Bumstead and Wheeler — Radio-active Gas. 109 



and this justifies the determination of k by the method pur- 

 sued above, without regard to the induced activity. 



In discussing the diffusion experiment, it is convenient to 

 take the initial instant (£=0) as that when the diffusion begins ; 

 we have then 



^ A X A 



dt r 



where \i is the diffusion constant for this gas through the 

 porous plate used, and thus, 



and 



as before, 



A = A e -(*+J*)« 



1 + x ~ k ) 



= a\— AE 



dt 



which under the assumed initial conditions gives 



X — K fJL { X — K , 



E 



7* 



and 



I = A. Ul + ~A e-(^^ _ _«_ '^JtiJL. e -M \ (4) 



L\ X — K— jX/ X K X K — jX ) 



and by properly choosing the value of /jl this should fit the 

 observations plotted in figs. 6 and 7. In order to get a first 

 approximation to /jl we may make use of the initial slope of 

 the curve : 



\dt/t=0 ° ; ( K—K—ii\ \-k} j X! 



and since k is so small in comparison with X, we have approxi- 

 mately 



As has been pointed out, the value of A , that is the activity 

 of the gas itself without the induced activity, is very imper- 

 fectly known. If we take the values indicated by the exten- 

 sions of the curves in figs. 4 and 5 (30 to 40 per cent of the 



