226 Prof. Rutherford on Dependence of the Current through 



gases and under different pressures. On studying the matter 

 experimentally, however, an unexpected source of error dis- 

 closed itself, which renders the method inapplicable in the 

 case of ionization by Eontgen-rays. With a very intense 

 beam of rays, the ions produced were not confined only to the 

 cross-section of the beam, but were produced far outside the 

 geometrical limit of the rays. By means of suitable lead 

 screens, with slits cut in them, care was taken that the rays 

 did not impinge on the surface of the plate close to which 

 they passed, so that the effect was not due to the secondary 

 rays emitted by the metal plate. 



The effect appears to be entirely due to the fact that the 

 air through which the rays pass becomes a source of secondary 

 radiation, which has the power of ionizing the gas in its 

 neighbourhood. The experiments of Rontgen and others 

 show that such an effect exists ; for a fluorescent screen is 

 lighted up outside the path of rays. The effect of these 

 secondary rays in producing ions between the plates was 

 readily examined by allowing the rays to pass near one plate/ 

 and then measuring the current between a metallic gauze 

 placed outside the path of the rays, and a parallel plate placed 

 some distance away. The experiment showed that with an 

 intense beam of rays, the amount of ionization produced 

 between the gauze and the plate was large enough to give an 

 error of several hundred per cent, in the determination of the 

 velocities, if the ionization was assumed to be confined to the 

 surface of one plate. 



This volume ionization, which is produced outside the path 

 of the rays, causes the current for a given P.D. to be much 

 larger than if the theoretical condition of surface ionization 

 only was fulfilled. 



It is possible that by using a bulb giving out soft rays, the 

 effect of this secondary ionization may be greatly diminished ; 

 but in any case it will always be a very serious source of error 

 in the method. 



As an instance of the value of the velocity deduced by the 

 application of this method, we will consider a numerical 

 example : — 



Distance between plates 10 cms. 



Mean width of stratum of rays 1 cm. 



The distance d lies somewhere between 9 and 10 cms. 



As an approximate value we will take d = *d cms. 



The current per square centim. was experimentally found 

 to be *0021 E.S. unit per square centim. when the plate A 

 was positive, and 32*5 volts between the plates. 



