SlKMli COXIliMl'DH.IKy .//»('. I.V(/.\ /.V /'//Js/cV ;/// Alt 



\\v li.iM- to invent some new ones for them. Tliis seems regret I, iMe. 

 liiit only l>ee.iuse our expectations were too liinh- 



Another eireiimstance leads us to another dilemma. Su|)pose that 

 we could circumvent that ditticulty alxxit the revoKinn electron, 

 which radiates part of its energy at each revolution and slides down 

 a spiral [lath into the nucleus; supjiose that we could find justifica- 

 tion for sayinn that no radiation occurs, that the electron like a planet 

 may revoke forexer in an ellipse. If two atoms colliiled, as in a K^i^ 

 they must ver\' frecpiently do, would not the electrons all he dis- 

 arranged, ilisorganized, Himj; over from one orbit into another.'' This 

 we should certainly expect; yet if it happens, no two atoms in a gas 

 can In.' exactly alike, ni>r can any atom retain its character for more 

 than a fraction of a second. If this is so, then the various sharpK- 

 det'metl properties of a gas must, each and every one of them, be 

 statistical properties- -not themseU'es properties of indi\idual atoms, 

 but the results of other properties of individual atoms, held in different 

 amounts by different atoms and all averaged together. In some 

 cases this is unobjectionable; the pressure and the temperature of a 

 gas are sharply definite properties, resulting from the mass and the 

 motion of the atoms, and the latter of these properties is not neces- 

 sarily the same for any two att)ms at the same moment nor for any 

 atom at different moments. But one wouki be reluctant to treat the 

 spectrum of a gas as such a prfiperty; according to all the traditions 

 of physics this is one of the properties of the individual atoms, liut 

 the sjjectrum is very constant, sharp, immutably defined; we must 

 therefore assume either that it de[)ends onh' on the number of elec- 

 trons in the atom and not upon their motion nor position, an idea 

 which would be difficult to carry through; or that the electrons are 

 ineluctably constrained to certain orbits or certain positions, so that 

 the atom retains its personality and its character. 



We have now inade the accjuaintance of two ideas whicli will be 

 exceedingly prominent in the second di\ision of this article. Tlu' 

 nuclear atom-nnKlel is of itself unstable; therefore stability mu>i be 

 enforced ut)on it by outright assumption, it must be made stable b\ 

 fiat. But this stability may nf)t be extended to all concei\ablc 

 arrangements or configurations of the nuKlel; it must be reserved 

 for one or a few, that the atom may possess a fixed character and a 

 personality. 



We now arrive at the phenomena by means of which these vagueK- 

 expressed ideas are t(j be sharpeiieil and hardened into detinite 

 doctrines. 



