed by the circulation attending a great revolution, they would gradually be re-established during the long later period of thermal loss. We perceive, in short, that heterogeneity in the circumstances is not stable, but must result in radioactive and gravitational stratification. Reversing our line of argument, we might justify our assumption of initial heterogey in recognition of the revealed surface structure of the earth.If these views are correct, it would appear that radioactivity mainly has been responsible for the stratification of the earth's outer materials. It has determined the origin of the radioactively rich and gravitationally light continental layer, of the isostatic layer of intermediate radioactivity and density, and of that more deep-seated layer which only at long intervals takes part in the great events of surface history: the major revolutions (" The Halley Lecture," 1924, pp. 31 et seq., and A. Holmes, Geol. Mag., July 1926). In short, it would appear to have fashioned those structural conditions which have been responsible for geological history and for the development of life upon the globe. Further evidence of stratification in the earth's great basaltic layer is revealed in the petrology of the oceanic islands. The island basalt-which we must regard as representing the same lava as composes the ocean floor or prevails immediately beneath it -is richer in all the radioactive elements, and at the same time lower in density, than are the plateau basalts. These island lavas reveal, in fact, the final differentiation of the substratum where it attains the surface of the globe; a differentiation referable to physical causes similar to those we have referred to above.As bearing on all our views of earth-history we would point out that the low radioactivity of eclogite directly affects estimates of geological time based upon the period required to bring about a major revolution. The length of previous estimates will require to be doubled.