CONTEMPORARY ADVANCES IN PHYSICS, XXIX 303 



{d) The phenomenon of alternating intensities in band-spectra serves 

 to reveal the spins of a few kinds of nuclei, and in a very interesting and 

 reliable way, but must be left for another occasion. 



There remain the methods which involve the use of a magnetic 

 field in one way or another, and some of which incidentally tell most of 

 what we know about the magnetic moments of nuclei, though nowhere 

 near so amply or so exactly as we should like. 



F'irst it must be said that the analogy between the vectors / and / 

 on the one hand, L and S on the other, which thus far has been so full 

 and helpful, breaks down completely when the atom is exposed to a 

 magnetic field of ordinary strength. Were the analogy perfect, an 

 atom in a state distinguished by the quantum-number F for its total 

 angular momentum would act as a rigid spinning body and would be 

 able to assume (2F + 1) discrete orientations in the magnetic field, 

 corresponding to{2F -{- 1) magnetic levels. This would be true of each 

 of the (2/ + 1) or (2/ + 1) states comprised in what I have been 

 calling a "cluster" with a common value of /, though the value of 

 F and hence of {2F -\- 1) would differ from one state to the next. 

 The magnetic levels would be distributed in groups, each corresponding 

 to a difTerent value of F. The numbers in the different groups would 

 be unequal. The total number for all the groups or states of the cluster 

 would amount, as the reader can figure out, to the product (2/ + 1) 

 X (2J+ 1). 



It is altogether probable that this is precisely what does happen in 

 magnetic fields so weak as not to separate the magnetic levels per- 

 ceptibly (their separation being then, it will be recalled, proportional 

 to the field-strength). Yet in fields strong enough to produce a 

 measurable effect, the disposition of the magnetic levels has only one 

 thing in common with this hypothetical distribution. Their total 

 number is precisely {21 + 1) (2/ + 1). They are, however, distrib- 

 uted in (2/ + 1) groups, each consisting of (2/ + 1) levels; as though 

 first of all the atoms were to forget their nuclear angular momentum 

 and remember only their electronic angular momentum, and were to 

 orient themselves in the field in the (2/ -f 1) different ways which were 

 prescribed for them (page 295) while the nucleus was still being neg- 

 lected; and as though then they were to remember the nuclear angular 

 momentum, and were to allow for it by adopting, in place of each 

 separate one of the (2 J +1) very different orientations, a group of 

 (2/ + 1) orientations differing only a little from it and from each 

 other. 



This rather animistic idea is not very far from the model commonly 

 conceived. It is supposed that in the strong magnetic field the nucleus 



