ON EVAPORATION AND DISSOCIATION. 
109 
§ G3. Nitrogen Peroxide .—This substance was prepared in the following manner. 
A mixture of arsenious oxide, sulphuric acid, and nitric acid was heated ; the vapour 
evolved was led through a U-tube containing arsenious oxide, and then through a 
drying-tube filled with phosphoric anhydride. It then entered a U-tube, provided 
with a bulb, which was placed in a freezing mixture. To avoid decomposition, the 
apparatus was constructed, as far as possible, in one piece. After redistillation the 
substance, which had a deep-red colour, was used for the following experiments. 
The vapour-density of nitric peroxide in presence of nitrogen was first determined 
by Playfair and Wanklyn (‘ Annalen,’ 122, p. 249). From their results they con¬ 
cluded that at 100° the gas consists chiefly of N0 3 , and at the ordinary temperature 
of N 3 0 4 . Muller (‘Annalen,’ 122, p. 15), and Deville and Troost (‘Comptes 
Rendus,’ 64, p. 237) determined its vapour-density by Dumas’ method, with very 
concordant results. Naumann (‘Annalen,’ Suppl., 6, p. 203) calculated from their 
data the percentage number of molecules decomposed at different temperatures, and 
found that at 26'7° it contained 80'04 p.c. of molecules of N 3 0 4 ; while at 135°, 1*31 
p.c. of the total number of molecules was left undecomposed. The curve constructed 
from Naumann’s numbers presents the usual double flexure of a dissociation-curve. 
At the boiling-point, under normal pressure, 21 - 7°, it is evident that less than 20 p.c. 
is dissociated. De la Fontaine (‘Archives Phys. Nat.’ (2), 28, p. 271) drew attention 
to the remarkable change of colour which nitric peroxide gas shows on heating; and 
as the liquid also grows lighter in colour when cooled, it appears probable that dis¬ 
sociation still proceeds in the liquid state (Naumann, ‘ Thermochemie,’ eel. 1882, 
p. 138). From these and similar results Willard Gibbs has constructed a formula, 
showing on thermodynamical principles the rate of dissociation of nitric peroxide. 
§ 64. Vapour-pressures of Nitric Peroxide. 
As liquid nitric peroxide quickly attacks mercury, it was impossible to determine its 
vapour-pressure by the usual method. Tolerably accurate results were obtained by the 
following device. An apparatus of the form given in the annexed drawing was 
employed. 
A large bulb A of about 50 ccs. capacity, very thin walled, yet strong enough to resist 
the pressure of the atmosphere, was sealed to a graduated glass tube of thermometer- 
bore, B, and was fused into an encasing bulb, C, the walls of which were much thicker. 
From the upper part of C a tube, D, projected vertically for some distance, and was 
then bent twice at right angles. The other extremity of this tube was enlarged at E; 
a T-piece, F, was sealed to the horizontal portion of the tube; the bulb A was filled 
with mercury, by exhausting it, and allowing mercury to enter. The bulb C was 
placed in a vessel with a perforated bottom, containing finely-pounded ice, which was 
heaped round the lower portion of the graduated stem, and continually renewed. As 
the bulb A forms an extremely delicate thermometer, in which 60° of the arbitrary 
