THE BEHAVIOR OF INCLUSIONS IN IGNEOUS MAGMAS 545 



when amounts are added in excess of those mentioned will a 

 new crystalline phase appear and new fractionation possibilities 

 enter. 



There are, however, certain mixtures in the system that will 

 immediately present a new phase upon addition of the slightest 

 amount of CaO, AI2O3, or SiOj. Such is the mixture {B) which 

 consists on consolidation of anorthite and wollastonite, one half of 

 each. The addition of either CaO or AI2O3 will bring in the new 

 phase, gehlenite, and of SiOz the new phase, tridjnmite, and each will 

 change the course of crystallization. This is because the composi- 

 tion chosen is a limiting case in the ternary system and is in reality 

 of only two components; and the addition of, say, lime carries it 

 out of the two component system. It will be noted that the mix- 

 ture considered contains the three oxides, though of only two com- 

 ponents, and on consolidation only two soHd phases are formed. It 

 might seem on first thought that this corresponds with the case of 

 the natural magmas, for these usually form solid phases fewer in 

 number than the oxides present. However, there is another factor, 

 namely, solid solution, that may give rise to this pecuHarity and 

 we shall find that it is to solid solution that the limited number of 

 phases formed from magmas is usually to be referred. 



Let us now examine a ternary system of oxides that has been 

 completely investigated and in which the factor of solid solution 

 enters. Such is the system CaO iMgOiSiOz studied by Ferguson 

 and Merwin."" A liquid of composition ^ , Figure 10, forms on com- 

 plete consolidation just two soUd phases, the olivine, forsterite, and 

 clinppyroxene of composition between diopside and MgSiOj. To 

 this liquid any amount of calcite up to 12 per cent of its weight 

 could be added without changing the mineralogy of the consolidated 

 product. This would still consist of olivine and clinopyroxene but 

 the pyroxene would be richer in CaO, that is, closer to diopside. 

 Only an amount of calcite in excess of 12 per cent would bring in 

 another phase, akermanite, together with the diopside and oUvine. 

 It is easy to see from an inspection of the figure that addition of 

 dolomite would likewise have no effect on the kind of phases crystal- 

 lizing unless more than about 16 per cent were added. 



'^Amer. Jour. Sci., Vol. XL VIII (1919), p. 109. 



