Discharge of Electricity through Gases. 323 



already on a former occasion drawn attention to the very suggestive 

 fact that the compounds in which the molecules are made up of two 

 or more carbon atoms (cyanogen, hydrocarbons) give the carbon bands, 

 while those containing only one atom (carbonic oxide) show the true 

 line spectrum of carbon. Angstrom and Thalen observe that the violet 

 line of carbon is generally visible in tubes filled with carbonic oxide, 

 but I have only observed this line in the negative glow, where it 

 undoubtedly is visible even when no trace of it can be seen in any 

 other part of the tube. I have also observed that tubes filled with 

 carbonic oxide show a band peculiar to the negative pole. Though 

 this observation is nearly ten years old, I have hitherto left it 

 unpublished together with other matter bearing on the same point, as 

 I had hoped to find time to complete my investigations and to write 

 a general description of the spectra of carbon compounds. 



I do not think that any of the other gases have been sufficiently 

 well investigated to be noticed here ; but some observations which I 

 have made at different times on the spectrum of chlorine confirm the 

 general conclusion which I have drawn. 



Messrs. De La Rue and Muller have noticed in some of their 

 experiments a sudden expansion of the gas when the discharge was 

 sent through it. They have proved that this expansion is not due to 

 increased temperature, and a breaking up of the molecules into atoms 

 seems to me to be its next natural explanation. 



While we thus have ample evidence of dissociation whenever an 

 electric current passes through a gas, we also find that whatever 

 increases independently the dissociation improves the conducting 

 power of the gas. Thus we know that a flame is a good conductor, 

 and Hittorf has, in a series of very interesting experiments, shown 

 that if we heat up the electrodes to a white heat, an electromotive 

 force of a few volts will send a current through the gas. 



The fact also discovered by Hittorf, that if a discharge is set up a 

 small electromotive force is sufficient to pass a current across, is also 

 easily explained by our theory, as the original discharge throws the 

 molecules into that state of dissociation which favours the passage of 

 the current. 



On the Influence of a Positive Electrode on the Negative Glow. 



I now pass to the description of a series of experiments which seem 

 to me to be only capable of explanation by the views brought forward 

 in this paper, and I should like therefore to consider them as crucial 

 experiments which have to be explained by any true theory of the 

 electric discharge. 



It was up to recently believed that the form and extent of the 

 negative glow was determined only by the shape of the negative 

 electrode, but was independent of the shape of the vessel or the 



