Properties of the Feldspars. 131 



been eutectic in character, the component (albite or anorthite) 

 which happened to be in excess would have crystallized out in 

 each case, causing a continual change in the composition of the 

 remaining glass until the eutectic proportion was reached and 

 the resulting charge would have contained only crystals of one 

 (or, in case of hysteresis, both) of the components and the 

 eutectic. Our curve is continuous and the resulting charges 

 homogeneous for all proportions of the components. Lane's 

 suggestion* that the triclinic feldspars from a eutectic series in 

 which the eutectic proportion is at or near Ab 2 An 3 is therefore 

 not borne out by our experiments. 



Laying aside the eutectic mixture, and passing over to solu- 

 tions of components which are miscible in many or all propor- 

 tions, we find a small number of examples, chiefly organic com- 

 pounds, which have been studied as types by Roozeboom, 

 Kiister, Bodlander, Garelli, Bruni, Yan Eyk and others, among 

 wmich our series appears to fall. 



Laws of Solutions. — From the physico-chemical standpoint, 

 the case we now have in hand closely resembles Krister's prob- 

 lem of 1891. f His measurements were made upon mixtures 

 of organic compounds of low melting point, while ours reached 

 a maximum temperature of 1532°, but we have, between albite 

 and anorthite, an exactly similar series of solid solutions the 

 melting points^ of which change in nearly linear relation to 

 the percentage of the two compounds which enter into their 

 composition. 



This simple linear relation was called by Kiister perfect 

 isomorphism, and he formulated the "Rule" which has since 

 borne his name, that the solidifying point of an isomorphous 

 mixture lies on a straight line joining the melting points of 

 the components and can be calculated from the percentage 

 composition of the mixture. If this line proved to be slightly 

 concave or convex, as it did in most cases, imperfect isomorph- 

 ism was assigned as the cause. To this rule an objection was 

 raised by Garelli§ and elaborated by Bodlander] — if the solid 

 solution behaves like other solutions, a small quantity of com- 

 ponent B added to component A can only lower the solidify- 

 ing point of A when the solid phase is richer in A than the 

 liquid phase. The reasoning is this (Bodlander): let x x (fig. 13) 

 be the vapor tension curve of component A in the liquid state, 

 y 1 the solidifying point (t,) of A, and z x the vapor tension 



* Lane, Journal of Geology, XII, ii, p. 83, 1904. 



f F. W. Kiister, Zeitschr. fur Phys. Chem., viii, p. 577, 1891. 



X Kiister measured solidifying points, but we have pointed out above that 

 such measurements lead to no positive result in liquids of such viscosity as 

 the feldspars in which equilibrium is not established during solidification. 

 Undercooling rarely appeared at all in Kuster's cases. 



J5 F. Garelli, La Gazzetta Chimica Italiana, xxvi, p. 263, 1894. 



|| Bodlander, Neues Jahrb. f. Min., Beilage, Bd. xii, p. 52, 1899. 



