801 
of Edinburgh, Session 1885—86. 
to the liberation of iodine; if so, boil until the colour disappears. 
If chlorates and nitrates are absent, no change of colour takes place 
in the bulb, until on adding two drops of hydrogen peroxide, whence 
if iodine be present, the liquid in the bulb immediately assumes a 
yellow, then a rosy red colour, and on boiling a violet vapour makes 
its appearance in the upper part of the tube, even when only minute 
quantities of iodine are present ; dispel the iodine gradually by care- 
fully adding hydroxyl, two drops at a time. The iodine being 
expelled, add two drops of dilute H 2 S0 4 and three or four drops of 
hydroxyl ; the bromine will now make its appearance, and hydroxyl 
solution must be added, two drops at a time, until it has all dis- 
appeared, which will be known by the clearness of the liquid in the 
distilling bulb, and will be also almost colourless. Very often a 
little chlorine can be detected coming over with the bromine, but 
this only occurs when the quantity is large, or when too much 
H 2 S0 4 has been added, and so making the solution strongly acid. 
In fact, this must be done if the experiment is to be successful in all 
cases, and great care must be exercised in adding the H 2 S0 4 , as well 
as the hydroxyl (H 2 0 2 ). If now, after the bromine has disappeared, 
a few drops of a solution of AgN0 3 be added to the solution remain- 
ing in the bulb, a white precipitate of AgCl will appear, showing 
the presence of a chloride. It is always better to divide the solu- 
tion into two parts, or take only a portion for the detection of a 
chloride. 
In every experiment I detected the iodine when present, either 
by the flash of red colour in the distilling bulb, the vapour (violet) 
in the upper part of the tube, or the blue colour with starch, even 
if the iodine present only amounted to the one-thousandth part of 
the substance present. 
In only one case, out of a great many experiments, was I unable 
to detect bromine, and that when the amount of bromine, compared 
with the iodine and chlorine, was very small. The detection of the 
chlorine was always sure, in some cases by the smell, but always by 
means of silver nitrate (AgN0 3 ). 
The following acids do not interfere with the test : — hydro- 
sulphuric, sulphurous, hyposulphurous, carbonic, phosphorous, and 
hydrofluoric acids. 
When the chlorine, bromine, and iodine are in combination with 
3 G 
VOL. XIII. 
