The Discharge of Electricity through Gases, 527 



theory which seemed to me to form a hopeful starting point for future 

 investigation. I have every reason to be satisfied with the way in 

 which that theory has been received by other workers in the same 

 field, and I think that, in spite of difficulties, which I do not wish to 

 underrate, it is generally considered to have a good chance of ultimate 

 success. I have since been able to extend my views, and am bold 

 enough to think that we may now form a fairly complete idea of the 

 most important features of the gas discharge. 



In 1882, that is, two years before my paper was presented to the 

 Royal Society, Mr. Griese* was led, by a study of the electric corb- 

 ductivity of the gases which rise from flames, to the identical 

 hypothesis which formed the basis of my paper. I much regret that 

 I have only recently become acquainted with Mr. Giese's work, as 

 I should have been glad to have drawn attention to it. 



We both assume that each molecule of a gas contains atoms which 

 carry equal and opposite charges ; and that these charges are the 

 same as those carried by the ions in electrolytes ; that, further, a 

 current of electricity through a gas can only be maintained by a 

 diffusion of the charged atoms through it. I have never claimed 

 much originality for this hypothesis, as the same idea must have 

 occurred to many others, and it would probably be difficult to trace 

 its early history. The hypothesis of definite charges only forms a 

 small part, however, of the theory which I have sketched out in my 

 previous paper. By itself alone it is not sufficient to account for the 

 observed facts. In order to distinguish the theory from others, 

 I shall call it the theory of " electrolytic convection," as it resembles 

 in many ways the kind of conduction in liquids which has been 

 described under that name by Helmholtz. The theory of electrolytic 

 convection offers obvious explanations of a number of different 

 phenomena, as has been shown by Giese for the discharge through 

 gases rising from flames, by myself for the discharge introduced 

 by strong electromotive forces, and by Elster and Greitel for a number 

 of other phenomena. According to this theory, a gas insulates as 

 long as there are no free ions present, but acts as a conductor as soon 

 as, through some cause or other, the molecules are split up into ions. 

 We meet with formidable difficulties, however, when we come to 

 discuss the transference of electricity between the solid conductor 

 and the gas ; for here we are to a great extent tied down by our 

 knowledge of electrolysis in liquids. We know that a difference of 

 potential of 2 volts between the electrodes is sufficient to decompose 

 water, and therefore to allow the interchange of electricity between 

 the metal and the ion of hydrogen or oxygen. .No theory of gas 

 discharges is admissible which cannot be brought into harmony with 

 these facts. 



* 'Wiedemann, Annalen,' vol. 17, p. 537 (1882). 



