on the fluoric Compounds. 65 
and by a very careful analysis, it was found that this subcar- 
bonate contained 31 per cent, of carbonic acid ; now, if the 
remainder of the salt be supposed to be pure potassa, and the 
calculations be made on this idea, it will appear that fiuate of 
potassa must consist of 125.9 of potassium and 54.74 of fluo- 
rine, and the number representing fluorine must be regarded 
as 32.6. 
The 18.15 gi^ains of fluate of potassa decomposed by sul- 
phuric acid afforded 38.5 of acid sulphate of potash. Suppos- 
ing acid sulphate of potassa to consist of two proportions of 
sulphuric acid 150, and one of potassa 90, they ought to have 
given 40 grains, and this loss is no more than might be ex- 
pected in the process of decomposition and evaporation. 
In some experiments that I made on the decomposition of 
the hydrates of potassa and soda, when decomposed by fluoric 
acid, the results indicated a number for fluorine a little lower. 
Thus 20 grains of hydrate of potassa were converted into 19.8 
grains of fluate of potassa, and 20 grains of hydrate of soda 
into 19.6 of fluate of soda ; but I do not place so much confi- 
dence in these results, as there always was great heat pro- 
duced during the action of the acid upon the hydrates; and 
probably a minute quantity of the hydrates might have been 
dissipated at the beginning of the process. 
It appears reasonable to conclude, as I have stated in my 
last paper, published in the Philosophical Transactions, that 
the number representing fluorine is less than half of that re- 
presenting chlorine, about 33. 
I endeavoured to ascertain the composition of fluate of am- 
monia, by adding together hydro-fluoric acid and solution of 
ammonia of known composition, and I found in this way, that 
MDCCCXIV. K 
