the Electrical Conduction produced by it. 147 



Similarly the ratio p x of the number of positive ions that 

 give up their charge to the external cylinder to the total 

 number is given by 



lh = W7t — (2) 



In the above equations it is assumed that the current of 

 air is uniform over the cross-section of the tube, and that the 

 ions are uniformly distributed over the cross-section ; also, 

 that the movement of the ions does not appreciably disturb 

 the electric field. Since the value of t ran be calculated from 

 the velocity of the current of air and the length of the elec- 

 trode, the values of the velocities of the ions under unit 

 potential gradient can at once be determined. 



The equation (1) shows that p 2 is proportional to V, — 

 i. e. that the rate of leak of the electrode A varies directly as 

 the potential of A, provided the value of V is not large 

 enough to remove all the ions from the gas as it passes 

 by the electrode. This was experimentally found to be the 

 case. 



In the comparison of the velocities the potential V was 

 adjusted to such a value that p 2 was about one half. This was 

 determined by testing the rate of leak at B with a saturating 

 electromotive force. The amount of recombination of the 

 ions between the electrodes A and B was very small, and 

 could be neglected. 



The uranium cylinder was then removed, all the other 

 parts of the apparatus remaining unchanged. An aluminium 

 cylinder was substituted for the uranium cylinder, and ,^-rays 

 were allowed to fall on the aluminium. The bulb and 

 induction-coil were placed in a metal box in order to care- 

 fully screen off all electrostatic disturbances. The rays were 

 only allowed to fall on the central portion of the cylinder. 

 The intensity of the rays was adjusted so that, with the same 

 current of air, the rate of leak was comparable with that 

 produced by the uranium. It was then found that the value 

 of p2 was nearly the same as for the uranium conduction. 

 For example, the rate of leak of B was reduced from 38 to 

 14 scale-divisions per min. by charging A to a certain small 

 potential, when the air was blown by the surface of the 

 uranium. When Rontgenized air was substituted, the rate of 

 leak was reduced from 50 to 18 divisions per min. under the 

 same conditions. The values of p 2 were '63 and *64 respec- 

 tively. This agreement is closer than would be expected, 

 as the bulb was not a very steady source of radiation. 



L 2 



