CLASSIFICATION OF IGNEOUS ROCKS 273 



modified by the inclusion of titanium oxide, which would merely 

 raise the upper limit. There are, however, good reasons for 

 placing it beside silica for classificatory purposes. The total of 

 the silica, alumina, and alkalies, the essential chemical con- 

 stituents of the more acid rocks is approximately complementary 

 to that of the iron, magnesium, and calcium oxides, and varies 

 in a similar fashion. 



The reason for these wide variations in the chemical com- 

 position of the classes is not far to seek. They must be 

 attributed to the system by which the oxides are combined to 

 form the minerals of the norm, and these are divided arbitrarily 

 between the salic and femic groups. The result is that a 

 considerable amount of silica is .combined with the ferrous, 

 manganous, magnesium, and calcium oxides, and counts on the 

 femic side, while a large percentage of lime is often transferred 

 in anorthite in the opposite direction. 



The amount of silica carried to the femic side by the metallic 

 bases is not proportional to the amount in which they are 

 present. In acid rocks they usually form more acid silicates, and 

 give a femic character to a larger proportion of silica than in 

 basic rocks. The position of lime is peculiar. If there be 

 enough alumina to form anorthite, the lime as well as the alumina 

 and silica combined with it are salic, but if no alumina be 

 available, both lime and silica are femic. Alumina has therefore 

 an influence on the classification of rocks far beyond the amount 

 in which it is present, for it may change from one side to the 

 other lime and silica equal to 17 of its own weight. 



The state of oxidation of the iron also materially influences 

 the result ; which is obviously undesirable, in view of the fact that 

 it is not always easy to ascertain the extent of the oxidation that 

 existed at the time of the'consolidation of the rock. In the calcula- 

 tion of the ideal constitution or norm, the ferric oxide is either 

 combined with ferrous oxide to form magnetite or, if there is not 

 enough ferrous oxide, remains uncombined as haematite. 1 If, 

 therefore, there is much ferric oxide, a large proportion, or, it may 

 be, the whole, of the iron takes no silica with it to the femic side, 

 while if the iron be all present as ferrous oxide it will take over 

 from "42 to "83 of its own weight of silica, so that the femic total 

 will be considerably increased. 



1 With the exception of any that may be needed to unite with soda and silica to 

 form " acmite " molecules. 



18 



