on different Combinations of Fluoric Acid. 3(33 
say that the whole of the silex cannot by this method be re- 
moved, but only the principal part. Their reason for this 
belief, is, that on repeatedly evaporating the salt after the addi- 
tion of ammonia and redissolving it, they have each time ob- 
served a residue of silex. If they employed metallic evaporating 
vessels, the results of my experiments do not agree with theirs; 
for making use of platina for this purpose, and adding an ex- 
cess of ammonia, I never detected traces of silex on evaporating 
the filtered fluat. But our results agree, if they employed glass 
or porcelain vessels, which fluat of ammonia has the property 
of corroding. 
I now proceed to the consideration of fluat of ammonia; but 
before I describe some of the properties of this fluat which I 
have observed, I shall briefly mention the means pursued for 
ascertaining the proportions of its constituent parts. 
The composition of subsilicated fluat of ammonia being known, 
that of the fluat (granting what is already advanced respecting 
its formation to be correct) may be inferred from the proportion 
of silex, that a given quantity of ammonia will precipitate. 18 
cubic inches of ammoniacal gas were condensed by ^ of a cubic 
inch of distilled water in a small glass tube over clean mer- 
cury. This ammoniacal solution was added to a clear filtered 
solution of subsilicated fluat of ammonia. A precipitate of silex 
was immediately produced. After several hours standing, this 
precipitate was collected on a filter, well washed and dried and 
heated to redness. It was pure silex, and weighed 1 .6 grains. 
This experiment, like all the preceding, was repeated, and the 
result confirmed. In both instances there was an excess of 
subsilicated fluat. The precipitations were made in a platina 
vessel, and the solutions were neither heated before or after 
