480 WHITMAN CROSS 



rected in the publication to be discussed below. These so-called 

 types were assumed by Loewinson-Lessing" to correspond more 

 or less closely to certain commonly known rock groups. 



The more elaborate discussion of the chemical relationships 

 of igneous rocks by Loewinson-Lessing, presented to the Inter- 

 national Geological Congress at St. Petersburg, in 1897, and 

 published two years later in the Compte- Rendu, deserves some- 

 what fuller consideration. 1 A brief statement of the author's 

 point of view is desirable before explaining the system proposed. 



In reviewing the applicability of various factors in producing 

 a rational system, Loewinson-Lessing asserts that the mineral 

 composition of a rock is a function of its chemical composition. 

 The exceptions to this rule admitted by him are of little impor- 

 tance. Then follows the further statement that the principle or 

 characteristic of mineral composition as a basis of classification 

 is faulty and unsatisfactory because it does not show the relative 

 abundance of the minerals in the various rocks. That is, how- 

 ever, as it appears to the writer, not the fault of the principle, 

 but of the manner in which it has been applied in existing sys- 

 tems. If mineral composition is a function of the chemical 

 composition, it is just as capable of expressing the constitutional 

 relations of rocks as the latter, if properly used. The real 

 objection to its application in the quantitative way, necessary to 

 this expression, is simply one of practicability. The problem is 

 too complex. 



As for his own system, Loewinson-Lessing starts from the 

 idea that eruptive rocks may be considered as silicate rocks and 

 classified as such. Whatever the facts as to predominance of 

 silicates in these rocks may be, it seems to the writer that this 

 conception is not complete as to its basis of fact, and is thus 

 inadequate to serve as a means of classification. Further funda- 

 mental propositions enunciated by the author are that: (1) sili- 

 cate rocks should be classified by the same artificial means as 

 the silicates themselves; (2) while rocks are not stoichiometric 



1 F. Loewinson-Lessing, " Studien uber die Eruptivgesteine," Compte-Rendu de 

 la VII session du Congres Geologique International, Russie, 1897, pp. 193-467. 



