CONTEMPORARY ADVANCES IN PHYSICS 233 



and thus from the measurement of do at any temperature we derive 

 the magnetic moment of the individual atom or molecule of the gas. 

 This formula ought to give the right order of magnitude for /x in any 

 case. Whether or not it will give exactly the right value, will depend 

 on the validity of one of the assumptions, which I now recall. This 

 particular formula is for the case in which the atoms have only two 

 permitted orientations in the field, the parallel one and the anti- 

 parallel one. Had we supposed that every inclination is a permissible 

 one, we should have arrived at (ll3)NfjL^!kT. Had we supposed a 

 number of permitted inclinations greater than two and less than in- 

 finity, we should have arrived at some intermediate value. So, I now 

 write as the general formula, 



volume-susceptibility x = bN/jr/kT, b = 1 to 1/3, (8) 



having placed on the left the name and the symbol by which is generally 

 known what I have been denoting by tan do, and on the right a factor b 

 of which the value will depend on the number — I will call it n — of 

 permitted orientations, but will fortunately never be outside of the 

 narrow range between unity and 0.33. 



Thus a rough estimate of an atomic moment may be made without 

 knowing the number of the permitted orientations. Very many such 

 estimates have been made, and they always give values of fj. quite 

 compatible with what we know in general about the structures of the 

 atoms. If we want to make an estimate truly accurate enough to 

 serve as a stringent test of theory, then we must take from the spectrum 

 of the atom, not only the spectroscopic value of magnetic moment with 

 which we are going to make the comparison, but also the angular 

 momentum of the atom which is what determines the number of 

 orientations. This causes us no extra trouble, for if we understand the 

 spectrum well enough to get the one we also understand it well enough 

 to get the other. Now when we look into the literature to see how 

 many such comparisons have been made, we suffer again a disappoint- 

 ment. It turns out that the noble gases and most other convenient 

 gases exhibit the magnetic moment zero. This is of course no fortuitous 

 bit of ill luck; it is the same thing, to wit a certain stable interlocking 

 of the various electronic orbits and rotations in the atom, which leads 

 on the one hand to a zero magnetic moment and on the other hand to a 

 relative smallness of the forces which make for chemical combination 

 and for condensation. Anyhow it is an inconvenience; but luckily 

 there are two convenient gases, oxygen and nitric oxide — O2 and NO — 

 which do have magnetic moments different from zero; and the test of 

 the theory is in these cases most satisfactory. The agreements be- 



