218 J. S. Diller — Mineralogical Notes. 



tion and an electro-magnet to obtain the latter mineral for 

 chemical analysis. The results are given under I, below. 



These figures appear to show the material analyzed to be im- 

 pure and it was thought advisable to obtain more of the rock 

 and endeavor to separate the dumortierite as far as practicable 

 from all impurities. Mrs. C. A. Bidwell kindly furnished a 

 sufficient amount of much better material in which the only min- 

 eral associated with the dumortierite was quartz. As dumor- 

 tierite is not acted upon by hydrofluoric acid the rock after be- 

 ing crushed to small particles, was digested in this acid for a 

 length of time sufficient to decompose most of the quartz. 

 The mass was then washed with water, dried and any quartz 

 that might still remain, separated by Thoulet's solution. After 

 thorough washing the material was examined with the aid of a 

 microscope and found to be free from gangue. Having been 

 ground exceedingly fine and dried at 104° C. for about three 

 hours the mineral was analyzed with the following results (II) : 



I. II. 



Si 2 2 31-52 27-99 



A1 2 3 63-66 64-49 



CaO trace 



MgO -52 trace 



Na 2 -37 



K 2 -11 



B„0 3 , 2-62 4-95 4-93 



P 2 6 0-20 



Ignition ... 1-34 H„0 ..1-72 



100-14 99-35 



Analysis II shows less impurity than the first specimen anal- 

 yzed. These results indicate either that dumortierite is not a 

 simple silicate of alumina as stated by Damour,* or else that 

 the material analyzed was a mixture of dumortierite with some 

 other compound. 



If we assume the formula of Damour to be correct and esti- 

 mate all the SiO a in the analysis as belonging to the formula 

 Al 8 Si s O ie , then there will be left unaccounted for a small amount 

 of Al 2 O s , H 2 aud B 2 3 and these are present in the propor- 

 tions represented by the formula A1B 3 6 . 2H 2 0. If this mode 

 of interpretation be correct then the mineral from Arizona 

 corresponds approximately to the formula 3Al 8 Si 3 18 AlB 2 6 . 

 2H 2 which requires Al 9 6 8 =65-2^Si0 9 =27'6#. B 2 3 =5-4^; 

 H 2 = l-8^ agreeing quite closely with the actual analysis. A 

 borate of alumina corresponding to the above formula is, we 

 believe, not actually known, and concerning its properties noth- 



* Bull. Soc. Min. d. France, vol. iv, p. 6. 



