136 NATURAL SCIENCE. Feb.. 



Neither is the idea of the segregation and settHng down or 

 deposition of certain crystals a very comprehensible one. It is hard to 

 deny that such a process might take place, but it seems highly doubtful 

 that any but very large crystals could undergo sinkage in a highly 

 viscous medium ; besides, it is not true that minerals separate 

 in the order of their basicity or of their specific gravities, a fact I 

 pointed out several years ago.- Leucite, for instance, is only formed 

 in the open chimney of a volcano, and depends, like some of the augite, 

 with probably sodalite, haiiyne and nepheline (in some cases only), 

 on the liberation of the alkalies and alkali-earths from their chlorides 

 and sulphates by the dispersal and escape of their acid radicle 

 vapours. These bases can then take up the silica from the residual 

 basic glass, thus forming leucite and magnetite. 



Again, we find that several mineral species become individualised 

 contemporaneously. In many rocks, felspar and magnetite or augite, 

 or again augite and leucite, were certainly separating at the same 

 time ; now if these sink together we shall have a rock of rather 

 curious composition. Brogger and Vogt seem to believe that 

 segregation takes place at certain points of a reservoir where the 

 conditions are more favourable to the crystallisation of the one or 

 the other mineral species. This much appears not improbable, and 

 would easily explain the peculiar iron ores described by Vogt, and 

 many rocks that have cooled in place, but it is hardly applicable to 

 igneous effusive rocks in general. 



Other suggestions that have been offered may be denominated 

 the "fusion hypothesis," and the "osmotic hypothesis." Professor 

 Sollas has demonstrated how, in some Irish igneous rocks, fragments 

 of other igneous rocks of different composition and earlier date have 

 been dissolved in a later magma. 3 Backstrom+ and Fromm^ have 

 likewise described the local changes produced by quartz and felspar 

 inclusions in basic rocks, but their observations are limited to the 

 immediate neighbourhood of the enclosures. The former author found 

 three kinds of augite in the enclosing rock. I have likewise shown 

 that in basic and andesitic rocks of Stromboli quartz is fluxed down, 

 forming with the residual basic glass an augite which becomes an 

 important rock constituent.*^ 



These facts in themselves are not of paramount value in the 

 question, but they are clues to an important line of investigation that, 

 when more carefully pursued, may afford valuable evidence as to 

 change of composition in a primitive paste by adulteration with 

 extraneous additions. 



Ricciardi, on the evidence of his careful and elaborate analyses of 



'^Scient. Proc. R. Dublin Soc, n.s., vol. v., p. 143, et seq. 



3 Rep. Brit. Assoc, 1893. 



* Bihang t. k. Svensk. Vet. Akad. Ilaitdl., vol. xvi., pt. ii., no. i. 



5 Zeitschr. deutsch. Geol. Gcs,, vol. xliii., p. 43. 



^Rep. Brit. Assoc., 1893. 



