200 G. HARKER. 



being more than one to three per cent. When the ammonium 

 carbonate was added hot to the solution and then allowed to stand 

 until quite cold, a larger proportion of fluorine was found in the 

 precipitate than when it was filtered hot, the precipitate con- 

 taining the fluorine coming down on cooling. This precipitate 

 contained a larger amount of fluorine when sodium potassium 

 carbonate was used than when sodium carbonate alone was 

 employed. In obtaining the results given by Wohler's method for 

 fluorine, sodium potassium carbonate was frequently employed, 

 which accounts for some of them being so low, but even when 

 sodium carbonate alone was employed, the highest percentage 

 obtained was 1269. 



On adding the amount of fluorine recovered from the precipitate 

 to the amount obtained from the solution after the addition of 

 ammonium carbonate, the total is nearly the same as that 

 obtained by Berzelius' method. Thus 14'2 percent, was obtained 

 from Brazilian topaz by Berzelius' method. In one experiment 

 6-92 per cent, was obtained from the solution and 6 50 from the 

 precipitate, total 13 42 per cent. In a second experiment 12-69 

 from the solution and 1*21 from the precipitate, total 13*90 per 

 cent. This seems to shew that Berzelius' method gives the total 

 fluorine present. The precipitates of calcium fluoride were tested 

 by conversion into calcium sulphate. The residue obtained on 

 treating the fused mass with water in Wohler's method, contained 

 all the silica and about two-thirds of the alumina and some soda, 

 and had approximately the constitution 3 Si0 2 , 2 Na 2 0, 2 A1 2 3 . 

 On boiling with water the soda and alumina were slowly removed, 

 but to what extent this could be carried out was not determined. 

 In the case of Berzelius' method, the residue contained all the 

 aluminium. The silica prevents the aluminium from passing into 

 solution and thus stops the loss of fluorine which occurs in 

 Wohler's method. The amount of silica necessary to do this 

 was found to be 31-6 per cent, in the case of the Brazilian topaz. 



Examination of the ammonium carbonate precipitate. — The 

 topaz was fused with sodium carbonate alone and the fused mass 



