Andersen — System Anorthite-Forsterite-Silica. 423 



This system was particularly difficult to work with on 

 account of the great viscosity of the melts. The preparation 

 of the mixtures was a time-consuming operation. The melt- 

 ing and grinding had to be repeated several times to insure 

 homogeneity, and a long time was required to crystallize the 

 glass.* The products obtained from quenchings were ex- 

 tremely tine-grained, even in mixtures containing no more 

 than 20 per cent free silica. 



No actual determination of the silica liquidus could be made, 



Fig. 2. 







































































CRIS1 



OBALITE 



a LIO 



UID 



























P 



























TR 



DYMITE 



a 



LIQ 



UID 









ANORTH 



te a 



LIQUID 















\y^ 









































TRIDYN 



ITE 



a an 



ORTHITE 









1200 























WT PER CENT 



but its general course near the eutectic was inferred from the 

 isotherms in the ternary diagram (fig. 10). 



The eutectic was located in the following way : The direc- 

 tion of the anorthite liquidus was first determined by quench- 

 ings on the mixtures 20 and 30 per cent silica in connection 

 with the known melting point of anorthite. Then quenchings 

 were made to determine the temperature of the eutectic on 

 the 30 per cent mixture. This temperature was found to be 

 1350°=b5°. From these observations the location of the 

 eutectic could be estimated at 49 per cent silica. A mixture 

 of this composition was then prepared and quenchings carried 

 out for a more accurate determination. The quenchings 

 proved that the mixture was on the silica side of the eutectic 

 (see Table 3), the eutectic temperature being determined at 

 1353°zb2°, and the point on the liquidus at 1364°±4°. These 

 points in connection with the approximate direction of the 

 silica liquidus determined the composition of the anorthite- 

 silica eutectic to be An 52, Si0 2 48. 



*The mixture An 51, SiO a 49, for instance, required eight days for com- 

 plete crystallization at a temperature around 1200°. 



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



28 



