138 Day and Allen — Isomorphism and Thermal 



I OoA«— Composition-* lOpB 

 Fig. 20. 



near the anorthite end of the 

 series its slight convexity is 

 unquestionably real. 



It should be added that Prof. 

 Iddings has found slight traces 

 of inhomogeneity (less than 1 

 per cent) in the slides of several 

 of our intermediate feldspars. 

 Crystals have been found which 

 were evidently of the earliest 

 formation and with one excep- 

 tion were more calcic than the 

 body of the charge, as Rooze- 

 boom's theory would lead us to 

 expect. The exception was an 

 occurrence of tiny plates of 

 Ab 4 An x discovered in a charge 

 of At^Aiij. The extremely 

 small quantity of the optically different feldspar, the fact that it 

 could not be found in all the slides and that in one case a less 

 calcic feldspar appeared, suggest that the inhomogeneity may 

 have been of other origin — perhaps merely a consequence of 

 the tremendous difficulty in mixing a homogeneous charge 

 where ultraviscosity precludes stirring, for example. The 

 chemical analyses of the solid and liquid phases, it will be 

 remembered, showed identical composition within the limits of 

 experimental error. 



It is clear that if Roozeboom's theory is valid, the line of 

 the melting points can not become perfectly straight unless the 

 f-curves for the solid and the liquid phases can be superposed 

 point for point throughout, i. e., are identical. This would 

 mean that the energy content per gr.-mol. of solid and liquid 

 phase was the same for all compositions, i. e. that all mixtures, 

 and the components separately should have the same melting 

 point, — a case which is known (Roozeboom, d- and /-camphor 

 oxime) but is certainly confined to optical antipodes. 



Another reason for supposing the case to be much less sim- 

 ple than a mere linear relation with equilibrium between solid 

 and liquid phases of identical composition, appears at once 

 from a direct application of the phase rule. A necessary con- 

 dition for equilibrium in any mixture is that the number of 

 phases exceed the number of components by two. If the solid 

 and liquid phases are homogeneous, the number of phases (count- 

 ing vapor) is only three and equilibrium can not obtain there. 

 Reviewing this discussion briefly : The triclinic feldspars 

 are solid solutions and form together an isomorphous series. 

 It is a sufficient condition for the latter that the curve of melt- 

 ing points is continuous (Bruni, loc. cit.). Like Kuster's 



