PRESIDENTIAL ADDRESS. 379 



particular mixture a smaller proportion of the original oxygen should appear as 

 steam in the products than would be the case with either the 2011., + 0, or the 

 CH.+O, mixture, which again is precisely what we find. 



In considering the question of the explosive combustion of hydrocarbons it is 

 important to distinguish between (1) the primary oxidation of the hydro- 

 carbon, which is an exceedingly rapid process and is probably completed during 

 the short interval between ignition and the attainment of maximum pressure, 

 and (2) certain probable secondary interactions where influence may extend far 

 into the subsequent cooling period; for it is only this latter which would be 

 effected by variations in the rate of cooling down from the maximum 

 temperature. Such secondary interaction may include (a) the reversible change 

 CO + OH^ i^ "* : (JOo -f H, and, in cases where carbon is deposited as the result of 

 the decomposition of primary oxidation products, the interaction of steam and 

 carbon C + OH, = CO + Hj. In this connection, I may draw attention to the 

 recently published work" of G. W. Andrew," one of my former pupils and 

 collaborators, on the ' Water Gas Equilibrium in Hydrocarbon Flames ' which 

 proves that in a system containing only CO^, CO, H,, H^O, rapidly cooling down 

 from the high temperatures prevailing in hydrocarbon flames, the equilibrium 



ratio --w^^fT- adjusts itself automatically with the temperature until a point 



COj X Hj 

 bjet-ween 1500° and 1600° C. on the cooling curve is reached (corresponding to a 

 value K=4-0 approximately), after which no further readjustment occurs. He 

 also found that this adjustment of equilibrium is not greatly influenced even 

 when relatively large quantities of methane and carbon are found in the tinal 

 products. 



I am able, from my own experiments, to confirm Andrew's conclusions in 

 all cases where the initial firing pressure is insuflicient to set up detonation ; 

 but, in cases where both detonation and separation of carbon occur, my results 

 undoubtedly indicate an appreciable intervention of the separated carbon dui'ing 

 the cooling period. There is nothing in my results, however, suggestive of an 

 appreciable intervention of methane. 



The fact that the primary oxidation of methane usually involves a direct 

 transition from CH^+Oj to CH,(OH),, which latter breaks up into, ultimately, 

 CO + H,-|-H„0, -without any deposition of carbon, opened up the possibility of 

 instituting a direct experimental comparison between the relative affinities of 

 methane and hydrogen for oxygen in explosions by exploding a series of 

 mixtures CH^ + O^+aHj in which the hydrocarbon and oxygen were initially 

 present in as nearly as possible equimolecular proportions, but in which x (the 

 vokune ratio of H^ to CHJ was varied between 2 and 8, and determining 

 (1) the oxygen distribution on explosion when x=2, and (2) the influence upon 

 such distribution of successive equal increments of x up to 8. 



Ever since Davy's experiments on Flame, the combustibility of hydrogen 

 has been erroneously considered to be superior to that of methane ; but, on the 

 other hand, my previous researches upon hydrocarbon combustion have shown 

 (1) that in slow combustion in borosilicate glass bulbs at temperatures between 

 300° and 400° C. methane, ethane, and acetylene are all oxidised at a much 

 faster rate than is either hydrogen or carbon monoxide ; and also (2) that in 

 exploding such mixtures as C,H^ + H, + 0„or G„B.„ + 2il2+ ^2 ^1^^ hydrocarbon 

 is burnt in pi'eference to hydrogen, facts which are at variance with any notion 

 of the superior aflanity of hydrogen for oxygen in flames. 



My further experiments upon the relative affinities of methane and hydrogen 

 for oxygen in explosives were carried out at high initial pressures in two 

 special steel bombs, namely, A, with a cylindrical explosion chamber 8 inches 

 long and 1 inch in diameter ( capacity = (irca 103 c.c), and, B, with a spherical 

 cavity 3 inches in diameter (capacity = c/rca 275 c.c). The ratio of wall surface 

 of the explosion chamber in the case of Bomb A would thus be about 2-75 times 

 that of Bomb B. The mean results obtained for the distribution of oxygen 



" Trans. Vhem. Soc. 1914, 105, PP- 444 to 456. 



