CONTEMPORARY ADVANCES IN PHYSICS, XXIX 317 



It is, however, an important point — a very important point — that 

 the different schemes imply different numbers of particles for a given 

 nucleus: {2A — Z) by the former, and simply A by the latter. I have 

 spoken of a notable rule contrasting the values of I for even mass- 

 number with those for odd mass-number. Notice now that if we 

 adopt the latter or proton-neutron scheme, the rule becomes: The 

 quantum-number of the nuclear angular momentum is a half-integer if the 

 number of particles in the nucleus is odd, and a full-integer {if ascertain- 

 able at all) if that number is even. With the proton-electron scheme, 

 this could not be said. 



This difference would give an advantage to the latter scheme, even 

 if the model were no more specific. However, the rules for quantizing 

 orientations of vectors of quantum-number 1/2 require that these 

 shall set themselves either parallel or anti-parallel (in the loose sense) 

 to one another. If / be built up out of such vectors, then necessarily 

 an even number thereof implies a full-integer value and an odd number 

 a half-integer value, and vice versa. 



This argument for the proton-neutron scheme is therefore strong, 

 though perhaps not so strong as it would be, were not the basis for the 

 test narrowed down by one of the curious empirical rules of the world 

 of atoms: it is found that mostly {2A — Z) is even when A is even, 

 and odd when A is odd. Fortunately there are some exceptions; 

 the famous one is iV", the chief isotope of nitrogen, for which it is 

 certain (from alternating intensities in the band-spectrum) that / is 

 full-integer (unity), whereas {2A — Z) is odd but A is even. This 

 is the only case of its kind, but there are something like ten in which 

 {2A — Z) is even but A is odd, and there is a hyperfine-structure 

 believed to correspond to a half-integer value of /; this seems especially 

 well established for two isotopes of tin and two of mercury. Another 

 and very powerful argument from alternating intensities in band- 

 spectra, unfortunately too long to be expounded here, supports the 

 belief that the nucleus of N^* has an even and not an odd number of 

 constituent particles. On the whole it is pretty likely that any 

 nucleus-model providing an even number of particles for an even mass- 

 number and an odd for an odd will always be preferred to any model 

 not having this feature. 



The field is now open for interpreting the observed values of / 

 by compounding proton-spins and neutron-spins (or proton-spins and 

 electron-spins), and trying to find reasons for the resultants which are 

 observed. It seems natural enough to have unity for the deuteron 

 (proton and neutron spins parallel), zero for He** and C^- and O*^ 

 (spins cancelling each other two by two) ; but farther along the list 



