MAGNETIC RESONANCE. II 389 



gible in effect, one's story would be relatively short and it would be 

 grossly inadequate. But here the physicist, true to the tradition of his 

 science, turns hindrance into help, and analyzes the distortions for the 

 knowledge they are capable of giving about the fields prevailing in the 

 sample. Thus whereas nuclear resonance is largely used for getting light 

 on nuclei, the electronic resonance is largely studied for the information 

 that it yields about the solid state. 



Another of the great contrasts is due to what are called the "anti- 

 parallel couplings" between electrons. Generally speaking (and this 

 means: conceding an occasional exception) any type of nucleus of non- 

 zero magnetic moment will display a detectable resonance if there are 

 enough of them in the sample. Were this so A\dth the electron, every 

 substance whatsoever would display electron resonance. Experience 

 shows that electron resonance is rare, usually conspicuous by its absence. 



This is because electrons may, and not only may but usually do, 

 pair off with one another in such a manner that the spin of such an 

 "anti-parallel" pair is zero and so is the magnetic moment. There is no 

 resonance for such a pair; and the customary absence of electron reso- 

 nance signifies that in most solids, all the electrons are joined two by 

 two into antiparallel pairs (this was known before magnetic resonance 

 was first produced). 1 will call such electrons "compensated"; in this 

 language, the substances in which magnetic resonance is to be sought 

 for are those with uncompensated electrons. Mostly these belong to 

 one or the other of two classes: the ferromagnetic bodies including the 

 anti-ferromagnetic, and the "strongly paramagnetic salts." But there 

 are a few other cases, and among these are those which are closest to the 

 (unattainable) ideal of the perfectly free electron subjected. 



THE NEARLY IDEAL CASES 



Nearest of all to the ideal case are presimiably the atoms which con- 

 tain uncompensated electrons and are available for study by the molecu- 

 lar-beam method. Outstanding among these is the hydrogen atom, whose 

 single electron must remain uncompensated because there is no other in 

 the atom. About or quite as good are the atoms of sodium, potassium, 

 and the other alkali metals, each of which contains a single uncompen- 

 sated electron not to speak of several which are compensated. Moreover, 

 these atoms are normally in a "ground state" in which the uncompen- 

 sated electron has no orbital angular momentum. This hints at a 

 complexity which is not always without influence on electron resonance, 

 and must be mentioned here at the price of a detour. 



