The Discharge of Electricity through Gases. 539 



siderable amount of energy to be required to convert the metal surface 

 partially into dust, and it seems doubtful whether the mere abs orption 

 of ultra-violet light is sufficient for the purpose. If the action of the 

 light is primarily a chemical one, the disintegration might be ex- 

 plained, like the discharge itself, as a consequence of this chemical 

 action. 



The fact that flames behave as diamagnetic bodies has long been 

 known, and could be explained if we were free to assume that ions 

 are diamagnetic. From the point of view of Weber's theory of dia- 

 magnetics, this seems probable, and I made a few experiments in 

 order to prove in other ways the diamagnetism of ions. I have not 

 hitherto obtained decisive results, but some effects of magnets on the 

 discharge described in my previous Bakerian Lecture seem to me to 

 point directly to such a diamagnetism.* 



From the point of view, then, of our theory, a gas is sensitive elec- 

 trically when it contains free ions. On the other hand, we must 

 conclude that when a gas is not sensitive such free ions are not pre- 

 sent. I do not see- how the insulating power of air at the ordinary 

 temperature is consistent with the presence of ions, however few in 

 numbers ; for, ultimately, a diffusion to the electrodes and a discharge 

 would necessarily take place. This seems to me to be fatal to J. J- 

 Thomson's view of the disruptive discharge.f 



The Continuous Discharge. 



Facts hitherto established. 



The first step towards the investigation of the continuous discharge 

 is the investigation of the relation between current and fall of 

 potential in a gas through which a current is passing. An important- 

 fact discovered by HittorfJ relates to the fall of potential in the 

 positive part of the discharge, that part which extends from the 

 anode towards the negative glow. Hittorf's law may be stated as 

 follows : — 



" In the positive part of the discharge the rate of change of potential at 

 a given pressure is independent of the current." 



The law was tested by Hittorf with currents which differed in in- 

 tensity nearly in the ratio of 1 to 50. 



As regards the negative glow, Hittorf chiefly experimented with 

 electrodes in the form of wires running parallel to the axis of his 

 tubes. The glow for weak currents only covers the ends of the 

 electrode, but as the intensity is increased it gradually extends back- 

 wards; during this stage the rise of potential near the kathode 



* ' Boy. Soc. Proc.,' vol. 37, p. 317 (1884). 



f ' Phil. Mag.,' vol. 15, p. 427 (1883). 



X ' Wiedemann, Annalen,' vol. 20, p. 705 (1883). 

 VOL. XLVII. 2 » 



