fenner: forms of silica 475 



a- into /3-cristobalite. The method chiefly employed in investi- 

 gation was that of heating-curves. 



Inversion of a-quartz into 13-quartz. By optical methods this 

 inversion had previously been placed at 575? The writer con- 

 firmed this, finding a well-defined break in the heating-curve 

 at 576? The change has been regarded as one from tetartohedral 

 to hemihedral symmetry of the crystals. 



Inversion of a-tridymite into $-tridymite. The hexagonal plates 

 of tridymite are birefringent at ordinary temperatures, even on 

 basal sections, but when heated they suddenly become isotropic. 

 The inversion point has been placed by F. E. Wright at about 

 120? 



By the use of heating-curves the writer has found what appear 

 to be two inversions, one at 117° and another, less pronounced 

 but quite sharp, at 162? The significance of this is not readily 

 apparent and more investigation is required. 



Inversion of a-cristobalite into (i-cristobalite. The temperature 

 at which the birefringent, low-temperature form changes to the 

 isotropic state appropriate to the external symmetry of the 

 crystals was placed by Mallard at 180? F. E. Wright, working 

 on spherulites formed by the devitrification of silica glass, found 

 that Mallard's point was too low and determined it as approxi- 

 mately 225? The writer, making use of cristobalite formed 

 in various ways, and employing both optical methods and heat- 

 ing-curves, obtained rather remarkable results. On different 

 preparations the point at which the break occurs on the heating 

 curve has been found as high as 274?0 and as low as 240? 1. On 

 the cooling curve the break occurs as high as 236? 6 and as low 

 as 209? 3. The first results obtained were much higher than 

 the figures of Mallard and of Wright, and the explanation which 

 suggested itself was that the abnormally high temperatures were 

 due to hysteresis or to the presence of impurities in solid solution, 

 but these possibilities were eliminated. It was found, however, 

 that the temperature to which the silica was heated in con- 



