of Edinburgh, Session 1885-86. 
647 
being given off during oxidation of the substance used to reduce the 
permanganate, the solutions of permanganate and of ferrous-ammo- 
nium sulphate were separately poured into long glass tubes, which 
having been completely filled and closed with the thumb, were 
opened under water. The liquids gradually mixed by diffusion, but 
no evolution of gas could be detected. Oxygen, however, being 
somewhat soluble in water (roughly 3 volumes in 100 volumes of 
water), it may be questioned whether, in this last experiment espe- 
cially, oxygen might not be given off, and be dissolved by the water. 
The permanganate and ammonio-ferrous solutions were separately 
shaken up with pure air, and after standing for some little time were 
mixed as before, and as before gave negative results. 
The next point I wish to discuss is in what order should the 
various solutions be added? In all my experiments in reference to 
this question I have employed sufficient distilled water to dilute the 
the nitrous acid in the nitrite, to the extent of at least 1 in 5000, 
and in all the same bulk of water and the same solution of the 
nitrite — nitrite of sodium — were employed. 
In this series, therefore — (1) I have added the nitrite first, then 
the acid, and immediately followed this by the permanganate; or 
(2) I have run in permanganate before the acid, and added this 
latter (a) slowly and (b) rapidly ; and (3) finally, I have added the 
nitrite last, both (a) rapidly and (b) slowly. All these methods 
give fairly approximate results, but all (with one exception) allow 
some slight decomposition of the nitrous acid, evidenced by the 
odour of nitric peroxide, and by the amount of permanganate de- 
colorised; this decomposition, judging from the odour and the 
results of the analysis, being especially marked where the sulphuric 
acid preceded the permanganate, the nitrite of sodium having been 
added first. The one exception was where the nitrite was slowly 
added, in a way I shall afterwards mention, to the previously acidified 
permanganate (3, a). 
Again, as anhydrous nitrous acid decomposes ammonia into nitro- 
gen and water according to the well-known equation, 
2NH 3 + N 2 0 3 = 2N 2 + 3H 2 0, 
does the presence of ammonia or its compounds give fallacious 
results with this permanganate process 1 Two beakers, each contain- 
