590 



JV. L. Bowen — Melting Phenomena of thi 



By applying similar reasoning to component B we get 



R 



a 



where Z 2 = latent heat of melting (molal) of B at T. 



These two equations express the simultaneous concentrations 

 on the liquidus and solidus at any temperature T in terms of 

 the latent heats of melting at T and the temperatures of melt- 

 ing of the two components (T t and T 2 ). 



Van Laar derives a more rigid relation which contains fac- 

 tors for the heats of mixing in both phases, but in its appli- 

 cation he finds it necessary to neglect these and the equations 

 then reduce substantially to the form given above. 



In the present instance T x and T 2 and also the values for x 

 and a?, (the mol-fraction of albite on the liquidus and solidus 

 curves respectively) for various values of T have been experi- 

 mentally determined. (See fig. 1.) It suffices, then, in order 

 to test the applicability of the equations in the case at hand, to 

 substitute in the general equations and see whether or not L x 

 and Z 2 remain nearly constant and of appropriate order of mag- 

 nitude. In the following table the calculated values are given : 









Table IV 









{T-273): 



X 



03! 



1-x 



1 — Xi 



Lab 



Lau 



observed:* observed: 



observed: 



observed: 



calculated: 



calculated 



1500 



•277 



•098 



•723 



•902 



12670 



28600 



1450 



•500 



•192 



•500 



•808 



12750 



30000 



1400 



•651 



•282 



•349 



•718 



12790 



29200 



1350 



•772 



•376 



•228 



•624 



12790 



30400 



1300 



•846 



•475 



•154 



•525 



12470 



28300 



1250 



•905 



•575 

 Mean 



•095 



•425 



12920 



27700 





12740 



29000 







Average deviation from mean, 



0-8$ 



2-9^ 



It will be seen from the table that the values for the latent 

 heats vary but little from a mean value, the random variation 

 being what might be expected from the possible errors of the 

 temperature measurements. Any systematic variation with 

 temperature, due to the difference of specific heats of liquid 

 and solid, would, by analogy with other silicates, be still smaller, f 



If the calculated mean molal latent heat of melting of anor- 

 thite is divided by the formula weight, the result is 29000 -^ 



*See fig. 1. 



fCf. W. P. White, Specific heats of silicates and platinum, tbis Journal 

 (4), xxviii, p. 345, 1909. 



