150 



TRANSACTIONS OF SECTION A. 



An assumption that has to be made at the outset is that, although one 

 filament as a whole is emitting electrons continuously according to the accepted 

 exponential temperature law, yet there are often one or more spots which 

 are emitting with exceptional power. Such spots in the case of a tungsten 

 filament are probably of chemical origin, due to the presence of local impurity, 

 and would be very sensitive to small changes of temperature. In the theory 

 to be developed it is supposed that the bombardment by positive ions of the 

 filament in the neighbourhood of such an emitting spot would greatly increase 

 the local electronic emission so long as the bombardment lasted. Very direct 



rAA/\A/WWVVWW\/\A 



I I 



Fig. 2. 



evidence of the existence of such selectively emitting spots on a tungsten 

 filament is afforded by the experience of manufacturers of hard valves. It 

 is customary in the factories to ' clean up ' the anode by passing a heavy 

 thermionic discharge through the valve when on the pump. The dissipation 

 of energy at the anode is regulated to such a point that the metal of the 

 anode is maintained at a cherry-red heat. During this process it is frequently 

 observed that one or more points on the anode are very much hotter than 

 the main surface, a fact which can only be explained on the assumption that 

 there is exceptionally powerful emission from the corresponding points on 

 the hot filament. 



Consider one of these spots on the filament. If a burst of electrons be 

 emitted they proceed towards the filament with a speed u given by 



= Ve ; or, ?< = \/ 2 V , e/w 



where e/m is the charge to mass ratio for the electron and V is the positive 

 grid potential with respect to the particular point on the filament considered. 



(It is assumed as an approximation that the filament is screened from 

 the anode by the grid, approximately true for the particular type of valve 

 used in the experiments, which had a fine-meshed grid.) 



The electrons will thus take a definite and calculable time to travel from 

 filament to grid under the moderate potential applied. On passing through 

 the grid, however, the electrons emerge into a strong electric field and assume 

 ionising speed. The negative ions produced follow the electrons to the anode, 

 but the positive ions pass back through the grid towards the filament with speed 



w, = v/2 V77/ot^ 



where e/m is in fhis case the charge to mass ratio of the atom or molecule 

 concerned. 



There will thus be a cloud of positive ions focussed on the filament and 



