252 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1923 



rial under a constantly increasing pressure. It may be argued that 

 the effect of temperature in this region may decrease the elastic 

 constants and hence also the velocity. But on any hypothesis the 

 temperature is not increasing rapidly as far down as this, and it 

 seems highly improbable that increasing temperature would have 

 sufficient effect on both the rigidity and the bulk modulus so that 

 with increasing depth the two velocities would remain nearly con- 

 stant over a range of 1,400 km. The constancy of the velocities is 

 truly a remarkable feature and, as will appear below, furnishes an 

 important clue for the solution of the problem of the earth's interior. 

 In recent times Goldschmidt 20 has postulated an arrangement of 

 the matter within the earth as follows: (1) An outer silicate layer 

 120 km. thick and of density 2.8; (2) a layer of dense silicates 

 (eclogite) extending to 1,200 km. depth with density varying from 

 3.6 to 4; (3) an intermediate zone of sulphides and oxides of density 

 5.6 and extending to 2,900 km.; and (4) a central core of nickel- 

 iron having a density about 8. The average density of this arrange- 

 ment is very close to the accepted value, and the moment of inertia, 

 although 3 per cent too low, can be considered in fair agreement. 

 Zoeppritz, Geiger, and Gutenberg, 21 and Mohorovicic, 22 and others, 

 have adduced evidence in favor of the existence of various shells 

 or layers in the earth, but the arrangement which most nearly 

 resembles that which we shall describe in the next section is that 

 given by Gutenberg 23 in a paper which came to our attention after 

 our paper was written. He obtains a density-depth curve which, 

 like ours, consists of four parts. The starting point of his deriva- 

 tion is the assumption that the core is of constant density 2.3 times 

 that of the layer above it, which also is of constant density. He 

 concludes that the density of the layer extending from 60 km. to 

 1,200 km. depth varies from 3% at the top to 4% at the bottom. In 

 this respect his estimate of the density change within the earth is 

 strikingly like that which we shall now describe, although ours 

 is based on the change of density due to compressibility and involves 

 assumptions quite different from those of Gutenberg. 



PROPOSED DISTRIBUTION OF MATTER WITHIN THE EARTH 

 OUTERMOST LATER 



The average density of the igneous rocks 24 at the surface is about 

 2.8. Allowing for a small amount of sedimentary rock, let us take 

 the surface density as 2.7- The density and basicity of the rocks 



» V. M. Goldschmidt. Z. Elektrochem., 28, 411. 1922. 



21 Nactar. Kgl. Ges. Wiss. Gottingen. 1912, p. 121. 



22 S. Mohorovicic. Beitr. z. Geophysik, 13, 217-240. 1913 ; 14, 188-198. 1914. 

 * B. Gutenberg. Phys. Zeit., 24, 296-299. 1923. 



M H. S. Washington. Bull. Geol. Soc. Am., 33, 3S8. 1922. 



