Andersejx — System AnortMte-Forsterite-Silica. 447 



position An 8, En 89, SiO a 3 has a crystallization curve of this 

 type. Table 9 shows the phase relations at the principal points 

 of the curve. All mixtures within the Held E F M have 

 similar crystallization curves. Those within the minor field i 

 M E crystallize exactly as described. For those within iFE 

 there is the difference that the second step of the crystalliza- 

 tion consists in simultaneous separation of silica and clinoen- 



Fig. 12. 















































Fc 



















































^^ 



















FORS 



TERITE 



^\ 



<s ^ 











ANORT 



HITE 





CLINC 



5-ENSTATITE 





x< 













LIQUID 



1300 



& 



LIQUID 





X 



V 





-^ 



y 

















\ 



^U- 



^-^ ANORTHITE 



FORSTE 



?ITE a 



LIQUID 























1200 







CLlNO-ENSTAT 

 1 



TE 



\ ANORTHITE 

 1 









MgSiO 



WT PER CENT 



Fig. 12. A section through the ternary system anorthite-forsterite-silica 

 at AE fig. 9. This figure is, therefore, a convenient projection upon a plane 

 of ternary relations, not a representation of a binary system. 



statite instead of anorthite and clinoenstatite. The final result 

 is always eutectic crystallization at N. 



The section anorthite-clinoenstatite. — In the preceding dis- 

 cussion mixtures along the conjugation line A E (fig. 9) have 

 been repeatedly mentioned. It is evident that the final solidi- 

 fication product of all these mixtures must be the same as if 

 they belonged to a true binary system anorthite-clinoenstatite. 

 The relations of the phases during the crystallization can, how- 

 ever, not be expressed in simpler form than in terms of the 

 ternary system, and a two-component diagram must be some- 

 what modified to express the same relations. Fig. 12 repre- 

 sents such a diagram based on the results of the quenchings 

 contained in Table 5. This diagram is, of course, simply a 

 vertical cross section of the ternary diagram along the conju- 

 gation line A E. As such cross sections sometimes may be 

 useful in the study of ternary systems the present diagram is 

 given as an illustration, although it does not actually show 

 relations that do not appear from the ternary diagram. 



