/ 



soMi iOM I MroK.iNv .inr.ixcis i\ rinsns ii ui 



pri'iipil.ilc upwaril rii.sli ut the lurrfiil-volt.iKi' tiirM- in 1"^. 4, from 

 the work of Davis aiui Cioucher.' 



At this point I will digress to speak very l)riftl\' of thu siuTcssioii of 

 i'\cnts which oinirs when the eleitron-syeani is much denser than 



Fig. 4 — Onset of ionization in mercury vapor at 10.4 volts (preceded by suhsidiary 

 etTects at 4.*> volts and 6.7 volts; see footnote'). (Davis and Goucher) 



we ha\e hitherto imagined. So long as the energy of the electrons 

 does not attain the resonance-potential of the gas, there is no reason 

 to expect any novel effects; the collisions will be perfectly elastic, 

 just as when the electrons were few. But when the atoms are thrown 

 into the "excited state" by impacts, there will be occasional cases 

 of an atom being struck twice by electrons in such quick succession 

 that at the moment of the second blow, it is still in the excited state 

 provoked by the first. Now, much less energs' is reqtiired to ionize 

 an atom when it is in the excited state than when it is normal; con- 

 secjuently when the electrons are so abundant that these pairs of 



• The sudden upturn at 10.4 volts is the swift rise of current at the onset of ioniza- 

 tion. The much less violent upturns at 4.9 and 6.7 volts are due to the electrons 

 e.xpelled from the metal parts of the apparatus by the radiation from the excited 

 atoms. In the lower curve, by modifying the apparatus, the latter upturns are 

 translated into downturns to distinguish them from the upturn which denotes 

 ionization. This distinction was not realized until I'M/, and in articles published 

 iK'tween 1913 and 1917 the lowest resonance-potentials of gases are given as their 

 ionizing potentials. Enormous improvements in the methods and technique of 

 measuring these critical P'Otentials, and recognizing of which kind they are, have 

 lieen effected since then. 



