244 BELL SYSTEM TECHNICAL JOURNAL 



always agrees with the "theoretical" value derived from the spectrum 

 of that substance when a gas. This is indeed the case with gases and 

 even with a certain number of solids, a large enough number to inspire 

 confidence in the theory. There are, however, numerous exceptions 

 among solids — a circumstance not to be wondered at, since an atom 

 incorporated in a solid is usually in a very different condition from an 

 atom freely wandering about in a gas. The like is true about that num- 

 ber «, the "number of permitted orientations of the atom in a field," 

 which was introduced near the beginning of the article. Either the 

 trend of the I-vs-H curve for a paramagnetic, or the trend of the 

 /u-vs-iJ curve for a ferromagnetic, enables us (if it has been sufificiently 

 well measured) to ascertain the value of n ; and in a surprising number 

 of instances, comprising iron, cobalt and nickel as well as various rare- 

 earth elements in chemical compounds, the curves prescribe the value 

 two, when the free atom according to its spectrum would display some 

 other value. Thus when the atoms are compacted together into a 

 solid, their proximity afifects them in such a way as to bring about this 

 result. 



Now the important point about this value two for n is, that it is the 

 value to be expected for an electron which is either isolated, or else 

 linked to its atom in such a way that it has no orbital angular momen- 

 tum. The contemporary theory of spectra includes, as one of its 

 essential elements, the postulate of the "spinning electron" — the pos- 

 tulate that each electron by itself is endowed with an intrinsic and 

 indestructible angular momentum and magnetic moment, of definite 

 known amounts, having nothing whatever to do with its orbital 

 revolutions.^ This angular momentum or "electron-spin" is of the 

 amount which requires n = 2, when it is not compounded with an 

 angular momentum of orbital motion or with angular momenta of 

 other electrons. The atoms in question behave, when compacted 

 into solids, as though this angular momentum of individual spinning 

 electrons were the only one left outstanding. 



This striking inference is greatly strengthened by measurements 



upon the one phenomenon in which that angular momentum, which 



according to atomic theory is always the companion of magnetic 



moment, comes to light. Imagine a cylinder of some paramagnetic or 



ferromagnetic substance, hanging freely from a suspension attached 



to one end. Suppose it to be unmagnetized at first; this signifies 



that the atoms (whether or not they are grouped into domains) are so 



oriented that the resultant of all their angular momenta, as well as 



* The reasons furnished by spectroscopy for making this postulate are much too 

 complex to be interpolated in this article: I refer to the first fourteen pages of "Con- 

 temporary Advances in Physics," XXIX, this Journal, 14, 285-321 (April 1935). 



