IV.] THE CHEMISTRY OF THE PRIMEVAL EARTH. 39 



the atmosphere of the non-gaseous nucleus. We may suppose 

 an arrangement of the condensed matters at the centre accord- 

 ing to their respective specific gravities, and thus the fact that 

 the density of the earth as a whole is about twice the mean 

 density of the matters which form its solid surface may be 

 explained. Metallic or metalloidal compounds of elements, 

 grouped differently from any compounds known to us, and far 

 more dense, may exist in the centre of the earth. 



The process of combination and cooling having gone on until 

 those elements which are not volatile in the heat of our ordinary 

 furnaces were condensed into a liquid form, we may here in- 

 quire what would be the result, upon the mass, of a further 

 reduction of temperature. It is generally assumed that in 

 the cooling of a liquid globe of mineral matter, congelation 

 would commence at the surface, as in the case of water ; but 

 water offers an exception to most other liquids, inasmuch as it 

 is denser in the liquid than in the solid form. Hence, ice floats 

 on water, and freezing water becomes covered with a layer of 

 ice, which protects the liquid below. With most other matters, 

 however, and notably with the various mineral and earthy com- 

 pounds analogous to those which may be supposed to have 

 formed the fiery-fluid earth, numerous and careful experiments 

 show that the products of solidification are much denser than 

 the liquid mass ; so that solidification would have commenced 

 at the centre, whose temperature would thus be the congealing 

 point of these liquid compounds. The important researches 

 of Hopkins and Fairbairn on the influence of pressure in aug- 

 menting the melting point of such compounds as contract in 

 solidifying are to be considered in this connection. 



It is with the superficial portions of the fused mineral mass 

 of the globe that we have now to do ; since there is no good 

 reason for supposing that the deeply seated portions have in- 

 tervened in any direct manner in the production of the rocks 

 which form the superficial crust. This, at the time of its first 

 solidification, presented probably an irregular, diversified sur- 

 face from the result of contraction of the congealing mass, which 

 at last formed a liquid bath of no great depth, surrounding 



