338 HISTORY OF GEOLOGY AND PALEONTOLOGY. 



great depths previously to any surface eruption of the magma; 

 whereas the younger and minute mineral elements probably 

 originated during the epoch of eruption. With the outflow 

 of a glowing rock-magma at the surface, and the escape of the 

 water vapours, the chemical constitution of the rock-material 

 is changed. The structure of the ground-mass is holocrystal- 

 line, hemicrystalline or glassy, according to the more rapid or 

 slower cooling of the magma. 



Rosenbusch sub-divides the rocks of eruptive flows into 

 palceovolcanic (porphyry, porphyrite, augite porphyrite, mela- 

 phyre, and picrite porphyrite) and neovolcanic (liparite, trachyte, 

 phonolite, andesite, basalt, etc.). Some of the older flows, 

 such as the diabase porphyrite and picrite porphyrite, resemble 

 granitic-porphyritic intrusive rocks so closely that they seem to 

 bear the same relationship to them which the typical intrusive 

 rocks bear to the plutonic or deep-seated masses. They may 

 be distinguished from true eruptive flows by the absence of 

 tuffs. 



The new classificatory scheme of Rosenbusch showed quite 

 clearly that he had been strongly influenced by the views of 

 MM. Fouque and Michel-Levy, and these two French petro- 

 graphers felt it incumbent to declare the position they assumed 

 towards it. In 1889, Michel-Levy discussed the work of 

 Rosenbusch in a special memoir, agreeing with many of its 

 principles, but disputing others. Regarding the sub-division 

 of the eruptive rocks into deep-seated masses, intrusions, 

 and flows, Levy points out that the intrusive group is quite 

 artificial and untenable, as intrusions may either take the 

 form of narrow vertical dykes or almost horizontal sheets or 

 "sills." He also contests the conclusion of Rosenbusch that 

 only one generation of the crystalline constituents took place 

 in the deep-seated rocks, a group which almost precisely 

 corresponded with the granitoid group of MM. Fouque and 

 Levy. 



In Michel-Levy's opinion, the geological aspects and associa- 

 tions of the eruptive rocks, as well as the geological age, have 

 too little connection with the structure of the rocks to provide 

 a good basis of classification. Michel-Levy cites cases where 

 rocks belonging to the "deep-seated group" of Rosenbusch, 

 e,g., granite, ophite, and gabbro, occur in the form of eruptive 

 sheets. According to Michel-Levy, the different types of 

 structure in eruptive rocks are due to variations of temperature 



