Electrons concerned in Metallic Conduction. 263 



electrons in solids, but that the value of K derived is inde- 

 pendent of this ratio, as may be shown on reference to the 

 formulae. We have used the value e/m = l'S6 10 7 . Drude 

 was of the opinion that the proper value was 1*5 10 7 for the 

 electrons in solids. But it has seemed preferable to select 

 the other value. Prof. 0. W. Richardson, by a beautiful 

 series of experiments, has established the nature of the ions 

 emitted by hot bodies, and accepting his view that the 

 negative ions are the negative electrons of conduction wjiich 

 have escaped from the forces retaining them in the metal, 

 it is necessary to take a value higher than this, Prof. 

 Richardson's standard value being l'*8 10 7 . We may say 

 that no other value generally accepted for e/m leads to values 

 of p so conclusive as those of the last table. 



A brief recapitulation of the other assumptions may be 

 desirable. In the first place we have supposed that the 

 carriers of electricity are negative electrons and not ions. 

 Moreover, the free positive electricity remains fixed to the 

 atom, and can only move by vibration with it. If the vibra- 

 tions are not sympathetic the motion of the positive electricity 

 and of the atoms containing it may be neglected, but if they 

 are the amplitude of vibration may be large, and thereby an 

 apparently negative value of K may be produced by ignoring 

 it. A negative value of K denotes the presence of such 

 vibrations, which, of course, cannot be predicted from the 

 absorption spectrum of the vapour of the metal, in which 

 the aggregates of atoms forming the molecules will be dif- 

 ferent. If the negative value of K is small the calculation 

 of p by the method given will not be seriously affected. 

 A ibrations of the saturated portion of an atom which has lost 

 electrons will obey, for internal vibration in the atom, the 

 laws of material polarization in doublets, and if the convection 

 of these doublets is not too large, may be ignored except in 

 the contribution to the dielectric capacity. On this basis the 

 optical constants may be found for a given law of velocity 

 among the electrons, and it is found possible to discriminate 

 between such laws in complete favour of that of Maxwell, 

 which is found to lead to results of definitely close agreement 

 with experiment. It may be claimed that this treatment is 

 more satisfactory, in that it is less vague than, for example, 

 that of Drude, which involves an introduction of frictional 

 forces whose origin is difficult to interpret. It would seem, 

 also, that negative electrons and unsaturated atoms caused 

 by their escape, are the only entities which ought to be 

 admitted. Richardson's experiments appear to the writer 

 to prove that other ions are not present in metals, for he has 



