Ferguson and Merwin — The Ternary System. 119 



in his investigation of it used a flux of sodium tungstate. 

 We have studied it in melts of our ternary system. We 

 selected a composition, CaO 24, MgO 7, Si0 2 69, which lies 

 within the silica field and melts completely at 1536 °C. 

 The tridymite and cristobalite present in the various 

 quenches made could not be separated and analyzed, 

 therefore the only evidence that they are little if any 

 affected by solid solution rests upon determinations of 

 refractive index. The values obtained agree within 

 ± 0-003 with those of purest natural crystals and with 

 Fenner's values for material formed in molten sodium 

 tungstate. Our observations are given below. 



Fig. 16. 



2200 . .2200 



2100! .2100 



2000. ""*"" — ^ Melt .2000 



1900. ^-^ .1900 



1800 . \ . 1800 



^\ 2CaO-MaO-2Si0 2 + KUlT 



1700. \. \ .1700 



or 2CaOSiO a + Melt \^ \ 



1600. "< \_1600 



1500. V. \ 1500 



1400 : 7* : 1400 



(300. .1300 



(i 2Co<0 SiO, + 2C«0 MojO 2Si0 2 



1 1 1 1 1 1 1 1 1 



2Ca0 Si0 2 2CaOMojO 2SI0, 



Fig. 16.— The binary system 2CaO.Mg02Si0 2 -2CaO.Si0 2 . wt. per cent. 



1. A charge containing only tridymite plates and glass 

 was prepared by heating the original material which 

 contained cristobalite for 16 hours near 1370° C, the 

 eutectic temperature being 1320° C. Portions of 

 this charge were then given the following treat- 

 ments. 



(a) Heated 5 hours at approximately 1530 °C. 

 The charge then contained much cristobalite. 



(b) Heated 5 hours at- 1515° C and then 10 hours 

 during which time the temperature fell to 

 1500 °C. The charge contained cristobalite; 

 no tridymite could be identified. 



(c) Heated 9y 2 hours at 1496 °C. Much tridymite, 

 no cristobalite identified. 



