64 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 65 



selves definitely, in greater or less numbers according to the strength 

 of the external field, until at saturation they are all aligned. 



The relation kT= const, for paramagnetic substances was found by 

 Curie to hold for Oxygen, Palladium, and certain salts, and over a 

 certain range of temperature for Magnetite ; but it has been amply 

 demonstrated that the law is not of universal application. Weiss and 

 Kamerlingh Onnes have shown (Journal de Physique, 1910) that 

 Iron, Nickel, and Cobalt are exceptions ; and that the susceptibilities 

 of Vanadium, Chromium, and Manganese are not increased by cool- 

 ing to the boiling point of Hydrogen (14 abs.), whereas according 

 to Curie's law they should increase about twentyfold. It is remark- 

 able also that, even down to such low temperatures, no transition into 

 a ferromagnetic state is observed in these elements. Again, the work 

 of Honda {loc. infra) has shown that there is a class of substances 

 whose susceptibilities even increase with rise of temperature. 



This last effect is to be expected from the magneton theory, 

 because sometimes magnetons which at lower temperatures are held 

 quasi-rigid by interatomic forces might at higher temperatures be 

 more free to add to the paramagnetic effect. We may say, then, that 

 the occurrence of molecular collisions will tend to make paramagnet- 

 ism obey Curie's law, but that the action of intermolecular and inter- 

 atomic forces would be expected to enhance paramagnetism as the 

 temperature rises, although the latter effect would not be marked 

 except during a change of molecular complexity, such as occurs 

 during gaseous dissociation and to a less extent during fusion and 

 volatilization (see §22). 



The law for diamagnetism — that it is independent of tempera- 

 ture — is not universally true either. As a striking example of this, 

 Bismuth becomes less diamagnetic as the temperature rises, and at 

 the melting point the change is so great that liquid Bismuth is the 

 least diamagnetic substance known. 



These departures from Curie's laws will seem less anomalous after 

 considering the vivid picture of the distribution of magnetic forces 

 within atoms, molecules, and lumps of solid matter, which this theory 

 affords. 



In beginning to apply the theory, we can see at once that the stable 

 group of eight magnetons (§6) is an almost ideally diamagnetic 

 system. It has no intrinsic magnetic moment, and on account of its 

 low magnetic energy is not easily given one. The way in which an 

 external field would tend to distort the group is well shown in 

 configs. 1, 3, 3a on plate 2 (§6) : this illustrates the paramagnetic part 





