﻿382 Penfield and Harper — Composition of JRalstonite. 



eral as far as possible with sulphuric acid, assuming that the 

 sodium was all in solution and the insoluble portion some 

 compound of aluminum. Water was determined by igniting 

 the mineral in a combustion tube, passing the vapor over dry 

 sodium carbonate and collecting it in a weighed chloride of 

 calcium tube. The fluorine could not be determined by 

 driving off the silicon fluoride. About twenty-five per cent 

 of fluorine was readily driven off; by continuing the decompo- 

 sition with the temperature of the sulphuric acid about 160°, 

 silicon fluoride was slowly, but, in the course of six or eight hours 

 never completely given off. The greatest amount of fluorine 

 which we were able to drive over by this method was about 

 thirty-one per cent. This is not at all in accordance with the 

 statement of Brandl,* who states that the decomposition com- 

 mences at 145° and is completed at 160°. We found it impos- 

 sible to make a determination according to the method described 

 by him. Our determinations are made according to the Berze- 

 lius method by fusing with mixed potassium and sodium car- 

 bonates and silica. They are probably too' low by about one- 

 half to one per cent, judging from test experiments which we 

 made on cryolite. Our determinations are as follows : 



Mean. Ratio. 



4-39 -183 1 



4-27-M6 '093 [• -278 l'OO 



•12-^-78 '002 ) 

 •03 

 24-25 -882 3*17 



39-9J 2-101 7*56 



18-73 



Mg 



4-46 



4-31 



Na 



4-25 



4-27 



K 



•12 



•12 



Ca 



•03 







Al 



24-23 



24-27 



F 



39-76 



40-05 



H 



18-72 



18-74 



91-70 



The ratio of (MgNa 2 K 2 ) : Al = l : 3 nearly. This ratio being 

 assumed as correct, the ratio of the fluorine necessary to unite 

 with the metals should be 11, whereas we only find 7*56. The 

 fluorine is therefore not sufficient to unite with the metals and 

 this is fully in accordance with the suggestion of Nordenskiold. 

 If the metals in our analysis, which are in excess of the fluorine, 

 are united to hydroxvl, which, as has been shown to be 

 the case in several instances, is capable of replacing fluorine, 

 it would be necessary, in order to make the ratio (Mg, ]STa 2 , 

 K 2 ): Al:(F+OH)=l':3:ll, to have 16-27 per cent of hy- 

 droxyl corresponding to 8*61 per cent of water; the remaining 

 10*12 per cent of water would then be regarded as water of 

 crystallization, and would correspond to two molecules, mak- 

 ing the formula of the mineral (Mg, Na 2 )Al 3 (F+OH) n . 2H 2 0. 

 Making this disposition of the water, our analysis would be: 



* Loc. cit. 



