104 BELL SYSTEM TECHNICAL JOURNAL 



Much work is now being devoted to the spectra of ionized atoms. 

 One instance, that of ionized potassium, will suffice to illustrate 

 the problem. The potassium atom differs from the argon atom in 

 three respects; the weight of its nucleus is slightly different, which is 

 probably inessential; the charge on its nucleus is 19/18 as great; and 

 it has a nineteenth electron outside of the three closed electron-shells, 

 comprising eighteen electrons, which by themselves constitute the 

 whole electron-system of the argon atom. When this outermost 

 electron is removed, we have a system which probably differs from 

 the argon atom in only one essential respect — that the central nucleus 

 has a somewhat larger attractive power and hence the three electron- 

 shells are somewhat more drawn inward. The spectra of ionized 

 potassium and of argon should therefore be very nearly alike. This 

 has been tested by Zeeman and Dik at Amsterdam; the result is very 

 satisfactory, and the difference between the simple and clearly- 

 arranged spectrum of potassium on the one hand, and the rich and 

 intricate spectra of ionized potassium and argon on the other hand, 

 is very striking. At Bonn, the spectrum of ionized rubidium is 

 being compared with that of krypton for the same purpose. 



The most extensive results, however, have been obtained by Fowler 

 with silicon. For some reason or other, silicon is a particularly easy 

 element from which to obtain spectra not only of the neutral and the 

 ionized atom, but also of the twice-ionized and thrice-ionized atom — 

 four distinct spectra, one from neutral silicon, the next from an atom 

 resembling aluminium, the next from an atom resembling magnesium, 

 and the last from an atom resembling sodium. These four spectra 

 can be observed in the stars and in the laboratory, some of the im- 

 portant lines from thrice ionized atoms having been photographed 

 by Millikan in the extreme ultra-violet. In their general type, they 

 resemble the spectra of the neutral atoms corresponding in structure 

 to the atoms which emit them. 



Data have also been made available for doubly-ionized magnesium 

 (by Paschen) singly-ionized magnesium, and for neutral sodium — three 

 atoms in which the nuclear charge is respectively 13e, 12e, and lie, 

 while in each of them the nucleus is surrounded by ten electrons and 

 there is an eleventh one much further out. This eleventh electron 

 being responsible for the spectrum and being relatively exempt from 

 perturbations due to the other ten, the spectra of these three atoms are 

 of the simplest and clearest type. The series-lines which in the 

 spectrum of neutral sodium are in the inaccessible infra-red are 

 moved up, in the spectrum of doubly-ionized aluminium, into the 

 visible region. Further study of spectra related to each other in this 



