280 Berzeliiis'ii Experiments to deter m ine the Composition [April, 



of opinion, however, tliat less confidence should be put in these 

 experiments than in others made by Dr. John Davy on fluosilicic 

 acid and its combinations with ammonia. These experiments 

 give the capacity of satmation of fluoric acid still greater than 

 the analysis offluate of lime by Sir H. Davy. I employed it not 

 only to determine the composition of some fluates, but likewise 

 to ascertain the composition of silica. However, the results of 

 Dr. Davy's experiments are inaccurate, as I shall show hereafter, 

 when treating of the composition of silica; but the error result- 

 ing I'rom it in the calculations cannot be immediately discovered, 

 because upon the fluates, it makes a difterence of only 2 per 

 cent, and with respect to the composition of sihca, • another 

 inaccuracy in a contrary way corrects the results of the calcula- 

 tion. However, as it was necessary for me to know with much 

 precision the composition of silica, and as I thought that the 

 analysis of fluosilicic acid gas might be serviceable for that 

 object, I wished to begin by studying the capacity of saturation 

 of fluoric acid. To avoid the objection of the presence of silica 

 in the fluate of lime, 1 chose for my first experiments fluate of 

 silver. 



To prepare it, I distilled, in a small platinum apparatus, fluate 

 of lime with sulphuric acid. The acid was received by water in 

 a platinum cup. I then added carbonate of silver to the liquid 

 till it was saturated. There fell a slight greyish precipitate which 

 I considered as fluosilicate of silver. The neutral liquid was 

 filtered through paper in a platinum funnel, as I had found that 

 it corrodes glass, and deposits in that case the greyish precipi- 

 tate of which 1 have spoken. The liquid was then evaporated 

 and the mass heated to redness. It melted, and continued 

 always to give out fluoric acid gas and oxygen gas, while metallic 

 silver was disengaged. This phenomenon continued as long as 

 the fluate remained exposed to the fire ; and I have reason to 

 believe that it is not owing to the presence of water. 



4*936 grammes of fused fluate of silver left, when dissolved in 

 water, 0'185 gramme of metallic silver. The liquid precipitated 

 by sal-ammoniac produced a quantity of muriate of silver, which 

 weighed, after being fused, 6'349 grammes ; that is to say, that 

 ICO offluate of silver had given 1 12"587 of muriate of silver. 



9'9'22 grammes of fluate of silver fused and redissolved in 

 water left 0"37(j gramme of metallic silver, and furnished 10'7465 

 grammes of muriate of silver. These two numbers are to each 

 other as 100 : 112-o7. 



These two experiments agree as nearly as possible. Accord- 

 ing to them, lOU of fluoric acid combine with 1021-8 of oxide of 

 silver, of which ^he oxygen is 70"4, which in this case ought to 

 be the number which represents the capacity of saturation of 

 fluoric acid. 



Filiate of Ban/tes. — A portion of fluate of soda slightly acid 

 was evaporated to dryness, and slightly heated, without, how- 



