Disintegration in a Vacuum Tube, 417 



•current through the tube was adjusted by a sliding liquid 

 resistance (R) of amyl alcohol and cadmium iodide. 



Glass taps T 1 and T 4 connected the vessel D to a Gaede 

 pump and a 10-litre bottle respectively. The taps T 2 and T 3 

 were joined by 8 cm. of capillary tubing : by these gas 

 could be admitted in small quantities into the apparatus. 

 The joints were either of blown gas or made airtight with 

 picein. The pressure was measured by a McLeod gauge. 

 Any deposit on the tube, cathode sheath, and exploring wire 

 was removed before each fresh discharge. 



In general, the cathodes were platinum wires, but in some 

 of the earlier work copper wires were used. These were 

 4 mm. in diameter and 5*75 cm. long. The platinum 

 cathodes were for the most part 1*35 mm. in diameter and 

 •of exposed length 6 cm. The ends of the cathodes were 

 hemispherical. The cathodes were weighed before and after 

 discharge on a balance weighing to 0*1 mg., it being arranged 

 that differences in weight were measured in every case in 

 terms of the rider. 



Eesults. 



The statement of results falls naturally into certain 

 •sections. 



Early Stages of Discharge. 



Experiments were carried out first with copper and later 

 with platinum cathodes on the following points : — 



(a) The effect of leaving the cathode in vacuo previous to 



use. 



(b) The effect of previous discharge from a cathode on its 



subsequent rate of disintegration. 



Those who have used the disintegration method of obtaining 

 thin metallic films have often observed that the rate of dis- 

 integration under ordinary vacuum-tube conditions seems 

 to be inappreciable in the early stages. J. J. Thomson has 

 •suggested that discrimination must be made between dis- 

 integration of the cathode and disintegration of the gaseous 

 layers of the surface ; and some experiments by Hodgson * 

 ;also suggest that this apparent delay in disintegration is 

 intimately connected with the evolution of gas from the 

 cathode in the early stages of discharge. An investigation 

 of the effect of previous use of a cathode on its subsequent 

 rate of disintegration seemed, therefore, to be of considerable 

 interest. 



* Hodgson, PJii/s. Zeit. xiii. p. 595 (1912). 



