218 JV. L. Bowen — The Ternary System : 



The Melting Point of Cristobalite. 



In the paper on the system MgO-SiO a * it was stated that 

 indications were found that the liquidus of cristobalite is very 

 steep close to the eutectic, clino-enstatite-cristobalite. The 

 statement was made in this tentative form because it was 

 considered possible that equilibrium had not been obtained 

 in these rather siliceous mixtures. Later work has, however, 

 shown that equilibrium between liquid and solid is readily 

 attained in these mixtures and that points on the cristobalite 

 liquidus can be fixed with precision. The points found in the 

 diopside-silica system, in the clino-enstatite-silica system and 

 several intermediate points in the ternary system all agree in 

 showing that the silica surface is very steep. 



It is true that a lessening of the gradient is shown at the 

 higher attainable temperatures (note spacing of the isotherms 

 as fixed by determined points fig. 6) so that the silica surface 

 may not mount as high as the steep gradient at lower 

 temperatures would suggest. Nevertheless it is clear that the 

 indicated melting point of cristobalite must be higher than 

 1625°, the value found by Fenner, f higher than 1685° even, 

 the value found by Endell and Rieke4 



The mixture MgSiO a 75 per cent, Si0 2 25 per cent held at 

 1625° for one hour gives glass and considerable cristobalite, an 

 amount which would require a raising of the temperature 40°- 

 50° to effect its solution. This estimate (40°-50°) may be made 

 with considerable assurance on the basis of much experience 

 with quenchings of mixtures in the silica field. Even at the 

 composition 25 per cent Si0 2 , then, the liquidus is well above 

 1625°. I do not know how to reconcile these determinations 

 with the evidences of melting found in cristobalite by Fenner 

 at 1625° and by Endell and Rieke at 1685° unless it is as Dr. 

 Fenner has suggested to me, that cristobalite has a variable 

 molecular constitution and a similarly variable melting-point 

 according to the conditions under which it is formed, a 

 peculiarity which he has shown to be true of its low-temperature 

 inversion. In the various mixtures used in the present 

 work, in which the cristobalite is crystallized simply by cooling 

 the molten mixture, there is no evidence of variability. 

 All the points on the silica surface are entirely consistent 

 with each other, a fact which probably indicates that true 

 equilibrium is obtained between the various molecules which 

 make up the solid phase, cristobalite. 



* Bowen and Andersen, this Journal (4), xxxvii, 487, 1914. 

 t Fenner, C. N., The Stability Relations of the Silica Minerals, this Journal, 

 (4), xxxvi, 381, 1913. 



}K. Endell and R. Rieke, Zs. anorg. Chein., lxxix, 239-259, 1912. 



