Diopside — Forsterite — Silica. 229 



Crystallization when perfect equilibrium obtains will first be 

 considered. 



In the case of two substances which form a eutectic, the 

 crystallization of any mixture can be fully described when the 

 two freezing-point curves have been experimentally deter- 

 mined ; or, if the two substances form a complete series of solid 

 solutions like the plagioclase feldspars, the crystallization of 



Fig. 15. 



Fig. 15. Showing three-phase-boundaries for the phases pyroxene, silica 

 and liquid (true scale). 



any mixture can be described if solidus and liquidus have been 

 experimentally determined. So in the present system the ex- 

 perimental determination of boundary curves, isotherms and 

 three-phase-boundaries makes possible a discussion of the crystal- 

 lization of any mixture. Such a discussion is not theoretical, but 

 is based directly on the experimental results. It is true that 

 with a charge of a given composition the exact course of crys- 

 tallization, necessary when complete equilibrium obtains, can not 

 in some cases be experimentally realized because the charge can 

 not be cooled slowly enough, but, by making quenching experi- 

 ments at various temperatures, any desired stage of the process 

 can be experimentally realized. The discussion of crystalliza- 

 tion when perfect equilibrium is attained is important because 

 a very close approach to this behavior would be furnished by 

 a large body of material when very slowly cooled and kept 

 thoroughly mixed. 



In order to make it possible to follow the geometrical rela- 

 tions it is necessary to distort the figure somewhat as has been 



