hillebrand.] CHLORINE, FLUORINE. 103 
should not l)c omitted; otherwise, a few tenths per cent of carbon 
dioxide can very well he overlooked. Moreover, the powder should 
first be stirred up with a little hot water, to remove all entangled air 
which might otherwise be mistaken for carbon dioxide. 
It has been already shown under Water (p. 45) how, in case of need, 
the determination of carbon dioxide can he combined with that of water 
by fusion with lead chromate or potassium chromate. This latter 
method must always be resorted to when the carbon of graphite or 
carbonaceous matter has to he 1 estimated. If carbonates arc present 
at the same time the result of the test includes the carbon from both 
sources, and a separate determination by the wet way of that of the 
carbonates is necessary. 
XXII. CHLORINE. 
To make sure of getting all the chlorine, it is best to fuse with 
chlorine-free sodium-potassium carbonate, or even sodium carbonate 
alone, first over the full burner, then for a moment or two over the 
blast, leach with water, acidify with nitric acid, and precipitate by 
silver nitrate without preliminary separation of silica. If 1 gram of 
material has been used no precipitation of silica need be feared on 
acidifying or on standing. 
In many cases it is quite sufficient to attack the powder by chlorine- 
free hydrofluoric acid and a little nitric acid, with occasional stirring, 
and after filtering through paper fitted into a large platinum cone or 
rubber funnel, to throw down the chlorine by silver nitrate. The 
presence of nitric acid is necessary, since otherwise ferrous fluoride 
reduces silver nitrate with deposition of crystallized silver. When 
coagulated b} r heating and stirring, the precipitate is collected on the 
filter, washed, dissolved by a little ammonia, and reprecipitated by 
nitric acid, when it can be collected in a Gooch crucible and weighed, 
or, if very small in quantity, on a small paper filter, which is then dried, 
wound up in a tared platinum wire, and carefully ignited. The in- 
creased weight of the wire is due to the metallic silver of the chloride 
which has alloyed with it. 
XXIII. FLUORINE. 
Fluorine can only be estimated by the method of Rose, care being 
taken to use sodium-potassium carbonate as a flux, and to avoid use 
of the blast if possible. For minerals rich in fluorine and low in silica 
it may be necessary to add pure silica before the fusion in order to 
effect complete decomposition of the fluoride jusl as with the alkaline- 
earth phosphates. To the hot aqueous extract several grams of 
ammonium carbonate are added and more on cooling. After twelve 
