Prof. J. J. Thomson on Conductivity of Hot Gases. 135 



the discharge under a variety of conditions, and found that 

 when liquid surfaces (mercury or water) were used in place 

 of the ball, also that when hydrogen, carbon-dioxide or 

 illuminating-gas were substituted for air, as a dielectric, the 

 ball-and-point effect, more or less modified, could still be 

 obtained. The investigation of these points, although it has 

 already led to some results of significance, is as yet very 

 incomplete. 



Physical Laboratory of Cornell University, 

 September 1890. 



XVI. Conductivity of Hot Gases. 



To the Editors of the Philosophical Magazine and Journal. 

 GENTLEMEN, Cambridge, January 19, 1891. 



IN the January number of Wiedemann's Annalen Dr. 

 Arrhenius publishes an account of some experiments on 

 the conductivity of hot vapours. His results differ from those 

 obtained by me (Phil. Mag. April and May 1890) in the 

 cases of the vapours of hydrochloric acid, hydriodic acid, and 

 ammonium chloride. According to his experiments these 

 vapours do not conduct appreciably better than air at the 

 same temperature, while I found that their conductivities 

 were very greatly in excess of that of air. The difference 

 between these results is, I think, easily explained on the view 

 given in my paper, that the conduction of electricity through 

 hot gases is due to dissociation. The method employed by 

 Dr. Arrhenius was to inject these substances into a flame; 

 thus the hydrochloric-acid gas, for example, would be sur- 

 rounded by a multitude of other gases, and especially by 

 hydrogen. The presence of a large excess of hydrogen would 

 retard the dissociation of a gas such as hydrochloric acid, of 

 which hydrogen is one of the products of dissociation ; for it 

 is a well-recognized principle in the theory of dissociation, 

 that it is retarded by the presence in excess of one of the 

 products of dissociation. This has been verified by Wurtz, 

 who found that an excess of PC1 3 stopped the dissociation of 

 PC1 5 . In Dr. Arrhenius's experiments the dissociation of the 

 hydrochloric acid would thus be very much less than it was 

 in mine, when the gas was heated by itself, and when the 

 dissociation was large enough to be detected by chemical 

 means. This, on the view that the conductivity is due to 

 dissociation, would be sufficient to explain the difference in 

 the results. I have found myself that the conductivity of 

 HC1 is very much reduced by dilution with hvdroaen, more 



