128 BELL SYSTEM TECHNICAL JOURNAL 



by various orientations of the spinning atom relatively to the field. 

 It will presently be seen that the evidence for the orientation-theory 

 is much more abundant and more nearly direct, in this case of mag- 

 netically-excited Stationary States, than in that former case of mul- 

 tiplets. This case in fact was the earliest to which an orientation- 

 theory was applied; but for it, some quite different form of theory 

 might have been developed for multiplets. Even here the data and 

 the theory are not entirely concordant; but the concordance is so 

 extensive, that the discord is sharply localized and identifiable. 



From the Second Part of this article (Section L) I quote the prin- 

 ciple that an electron (of mass ju) revolving in an orbit with angular 

 momentum P is equivalent to a magnet of which the magnetic moment 

 M is proportional to P, being 



M = eP/2yiC (7) 



Both P and M are vectors normal to the plane of the orbit and hence 

 parallel to each other. If several electrons are revolving in divers 

 plane orbits about the same nucleus, their separate angular momenta 

 may be summed vectorially into a vector which is the angular mo- 

 mentum of the entire system, and their separate magnetic moments 

 may likewise be summed vectorially into a vector which is the mag- 

 netic moment of the entire system; and these two summation-vectors 

 will be parallel to one another, and related by the foregoing equation. 

 Hence a rigidly-connected revolving framework of electrons — if such 

 a thing there be — may be treated like a single electron, insofar as the 

 ratio of magnetic moment to angular momentum is concerned. Wher- 

 ever in the course of this article we have envisaged electrons, kernels, 

 or atoms revolving with angular momenta prescribed as integer 

 multiples of h/2Tr or of | h/2Tr, we might have imagined these as mag- 

 nets with magnetic moments prescribed as integer multiples of eh/4:TrfjLC 

 or of I eh/4:TiJ.c.^ This is not necessary ; though the relation between 

 angular momentum and magnetic moment is derived directly from an 

 equation valid for perceptible electric currents, it might not be true 

 for individual electrons. Nevertheless we shall arrive at striking 

 results, by supposing that it is. 



When a magnetic field is applied to a multitude of radiating atoms, 

 most of the lines of their spectrum are replaced by groups of several 

 lines each, or "split up" into several components, as the phrase is. 

 This signifies that each of the Stationary States of each atom is ap- 

 parently replaced by several. One may infer that when an atom is 



'The quantity eh/'iwuc, the presumptive magnetic moment of an electron circu- 

 lating in an orbit of angular momentum h/Irr, is known as the Bohr magneton. 



