﻿470 Dr. Norm an Campbell on 



could be obtained. At lower pressures noticeable ionization 

 by collisions began before saturation was attained. It was 

 found, however, that within the range from 100 to 700 mm. 

 the current was very accurately proportional to the pressure, 

 so that the interpolation to lower pressures does not introduce 

 much uncertainty. And yet the greatest uncertainty in the 

 experiments is connected with n', for the values obtained 

 on different occasions varied by as much as 4 per cent. 

 (Allowance was made, of course, for the decay of the 

 polonium ; the accepted value X = 139 (days) -1 agreed well 

 with the observations.) This uncertainty is probably due to 

 the method of measurement adopted. After the experiments 

 were finished— rather late in the day — a more complete in- 

 vestigation into the properties of the high resistances was 

 undertaken. The results will be published shortly. It was 

 found that the temperature coefficient was higher than had 

 been suspected, and that there were sources of error con- 

 nected with the lower resistance (by means of which n was 

 measured) which did not affect the higher resistance (which 

 had been investigated before). If this inquiry had been made 

 earlier a greater accuracy could certainly have been attained, 

 but nothing has been found to throw doubt on the results. 

 Indeed^ the results of the measurements themselves may be 

 judged to be good evidence of the value of the method 

 em ployed. 



wj, the number of ions liberated at the electrode per second, 

 or, in these experiments, the number of delta rays excited 

 there 5 Can be ascertained by measuring the saturation current 

 at the lowest pressures attaii<able (<'001 mm.). In the 

 earlier observations this quantity was measured several times 

 daring the process of finding the relation between i and y, 

 and the values obtained on a single day sometimes varied as 

 much as 2 per cent. Moreover, the currents at higher pres- 

 sures varied with w,, leading to some irregularity in the 

 results obtained. (Thus in Tables XV. and XVIII. it will 

 be seen that the points do not lie accurately on a smooth 

 curve.) The variation was traced to the presence of an air- 

 film on the surface of the electrodes, which increased the 

 emission of delta rays from theim If the pressure was kept 

 below 0'5 mm. the air-film was gradually destroyed; during 

 three days when the pressure was kept as low as possible 

 continuously, rij fell by 14 per cent. (The same change was 

 noticed in the first paper on delta rays, p. 298.) Later, 

 therefore, the pressure was never reduced below 0*5 mm. 

 during the course of a set of experiments, and under these 

 conditions there was no irregularity in the curves. n Y was 



