198 G. HARKER. 



Second series. — The substance was decomposed without previous 

 fusion with alkaline carbonates. The calcium fluoride in the first 

 four of these experiments was from crystals of fluor-spar ground, 

 and evaporated to dryness with hydrofluoric acid to get rid of 

 silica and then ignited. On testing by conversion into calcium 

 sulphate they gave 99-8 per cent, of calcium fluoride. 



Amount taken (1) -1565 grams. Percentage of fluorine found 45*90 

 (2) '2722 „ „ 47-08 



(3) '5212 „ „ 47-21 



(4) -2900 „ „ 46-64 



(5) -1736 „ „ 46-31 



(6) -1787 „ „ 45-95 



The theoretical percentage was 48-62 and the average difference 

 2-25 per cent. These results were corrected for the solubility of 

 potassium silicon fluoride which was found to be -0016 gram, in 

 50 ccms. of a 50 per cent, alcohol solution. Nos. 5 and 6 were 

 precipitates of calcium fluoride obtained by Berzelius' method from 

 the Mudgee topaz. 



Cryolite similarly gave 52*17 per cent, when previously fused 

 with alkaline carbonates, and 52*62 without previous fusion, and 

 when treated by Berzelius' method 55*34. Apatite gave 2*80 by 

 the first method and 2-81 by the second. 



Application of the tetra-fluoride method to topaz. — The Mudgee 

 topaz by this method gave only 1 3 per cent, of fluorine which was 

 much too low. When repeated on the Brazilian topaz a similar 

 low result was obtained, viz. 10*9 percent. It seems therefore 

 that this method works well for minerals not containing silica. 

 That it is the presence of silica in the topaz which causes the 

 error was shewn by adding silica to calcium fluoride or topaz 

 before fusing, when in all cases a much lower result was obtained,, 

 the error being larger as more silica was added. Thus calcium 

 fluoride gave 36*54 and 35*06 per cent, of fluorine and topaz 7*05 

 and 4*26 per cent., results much lower than those obtained when 

 silica was not added. 



