336 



PHENOMENA, ATOMS, AND MOLECULES 



Fig. 7. 



(a) Irregular evaporation rate for ions from an unaged filament, (b) Uniform 

 variation of evaporation rate from a sufficiently aged filament. 



Control of \ia 



In these experiments \ia was varied about one thousand -fold from 

 ^lo^^ to 10^'* atoms cm'^ sec."^. To maintain any particular pressure a 

 large Dewar flask containing kerosene vigorously stirred surrounded the 

 tube. The bath was heated electrically or cooled belovv^ room temperature b}- 

 use of a coil containing liquid air. In preliminary experiments only the 

 appendix H was immersed in the bath, the rest of the tube being at room 

 temperature. It was then observed that the galvanometer readings following 

 ballistic kicks, especially at high ^'s, did not return immediately to zero, 

 but showed a lag of several seconds. This lag was found to be due to the 

 burst of atoms suddenly liberated from filament A. These apparently did 

 not come instantly to equilibrium with the walls of the tube and there 

 resulted a pressure increase of short duration which was registered on 

 filament B. This effect was eliminated by freshly coating the walls of the 

 tube with caesium before any daily series of experiments and by immersing 

 the whole tube in the bath liquid. This served to cover and keep covered 

 any spots on the walls of the tube which were not acting as true caesium 

 surfaces. 



FilMnent temperatures 



The filament temperatures were found from the known diameter and 

 the measured value of filament current. The temperature scale for tungsten 

 and the tables given by Jones and Langmuir ^ were used. 



Below /— ' 550°, even with the long filament used, cooling of the central 

 part of the filament A became appreciable, because of conduction from 



^H. A. Jones and I. Langmuir, G. E. Rev., 30, 310, 354 (1927). 



