REPORT OF THE CHIEF ASTRONOMER 789 



SESSIONAL PAPER No. 25a 



with the writer's general theory. The theory is evidently opposed to Rosenbusch's 

 Kern hypothesis, or any other which postulates (post-Keewatin) primary magmas 

 other than the basaltic. The writer is elsewhere presenting the evidence for 

 the belief that the alkaline rocks are all more or less differentiated syntectics.* 

 Their alkalies are regarded as having been derived from primary basalt; less 

 often from masses of the earth's acid shell, which has been assimilated by 

 primary basalt; or, in still less degree, from assimilated sediments. The princi- 

 pal cause for the special concentration of alkalies is found in the assimilation of 

 limestones and dolomites, or other calcareous sediments. The solution of a few 

 other types of rock may produce the same effect. Or, finally, it is conceivable 

 that the addition of foreign gases to the magmatic solution may give the proper 

 conditions for the concentration of alkalies in limited masses of rock. 



This conclusion was first reached after an inductive study of the field 

 association of nephelite syenite. It was found that, with very few exceptions, 

 nephelite syenite and its effusive equivalent, phonolite, only occurred where sub- 

 alkaline magma had cut important limestone or dolomite formations. In 

 general, the original subalkaline magma was of basaltic composition. This is 

 conceived to have been fluxed by the solution of the carbonate. The new lime 

 or magnesia must fix silica to the extent of several times the weight of either 

 base. The molecule thus formed is normally a pyroxene, which, with the like- 

 wise early-formed magnetite, olivine, etc., will tend to settle out of the magma. 

 It is not essential to determine whether gravity acts before or after the actual 

 crystallization of ferric, ferromagnesian, and cafemic components. The residual 

 magma is necessarily higher in alkalies than the primary basalt. The fixing of 

 silica by the new lime and magnesia means a desilication of the rest of the 

 magma, and nephelite or leucite forms instead of feldspar. Since little alumina 

 enters into the sunken components, this oxide may be in excess, and corundum 

 will ultimately crystallize in the residual magma. Meanwhile, the carbon dioxide 

 set free from the dissolved carbonate tends to rise through the magma. It may 

 possibly carry with it soda and potash in combination as the alkaline carbonates. 

 These familiar fluxes could rapidly enrich the upper part of a lava column with 

 either or both alkalies and thus furnish a leucite basalt, a leucitite, or a 

 nephelinite, from which the ferromagnesian and cafemic constituents of nor- 

 mal basalt have not had time to settle out. If the separation becomes perfect 

 a phonolite or leucocratic nephelite syenite is the salic pole, while limburgites 

 form the femic poles. 



The hypothesis cannot be fully presented in this report but perhaps enough 

 has been stated to show its nature. It explains the remarkably common asso- 

 ciation of alkaline rocks with calcareous and magnesian sediments; the desilica- 

 tion of primary magma, as indicated by the presence of nephelite and leucite 

 in many alkaline types; *the common supersaturation of alkaline magmas with 

 alumina, resulting in the crystallization of corundum; the common occurrence 

 of primary calcite, cancrinite, melanite, melilite, scapolite, wollastonite and 



* Bull. Geol. Soc. Amer., Vol. 21, 1910, pp. 87-118. 

 25a — vol. iii — 51£ 



