Ionization of Platinum by Cathode Rays. 289 



or to the instrument for measuring current (a Dolezalek 

 electrometer and xylol-alcohol resistance of 10 9 ohms). P was 

 heated by a small transformer with well-insulated secondary. 

 W by a storage-battery insulated on paraffin blocks. 



Fio-. 1. 



The whole of the apparatus drawn was, of course, enclosed 

 in a vacuum-tight vessel which was constructed wholly of 

 glass ; there were no ground joints or sealing-wax joints. 

 It was connected to a Gaede mercury-pump and McLeocl 

 gauge through a U-tube which could be cooled in liquid air 

 in order to exclude all vapours from the essential part of 

 the apparatus ; during the period in which attempts were 

 made to free the platinum from impurity the U-tube was 

 kept cooled continuously. 



In most of the experiments P (disconnected from the 

 transformer) was connected to the electrometer. A potential 

 difference V\ was maintained between B and W, W being- 

 negative, so that the electrons from W entered A with the 

 speed V x . A potential Y 2 was maintained between P and B,. 

 B being positive, so that the electrons, entering A with the 

 speed V T struck P with the speed V^— "V 3 . A third potential 

 V 3 was maintained between P and A, A being positive, to* 

 ensure that no electrons leaving P were reflected from A 

 back to P; it was found that 4 volts was sufficient for this- 

 purpose, and that the observations were independent of the 

 value of V 3 so long as it lay between 4 and "20 volts. If it 

 were very much greater some of the rays entering A were 

 attracted to A and did not fall on P. Any electrons falling 

 f'rom P on T will not be reflected, for T was always at a 

 potential at least 20 volts higher than P. 



The speed of the rays falling on P was varied by 

 changing V 2 . This method is certainly preferable to that 

 of changing V l5 because the number of rays entering A and 

 falling on P is changed if V l is changed, while it is prac- 

 tically unaffected if V 2 is changed, so long as V 2 is always 



Phil. Map. S. 6. Vol. 2$. No. 1G4. .1 ug. 1914. U 



