274 !> H. A. Wilson. [June 3, 



The negative leak in air at constant temperature in general increases 

 with the pressure. This is shown to be due to ionisation of the air 

 produced by the collisions of the negative ions coming from the wire 

 with air molecules. If the P.D. used is small, no ionisation by 

 collisions occurs, and then the negative leak is independent of the 

 pressure at low pressures. 



If n a ions leave the wire, the number n^ of ions reaching the cylinder 

 is shown to be given approximately by the formula 



, T -NE b -NE/J&, 6 



" ~~ 



In this formula 



V = potential difference between wire and cylinder, 

 b = radius of cylinder, 

 a = radius of wire, 

 p = gas pressure, 

 N = maximum number of negative ions produced by one negative 



ion in going 1 cm. at unit pressure, 



E' = potential through which a negative ion must fall to enable 

 it to ionise an air molecule. 



Professor Townsend has shown that the number a of negative ions 

 produced by one negative ion in going 1 cm. is given approximately 

 by the formula 



-NEji 



a = N/?~ X , 



where X is the electric intensity. This formula of Townsend's is used 

 in deducing the above expression for %/%. 



It is shown that N varies nearly inversely as the absolute tempera- 

 ture of the gas through which the negative ion moves. 



The variation of the negative leak with the temperature is investi- 

 gated theoretically on the assumption that the liberation of negative 

 ions or corpuscles at the surface of the platinum is analogous to the 

 evaporation of a liquid, and the formula 



x = A J~e c-Q/2* 



is obtained.* In this formula 



x = negative leak per sq. cm. of platinum, 



== absolute temperature, 



Q = energy in gramme calories required to produce 1 gramme 



molecular weight of ions, 

 A = a constant. 



* A formula of this type was first used by the author to calculate the energy 

 necessary for the production of ions from the temperature variation of the leak 

 from hot platinunv.in 1901. See ' Phil. Trans.,' A, 1901, p. 430. 



