1902.] Electrical Conduction in Gases, and Luminosity. 



103 



dissociate — i.e., oxygen would be given off, but would be in part 

 reabsorbed, so that the silver would never be entirely free from oxygen ; 

 the change, in fact, would oscillate about a point of equilibrium, and 

 would never be total in the direction corresponding to the complete 

 deoxidation of the silver. May not the bombardment of the glass 

 described by Sir Wm. Crookes have been effected by the displaced 

 oxygen molecules, and may not the surface heating of the pole have 

 been an outward and visible sign of the return of the truants to their 

 silvery home 



A diamond pole, such as Sir Wm. Crookes speaks of, might behave 

 in a similar manner in presence of a minute amount of oxygen ; in 

 fact, the argument applies generally to the occlusion of hydrogen and 

 other gases by electrodes. 



An argument which I think will sooner or later be regarded as of 

 weight in favour of the view that the phenomena are electrolytic in 

 their origin is afforded by the luminous manifestations in vacuum tubes. 

 These can scarcely be either mere collision effects, or mere heat effects. 

 It has long seemed to me that luminosity and line spectra are the 

 expression — the visible signs — of the changes attending the formation 

 of molecules from their atoms, or, speaking generally, that they are con- 

 sequences of chemical changes, a chemical change being one which involves 

 .an alteration of molecular composition, or it may be of molecular 

 configuration, as it is conceivable that even changes involving but the 

 formation of isodynamic (tautomeric) molecules — changes in molecular 

 structure unattended with change in molecular size — may give rise to 

 such manifestations. Dealing with the question of the luminosity of 

 hydrocarbon flames, I expressed this view in 1895 in the terms given 

 below.* 



* " Interest should be revived in the subject by the recent observations of V. 

 B. Lewes (' Hoy. Soc. Proc.,' 1895, vol. 57, p. 450) on the decomposition of acetylene 

 at high temperatures, which have led him to suggest that this compound plays the 

 chief part in promoting luminosity in hydrocarbon flames, and that the heat 

 liberated during its decomposition endows the carbon particles produced from it 

 with an incandescence far higher than corresponds to the temperature of the 

 flame. It appears to me that while accepting this as a partial explanation, it is 

 unnecessary to suppose that luminosity is consequent on the production of presum- 

 ably solid carbon particles ; the conversion of acetylene, or other hydrocarbons, at 

 high temperatures into hydrogen and carbon, although a decomposition in the 

 ordinary sense of the term, is doubtless a change involving the interaction of several 

 molecules, and the consequent formation of carbon molecules of a high order of 

 complexity together with hydrogen molecules ; and it is exothermic because the 

 energy liberated in the combination of the carbon atoms among themselves, and 

 also of the hydrogen atoms, is far in excess of that absorbed in the decomposition 

 of the hydrocarbon molecules. The initial luminance may therefore well be that 

 of the molecules at the moment of formation prior to condensation, although the 

 continued incandescence of the carbon molecules after their reduction to the solid 

 state probably contributes largely to maintain luminosity, giving to hydrocarbon 

 flames their special value, perhaps. But this is to explain the phenomena in terms 



