30tt Shepherd, Rankin, Wright — Binary Systems of 



Time Flux Temperature Remarks 



12 hours NaCl 800° Unchanged 



72 " " 800° Little changed 



48 " CaY 2 6 1000° Cyanite decomposed but no sil- 



limanite formed 



Cyanite is thus more readily decomposed than andalusite, but 

 shows no sillimanite which can be determined with the micro- 

 scope. 



We have shown that at about 1300° both andalusite and cya- 

 nite change, the one into sillimanite and the other, from 

 Yernadsky's density determination, probably into sillimanite, 

 though badly formed. It is now in order to determine whether 

 sillimanite, which is stable at high temperatures, will change 

 into cyanite or andalusite at low temperatures. 



Sillimanite shows no heat effect between 1100° and 1500°. 



Time 



Flux 



Temperature 



Remarks 



168 hours 

 96 " 



None 



600° 

 1100° 



Unchanged 

 Unchanged 



Heated four days at 1000° with a small amount of borax, the 

 crystals show slight attack, but are not destroyed, and no 

 andalusite or cyanite appeared. Heated 48 hours at 1100° 

 with CaY 2 6 , the powdered crystals were slightly attacked 

 and new crystals of sillimanite formed. Working in steel 

 bombs with various aqueous solutions, negative results were 

 obtained. 



Time 



Flux 



Temperature 



Remarks 



7 days 



6 " 



7 " 



hi NH 4 F 



10$ NaCl 

 10$ KBr 



400° 



450° 

 400° 



The original grains are pitted, 

 but no new crystals formed 

 Unchanged 

 Slightly attacked 



Heated thirty days in a long steel tube which allowed a con- 

 tinuous current of hot water (250°) to pass over the mineral 

 and then to a cooler part of the tube, no alteration was pro- 

 duced. 



Thus at low temperatures no conditions were found under 

 which sillimanite tended to change into andalusite or cyanite. 

 Solutions of sillimanite in albite and borax gave always sil- 

 limanite. When fused sillimanite is rapidly cooled (quenched); 

 it always crystallized as sillimanite. 



It would be a waste of time to tabulate all of the metatheti- 

 cal reactions by which we have sought to produce these two 

 unstable forms. After overcoming the great difficulties of 

 finding a suitable vessel which will withstand the pressure of 

 aqueous solutions at high temperatures, and which will not be 



