656 HENRY S. WASHINGTON 



Such products are, however, very exceptional, and are only 

 to be expected in cases of very complete differentiation. In 

 general we would only look for sharp upward inflexions of the 

 (FeO) and MgO curves, which would yield, with the slowly 

 dropping silica and the high CaO, a pyroxenite rich in magne- 

 tite. This is just the character of the jacupirangite of Magnet 

 Cove (I), and of those of Brazil and Alno, and I have indicated 

 its connection with the others accordingly by the dotted lines to 

 the left. 



Inasmuch as the specimens of this come from a small iso- 

 lated mass outside the main area, its relations to the other types 

 are uncertain, and its diagrammatic position has been given on 

 the basis of its silica content. It seems to be probable that if, 

 as is likely, such a rock is connected genetically with the others, 

 its abscissal position should be considerably more to the left. 

 As this rock is met with in the immediate vicinity of the main 

 area, and is a theoretically possible differentiation product of the 

 magma, it seems reasonable to assume that the section at Magnet 

 Cove cuts the mass some distance above the center, and that 

 below the biotite-ijolite is a core of jacupirangite, as previously 

 supposed. 



It is obvious from the theoretical discussion, as well as from 

 observations here and at similar resfions, that the relations toward 

 the basic end are far more complex than at the acid end. This 

 arises from the fact that the oxides involved here are capable of 

 more numerous mineralogical combinations, and also because 

 elements which are only present to a small extent in the body 

 of the magma may here assume proportions of great impor- 

 tance. The fact that these extreme basic differentiation prod- 

 ucts are far more common than the purely siliceous ones may 

 be ascribed to the greater fusibility of magmas of a basic charac- 

 ter, and the consequent possibility of differentiation among 

 them at temperatures when the more acid end of the series is 

 solid. 



In this connection attention may be called to the fact, 

 analogous to the segregation of TiOg and PgOg at Magnet Cove 



