134 BELL SYSTEM TECHNICAL JOURNAL 



The only ways hitherto used to accommodate the atom-model to 

 this surprising and inconvenient factor g are tantamount to assuming 

 that it enters into the relation between angular momentum M and 

 magnetic moment P which was derived in Section L and written down 

 here as equation (7) ; which relation is accordingly modified without 

 discoverable reason into 



M/P = ge/2ixc (18) 



a very unsatisfying procedure. Lande found it possible to mitigate 

 this process somewhat and at the same time produce a partial ex- 

 planation of the formula quoted in the First Part of this article, 

 whereby g is related to the factors K, R and / which, in the atom- 

 mcdel of the two whirling parts, measure the angular momenta of 

 valence-electron and residue and entire atom respectively in terms of 

 the common unit h/2ir. This explanation involved the postulate 

 that g = l for the valence-electron and g = 2 for the residue. It would 

 therefore be necessary to justify, or to postulate without justification, 

 not a multitude of such relations as (16) with a multitude of unforeseen 

 values of g, but only a single such relation with a single unforeseen 

 value of g. This is bad enough, but not so bad as if it were inevitable 

 to assume that M/P may have a dozen different values in different 

 cases. 



It is now the occasion to recur to the extraordinary expermients 

 of Gerlach which disclose the magnetic moments of individual atoms 

 and verify the supposition that certain orientations are permitted 

 and others inhibited. These experiments having already twice been 

 mentioned in this series of articles, I shall spend no more space upon 

 the method than is necessary to say that atoms in a narrow stream 

 are sent flying across an intense magnetic field with a strong field 

 gradient, by which they are drawn aside. Were the atoms tiny 

 magnets oriented randomwise in all directions, the beam would be 

 broadened into a fan ; one edge of the fan would be the path of atoms 

 oriented parallel to the field, the other edge would be the trajectory 

 of atoms oriented anti-parallel to the field, while the space between 

 the edges would be filled by the orbits of atoms pointed obliquely 

 to the field. Actually Gerlach observed not the whole fan, but two 

 or several separate diverging pencils of atoms, and between them 

 vacant regions traversed by none. Certain orientations of atoms to 

 field were unrepresented in the beam. Here for the first time there is 

 direct evidence of discrete Stationary States, of quantum permissions 

 and quantum inhibitions, not deduced from observations upon transi- 



