lyo Transactions of the Canadian Institute. [vol. ix 



pressure of the air in the apparatus. But as a result of the later experi- 

 ments performed with the carbon electrode it is clear we must add an- 

 other and more important cause for the fall, namely, the decrease in the 

 secondary radiation throughout the experiment as the gas layer on the 

 electrode became less and less dense. 



The experiment just described with the carbon electrode shews that 

 when the gas layer was removed, there was no fall in the rate of charging 

 of the carbon electrode, as the potential on the polonium was increased, 

 but instead a gradual increase was obtained. 



The question arose, then, whether the brass electrode would act in 

 the same manner as the carbon electrode if the gas layer was first re- 

 moved from it. To answer this question the brass electrode was kept 

 in vacuo a long time and the experiment with it was repeated. Contrary 

 to expectation it was found that there was still a very slight decrease in 

 the rate of charging of the electrode after the potential of the polonium 

 was raised above 40 volts but this decrease was not nearly as great as 

 before. 



The reason for the final slight difference between the behaviour of a 

 carbon and of a brass electrode appears then to be due to a difference be- 

 tween the a ray reflecting power of carbon and brass at different volt- 

 ages, the brass reflecting more « rays than the carbon as the voltage was 

 increased. 



If the work done by Geiger and Marsden*^ on the reflection of « rays 

 is taken to be applicable to the present experiments, it would seem that 

 a particles can not be reflected in sufficient numbers to account for this 

 difference. However, the experiments described above point definitely 

 to the reflection of a rays as the cause of the slight difference in the be- 

 haviour of the carbon and the brass plates under the bombardment by 

 a rays. Moreover, it can easily be shewn that with the fields used 

 variations in the speed of the a rays amounting to 1.7% must ensue. 

 It is just possible that this variation might be sufficient to cause such a 

 change in the amount of '^ radiation reflected from the carbon and 

 brass electrodes as to contribute, in part at least, to the above effect. 

 It would appear, therefore, that additional experiments should be made 

 on the reflection of <* rays of different velocities at surfaces subjected 

 to low gas pressures before the explanation offered above of the effect 

 observed is set aside. 



VII. Summary of Results. 

 I. It has been shewn that there is a secondary radiation produced 

 when alpha rays fall on a brass or a carbon plate. 



*' Geiger and Marsden, Proc. Roy. Soc. Ser. A, 82 July 31, 1909. 



