SYSTEMATIC PETROGRAPHY 461 



based upon the " fundamental characters " whose application has 

 been reviewed, 



In practice, Fouque and Michel-Levy give the first importance 

 to the colorless constituents, quartz, feldspars, feldspathoids, etc., 

 to produce series within which the ferromagnesian minerals are 

 used to make subdivisions. The existence of certain phenocrysts 

 is recognized, in the naming of rocks, in a few cases only. 



The petrographic system for eruptive rocks elaborated by 

 Fouque and Michel-Levy in 1879 has remained the system of 

 France to the present time, with but slight change. 



In 1889 Michel-Levy compared the results of their system 

 with those of the Rosenbusch system, soon to be discussed, in a 

 work entitled Structures et classification des roc he s eruptive s. 1 This 



'Paris, 1889, pp. 93. 

 discussion presented no new propositions of note excepting a 

 plan of expressing the structure and mineral composition of any 

 given rock by a formula. The principles which must govern 

 the classification of eruptive rocks are concisely stated as follows : 



It is neccessary to base the classification and nomenclature of rocks upon 

 positive facts, independent of all hypothesis. Modern petrography possesses 

 the means to accomplish this end, since the principal structures or modes of 

 association of the minerals are well known and the minerals themselves may 

 be determined with precision. It is, then, exclusively structure and mineral 

 composition which must be relied upon in the classification and nomenclature 

 of rocks. 1 



It is to be remarked, once more, that chemical composition 

 is not taken into account by Michel-Levy, either directly or 

 indirectly, since the partial mineral composition used by him in 

 classification is clearly not an expression of the chemical com- 

 position of the magma nor of any definite part of it. The sub- 

 stance classified is not the rock but merely that variable portion 



'"Notre conclusion, .... est qu'il faut baser la classification et la nomenclature 

 des roches sur des faits independants de toute hypothese, et de nature positive. 

 La petrographie moderne dispose de moyens sufnsants pour atteindre ce but sans hesi- 

 tation : on est d'accord sur les principales structures d'association des mineraux des 

 roches ; on sait determiner ces mineraux avec precision. C'est done exclusivement 

 sur la structure d'association et sur la composition mineralogique que nous persisterons 

 a nous appuyer pour classer et nommer une roche." Loc. cit., p. 87. 



