298 



SCIENCE 



[N. S. Vol. XLVI. No. 1187 



would deny the validity of this assumption, 

 let us consider the typical compounds of 

 old-fashioned organic chemistry in regard 

 to whose molecular structure we already 

 know much — at the very least we may 

 speak definitely of the relative positions of 

 the atoms within their molecules. Among 

 such compounds we find the striking phe- 

 nomenon of isomerism. Numerous isomers, 

 substances of precisely the same chemical 

 constituents and differing only in the rela- 

 tive order in which the atoms are placed 

 in the molecule, have been prepared. In 

 the case of complex substances, if it were 

 worth while, millions of such isomers 

 could be prepared. Yet these isomers will 

 keep for years, and probably would for 

 centuries, without changing into one 

 another. In these inert organic compounds 

 the atoms are so persistently retained in 

 definite positions in the molecule that in 

 one part of the molecule atoms may be 

 substituted for other atoms and groups for 

 groups, sometimes through reactions of 

 great violence, without disturbing the ar- 

 rangement of the atoms in some other part 

 of the molecule. It seems inconceivable 

 that electrons which have any part in de- 

 termining the structure of such a molecule 

 could possess proper motion, whether or- 

 bital or chaotic, of any appreciable ampli- 

 tude. We must assume rather that these 

 electrons are held in the atom in fixed equi- 

 librium positions, about which they may 

 experience minute oscillations under the 

 influence of high temperature or electric 

 discharge, but from which they can not de- 

 part very far without altering the struc- 

 ture of any molecule in which the atom is 

 held. 



Let us therefore consider whether the 

 physicists on their part offer any irre- 

 futable arguments in favor of an atomic 

 model of the type of Bohr's. In an atom 

 of the simplest type, composed of a single 

 positive particle and a single electron, if 



these fail to merge with one another until 

 their centers are coincident — and it is uni- 

 versally assumed that they do not so merge 

 — only two explanations are possible: 

 either the ordinary law of attraction be- 

 tween unlike charges (Coulomb's law) 

 ceases to be valid at very small distances, 

 or the electron must be in sufSciently rapid 

 motion about the atom to offset the force 

 of electric attraction. The first of these 

 explanations is the one which I have 

 adopted. The second, which has been 

 adopted largely because it appears to save 

 Coulomb's law, is the one which has led to 

 Bohr's atomic model, in which the electron 

 revolves in definite orbits about the central 

 positive particle. Now it has frequently 

 been pointed out, and indeed it was well 

 recognized by Bohr himself, that this model 

 is not consistent with the established prin- 

 ciples of the electromagnetic theory, since 

 in the classical theory a charged particle 

 subjected to any kind of acceleration must 

 radiate energy, while, according to the 

 Bohr hypothesis, radiation occurs only 

 when an electron falls from one stable orbit 

 into another. Since, however, the equa- 

 tion for electromagnetic radiation is one 

 of the more abstruse and less immediate 

 deductions of the classical theory, it might 

 be possible by slight modifications of the 

 fundamental electromagnetic equations to 

 reconcile them with the non-radiation of 

 the orbital electron. I wish therefore to 

 point out a far more fundamental objec- 

 tion to the theory of the revolving electron, 

 due to the fact that Bohr has been forced 

 to assume that this revolution must con- 

 tinue even down to the absolute zero of 

 temperature.^ 



If, in Fig. 1, the circle represents the 

 orbit of an electron B revolving about the 

 positive center A, and if C represents a 

 charged particle in the neighborhood, then 

 if the electron exerts any influence what- 



