to produce Discharges in Gases at Low Pressures. 491 



5. Carves drawn to sliow the relation between diameter 

 and minimum sparking potential are of the same nature, but 

 the positive drops sharply for the two smallest diameters. 



6. The results obtained at pressures down nearly to the 

 critical pressure may be explained on the theory o£ 

 ionization by collision. Below that, the results are entirely 

 in harmony with the supposition that at these pressures some 

 kind of radiation is produced by the impact of negative ions 

 on the electrode under the action of a fairly strong force. 



The Effect, of a Continuous Discharge upon Sparking 

 Potentials at Low Pressures. 



Some time ago a series of experiments was carried out by 

 J. A. Brown to determine the relation between the sparking 

 potential and the potential required to maintain a current in 

 a gas *. The^e experiments show that above the critical 

 pressure, the potential required to maintain a current is less 

 than the sparking potential, the difference increasing with 

 the current, while below the critical pressure the effect is 

 the reverse. The collision theory, which explains perfectly 

 so many phenomena above the critical pressure, indicates 

 that the potential required to maintain a current should 

 be less than the sparking potential, regardless of the pressure, 

 except for very small currents. Some other factor evidently 

 comes into prominence at low pressures, in addition to the 

 ordinary process of ionization by collision, and it is suggested, 

 as an explanation of the relatively greater rise of the main- 

 tenance potential below the critical pressure, that the gas 

 is heated by the discharge and part of it driven out from 

 between the electrodes. The effect of such a diminution in 

 the quantity of gas between the electrodes would be to lower 

 the maintenance potential above the critical pressure, and 

 raise it below the critical pressure. In the former case, the 

 difference which should theoretically exist between the 

 curves would be increased ; but in the latter case the effect 

 would be exactly the opposite, and if large enough would 

 more than balance the difference demanded by the collision 

 theory, so that the maintenance curve would be above that 

 for the sparking potential. The present experiments were 

 undertaken in order to test directly for the existence of any 

 sir/h expulsion effect. 



R. V. Earhart has since published a paper embodying the 

 results of experiments concerning the effect of temperature 

 upon the potential required to produce a luminous discharge, 



* Plul-osoplaieal Magazine, September 1906. 

 2 Lv 2 



