96 ROYAL SOCIETY OF CANADA 



The effect it will be seen is a small one and the time required to 

 make the readings was long. The readings, however, were quite regular, 

 and with the precautions and care taken there seems no reason for inter- 

 preting the results otherwise than as representing a charging action pro- 

 dv.oed by the radiation emitted by the insulated potassium salt on the) 

 tray. Observations were made under precisely the same conditions with 

 the tray empty, but these failed to show any such charging action as that 

 obtained with the salts. These experiments with the potassium cliloride 

 were made, as stated above, with the apparatus shown in Fig. Til, but 

 they were carried out before the apparatus was used in the uranium 

 nitrate experiments, and consequently there was no possibility of a con- 

 nection between the charging action observed with the potassium salt 

 and any contamination of the apparatus through contact with the 

 uranium nitrate. This charging action with a salt of potassium was 



rmffllStt*:tttr4:rr:i 



rif;: ^j^::^ntgj!îRtrittimta;4 àv-tffin:<ai.:it'iÛhC;t^rr;tJ<i'^ i: 



Fig. VIII. Fig. IX. 



also observed with the apparatus shown in Fig. IV. In this set of ob- 

 servations the salt used was the sulphate of potassium. Two sets of 

 measurements were again made. In the one the potential on the needle 

 was positive and equal to 243 volts and the pressure of the residual air 

 in the apparatus was .08 mm. In the second set the needle was kept at 

 a negative potential of 243 volts, and the pressure of the air was main- 

 tained at .06 mm. In these two latter measurements it was not found 

 possible to lower the pressure of the gas below the readings stated, al- 

 though the Gaede pump used in the exhaustion was kept in action con- 

 tinuously. It is possible that the vessel containing the salt had somoet 

 minute opening in its walls, but although careful search was made for 

 such an opening none was detected. The readings taken in these obser- 

 vations are given in Tables YII and VIIT, and curves plotted from them 

 are given in Figs. VIII and IX. 



