294 BELL SYSTEM TECHNICAL JOURNAL 



tripled by virtue of the presence of two or three kinds of molecules 

 differing only in nuclear masses. 



More recondite is another influence of nuclei on spectra, which is 

 due neither to their charge nor to their mass. It often happens that 

 what appears with an ordinary spectroscope to be a single line is 

 resolved by an excellent instrument into several, although the earlier 

 theory affirmed quite decisively that it should be single and simple. 

 By "the earlier theory" I mean one which was substantially like the 

 atomic theory of today, except that it involved the assumption that 

 the field whereby the nucleus acts upon its attendant electrons is 

 purely an inverse-square electrostatic field. If we suppose that in 

 addition to this there is a magnetic field — that the nucleus is not only 

 a charged body, but also a minute magnet acting upon or (to use a 

 commoner term) "coupled with " the orbital electrons by the magnetic 

 as well as by the electric field — then the subdivision of these apparently 

 simple lines into clusters begins to become intelligible. It is well 

 known that spectrum lines are split into clusters by the action of an 

 external magnetic field — the Zeeman effect; it is natural to expect a 

 magnetic field applied to the orbital electrons from the center of the 

 atom to have somewhat the same effect as one applied from without, 

 and to produce these permanent splittings, which are known as 

 "hyperfine structure." Magnetic moment is attended with angular 

 momentum, inasmuch as magnetism is due to whirling of electric 

 charge; and some physicists prefer to regard the latter as primary, 

 and to say that the subdivision of the lines is due to some unspecified 

 kind of an interaction between the angular momenta or the "spins" of 

 the nucleus and the orbital electrons. To the ones, the hyperfine 

 structure yields the spin of the nucleus; to the others, its magnetic 

 moment. These are intricate questions, to which it will be necessary 

 to devote much space. 



The nucleus is a magnet; the incessant circlings of each electron in 

 its orbit constitute another magnet, a charge revolving in a closed 

 path being equivalent to a current flowing in a closed circuit; and 

 finally, it has proved essential for spectrum analysis to assume that 

 each electron is in itself, quite apart from its motion, a magnet. The 

 magnetic moment of the atom as a whole is the resultant of these 

 three component moments, or rather groups of moments, since there 

 may be many electrons and many orbits to a single atom. Now, 

 this resultant may be measured, for instance by the method of Gerlach 

 and Stern, in which a stream of atoms is deflected by a non-uniform 

 magnetic field; and if there is ground for believing that one knows 

 what part of the resultant is due to the electronic moments, then one 



