Effect of Gases on Contact Difference of Potential. 163 



wire and the surrounding cylinder under the conditions 

 investigated. 



Little need be said of the actual experimental arrange- 

 ments, which were not for the most part o£ a novel character 

 in thermionic work. For the production and maintenance 

 of a high vacuum a mercury vapour pump backed by an oil 

 box pump in series with a Geryck piston pump was used, and 

 this could be supplemented when desirable or convenient by 

 a liquid air cooled charcoal attachment. The tubes were 

 mounted so that a vacuum furnace* could be placed over 

 them when desired, so as to get the gas out of* the tube walls 

 and electrodes. "When desired pure hydrogen could be 

 admitted by gently heating an attached palladium tube with 

 a Bunsen burner and the pressure read off on a McLeod 

 gauge. The pressures of hydrogen used were all so small 

 that they did not affect the shapes of the curves to any 

 ascertainable extent by interference of the gas molecules with 

 the motions of the electrons. 



The simplest conditions with a view to the theoretical 

 interpretation of the results would be obtained from the 

 displacements of the curves for the emitting wire maintained 

 at a constant temperature in the different experiments. 

 This, however, involves a good many practical difficulties 

 under the experimental conditions. The problem of main- 

 taining a wire accurately at a constant temperature in 

 presence of varying amounts of gas is one of great difficulty. 

 In the case of platinum for example the effect of absorbed 

 hydrogen on the resistance and the cooling effect of the 

 hydrogen atmosphere vitiate the resistance method of holding 

 the temperature constant. It is doubtful whether the methods 

 of optical pyrometryas applied to thin filaments are sensitive 

 enough for this purpose. Another difficulty arises from the 

 very considerable effect of hydrogen in changing the emission 

 constants of the metal. If a standard temperature is set up 

 for some particular pressure, it may be that on changing to 

 another pressure the emission will be so high that the shape 

 of the characteristics is chanoed owino- to the effect of the 

 self-repulsion of the electrons or, on the other hand, if may 

 be so low that the instruments available are not sensitive 

 enough to make measurements of it. The last is hardly an 

 insuperable objection as it may be overcome by changing 

 Over to more sensitive instruments ; but such changing over 

 involves much time at the best, and rapidity of operation is 

 probably the most important factor contributing to success in 

 experiments of this kind. 



* Richardson, ' Emission of Electricity from Hot Bodies,' p. 15, 



M2 



i 



