﻿F. W. Clarke — Lithia Micas. 357 



the alkalies from 150 grams of mineral gave a spectroscopic 

 trace of rubidium. The Paris lepidolite showed traces of both 

 metals, while the material from Hebron, Norway and Auburn 

 was much richer. On these lepidolites approximate estimations 

 were made, as follows : 





Hebron. 

 Granular. 



Norway. 

 White. 



Auburn. 

 Granular. 



Auburn 

 Border. 



K 2 



11-44 



8'82 



10*51 



8-03 



Kb 2 0) 



'11 



3-73 



1-29 



2-44 



Cs 2 \ 



•08 



•45 



•72 



12-21 12-63 12-25 11-19 



In these determinations the caesium was separated as stan- 

 nic chloride, and the other two metals were computed indirectly 

 from the amount of chlorine in the mixed chlorides. The 

 results have only a comparative value, and no sharp accuracy 

 can be imputed to them. The methods of separation for the 

 three metals are still far too defective. 



The foregoing analyses of the Maine lepidolites, covering 

 several distinct types of the mineral from five different locali- 

 ties, indicate a great constancy of composition. The essential 

 identity of the mineral with that from Eozena and Cornwall is 

 also shown by the several published analyses of the latter, 

 while the lepidolite from Juschakova is slightly different, in 

 that it contains some manganese replacing aluminum, higher 

 fluorine and a little chlorine. In most of the analyses the 

 water and fluorine appear to vary reciprocally, suggesting the 

 ordinary replacement of the latter element by hydroxy]. With 

 this slight assumption the formula for lepidolite may be written 

 thus: 



Al 2 LiKF 2 Si 3 9 ; 

 a formula which has long had general acceptance, but which 

 now rests upon the surer basis of a wider range of analytical 

 data. It corresponds to the following theoretical composition : 



Found. (Riggs). 



48-80 to 51-52 * 



24-99 28-80 



3-87 4-98 



(NaKRbCs) a O 12-07 13-68 



(FH 2 0) 6-69 8-25 



Most of the variations are no greater than we should expect 

 to find in material so difficult to secure in absolute purity as 

 lepidolite. The granular structure of that species is peculiarly 

 favorable to the presence of inclusions, as for example of albite, 

 to which probable impurity some of the soda shown by the 

 analyses may be due. The greatest difference is in the case 

 of fluorine; although some of the foreign analyses, notably 

 Eammelsberg's analysis of the Juschakova mineral, contain very 

 nearly the full theoretical amount. 





Calculated. 



Si0 2 



49-18 



A1 2 G 3 



27*87 



Li 2 



4-09 



K 2 



12-81 



F 



9-84 



