490 Wellisch and Bronson — Distribution of the Active 



and case activities are determined by means of ionization 

 which is differently distributed. This has been shown experi- 

 mentally to have a negligible effect on the relative values of 

 the percentage cathode activities for different potentials, the 

 determination of which was really the object of the experi- 

 ment. As mentioned above, it was important to make an accu- 

 rate determination of the maximum activity on the case. An 

 error in vhe determination of this activity leads to approxi- 

 mately the same percentage error in the determination of the 

 percentage cathode activity ; this error in all our experiments 

 is probably less than 2 per cent. 



The plotting of a curve such as that of fig. 2 is obviously 

 only possible if the percentage cathode activity is independent 

 of the amount of emanation used in the experiment. This fact 

 was demonstrated repeatedly during the course of our inves- 

 tigations ; as an instance of the wide range over which this 

 holds it may be mentioned that the percentage cathode activity 

 was the same with 900 volts on the case when the amounts of 

 emanation in the vessel were such as to afford equilibrium rates 

 of 5*32 without any added capacity and 16'1 with the added 

 capacity, i. e. rates which are roughly in the ratio 1 : 65. When 

 we are working in the earlier part of the curve, i. e. with 

 relatively small potentials, this independence no longer holds 

 good ; but for the higher potentials the effect of the amount 

 of emanation within a large range becomes negligible. 



The curve bears a striking resemblance to the curves which 

 have been experimentally determined for ionization by a-par- 

 ticles. Many observers have drawn attention to the "lack of 

 saturation" which is a marked feature of such curves ; in par- 

 ticular, this phenomenon has been the subject of special inves- 

 tigation by Bragg,* Kleeman,f Moulin,;}; and Wheelock.§ 

 Reference is made to this point later (section 7). The similar- 

 ity was so striking as to suggest a more detailed study of the 

 ionization curves for emanation in equilibrium and with differ- 

 ent applied potentials. 



It was found that within the limits of error the ratio of the 

 two ionization currents obtained for two potentials not too 

 low| was, like the percentage cathode activity, independent of 

 the amount of emanation in the vessel ; and, moreover, that 



* Bragg, Phil. Mag. (6), xi, p. 466, 1906. 



t Kleeman, Phil. Mag. (6), xii, p. 273, 1906. 



% Moulin, Compt. Rend., cxlviii, p. 1757, 1909; Le Radium, vii, p. 350, 

 1910. 



§ Wheelock, this Journal, xxx, p. 233, 1910. 



|| The potentials should be sufficiently large to prevent volume recom- 

 bination, as distinguished from columnar recombination, i. e. should be 

 sufficient to saturate a uniform distribution of Rontgen-ray ions equal in 

 number to those produced by the a-radiation. 



