SOME CONTEMPOKAKY ADVANCVS l\ I'llVSICSII I.M 



we have to take the experimental data as we fwul lliem, and analyze 

 tlietn as Inist we ntav. not with too jjreat an experlatioii of penetrating 

 to the properties of the ultimate atoms; and yet, as we shall see, the 

 analysis does in certain cases penetrate unexpectedly far. 



'.\. TlIK KH)\V OK Kl.IiCTRONS A( ROSS UliNSE AlR, XlIROGliN, 



Hydrocen and Similar Gases 



The celebrated series of researches by Professor Towiisend of 

 Oxford and by his pupils, commenced in l'.U)2 and continuing through 

 the present, relate chiefly to such gases as iiydrogen, nitrogen, oxygen 

 and the familiar mixture of the last two which we breathe; and chiefly 

 to these gases at densities much greater than we have hitherto con- 

 sidered — densities corresponding to such pressures as a thousandth 

 or a hundredth of an atmosphere, therefore so great that an electron 

 crossing over from a cathode to an anode a few centimetres away must 

 collide with scores or hundreds of atoms. If a stream of electrons is 

 poured into perfectly pure helium of such a density, we must not look 

 for a sudden onset of ionization when the voltage between cathode 

 and anode is raised just past 24.5, for the reason illustrated by those 

 experiments of Compton and Benade — the electrons lose energy in 

 all of their collisions, even the elastic ones, and arrive at the anode 

 not with the full energ>' corresponding to its potential but with this 

 energy- diminished by what they lost on the way. In the familiar 

 diatomic gases, the electrons lose much more energy in their ordinary 

 collisions. I did not speak of these gases in the foregoing section, 

 because experiments of the very same type as those which show the 

 sharp distinction between elastic impacts and inelastic impacts in the 

 noble gases and give the sharply-defined values of the resonance- 

 potentials of these gases, yield comparatively vague and ill-defined 

 data, when they are performed on hydrogen or air. In these gases, 

 above all in active gases like oxygen or iodine, it is unlikely that any 

 of the impacts, whether the electrons be moving rapidly or slowly, 

 are truly elastic' 



' However, Foote and Mohler have obtained quite undeniable evidence of critical 

 potentials, at which the loss of energy by the impinging electron is much greater 

 than it is just below these potentials. The electron can transfer energy to (and 

 receive energy from) a molecule in more difTerent ways than to (from) an atom; such 

 as by setting the molecule into rotation, or putting its constituent atoms into vibra- 

 tion relatively to one another. There is also the mysterious fact of "electron affin- 

 ity" — an electron may adhere firmly to a non-ionized molecule. Numerous measure- 

 ments of the rate at which electrons progress through a gas (a field of research which 

 I have not space to consi<ler here) indicate that at field strengths such as prevail in 

 these experiments, adhesion of electrons to molecules is rare and transient. 



