NO. II STRUCTURE OF THE ATOM PARSON 59 



In comparison with this we have for the moment of the Iron atom 

 in the metal at saturation the value 2X io~ 20 only, and for the Iron 

 atom in salts or for the Oxygen atom considerably smaller values. 

 Also the moment of Weiss' magneton is 1.85 X io -21 . 



If, therefore, the magneton under discussion were required to 

 correspond to an empirical unit of magnetic moment, as does Weiss' 

 magneton,, this result would be fatal to it : but with magnetons which 

 are movable current circuits, the result is exactly what we would look 

 for. On the present theory, it can be definitely affirmed that no arom, 

 however many magnetons it contains, can have a moment greater than 

 that of one magneton, and that the moment of most atoms will be 

 very much less than this, because the force between the magnetons 

 in an atom will always tend to orient them so as to make their result- 

 ant moment zero. The group of eight, with a very low magnetic 

 energy, has no moment : the free, or valence, magnetons will always 

 tend to lie in configurations of no moment; and indeed the only 

 atoms that could be imagined to have a moment as great as that of the 

 magneton are the atoms H, Li, Na, K, Rb, Cs, of the constitution 

 wy+i, which contain only one valence magneton each. Further, 

 what has just been said applies only to isolated atoms : in polyatomic 

 molecules, or in the liquid or solid state, the moment of the atom will 

 be still further reduced by the mutual actions of the magnetons of 

 different atoms. To take an illustration, the isolated H atom should 

 have the moment of one magneton, but the H 2 molecule can have the 



configuration ( c^»s> I < =^> , with no moment, while in HC1 



or HI all the magnetons are in groups of eight (for most of the time) . 

 Nothing, I believe, is known about the magnetic properties of any 

 substances as monatomic gases except Helium and Argon (see §2), 

 and perhaps Mercury. The investigation of them would present 

 exceptional difficulties, but it would be a valuable test of much of this 

 magneton theory, and will be undertaken at the earliest opportunity. 

 Meanwhile, the view here taken that the moment of the Iron atom, 

 even, is a comparatively small difference effect has everything to 

 recommend it. The great dependence of the moment upon tempera- 

 ture and upon the mode of chemical combination makes it clear that 

 it is a very delicately balanced effect which is due beyond all doubt to 

 certain favorable configurations of those portions of the atom which 

 are responsible for the forces exerted on other atoms. It is most 



