4 MEANS FOR DETERMINING THE PURITY OF DRUGS, ETC. 
iodide, a crystalline precipitate of bitartrate of potassa is thrown 
down, and the supernatant liquid gradually becomes colored as 
when diluted sulphuric acid is used, and from the same cause. 
Chloride of barium but slightly if at all changes the solution, 
and proto-sulphate of iron not at all. If a whitish or greenish 
white precipitate should be occasioned by a solution of proto-sulphate 
of iron, which effervesces on the addition of a diluted acid, the iodide 
contains carbonate of potash, which it sometimes does to the extent 
of 15 per cent, or more when it has been crystallized from a saturated 
solution of that carbonate. In such a case turmeric paper will be 
colored brown by its solution, and reddened litmus paper restored to 
blue. If ten grains of the iodide in solution in water, be mixed with 
an excess of a solution of nitrate of silver, a lemon yellow precipitate 
of iodide of silver is thrown down, w T hich, when washed and dried, 
weighs 14.1 grains if the iodide is pure, and if the precipitate be 
treated with solution of ammonia, the clear ammoniacal liquid should 
yield no precipitate when an excess of nitric acid is added. If less 
than 14.1 grains of the yellow precipitate is obtained, and a white 
precipitate falls on the addition of nitric acid as above, chloride of 
potassium or sodium has been present. If the precipitate by nitric 
acid becomes yellowish by washing, bromide of potassium was pro- 
bably present. When bromide of potassium exists as an adulteration 
it may be detected thus: to a solution of the suspected iodide add an 
excess of sulphate of copper, and pass a current of sulphurous acid 
through the solution till it ceases to cause a precipitate of protio- 
dide of copper; filter the liquor from the iodide of copper, add chlo- 
rine water and ether and agitate the mixture. If bromide of either 
potassium or sodium was present, the supernatent ether in the test 
tube will be colored reddish yellow. By then washing, drying and 
"weighing the iodide of copper, the amount of iodine, and from this, 
of iodide of potassium present, can easily be ascertained by calcula- 
tion, as 10 grs. of pure iodide of potassium should yield nearly 11.5 
grs. of iodide of copper. 
When nitrate of soda, nitrate of potassa, or other salt not acted 
of by nitrate of silver or ammonia, is present, the fact will be in- 
dicated by the iodide of silver test above, not indicating the full 
amount of iodide, and no precipitate being yielded by nitric acid 
from the ammoniacal washings of the iodide. Water may exist 
to the extent of several per cent, especially if carbon.ate. of potassa 
