314 BELL SYSTEM TECHNICAL JOURNAL 



The experiments were performed by Stern and his school. They 

 are extremely delicate, owing to the excessively small value of the 

 magnetic moment and hence of the deflection. Indeed, they might 

 have been beyond instead of within the limits of the possible, were it 

 not for two distinctive features of hydrogen: the just-mentioned doub- 

 ling of the moment due to the parallelism of the spins of two nuclei, 

 and the fact that a beam of slow-moving molecules can be utilized, 

 because hydrogen does not have to be heated in a furnace in order to be 

 vaporized, but will emerge as a molecular beam through a hole in the 

 wall of a chamber no warmer, indeed very much colder, than the 

 temperature of the room itself! For the magnetic moment of the 

 proton, Frisch and Stern give 2.5 (eh/Airmpc) with an uncertainty of 

 "at most 10 per cent"; for that of the deuteron, Estermann and Stern 

 say that the value is between 0.5 and 1.0 times {ehj^irmpC). 



We now turn to the ensemble of the estimates of nuclear angular 

 momenta and of magnetic moments, and the laws and rules which 

 they seem to obey. 



Estimates of I have by now been made for some fifty-five elements; 

 but this is not an adequate statement to make, for the nuclear angular 

 momentum is one of those qualities — as I shall presently stress — ■ 

 which may vary from one isotope to another of a single element, and 

 there are already several cases in which values of / have been reliably 

 assigned to two or more different isotopes. The great majority are 

 derived from optical analysis of hyperfine-structure; four or five 

 from magnetic-deflection experiments; about ten from alternating 

 intensities in band-spectra, most of these last being checked from the 

 hyperfine structure of line-spectra. 



The values are of very unequal merit, some being derived inde- 

 pendently and concordantly from several different properties of 

 hyperfine structure (pages 301-302), some being further sustained by 

 deductions from band-spectra, while others are guesses based on a few 

 rough observations of intensities or intervals. The unreliability of 

 these last is of a type not to be described by giving a most-probable- 

 value coupled with a probable error. A critical and analytical review 

 of the lot by a neutral expert is badly needed. 



The first thing which strikes the eye on viewing a tabulation is the 

 immense preponderance of half-integer values of spin; but this is only 

 one sign of the most important of the rules of nuclear momenta, which 

 is one of the most important of the rules of nature. I recall that 

 the "mass-number" A of an atom is the integer nearest to the value 

 of its mass expressed in terms of one sixteenth the mass of the common- 

 est oxygen atom as unit. The rule, then, is: 



