226 JOSEPH P. ID DINGS 



present. The silicates capable of forming from CaO, MgO and 

 FeO are orthosilicates and metasilicates ; olivine, 2(Mg,Fe)0. 

 Si0 2 , with its theoretically possible extremes 2FeO.Si0 3 

 (fayalite) and 2MgO.Si0 2 (fosterite); orthorhombic pyroxenes, 

 enstatite, MgO.Si0 3 , and hypersthene, (Mg,Fe) O.SiO s , and the 

 monoclinic pyroxenes, diopside, CaO.Mg0.2Si0 2 , and salite p 

 CaO.(Mg,Fe)0.2Si0 2 . It is possible to introduce into the 

 composition of a magma such as already postulated one or more 

 of these molecules in such proportions as to make the sum of 

 all equal to ioo. But it is clear that the development of a 

 silicate of any of these kinds in the magma would affect the dis- 

 tribution of the silica among the other molecules, that is, it 

 would affect the relative porportions of albite and nephelite or of 

 albite and quartz. It is possible to discover by simple algebra 

 what would be the range of percentages of silica and the range 

 of the alkali-silica ratio for a series of rocks composed wholly of 

 albite and any one of the calcium, iron, magnesium molecules just 

 mentioned. This range will be expressed by a curved line on 

 the diagram extending from the point A to the point correspond- 

 ing to the locus in the diagram of the pure molecule under con- 

 sideration. Several of these lines are indicated on Diagram I. 

 The most extreme case, most favorable to the presence of albite 

 is that in which it occurs in combination with 2FeO.Si0 2 , 

 orthosilicate of iron. The range is indicated by line AFa, which 

 shows the alkali-silica ratios necessary in order that a rock 

 should consist wholly of albite and Fe 2 Si0 4 . Any rock occur- 

 ring above this line and containing only the constituents Na 2 

 = A1 2 3 , FeO,Si0 2 would have relatively too little silica to 

 form albite out of all the soda and alumina and would contain 

 some nephelite. That is, all rocks of this kind occurring above 

 the line AFa would carry nephelite, even those having alkali- 

 silica ratio =.03, if Si0 2 = 31. Conversely, all rocks occurring 

 below this line would have relatively too much silica to convert 

 all the soda and alumina into albite and there would necessarily 

 be quartz, if FeO alone formed an orthosilicate, which is the 

 case in hand. Quartz would thus be a necessity in rocks with 



