Excited Activity of Thorium. 633 



the tube and the intensity of the discharge than on its 

 source, or the connexions of the disk. When, however, the 

 disk was made the cathode, the effect was much greater, 

 being as much as 90 per cent, in 10 minutes for the coil 

 discharge, and 55 to 60 per cent, in 10 minutes for the 

 Wimshnrst discharge. 



The direct incidence of cathode rays is known to cause 

 volatilization with some metals, as found by Dr. H. A. Wilson, 

 who observed that a thin silver film was removed from a 

 glass surface when a strong beam of cathode rays fell upon 

 it. This, however, is not a sufficient explanation of the pre- 

 sent action, for in some cases the active side of the disk was 

 turned away from the cathode, the inactive side being exposed 

 to the rays : in others the rays were deflected off the disk by 

 means of a strong magnetic field. In the former case the 

 loss of activity was undiminished, being often as much as 

 50 or 60 per cent, in J hour ; in the latter ths loss, though 

 ss, was not prevented, being from 10 to 20 per cent. Also 

 when the pressure was so high that only very slight fluores- 

 cence appeared on the glass walls, a loss of activity still 

 occurred. 



The most probable cause Avhich next suggests itself, is an 

 action analogous to the sputtering of a metallic cathode during 

 the passage of the discharge, and this is supported by the large 

 effect obtained with the active disk as cathode. To test this 

 hvpothesis, a polished disk of cadmium (a metal which gives 

 considerable sputtering) was substituted for the active disks in 

 the di-charge-tube. When this was connected to the anode 

 tor the coil-discharge, a fairly thick film was obtained on the 

 wall- of the tube in 1 hour, the metal being sputtered off both 

 towards the cathode and back on to the glass stopper. With 

 the cadmium disk as cathode, a noticeable deposit was obtained 

 in lo minutes, though considerably less than before. With 

 the Wimshnrst discharge no deposit was obtained from the 

 anode in 35 minutes, and only a very slight one from the 

 cathode after 25 minutes. The effect cannot, therefore, be 

 regarded as corresponding at all closely to the loss of activity 

 from the active disk-. At the same time, it is very probable 

 that in the latter case more than one cause may be operating, 

 and that when the active disk is made cathode the sputtering 

 either of the active matter itself or of the metal on which it 

 i- deposited is the predominant factor, while in other cases 

 some other cause is more important. The very small amount 

 of -puttering obtained by various observers with aluminium, 

 even as cathode for high potential discharges, makes it more 

 probable that it is the active matter only which goes off. 



