234 Research Staff of the G. E. C, London, on the 



all the positive ions go to the grid, and i g ji e measures the 

 number oil ions produced by each electron, the electrons, or 

 some of them, must be pursuing long curved paths in the 

 vessel. Nor is this result surprising, for the grid, being 

 negative to the cathode, cannot receive electrons directly 

 from it ; the electrons must wander about the vessel until 

 they come into the very small space occupied by the strong 

 Held round the anode and fall into it. It is doubtless for 

 this reason that the glow potential is so much sharper than 

 in a triode of ordinary construction (where the area of the 

 anode is comparable with that of the walls) and the condition 

 in the neighbourhood of the glow potential un stable. For 

 the instability doubtless arises from the fact that a slight 

 increase in i n produces an increase in i p , which again produces 

 an increase in i n by further neutralizing the space charge. 

 This condition is the more likely to lead to instability the 

 greater the increase in i p produced by a given increase in 

 i n , and this increase is greater when the electrons pursue 

 very long paths than when they proceed direct from cathode 

 to anode. (Compare fig. 5.) 



And here it may be remarked that the discharge in the 

 neighbourhood of the glow potential is unstable when the 

 walls are maintained at a definite potential, as well as when 

 they are insulated. It is not due, as was thought at one time 

 when the very large value of i g ji e was not appreciated, to a 

 redistribution of the charge on the walls of the glass. 



On the other hand, the fact that i g \\ e does not increase at 

 higher pressures, even though the energy of the electrons is 

 sufficient to enable them to make many more than two ions, 

 is probably due largely to recombination of ions in the gas, 

 which causes i g to be less than i p . This matter will be the 

 subject of later discussion. 



5. One further question may be asked. Why is i g \i e so 

 much less below the glow potential than above it ? Doubt- 

 less because, below the glow potential, a much greater 

 proportion of ions reaches the cathode and not the grid, so that 

 i g is again less than i p and i e greater than i n . For in this 

 stage of the discharge an increase in i n is conditioned by an 

 increase in the number of positive ions arriving at the 

 cathode ; if the field were distributed so that most of the 

 positive ions reached the grid, the space charge would not 

 be neutralized and the discharge would cease. It is only 

 when the thermionic current is saturated that the cathode 

 can spare electrons, as it were, to reach the grid. The 

 problem would be interesting, but difficult, to treat mathe- 

 matically; the theory of the space charge, as developed 



