130 



Day and Allen — Isomorphism and Thermal 



which they form the ordinates, while the percentage composi- 

 tions of the different feldspars form the abscissas (fig. 12), we 

 discover, within the limits of accuracy of possible measure- 

 ment at these temperatures, a nearly linear relation : the melt- 

 ing point varies very closely with the composition. We have 

 no maximum, no minimum, no branching of the curve, but 

 from each fusion there separates a solid phase of the same 

 composition as the vitreous matrix. In Ab,An 5 it will be 

 remembered that this was proved by the separation and analysis 

 of the two phases ; in AbjAn 2 partial crystallization was 

 accomplished in the tirst cooling and the remainder in a subse- 



An 

 Ab 



1600° 







































































































































































1400 ■ 



















































































1300 - 































*>», 



« 













































"^ 







1200 









































k ^, 











































1100° 









T^ 







* 







— *r 











K 



K 













An AbiAn 5 AbiAar 



100 84-1 68-0 



159 32-0 



Percentage Composition. 



AbiArii 

 51-5 



48-5 



Fig. 12. 



65-3 



73-9 



An- 





 100 



quent reheating and cooling, the two groups of crystals prov- 

 ing optically identical ; a small quantity of Ab 4 An 1? which 

 admits of absolute identification optically, was crystallized out 

 of a melt of that composition and readily recognized. More- 

 over, evidence to show that the same phase always separated 

 was likewise presented. 



Stated in this way, the relation appears to be a simple addi- 

 tive one in which liquid and solid phases of like composition 

 are stable in all proportions of the components and behave like 

 a series of separate feldspars. But as soon as we consider it 

 with reference to the laws of solution and the phase rule, it 

 can not be explained in this simple way. 



First of all, the phase rule tells us at once that we can have 

 no true compound here between the components albite and 

 anorthite, for such a compound would mean one more com- 

 ponent and an additional phase in every solution before equi- 

 librium could be established. Moreover, if the mixture had 



