The author discusses the old hypothesis that the earth represents meteoritic material. With Jeffreys he assumes the planet to have been initially gaseous, with a mean temperature well above the boiling point of iron, namely 3000° Centigrade. A large part—perhaps much more than half—of the original body of gas may have been lost to space because of volatility at the high temperature. Condensation of the residual gas to liquid and differentiation in both the gaseous and liquid states occurred, but by processes significantly different from those pictured by Jeffreys. The suggested course of development gives an existing earth which has the following succession from surface to center: a true crust, the lithosphere; a thicker, vitreous asthenospheric shell; a still thicker, crystallized mesospheric shell; and a nickel-iron core, probably fluid and possibly behaving much like a gas.
The paper summarizes the principal observed facts about the nature of meteoritic stones and irons, and the reasons for supposing them to be fragments of a disrupted parental planet. A cause for this disruption and therewith explanation for some of the general characteristics of meteorites are briefly discussed. After comparison between earth and parental planet, the imagined evolution of the earth is tested by reference to the terrestrial discontinuities, mean density, moment of inertia, radioactivity, and plasticity. Rough estimates of temperatures in depth and degrees of strength of the materials in depth are deduced.