542 
THE PHARMACEUTICAL JOURNAL AND TRANSACTIONS. 
[January 6,1S?2. 
the accompaniment of frauds, so frequently encoun¬ 
tered in connection with high-priced products. It is 
important, therefore, that pharmaceutists should he 
acquainted with the means of exposing these falsifi¬ 
cations, so that they may he able to supply to their 
customers a pure bromide of potassium; and to this 
part of the subject M. Falieres devotes the first part 
of his monograph. 
Among the salts that have been detected in com¬ 
mercial bromide of potassium are iodide of potassium, 
chloride of potassium, potash in a free and carbo¬ 
nated state, sulphate of potash, nitrate of soda and 
bromate of potash. Considering its high price, it is 
not probable that iodide of potassium is fraudulently 
introduced into the bromide; but its presence is pro¬ 
bably due to the fact, that the bromine of commerce 
usually contains traces of iodine. Small quantities of 
iodine may be detected by treating a solution of the 
bromide of potassium with chlorine water or bromine 
water. The iodine which is set free gives a blue 
colour with the mucilage of starch, and rose or ame¬ 
thyst violet with benzine, chloroform or sulphide of 
carbon. But M. Falieres gives the preference to a 
process employed by M. Bouis to detect iodine in 
mineral waters. It consists in introducing some 
solution of the bromide of potassium under examina- 
nation into a small closed tube, then adding a few 
drops of solution of perchloride of iron and boiling 
the mixture. The perchloride of iron is without 
action upon the bromide of potassium, while it com¬ 
pletely precipitates the iodine from the iodide. Con¬ 
sequently, if a piece of starched paper be introduced 
into the tube, the blue tint of iodide of starch appears 
if the solution contains only traces of iodine ; but it 
is a better plan to distil a small quantity of the 
solution and seek for the iodine in the distillate. 
M. Falieres proposes to make this method more 
practical still, by simply dipping a small slip of 
ordinary white paper into the solution containing 
the bromide of potassium and perchloride of iron, 
when, if the liquor contains iodine, the paper will 
become blue. 
M. Falieres has noticed in bromide of potassium 
the presence of quantities, often considerable, of 
chloride of potassium. M. Adrian, also, has ascer¬ 
tained that of ten specimens obtained from the prin¬ 
cipal manufactories for the supply of pharmacy, one 
only was suitable for employment in medicine. In 
the other nine he found variable quantities of the 
chloride, amounting sometimes to 30 per cent, of the 
entire weight. He has remarked besides, that one 
of the most beautiful specimens, judging by the re¬ 
gularity of the crystals, was also one of the most 
impure. This fraud is a very grave one, as it can¬ 
not be doubted that it interferes seriously with the 
therapeutical action of the bromide, and it is im¬ 
portant, therefore, that pharmaceutists should be 
able to determine in a precise manner the propor¬ 
tion of the chloride ; but the estimation is one that 
presents great difficulties. 
It is well known that chlorine, bromine and iodine, 
and their compounds, present such close analogies 
that no good process has yet been discovered for 
their separation. Of chlorine and bromine particu¬ 
larly, the solubility and insolubility of the combina¬ 
tions which they form are so nearly alike that it is 
very difficult to accomplish their separation by 
means of an insoluble precipitate. The application 
of the numerous gravimetric methods presents so 
many difficulties, that the analysis of a mixture of 
chlorine and bromine, and even of iodine, is doubt¬ 
less one of the most delicate of operations, requiring 
great skill in the manipulation. 
In 1868, M. Baudrimont published* what he 
termed the “indirect process” for the detection of 
the presence of chloride in commercial bromide of 
potassium. It was based upon a fact well known to 
chemists, that a given weight of chloride of potas¬ 
sium decomposes a much larger proportion of a 
titrated solution of nitrate of silver than the same 
weight of bromide of potassium. For instance, 
1 gram of the bromide requires for its complete 
precipitation as bromide of silver only 1*427 gram 
of the nitrate of silver, whilst it requires 2 279 
grams of the same nitrate to entirely convert one 
gram of chloride of potassium into cliloride of 
silver. 
In February, 1869, M. Falieres communicated to 
the Societe de Pharmacie at Bordeaux a process of 
volumetric analysis for a mixture of cliloride and 
bromide of potassium by means of a titrated solu¬ 
tion of nitrate of silver,—an idea which he then 
thought to be entirely new, but for which he after¬ 
wards yielded the palm of priority to M. Baudrimont. 
But the process of M. Falieres constitutes a con¬ 
siderable advance, and will be very useful in ascer¬ 
taining the degree of purity of bromide of potassium. 
Taking advantage of the recognized fact that the 
before-mentioned proportions of nitrate of silver are 
required for the conversion of bromide and chloride of 
potassium into bromide and chloride of silver re¬ 
spectively, M. Falieres points out that in a mix¬ 
ture of ’9 of bromide and *1 of chloride, the quan¬ 
tity of nitrate of silver requisite for their complete 
conversion would be— 
1-427 X 0-9 -f 2-279 X 01 = 1 5122. 
Consequently, if the bromide contains 10 per cent, 
of chloride, it would be necessary, after having 
treated the mixture with P427 of the nitrate of 
silver, still to add 0"0852 of the same reagent. It is 
evident, therefore, sa} r s M. Falieres, that if '852 of 
nitrate of silver be dissolved in 100 cubic centi¬ 
metres of distilled water, and after having treated 
1"0 of bromide of potassium with 1127 of silver, it 
still requires for complete precipitation 5, 10, 20 or 
30 cubic centimetres of the titrated solution, it is 
because the bromide of potassium under analysis 
contains 5, 10, 20 or 30 per cent, of chloride. 
The details of the process, as given by M. Falieres, 
are as follows:—First, it is necessary to ensure— 
by means of perchloride of iron, salts of baryta 
and strong sulphuric acid in excess—that there are 
no other salts present, such as iodide of potassium, 
carbonate and sulphate of potash, or nitrate of 
soda, which would disturb the results of the analysis. 
Then dissolve, in a stoppered bottle, 1 gram of the 
bromide of potassium in 30 or 40 grams of water, 
and add a solution containing P427 gram of ni¬ 
trate of silver. When the precipitate has settled 
to the bottom of the bottle, add the bromometric 
liquor drop by drop by means of a Gay-Lussac 
burette. If the bromide be pure, the addition of a 
drop of tliis liquid will cause no turbidity; if, on 
the contrary, a precipitate is formed, the volume of 
the test liquor employed will show the quantity of 
chloride present. This operation is reported to be 
easily performed, and to yield good results. 
* Journal de Pharmacie et de Chimie, vii. 411. 
