Andersen — System Anorthite-Forsterite- Silica. 417 



(010). Optical properties :* 7 = 1-670 ; = 1*651 ; a = 1-635 ; 

 T a = 0-035 ; 2 V = 85°16 / . Optical character positive. 

 Optical orientation c = /3 ; a = 7 ; b = a. 



The crystals formed in the anorthite-forsterite-silica mixtures 

 are often well developed, always short prismatic after the oaxis 

 or thick tabular after (010). The combinations are simple : the 

 prevailing forms seem to be (120) (010) (021) and (101) ; but 

 no reliable measurements could be made. 



The best developed crystals were those included in glass 

 from which, however, they could not be separated. In some 

 of the more viscous melts the crystals were very small, appear- 

 ing as rounded grains, easily determinable, and distinguished 

 from clinoenstatite by their high refractive index, strong double 

 refraction and the lack of twinning. The refractive index was 

 approximately determined in several preparations at 1*65. 

 Optical character positive. Axial angle large, nearly 90°. 

 Distinct cleavage lines were generally seen except on small 

 crystals included in glass. Small glass inclusions, often in 

 symmetrical arrangement, were very common in crystals from 

 various mixtures. 



Gristobalite and Tridymite. 



Cristobalite and tridymite are the two modifications of 

 silica stable within the range of temperature of the silica 

 liquidus in the anorthite-forsterite-silica system. C. N. Fennerf 

 has shown that these two forms are enantiotropic, with an in- 

 version point at 1470°, cristobalite being the high temperature 

 form. The inversion is very sluggish and either of the forms 

 may remain unchanged for a long time in temperature regions 

 where it is unstable. 



The melting point of cristobalite has been determined by 

 Fenner at 1625° and by Endell and R\q\lq% at 1685.° In the 

 J\fgO-Si0 2 system,§ however, there were indications that the 

 melting point might be considerably higher, and in a later 

 work on the diopside-forsterite-silica system Bowen| has shown 

 that the melting point of cristobalite must be higher than any 

 of the values mentioned. This result is fully confirmed by the 

 present work on the anorthite-forsterite-silica system. (See 

 figs. 1, 2 and 10.) A possible explanation of the discrepancies 

 between the different results is indicated by Bowen on the 

 basis of the hypothesis suggested by Fenner^f for the explana- 



*N. L. Bowen and Olaf Andersen, loc. cit., p. 498. 



fThis Journal (4), xxxvi, 331, 1913. 



JK. Endell and R. Rieke, Zs. anorg. Chem., Ixxix, 239, 1912. 



§ N. L. Bowen and Olaf Andersen, loc. cit., p. 496. 



|| N. L. Bowen, this Journal (4), xxxviii, 218, 1914. 



IT Loc. cit., p. 372. 



