CLASSIFICATION OF IGNEOUS ROCKS 66 J 



The remaining oxide molecules are to be distributed among 

 the minerals with fixed molecules, whose quantity, however, is 

 undetermined, by assigning to each mineral its proper oxide 

 molecules in the proportions in which they occur in the ideal 

 mineral molecule, as, for example; for orthoclase, iK 2 0. 

 iAl 2 3 .6Si0 2 , and further, by indicating the number of mole- 

 cules of each mineral by an algebraic symbol or letter. Thus, if 

 there are x molecules of orthoclase, there must be assigned to 

 them ^K 2 O.^Al 2 3 .6^Si0 2 . The sum of all these assigned 

 oxide molecules must equal the total amount of a particular 

 chemical component in the rock after deducting that belonging 

 to the minerals whose composition and quantity have been deter- 

 mined. 



If there are more variable quantities than equations, that is, 

 in general, more kinds of minerals than chemical components, it 

 is necessary to reduce the number of unknown quantities by fix- 

 ing the relative amounts of several minerals, or by stating their 

 actual amounts. 



THE CALCULATION OF ALFERRIC MINERALS. 



As pointed out in the discussion of the classification of 

 igneous rocks on a basis of salic and femic minerals, the reasons 

 for omitting" the alferric minerals are their variable and complex 

 composition, our inadequate knowledge of the chemical charac- 

 ter of the amphiboles and micas, and the inconstancy of their 

 crystallization from magmas of any given chemical composition. 



The questions naturally arise, How may these minerals be 

 introduced into the calculation of the mineral composition of 

 rocks from rock analyses ? And what modifications of the norm 

 would follow their introduction ? 



These questions, though not important for the classification 

 of rocks according to the system here proposed, are of interest 

 because of their relation to the general problem of chemico- 

 mineralogical classification and the possibilities of its assuming 

 a more elaborate form than that given it by us. The discussion 

 of them emphasizes the relations between the aluminous ferro- 



