ON EVAPORATION AND DISSOCIATION. 
119 
this substance show dearly that the density increases more and more rapidly with fall 
of temperature, until at 50° the density is more than 55, that of the molecule C 4 H 8 0 4 
being 60 ; and some isolated determinations by a method, of the reliability of which we 
are in some doubt at present, appear to indicate that at 20° the vapour-density may be as 
high as 70, and there is no doubt that the trend of the curve representing the relation 
of saturated vapour-density to pressure or temperature would, if prolonged and if no 
sudden change of direction occurs, give vapour-densities at low pressures and tempera¬ 
tures considerably higher than 60. We have shown in a paper, recently presented 
to the Society, that alcohol, and also ether, which may be regarded as typical of 
liquids in general, do not show this increase in the density of their saturated vapours 
with fall of temperature and pressure. We are at present engaged in a study of the 
thermal properties of acetic acid, which we hope will definitely decide this point. If 
the vapour-density of acetic acid is really higher than 60, at low temperatures no 
structural formula can represent the method of combination of the molecules. 
§ 81. The formation of ammonium chloride is usually believed to be attended by 
the change of valency in nitrogen from triad to pentad, and to be accompanied by the 
breaking down of the molecule of hydrogen chloride, thus : — 
N 
fH 
H I H 
H-f HC1=N H; 
H 
H 
l Cl 
and the existence of the extremely stable phosphorous pentafiuoride is a strong 
argument in support of this view. 
But if it be held, there appears no reason why we should not place in the same 
class potassium hydrogen fluoride, KHF 3 , and hydroplatinichloric acid, H 3 PtCl 6 , and 
other similar compounds. 
On the other hand, such compounds may be placed in the class of “ molecular 
combinations; ” but if this be done it must be conceded that such compounds are not 
always to be distinguished from “ atomic combinations ” by any ordinary tests of 
stability. 
Now ammonium chloride, although it agrees with acetic acid and with nitrogen 
peroxide in its behaviour as regards vapour-pressures, differs from them in respect of 
its dissociating into dissimilar molecules. It is thus evident that the agreement in 
behaviour of these three substances with that of ordinary solids and liquids cannot 
be ascribed to the similarity of the molecules produced by dissociation. 
§ 82. As nitrogen peroxide and acetic acid in their dissociation do not require a 
breaking down of a molecule such as water or ammonia, and as ammonium chloride 
agrees with them in its behaviour as regards vapour-pressure ; and as this agreement, 
as has already been shown, does not depend on the amount of dissociation, for there 
are some bodies in the other class totally dissociated, and some partially dissociated : 
