SOMli ii'\;/ w;i'A'./A'J' .;/)/•. /.V( / s /\ /■//> ^/( ^ i; 1."* 



that function is precisely the expoiienli.il runction appearing in (•">). 

 [•".ssenlially the theory is rejiuced to this postulate: the nuinher of 

 molecules ionized l>y an electron in a centimetre of its path depends 

 only u(X)n the eneryj),- it accjuires from the field in its free ni^lil from 

 one collision to the next. If in this form the theory still cannot 

 give satisfaction, the next step will he to alter the original assuiyp- 

 tion that the electron comes practically to a dead stop in e\er>- col- 

 lision. In dealing; with the nohle gases and the metal \apnurs, the 

 facts alxiut elastic collisions which I ha\e alread\' oulliiuvl pro\e 

 that this assumption should not be made at all. It is clear ilial this 

 is another prolilein for the future Bolt/mann! 



Meanwhile, one of the cardinal features of the TowummhI experi- 

 menis is the fact that the\' display the .gradual advent of the trans- 

 formation of the maintained currents which we have hitherto con- 

 sidered, into the self-maintaining discharges which are the familiar 

 and the spectacular ones; and we now ha\e to examine the agencies 

 of this transformation. 



4. TnK Discn.\R<.F. Begins to CoxTRinuTE to the F,i.ECTRf)N- 

 Stre.xm Which M.mnt.mns It 



(■reatly though the current of primary electrons from the cathode 

 to the anode may be amplified by the repeated ionizations which I 

 have described, there is nothing in this process which suggests how 

 the discharge may eventually be transformed into a self-maintaining 

 one like the glow or the arc. The free electrons may ionize ever so 

 abundantly, but as soon as the supply from the cathode is suspended 

 by cutting ofT the heat or the light, the last electrons to be emitted 

 will migrate off towards the anode, and whatever electrons they 

 liberate will go along with ihem, leaving a stratum of gas devoid of 

 electrons in their wake; and this stratum will widen outwards and 

 keep on widening until it reaches the anode, and then the discharge 

 will be ended. Something further must happen continually in the 

 gas through which the electrons are flowing, something which con- 

 tinually supplies new free electrons to replace, not merely to supple- 

 ment, the old ones which are absorbed into the anode and vanish 

 from the scene. 



We have alreacK' noticed one sort of e\ent contiiuialh' hai)peiiing 

 in such a gas as helimn traversed by not-too-slow electrons, which 

 might conceivably develop into a mechanism for maintaining the dis- 

 charge; for, when an atom of the gas is put into the "excited state" 

 by a blow from an electron, it later returns into its normal state, and 



