A.— MATHEMATICAL AND PHYSICAL SCIENCES. 33 



furnished by X-ray spectra and by the variations of the chemical and 

 physical properties of the elements according to their positions in the 

 periodic classification. Bohr's well-known table of electron orbits (1922) 

 was built up by taking account of these properties and considering the 

 formation of atoms by the successive capture and binding of electrons. 

 The orbits themselves were distinguished by the quantum numbers 

 «* (A; < n), and it was assumed that the orbits of the earlier bound electrons 

 were essentially unchanged when another electron was introduced into 

 the system. The classification was notably successful in accounting for 

 the appearance, at certain stages, of elements which deviate in their 

 properties from corresponding elements of the previous periods. 



These ideas of Bohr have been remarkably developed in recent years, 

 especially through the work of Main Smith'^^ and Stoner,-* who inde- 

 pendently arrived at similar conclusions, chiefly from the consideration 

 of chemical and physical properties respectively. Whereas Bohr had 

 considered only the distribution of electrons among sub-levels defined by 

 the quantum numbers n^, Main Smith and Stoner suggested a distribution 

 among all the sub-levels which were known to exist from X-ray spectra, 

 and which were characterised by the three quantum numbers n^j. In 

 partial justification of the proposed new distribution of electrons in the 

 different groups, Stoner pointed to the remarkable analogy between the 

 accepted classification of X-ray levels and the terms of the optical doublet 

 spectra of the alkali metals which had previously been discussed by 

 Lande. This correspondence between X-ray and alkali terms may 

 perhaps be most conveniently shown in the form adopted by Wentzel, 

 as in Table VII. 



The combinations between the X-ray terms are governed by precisely 

 the same selection rules as those between the optical terms, and it can 

 scarcely be doubted that the two types of doublets have a similar origin. 

 This analogy was regarded by Stoner as strong justification for assuming 

 a relation between the inner quantum number and the number of electrons 

 in a sub-group. His reasoning was as follows. In the case of alkali 

 doublets, if the inner quantum numbers are given the values shown above 

 {k and k~l), the number of components into which each term is split in 

 a weak magnetic field, as revealed by observation, is double the inner 

 quantum number {e.g. 2 for S, 2 for Pj, and 4 for P.,). Since, therefore, 

 there are 2j possible states of the atom corresponding to each energy 

 level %j (distinguishable, however, only when there is an external magnetic 

 field), it seems reasonable to suppose that the number of ' possible and 

 equally probable ' orbits % also is 2j ; i.e. that there are 2j electrons in 

 a complete sub-group n^j. Although the doublet term values seem to 

 depend primarily on the outermost electron orbits, Stoner suggested that 

 the rule might be of general application. The distinction between orbits 

 having the same values of ?i,^ was ascribed to differences in orientation 

 with respect to the atom as a whole. 



These considerations led Stoner independently to an electron distribu- 

 tion similar to that proposed by Main Smith and differing in important 



" Chemistry and Atomic Structure, London, 1924 ; Review of Chemistry and 

 Industry, March 28, 1924. 



26 Phil. Mag., vol. 47, p. 719 (1924) ; vol. 49, p. 1289 (1925). 



1926 ^ 



