468 



inuiiber of ionisaüüiis per tiirie-unil will be direclly proportional to (/. 

 Later Fkanck and Hekïz '), just as C D. Ohh-d '), in connection 

 with N. Bohr's theory have assnmed that uiielastic collision can also 

 alreadj' take place before the tension of ionisation has been reached, 

 in which then removal of one of the electrons of the atom to a 

 path Ijing more outward, takes place. On return to the normal 

 path this energy can then be emitted. Yet the result of the elementary 

 theory will be approximated by three cases: 



I. Through absorption of the radiated energy by neighbouring 

 atoms (Compton) '). 



II. By increase according in quanta of the energy of slow elections 

 on collision with dislocated atoms ("collisions of the second kind" 

 in the theory of 0. Ivlp^in and S. Rosseland ; cf. also ^ 5). 



III. (In a slight degree) through the renewed collision between 

 dislocated atom and (rapid) electron, before the former has lost 

 energy by radiation (K. J. van der Bijl ") ). 



In agreement with our conclusion from equation (4) comniunica- 

 tion 1 it may therefore really be expected that in approximation the 

 relation a . g = const, will hold for each of these gases separately, 

 so long as the nature of the bearers is not subjected to aiiy charac- 

 teristic moditicatiou. For in this case the energy-compensation ensuing 

 from I — III will always be the same percentage. This compensation 

 must, however, be very different for different gases. So that, the 

 tensions of ionisation also being so greatly divergent, we are led to 

 accept the obvious conclusion that the value of ag will be different 

 for different (noble) gases. We shall revert to this when discussing c. 



That for the rest deviation is to be empirically observed ad b 

 between a calculation based on formula III and observation (chiefly 

 as a consequence of errors of observation), may appear from the 

 following example (argon); though foi'mula III would lead us to 

 expect that the value : 



Ag\/Ml 

 would be constant, a consideiation of the values published by Rüt- 

 TENAUER shows that in table 5 e.g. the "constant" which amounts 

 to about 2,3 . 10-^^, for one and the same noble gas in a definite 

 case (in which the pressure varies from 0,5 to 0,64 mm., the cuirent 

 density from 1,49 to 1,21 A/cm', and the bore of the tube from 



^) J. Franck and G. Hertz. Phys. Zeitschr. 20, 133 ('19). 



') C. D. Child. Phil. Mag. (6) 278 ('14). Phys. Rev. (2) 15, 33 ('20). 



•') K. T. Compton. Phys. Review (2) 15, 476, 1920. 



^) K. J. VAN DER Bijl. Phys. Rev. 10, 546 ('17). 



