492 REPORTS ON THE STATE OF SCIENCE. 



45 per cent, of that required at the second stage of the ethylene scheme 

 — in none of the experiments was it found possible to detect or isolate 

 any of the monohydroxy derivatives which are, ex hypothesi, initially 

 formed. This was perhaps hardly to be expected in the cases of 

 ethylene or acetylene, where the monohydroxy derivative would be 

 extremely unstable. In the case of ethylene, however, acetaldehyde, 

 which is known to be readily formed by molecular rearrangement from 

 vinyl alcohol, was isolated in certain experiments. But failure to 

 detect the formation of ethyl alcohol during the slow oxidation of ethane 

 did at first sight seem a serious obstacle to the acceptance of the 

 ' hydroxylation ' theory. This difficulty was, however, largely removed 

 when it was subsequently found that ethyl alcohol is oxidised at 

 a much faster rate than is ethane under like conditions — a circumstance 

 which seems to warrant the view that the effect of the initial ' hydroxy- 

 lation ' of the hydrocarbon is to render the molecule much more sus- 

 ceptible to further attack. Finally, when Drugman, working in the 

 writer's laboratory, obtained direct proof of the large formation of 

 ethyl alcohol as the result of the interaction of ethane and ozone at 

 j 00°, the difficulty referred to entirely disappeared. Moreover there is 

 strong indirect proof of the initial formation of monohydroxy deriva- 

 tives during slow combustion in the fact that whereas the rates of 

 oxidation observed with mixtures containing two molecules of hydro- 

 carbon to one molecule of oxygen were hardly, if at all, inferior to those 

 observed with equimolecular mixtures, the process was always much 

 retarded by any further addition of oxygen beyond the equimolecular 

 ratio. It may also be urged that the hydroxylation theory 

 readily explains the large formation of carbon dioxide in the bulb 

 experiments under conditions which would entirely preclude the idea 

 of its formation by the direct oxidation of the monoxide or as the result 

 of interaction of the monoxide with steam. 



The next step in the inquiry was to ascertain whether the presence 

 of moisture is essential to hydrocarbon combustion. A series of careful 

 experiments on mixtures of the three typical hydrocarbons respectively 

 with oxygen in equimolecular proportions thoroughly dried by long 

 contact with redistilled phosphoric anhydride, under conditions which 

 (as was proved) inhibited the formation of steam from electrolytic gas, 

 gave wholly negative results. If anything, combustion occurred rather 

 more readily with the dried gases than in the case of the corresponding 

 experiments with undried mixtures. Therefore, whilst the conclusion 

 finally advanced concerning the mechanism of hydrocarbon combustion 

 agrees with the view originally put forward by H. E. Armstrong, 

 in so far as the nature of the intermediate products is concerned, it 

 differs from his in supposing that the oxygen is directly active. 



With the extension of the investigation to hydrocarbon flames and 

 explosions, including ' detonation ' and explosions under high initial 

 pressures, it became increasingly evident that the mechanism of com- 

 bustion is essentially the same above as below the ignition-point, in so 

 far as the result of the initial molecular encounters between hydrocarbon 

 and oxygen is concerned. At the higher temperatures of flames 

 secondary thermal decompositions undoubtedly come into operation at 



