30 



second swings by the formula deduced by Langevin, i.e., 

 where 6 is the charge communicated to the electro- 



6 = 



1+m 



meter, and e and e^ are the first and second swings, and rri. 

 a factor which in this case had a value of '65. Even with 

 the sliding condenser it was not possible to eliminate small 

 effects due to rise of potential in the system, D, when the 

 gauze, G, was electrified. 



To make sure that this effect as well as leakage was 

 eliminated by subtracting readings with the semaphore on 

 and off a separate experiment was performed. The pin, L, 

 was set so .that the field was on all the time the chamber was 

 exposed to the a radiations. Results of experiments with the 

 semaphore on and off were subtracted. The curve obtained 

 showed close agreement with that obtained with the pendu- 

 lum removed (and the saturation curve obtained in the ordi- 

 nary manner) and formed a valuable check upon the relia- 

 bility of the whole apparatus. 



In this form of experiment ordinary recombination is 

 not eliminated, and to allow for it we must consider the ar- 

 rangement of the ions in the chamber, as they cross it under 

 the influence of the field. As shown by Langevin, if in the 

 first place the ions be assumed to be uniformly distributed 



Fig. 9. 



and the velocity of the positive and negative be v cm. per sec. 

 under unit P.D. (per. cm.), the time taken for an ion to move 



across the chamber from electrode to electrode = — where L 



V 



is the distance between electrodes. If time be reckoned 

 from the instant the field is applied, after an interval of time 



^ recombination will have ceased. The number of ions lost 



