Alumina with Silica, Lime and Magnesia. 303 



the inversion takes place so slowly as to completely veil the 

 character of the corresponding heat change. 

 A very pure andalusite from Hill City, S. D. : 



Heated 



Temperature 



Eemarks 



48 hours 



1100° 



Unchanged 



62 " 



1150° 



a 



68 " 



1150° 



a 



28 days 



900° 



a 



4 " 



1500° 



Much altered 



results. 







Time 



Flux 



Temperature 



12 hours 



NaCl 



800° 



72 " 



a 



a 



168 " 



a 



a 



48 " 



CaV 2 6 



1000° 



48 " 



a 



900° 



216 " 



a 



900° 



If the transformation is reversible, the change is slow even 

 at temperatures much above the supposed inversion point. All 

 natural andalusite is contaminated with mica which masks 

 the reaction. Using a flux in the hope of getting the trans- 

 formation at lower temperatures, failed to yield satisfactory 



Eemarks 

 Little altered 

 Andalusite still abundant 

 Andalusite still abundant. No 



sillimanite 

 Andalusite still present. No sil- 

 limanite 

 Little changed. No sillimanite 

 Little changed. No sillimanite 



At higher temperatures the andalusite is decomposed with- 

 out forming sillimanite. Andalusite heated seven days at 400° 

 in a bomb containing 10 per cent NaCl solution snowed no 

 change. The result indicates that andalusite changes to sil- 

 limanite at high temperatures, but with considerable diffi- 

 culty. The reverse change, sillimanite-andalusite, does not 

 occur under any conditions which we have yet tried. 



Cyanite, — Like andalusite, cyanite is much contaminated 

 with mica, rendering satisfactory thermal study of the natural 

 mineral difficult. It is not possible to separate it completely 

 by purification with aqua regia and cold HF. Its specific 

 gravity is 3*5 — 3*7. Hardness = 4-5 or 6-7, depending on 

 the direction with respect to the prism axis. Vernadsky 

 found the mineral changing to sillimanite above 1300°. We 

 have found the change to be slow, though more rapid than 

 the change from andalusite to sillimanite. 



Time 



Flux 



Temperature 



Eemarks 



^ hour 



None 



1500° 



Decomposed, but no sillimanite 

 identifiable 



7 days 



■ a 



1150° 



But little changed 



2 " 



tt 



1150° 



But little changed 



2 " 



a 



1000° 



Unchanged 



28 " 



a 



900° 



Little changed 



Borax 



decomposes the mineral. 





