Structure of the Earth's Crust. 23 



from its not continuing a surface of fluid equilibrium ; and 2ndly, 

 because I suppose the nucleus to shrink in cooling more than its 

 solid envelope. The fii'st of the assumptions alluded to is con- 

 tained in Section IV. of my Second Memoir on Tei-restrial Phy- 

 sics*. It does not, of course, imply that when the shell was very 

 thin it was not capable of modifying its shape in conformity 

 with that of the fluid matter beneath. But it should be remem- 

 bered that the changes of shape in the surface of the fluid mat- 

 ter which have been discussed by me are such as would neces- 

 sarily be continuous and extremely slow, from their dependence 

 on the slow refrigeration of the whole earth, and gradual deposi- 

 tion of new strata of solidified matter upon the inner surface 

 of its crust. No abrupt changes of shape in the figure of 

 the fluid nucleus could take place from such causes, while the 

 viscous and plastic matter which may happen to be passing into 

 the sohd state would be directly influenced by its immediate 

 contact with the nucleus. It is thus I have been led to the con- 

 clusion, that the ellipticity of the strata of equal density in the 

 shell would not follow the law prevailing in those of the nucleus, 

 or of the entire mass when it existed in a state of fluidity, namely, 

 that of an increase in proceeding from the surface towards the 

 centre. I have only recently ascertained that M. Planaf had 

 arrived at a conclusion very similar to mine, not long after the 

 publication of my researches. My conclusion is, that the least 

 ellipticity which the inner strata and the inner surface of the 

 shell can have, could not be less than that of the outer surface. 

 M. Plana's is, that all the strata would have the same ellipticity 

 as the outer stratum. The difi"erence between my result and 

 that of the distinguished Sardinian mathematician arises from 

 the circumstance that he does not appear to have considered the 

 change of volume which the fluid matter may undergo in passing 

 to the sohd state. This I have attempted to do, because all the 

 experimental knowledge which we possess relative to the solidi- 

 fication of fused rocks, points to a diminution of volume as an 

 accompaniment of their gradual transition to a state of solidity. 

 It is true that rapidly-solidified and scoriaceous lava is lighter 

 than the same substance in a fluid condition, and it is therefore 

 observed to float upon the latter during the progress of volcanic 

 eruptions ; but observers are also well aware that slowly-solidi- 

 fied and compact fragments of the same rock will gradually pe- 

 netrate and sink through the viscous surface of a molten stream. 

 Still more decisive are the results separately obtained by MM. 

 Bischof, Ch. Deville, and M. Delesse, regarding the coefficients of 

 cubical contraction of the crystalline rocks in passing from the 



* Phil. Trans. 18.01, j). .525. 



t Aslronomische Nachrichten, No. HW. vol. xxxvi. p. .Slf). 



