142 PROFESSOR H. B. DIXON ON THE RATE OP EXPLOSION IN GASES. 
Since steam is known to be partially dissociated under atmospheric pressure at the 
temperature of the oxy-hydrogen flame, it does not seem improbable that dissociation 
would also occur at the higher pressures and higher temperatures of the explosion- 
wave. The researches of MM. Berthelot and Vieille, and those of MM. Mallard 
and Le Chatelier, on the pressures registered in an explosion of gases, have led 
these investigators to the conclusion that the specific heat of steam rapidly rises 
with the temperature. The deficiency of “ available ” pressure, which Bunsen first 
observed in the explosion of gases and attributed to incomplete combustion, they 
consider to be due to an increase of specific heat. Such an increase in the specific 
heat of steam with rise of temperature would explain the divergence between the 
observed and calculated rates of explosion of hydrogen with oxygen, and with nitrous 
oxide. And conversely, it appears to me that the results of the French experi¬ 
menters would be equally well explained by the temporary dissociation of steam in 
their explosions. 
One fact appears, at first sight, to be opposed to the hypothesis of incomplete 
combustion. The addition of oxygen, one of the products of the dissociation of 
steam, retards the explosion of electrolytic gas more than the addition of an equal 
volume of nitrogen—an inert gas. Wurtz found, in the case of phosphoric chloride, 
that the amount of dissociation was greater when the chloride was volatilised in air 
—an inert gas, than when it was volatilised in the presence of the lower chloride— 
one of the products of its decomposition. Arguing from this observation, one might 
have expected that excess of oxygen would tend to hinder dissociation, and therefore, 
if the combustion were really incomplete, that the rate should be faster when oxygen 
was used as a diluent instead of nitrogen. But the condition of equilibrium under 
which Wurtz determined the vapour density of phosphoric chloride is not the 
condition of the wave-front of an explosion. If we imagine in the wave-front a 
molecule A dashing against a molecule B with such extreme velocity that the com¬ 
pound AB is not formed, the two molecules may be supposed to rebound, exchanging 
energies, and the motion of B may be communicated to the molecules in front 
continuing the explosion. This is a case where dissociation of AB occurs in the 
wave-front. Now if a number of B molecules are present in excess, A will shortly 
combine with one of them, behind the wave-front, and the dissociation wall not be 
permanent. On the other hand, if there are a number of inert molecules C present, 
dissociation may occur as before in the wave-front, but the uncombined molecules 
may be prevented from combining behind the wave by the inert molecules present. 
The propagation of the wave does not depend on what occurs after it has passed; 
the rate of the v'ave will therefore be independent of the permanency of the dis¬ 
sociation occurrinj?' in it. 
o 
That the combustion of pure electi'olytic gas is not wholly complete in the explosion- 
wave has been proved by collecting the residue and exploding it."'" In the propagation 
* “ lucompleteness o£ Combustiou in Gaseous Explosions,” by H. B. Dixox and H. W. Smith. 
‘ ManclLCster Memoirs ’ [IV.], vol. 2. 1888. 
