454 Andersen—System Anorthite— Forsterite-Silica. 



Remarks on spinel-bearing rocks. 



With the object of explaining the early crystallization of 

 spinel in rocks Harker* has suggested in a tentative way that 

 spinel melts were only partly miscible with silicate melts. 

 Under the experimental conditions considered in the present 

 study this does not hold true. We find no necessity for pos- 

 tulating such special limiting conditions. Spinel separates 

 from homogeneous melts and the same melts also yield anor- 

 thite, forsterite, etc., that is, there appears to be no essential 

 difference between the mode of formation of spinel and that of 

 the silicates. We may therefore conclude that the spinel 

 found in rocks is formed by crystallization from a homogeneous 

 magma. 



Summary. 



The experimental methods are first briefly described and the 

 results stated in Tables 1-5. 



The following solid phases were observed : Anorthite, 

 forsterite, cristobalite, tridymite, clinoenstatite and spinel. 

 The thermal and optical properties of these phases are de- 

 scribed. 



Among the binary mixtures those of anorthite and silica 

 form a simple eutectic system, those of forsterite and silica 

 form a system with an unstable compound, clinoenstatite 

 (MgSi0 3 ) and those of anorthite and forsterite form no true 

 binary system, as spinel is the primary phase in some of the 

 mixtures. 



In a theoretic discussion the general qualities of some ter- 

 nary systems without solid solution are reviewed with refer- 

 ence to temperature-concentration diagrams. The quantitative 

 relations of the diagrams are pointed out and types of crystal- 

 lization curves discussed. 



As some of the ternary mixtures anorthite-forsterite-silica 

 contain spinel as a primary phase, the relations of the whole 

 system cannot be properly expressed more simply than in 

 terms of the four-component system CaO-MgO-Al 2 3 -Si0 2 . 

 After a brief discussion of the spinel field the ternary part of 

 the system is described. The system has two quintuple points, 

 one of which is a ternary eutectic, the other an alteration 

 point. Of the boundary curves one is an alteration curve, the 

 others ordinary melting curves. The crystallization curves in 

 different sections of the diagram are discussed. The resorp- 

 tion of forsterite as a result of cooling with equilibrium 

 between liquid and solid is pointed out. Some cases of crys- 

 tallization with lack of equilibrium are also discussed. 



It is shown how the results may be applied to actual rocks, 

 more especially the olivine-bearing rocks. 



Geophysical Laboratory, 



Carnegie Institution of Washington, Washington, D* C, Jan. 25, 1915. 



* Nat. Hist, of Igneous Eocks, p. 200. 



