on Ionization produced by [ot Platinum in Gases. 405 
to something shot off from the cathode, and not merely 
diffusing about in the neighbourhood of that electrode. 
The relation between the imparted emissibility and (1) the 
duration of the discharge with constant current and (2) with 
the intensity of the discharge current when this was allowed 
to pass for equal intervals of time, was next examined. The 
imparted effect was found to be very nearly proportional to 
the first power of the duration of the current and to the 
second power of its strength. 
It was shown in the previous paper that the initial positive 
leak from a hot platinum wire falls off with time in such a 
way that the current y is given at time ¢ by the expression 
g—Ne; 
where A and & are constants, A being proportional to the 
amount of substance present giving rise to the ions, and k 
being the velocity constant of the reaction in which this 
substance decomposes. A more rigorous examination of this 
formula has shown that it only applies over a limited range. 
The positive leak appears always to decay asymptotically with 
the time, although it does not follow an exponential law. = It 
is probable that in general it is produced by the decay of 
more than one substance, in which case it would be expressible 
as the sum of a number of terms of the above type. 
Another way of demonstrating the same thing is to 
measure the total quantity Q of electricity received by the 
electrometer at various times. Since on is equal to the 
current y we have 
if the above simple formula holds. Whence 
Q=Q,(1— Fe 
if Q, is the charge received by the electrometer atter an 
infinite time has elapsed. In seeking to verify this formula, 
curves such as those shown in fig. 2 were obtained. It was 
impossible to fit these curves by so simple a formmnla as the 
above; the theoretical curves become much flatter after a 
few minutes have elapsed, when the two coincide nearer the 
origin. 
It may perhaps be thought that the fact of the leaking 
current not dying away rapidly enough, 7. e. that after the 
lapse of, say, ten minutes quite a considerable fraction of 
the imparted effect still remains on the wire, would be a 
serious difficulty in measuring the imparted emissibility. 
As a matter, of fact this difficulty did not arise, since the 
Phil. Mag. 8. 6. Vol. 8. No. 45. Sept. 1904. 2H 
