Andersen — System Anorthite- For 'sterite- Silica. 13 9 



discussed a little more in detail, particular emphasis being laid 

 upon crystallization by cooling with maintenance of equili- 

 brium. 



The general character of the true ternary part of the dia- 

 gram is that of a system with no ternary compound, one binary 

 compound, MgSi0 3 (point E, fig. 9), unstable at its melting 

 point, and two quintuple points of which the one (M) is an 

 alteration point (An : Fo : En) and the other one (N) a ternary 

 eutectic (An : En : Tr). 



The system is not essentially different from that described 

 as Type 4 in the previous general discussion. The only devia- 

 tions from the type are the occurrence of the spinel field and 

 the sub-division of the silica field into two, the field of tridy- 

 mite and that of cristobalite. 



In the following the determination of the quintuple points, 

 boundary curves and isotherms of the ternary part of the sys- 

 tem will be briefly reviewed. The relations of these points 

 and curves will appear from the discussion of the course of 

 crystallization in different types of mixtures. 



The alteration point. (Point M, fig. 9.) — The point M rep- 

 resents the junction of the three fields of anorthite, forsterite 

 and clinoenstatite and corresponds to the point o, fig. 7. At 

 this point the three solid phases mentioned are able to exist in 

 equilibrium together with the liquid. As a result of any 

 energy change in the system there will be a reaction between 

 forsterite, clinoenstatite and liquid. This reaction will be dis- 

 cussed further on. 



The temperature of the point M was first determined by the 

 heating curves described on p. 409. As the quenchings proved 

 that no eutectic An : En existed, the only way of explaining the 

 heat changes was to assume that they corresponded with the 

 temperature of the alteration point M. The quenchings cor- 

 roborated the results of the heating curves, all determinations 

 together fixing the temperature of M at 1260° d= 2°. 



The composition of M was first roughly extrapolated from 

 different quenchings in the three fields of anorthite, clinoen- 

 statite and forsterite, and then several mixtures near the com- 

 position found by extrapolation were successively examined as 

 to their primary phase and the temperature of their liquidus 

 points. In this way the composition of M was determined at : 



An 55, Fo 25-5, Si0 2 19-5.* 



The ternary eutectic anorthite- clinoenstatite-tridy mite. 

 (Point N, fig. 9.) — All mixtures near the ternary eutectic 

 formed very viscous melts. Heating curves run to determine 

 the eutectic temperature on mixtures presumably near the 



*Si0 2 54-20 ; A1 2 3 20-16 ; CaO 11 -05 ; MgO 14'59. 



A.M. Jour. Sci.— Fourth Series, Vol. XXXIX, No. 232.— April, 1915. 

 29 



