ORIGIN OF AUGITE ANDESITE 419 



writer believes in magmatic splitting under other conditions than those 

 at the crater of a basalt volcano. Either hypothesis will, of course, 

 recognize augite andesite as a derivative from olivine basalt; the 

 balance of probability is here attributed to the hypothesis of fractional 

 crystallization. 



SUMMARY AND CONCLUSION 



The purpose of the writer has been to state the results of correlating 

 many scattered items of fact derived from experimental and field 

 studies. The correlation seems strongly to support the early views 

 of Scrope, Darwin, and others as to the efficiency of fractional crystal- 

 lization in the formation of igneous rocks. It is only quite recently 

 that this general hypothesis could be put on a quantitative basis. 

 Even now there are needed many additional physical and chemical 

 deierminaiions before ihe hypodiesis can reach iis full measure of 

 conviction for the petrologis:. Nevertheless a compilation of die 

 already estabHshed facts seems to show that the idea of rock-differ- 

 entiation by the gravitative separation of certain minerals gains 

 greatly in meaning, force, and usefulness when applied to actual 

 rock-types and to actual petrographical provinces. 



The hypothesis explains the origin of a considerable number of 

 igneous rock-types. Augite andesite and many olivine-free basalts 

 form what may be called one pole of the differentiation of olivine 

 basalt. Picrite, limburgite, many peridotites and other ultra-basic 

 types form the other polar group of differentiates. The conditions 

 for the differentiation in the typical and general case, involve, in 

 each case, a somewhat prolonged residence of the primary basalt in a 

 volcanic vent in which the temperature varies from about 1200° to about 

 1050° C. The phenocrysts formed in the lava at these temperatures, 

 must slowly but surely sink. They then collect in the lower part of the 

 lava-column. While still undissolved, they may be erupted along with 

 the fluid lava in which they rest, giving ultra-basic porphyritic lavas; 

 or, as seems more probable, they are slowly dissolved in this lower, 

 hotter part of the lava-column, forming one or more ultra-basic layers 

 which, on injection, crystallize into peridotites or, following extrusion, 

 develop picritic or limburgitic rocks. 



The hypothesis is backed up by a comparison of the average olivine 

 basalt and average augite andesite of the world; by a comparison of 



