40 SECTIONAL ADDRESSES. 



both equal to n, k of an electron already present), will contain deep-lying 

 terms for which r=Rii, and I has the several values, IL— A;+^, 

 iL-^l+t, . . . L+i-i/ _ 



The theory has been applied in great detail by R. H. Fowler and D. R. 

 Hartree^'' to the spectrum of ionised oxygen (0 II), and the terms to be 

 expected on the theory have been found to be in very satisfactory agree- 

 ment with the numerous regularities deduced from the analysis of the 

 observed spectrum. 



By the application of these methods it has been possible, as indicated 

 in Table IX, to determine the probable ground terms of elements for which 

 the spectra have not yet been classified, or even for which no spectroscopic 

 observations are available. In cases where the arrangement of electrons 

 is most doubtful, such as osmium and iridium, on account of the near 

 equality of energies of different configurations, alternative arrangements 

 and ground terms are suggested. The precise arrangement in such 

 atoms can be decided only by further discussion of the spectroscopic data. 



Grimm and Sommerfeld^^ have drawn attention to two general rules, 

 namely : (1) All elements with completed sub-groups (n^.^..) of electrons 

 have ground terms with j=0. (2) The element which immediately 

 precedes or follows one that completes such a sub-group has a value 

 of j for the ground term identical with the inner quantum number (k') 

 of the sub-group to which the last bound electron belongs ( = k' for odd, 

 and k'—h for even multiplicities). Mg (12), for example, has ^S^ for the 

 ground term, while the'preceding element, Na (11), has °S| corresponding 

 with 3ii, and the following element, Al (13), has ^F^ corresponding with 

 3.2J. In these simple spectra we also have l=k, but, as already indicated, 

 this does not hold for complex spectra. A partial exception to thfe second 

 rule is apparently found in comparing Rh (45) with Pd (46), but it is to 

 be noted that there is a discontinuity in the succession of configurations 

 at this part of the table. 



There can be little doubt, however, that the foregoing table of electron 

 arrangements and ground terms is substantially correct, and it may 

 accordingly be utilised in the consideration of such questions as that of 

 chemical valencies. In the paper above mentioned, Grimm and 

 Sommerfeld have discussed the closing of sub-groups in relation to 

 valencies, and have concluded, among other things, that the completion 

 of the shell of two electrons must be taken into account, as well as that of 

 the shell of eight which occurs in the inert gases. Attention is also 

 directed to the elements copper, silver, and gold, which are of special 

 interest because they immediately follow nickel, palladium, and platinum, 

 respectively, these being the last of the ' triad ' elements of Group VIII 

 for which completed shells of eighteen electrons might have been expected. 

 The position with regard to silver is quite clear, because it has been found 

 that palladium has a ground term ^Sq, lying considerably below other 

 levels, and implying a completed shell more or less resembling that of 

 the inert gases. Thus the silver atom, with one additional electron, 

 behaves essentially like the alkali metals, which have underljdng com- 

 pleted groups or inert gas shells ; silver is correspondingly always 



34 Proc. Roy. Soc, A, vol. Ill, p. 83 (1926). 



35 Zeit.f. Pkys., vol. 36, p. 38 ; Nature, vol. 117, p. 793 (1926). 



