ivith Electron Currents in Different Gases. 



429 



a heating current of 1'05 amps. (OAA 1 A 2 A 3 ) when the 

 driving potential reaches 22 volts (A) the current suddenly 

 jumps from 130 microamps. to (A) 1100 microamps. At the 

 same time the resistance of the gap falls ; so that a potential 

 difference of only 21 volts is required to maintain the larger 

 current. On increasing the potential after this condition has 

 been established, the current increases uniformly along the 

 line AjAa- Above 35 volts (A 2 ) another kind of instability 

 appeared, the final reading (A 3 ) being considerably under 

 that observed when the potential (45 volts) was first applied. 

 This instability is probably due to the discharge wandering 

 inside the tube, but we have not paid much attention to it, as 

 it does not seem to be so interesting as that observed at the 

 lower voltages. On lowering the potential from 45 volts the 

 uniform part of the curve A 2 rL x was found to extend to the 

 left beyond 21 volts ; but when the applied potential fell to 

 17 volts ( A : ) the current suddenly dropped on to the lower 

 ■curve again, and the potential difference rose to 19 volts 

 owing to the increased resistance of the gap AB — >C 



The potential differences were tapped off a sliding rheostat 

 in series with a battery of 200 volts with its middle-point 

 earthed. The line ran to the negative end of the filament 

 through a 5000 ohm resistance to prevent the development 

 of an ordinary high-current arc. The potentials were mea- 

 sured by a voltmeter between the negative end of the filament 

 and the earth, to which the anode C was also connected 

 (through the microamperemeten). 



The other curves B, 0, D hardly call for individual comment. 

 The chief point to note is that the potential difference at which 

 the current jumps from one curve to the other drops uni- 

 formly to lower values as the filament temperature is raised 

 and the number of electrons available becomes greater. 

 This is true whether the jump is from the low- to the high- 

 current curve, or vice versa. 



The magnitude of the " kick " corresponding to the passage 

 from O A to A : A 2 and vice versa, as well as the potential 

 differences under which these changes take place, depends a 

 good deal on the pressure of the vapour as well as on the 

 temperature of the filament. At very low pressures the kick, 

 as measured by the ratio of the currents at A x and at A, is 

 small, but sets in at relatively high potential differences ; 

 as the pressure increases the ratio of the currents increases 

 to a maximum and suhsequently falls off until at pressures 

 comparable with a centimetre of mercury the kick is scarcely 

 noticeable. The potentials at which the kicks take place 

 appear to drop uniformly towards a limit as the vapour 



