in the Discharge of Electricity through Gases. 281 



The theory readily explains why it requires a greater 

 electric force to send a spark through a thin layer of gas than 

 through a thicker one, provided the thickness of the thin 

 layer is less than a certain value which is inversely pro- 

 portional to the pressure. For in order that the ions should 

 produce dissociation, it is necessary that they should strike 

 against the molecules of the gas, or if the dissociation is due 

 to rays produced by the collision, that the layer should be 

 thick enough to absorb the radiation ; but as soon as the 

 thickness of the layer becomes comparable with the mean free 

 path of an ion, the chance of their striking against a molecule 

 before reaching the electrode begins rapidly to diminish; so 

 that the ability of the moving ions to produce dissociation 

 will rapidly diminish when the thickness of the layer falls 

 below a certain value. And although the thickness of the 

 layer when the electric force begins to increase is large com- 

 pared with the free path of a molecule moving through the 

 gas, it is not so compared with the free path of a corpuscle; 

 i. e., one of the small ions which are found in the cathode rays, 

 and which we have reason to believe play an important part 

 in all cases of electric discharge. 



On this view of the electric discharge, the presence of a 

 small number of ions is required to start the discharge. The 

 experiments of Linss *, and the recent ones of Elster ?• 

 Geitel t, seem to show that such ions are present in 

 under ordinar} r conditions ; the first spark, too, is m 

 irregular in its behaviour than subsequent ones, an^ 

 in accordance with the view that the conditions attending its 

 formation are capricious and almost accidental. 



Let us now consider the more complicated phenomena 

 attending the passage of the discharge through a gas at low 

 pressure. The production of ions may take place through- 

 out ihe tube, or it may be localized in certain places. There 

 must, however, be at least two places where ionization takes 

 place. For on this theory the ionization at a place A is 

 produced by ions brought into A by the electric field, so that 

 there must be some other centre or centres of ionization to 

 produce the ions which are brought into A by the electric field. 

 In the ' Philosophical Magazine ' for March 1899 I showed 

 how, from the study of the distribution of electric force in a 

 discharge-tube, we could determine the places of maximum 

 ionization; the result of such an investigation leads to the 



* W. Linss, Meteor. Zeits. iv. p. 352 (1887). 



t ' Terrestrial Magnetism and Atmospheric Electricity,' iv. p. 213 

 (1809). 



