238 CARNEGIE INSTITUTION OF WASHINGTON. 



Carbon dioxide would be held some appreciable time before oxy- 

 gen, and still longer before nitrogen, and all these a notable time 

 before the vapor of water. The inference is that the initial atmos- 

 phere was very rich in carbon dioxide, for an abundant supply was 

 correlated with a superior power of retention. 



The amount of oxygen in the earh^ atmosphere is more uncertain 

 from doubt as to a competent source of supply. Crystalline rocks 

 and meteorites are not known to contain it in a free state. As 

 above remarked, it occurs among volcanic gases, but it is not known 

 that it comes from the deep interior. It is detected in the sun and 

 not improbably existed in the nebula, from which it might have 

 been gathered shortly after the accretion of carbon dioxide began. 

 The safer inference seems to be that it was not very abundant rela- 

 tiv^ely in the early atmosphere. This inference may be entertained 

 the more freely because it seems to give the better working hypoth- 

 esis, for the present large proportion of oxygen may be assigned to 

 the reduction of carbon dioxide by plant action, and the present 

 proportions and those of geologic history seem to come out best on 

 this basis. For the primitive atmosphere there is theoretical need 

 for onl}' enough oxygen to support the primitive plant life until it 

 could supply itself, after which it would produce a surplus. 



The amount of nitrogen occluded in rocks and meteorites is rela- 

 tively small, and it was perhaps a small constituent of the earlj^ 

 atmosphere. Owing to its chemical inertness, it may be supposed 

 to have been increasing ever since, and thus to have attained its 

 present dominance. A similar history may be assigned to the other 

 and even more inert elements, argon, neon, zenon, krypton, and he- 

 lium, of which the supplies seem to have been always very limited. 



After the earth acquired the power of holding water-vapor, the 

 supply being abundant, accession doubtless went on for a time as 

 fast as the capacit)'^ to hold increased. 



The problem of vulcanism assumes a quite new aspect under the 

 planetesimal hypothesis, if very slow accretion without ver>' high 

 temperature be assumed. It has been taken for granted in the pre- 

 ceding statement that there was volcanic action. It is necessary, 

 therefore, to consider how volcanic action may have arisen, and this 

 involves the more radical question how the high internal tempera- 

 tures of the earth ma}- have arisen if the earth did not inherit its 

 heat from a molten condition arising from a gaseous origin. 



The total amount of heat produced by the infall of the planetesi- 

 mals would undoubtedlv be more than sufficient to melt the whole 



