76 Prof. Thomson, On the Conductivity of Gases 



bourhood of the discharge. There is considerable range of 

 pressure in which the gas exposed to the 'Entladungsstrahlen' is 

 a conductor, while the gas not so exposed remains an insulator. 

 I conclude therefore that the ' Entladungsstrahlen ' produce con- 

 ductivity in the gas through which they pass. 



The apparatus was filled with hydrogen and then with carbonic 

 acid gas. It was found that the leak when the tube was filled 

 with hydrogen was much more rapid than when it was filled with 

 air, while the leak through carbonic acid was very much slower 

 than through air. This is in accordance with the observations of 

 E. Wiedemann and Hoffmann that the effects of the Entladungs- 

 strahlen are much greater in hydrogen and much less in carbonic 

 acid than in air. 



If the length of the arc in the horizontal discharge tube was 

 kept constant, and the terminals moved along the tube so that in 

 one position the terminals were clear of the opening from the 

 horizontal to the vertical tube, while in another position one of 

 the terminals was above the opening, the leak in the cylinder 

 exposed to the Entladungsstrahlen was greater in the second 

 position than in the first; showing that the Entladungsstrahlen 

 are emitted more copiously from the neighbourhood of the electrode 

 than from the rest of the spark. 



As the use of the alternating current from the transformer 

 made it impossible to separate the effects at the positive from 

 those at the negative electrode, the transformer was replaced by 

 an induction coil and the spark gap was moved along in the top 

 tube so that different portions of the discharge came over the 

 opening into the tube containing the cylinders. It was found 

 that the leak was greatest when the negative electrode was over 

 the opening, that there was a considerable leak when the positive 

 column was over the opening, but that there was no leak when 

 the dark space between the positive column and the negative 

 glow was in this position. From the experiments of Graham on 

 the distribution of potential along the discharge we may conclude 

 (see Phil. Mag., Mar. 1899) that at the cathode and close to the 

 anode ionization of the gas is taking place, while recombination 

 of the ions occurs along the positive column, and very little of 

 either ionization or recombination in the dark space between 

 the positive column and the negative glow. Thus the ionization 

 and recombination of a gas are accompanied by the emission of 

 Entladungsstrahlen. These, I think, are analogous to the secon- 

 dary rays which Rontgen has shown are emitted by a gas exposed 

 to Rontgen radiation in which ionization is taking place ; and to 

 the secondary rays which are produced when Rontgen rays fall 

 on a metallic surface. I have shown in a paper read before the 

 Society (Proc. Camb. Phil. Sod., Vol. ix. p. 393) that the sudden 



