306 BELL SYSTEM TECHNICAL JOURNAL 



force and a resulting deflection of the flying atom. Force and deflec- 

 tion will increase with the z-component of the magnetic moment of the 

 atom, and atoms with different values of this component will go in 

 different directions. 



First we disregard the influence of the sodium nucleus. The sodium 

 atoms in experiments of this type are always in their normal state, 

 for which I recall that / = and j = s: the angular momentum and the 

 magnetic moment are exclusively those of the spin of the valence- 

 electron, the former having quantum-number 1/2 and magnitude 

 VK2 + 1) ih/2T), the latter being parallel to the former and having 

 magnitude g{ef2mc) times the magnitude of the angular momentum, 

 with 2 for the value of g. Having this value, the angular momentum 

 of the atom may orient itself in either of the two ways which are crudely 

 (page 292) called "parallel" and "antiparallel" to the field, though 

 it is better (page 293) to think of the two permitted inclinations to the 

 field-direction as being arc cos hH^ih + 1) and arc cos ( — I/VK2 + !))• 

 Half of the atoms are so oriented that the ? onent of their mag- 



netic moment is |g(e/2mc)(/i/27r), the othe y^o that the 2-com- 



ponent has the negative of this value: th ^]' 1 is split into two, 

 diverging oppositely and symmetrically f "'i he axis of x. The 

 detection of this splitting is the Gerlach-St€ >"■ periment. 



Now we suppose that the sodium nucleus n angular momentum 



of quantum-number I, as a result of which t vo orientations afore- 

 said are not really two, but actually are twi 'ps of (2/ -f 1) not- 

 very-different orientations apiece. The pi 1 m is, to refine the 

 method sufficiently to bring out the fact (if it is .act) that each of the 

 two apparent beams aforesaid is actually a ci ')se group of several, 

 and to count the several. 



It is necessary to lengthen out the path of the atoms in the inhomo- 

 geneous magnetic field, since the longer their exposure to the deflecting 

 agent lasts, the farther the separate beams are drawn apart; this 

 means magnifying the scale of the apparatus and the volume which has 

 to be kept evacuated, and carrying to a very high pitch the geometrical 

 accuracy of its design, since the initial not-yet-separated beam must be 

 exceedingly narrow and must be shot forth in a very-exactly-adjusted 

 direction from its source into the field. It is essential also to reduce 

 the broad distribution-in-velocity which the atoms owe to the fact that 

 they come out of a furnace (in which sodium is being vaporized) with 

 the random velocities of thermal agitation appropriate to the tempera- 

 ture of the furnace, and wh h would more than suffice to merge the 

 beams which it is now desir to separate. One gathers that even at 

 present it would not be po. ible to make the wished-for separation. 



