118 
PROFESSOR W. RAMSAY AND DR. S. YOUNG 
Again, aldehyde-ammonia is usually viewed as a dihydric alcohol, in which one 
hydroxyl group is replaced by an amido group, and this again involves the breaking 
OH 
down of a molecule of ammonia, its formula being CH 3 CH jv-jj • Lastly, the com¬ 
bination of phthalic or succinic anhydride with water involves the breaking down of a 
molecule of water, thus :— 
C 3 H 4 (C0) 3 0+H.0H = C 2 H 4 (C00H) 2 ; 
C 6 H,(CO)o()+H.OH=C 6 H 4 (COOH) 2 . 
In all these cases the behaviour in the still differs from that in the barometer-tube. 
§ 79. The three remaining substances, nitrogen peroxide, acetic acid, and ammonium 
chloride, give identical vapour-pressures by both methods ; and in this their 
behaviour does not differ from that of stable solids or liquids. It is, therefore, to be 
surmised that these substances must show some essential difference in constitution 
from the others, and we shall consider them in detail. 
Nitrogen peroxide, when dissociated, may have one of two formula}, either 
N |-j or N <^ ( y It is indifferent for our purpose which is accepted ; but, taking the 
first as correct, there are three ways of representing the union of two molecules 
of N0 2 to form N 2 0 4 . First : Simple union of the nitrogen atoms thus 
q^N — N^q. Second : Union through the oxygen atoms :—0 = N — 0 — 0 — N=0. 
Third (and the hypothesis has been recently advanced yb Divers) : Union of a 
nitrogen atom of the one molecule with an oxygen atom of the other, the result being 
a nitrate of nitrosyl, q^N—0 —N=0. 
Now, none of these methods of representation involves the breaking down of a 
molecule, but only a partial molecular re-arrangement, such as takes place in the 
aldehyde group, CHO, which does not involve an alteration in position of the atoms. 
§ 80. Passing to acetic acid, it has been suggested by Playfair and Wanklyn that 
the abnormal vapour-densities point to a molecule of the formula C^HgOj., and the 
formula deduced by Willard Gibbs from the observations of these experimenters and 
others is based on this assumption. If this view is correct, the only structural 
formula applicable is 
ch 3 -c 
ch 3 -c 
JOH 
10 
[6 
|0H 
where the oxygen of the two carboxyl groups forms the connecting link. Here again 
there is no breaking down of a molecule. 
But experiments which we have made on the density of the saturated vapour of 
