324 Shepherd, Rankin, Wright — Binary Systems of 



On heating crystal aggregates obtained from the Si0 2 melts, 

 also from Si0 2 glass and from inverted quartz powder, it was 

 observed that at about 175°, or slightly higher, they become 

 isotropic and remained so at still higher temperatures. This 

 behavior proved them to be cristobalite and not tridymite, as 

 had been heretofore supposed. 



Several preliminary tests of the heat change involved in this 

 immersion have been made and found to be clearly marked. 

 The investigation of the stability relations between the crystal 

 quartz, tridymite and cristobalite has not yet been completed, 

 and need not therefore be discussed at this point. 



AlJSW h : Sillimanite ; Al % O t 62'85 per cent, SiO, 37' 15 

 per cent. — This compound crystallizes from the melt with 

 great rapidity on cooling, and as a result the preparations 

 available for the optical work are in general too fine-grained 

 for accurate determination. The crystallites are fibrous and 

 lath-shaped in habit, and, like the natural mineral sillimanite, 

 are often in close parallel groups. The optical effect is, there- 

 fore, usually that of an aggregate rather than of a single indi- 

 vidual. End views of the fiber bundles show that each lath is 

 prismatic in shape with a prism angle of approximately 90°. 

 In the center of nearly every section a minute inclusion is 

 present, often in the shape of a cross, the arms of which are 

 parallel with the sides. In this respect the sections resemble 

 in a way the chiastolite variety of andalusite. The end sec- 

 tions are weakly birefracting and extinguish parallel with the 

 diagonals. The refractive indices, which were measured by 

 the immersion method, are noticeably lower than those of pure 

 natural sillimanite, a fact for which no explanation has yet 

 been found. a=l-638±'003; /3=l-642±-003;7 = l-653±-003. 

 Direct measurements of the birefringence were made and 

 averaged roughly j — a about '014, but they were not satisfac- 

 tory owing to the fibrous character of the material and con- 

 sequent lack of transparency for good thickness measurements. 

 For the same reason the optic axial angle could not be meas- 

 ured satisfactorily. Judging by its general appearance, 2E 

 lies between about 40° and 75°, but unfortunately it was not 

 possible to obtain a more definite value with the material at 

 hand. The optical character is positive, and the acute bisec- 

 trix c lies in the direction of elongation of the fibers. On one 

 preparation of the composition A1 2 3 60, Si0 2 40, the crystalliza- 

 tion was somewhat coarser, and there the optical axial angle 

 was measured with the double screw micrometer ocular and 

 the value 2y=45°±4 obtained. Dispersion of the optic 

 axes w r as not observed, and if present is slight. In this prepa- 

 ration well-marked pinacoidal cleavage was also observed 

 parallel with the plane of the optic axes. The same cleavage 

 is characteristic of natural sillimanite. 



