CONSTITUTION AND TEMPERATURE ON MAGNETIC SUSCEPTIBILITY. 87 



evidence would be a direct test of the value of H c .* The above deduction is an 

 interesting confirmation of the suggestion made at the end of Part II. , p. 143, with 

 regard to the intensity of the local molecular field in diamagnetic crystalline 

 substances. 



In a crystalline substance which shows natural double refraction we may regard the 

 molecules as held in position in the crystalline structure by this intense local field, and 

 in that case the double refraction of the medium would be a consequence of the 

 orientation of the molecules due to the operation of this field when the substance 

 crystallizes. Now the magnetic double refraction induced in a liquid is proportional 

 to the square of the field intensity.! Let us assume that this law holds good for much 

 larger fields than the largest we can apply in the laboratory 4 If we take 5 x 10 4 gauss 

 as the maximum value of this external field then the magnetic double refraction of the 

 liquid will be represented by Cx2'5xl0 9 , where C is a constant independent of the 

 applied field. If we could apply a field of 1 7 gauss the magnetic double refraction 

 would be C x 10 11 , i.e., 40,000 times as great. This is of the order of magnitude of the 

 ratio of the double refraction of quartz to liquid nitrobenzene, the latter being 

 subjected to a field of 3 x 10 4 gauss. 



Conversely, if we assume that on crystallization there is an internal local molecular 

 field, which if interpreted as a magnetic field is of the order 10 7 gauss, we can see how 

 under its influence the molecules would become orientated to such an extent as to give 

 rise to a crystal possessing the high natural double refraction of quartz. 



All elements and compounds which show on fusion a small percentage change of x 

 must possess a molecular field whose local value, if interpreted magnetically, is of the 

 order 10 7 gauss. This value is also supported by observations on the artificially 

 induced and natural double refractio:is of other liquid and crystalline media. 



It is known that most uniaxal crystals have a double refraction comparable with 

 that of quartz, and of a much higher order of magnitude than that which has been 

 induced in a liquid by the largest magnetic field at our disposal. Thus the 

 abnormally high double refraction of Iceland spar is about 100 times that of quartz. 

 On the other hand, the double refraction of ice is only about y^o of that of quartz. 



* It will be shown later that the mechanical interpretation of the large local molecular field is the 

 internal stress which accounts for the rigidity of a crystalline medium. Now it has been shown by 

 HUMPHREYS (' Astrophys. Journal.,' vol. 35, p. 268) that there is a direct relation between the pressure 

 shift of spectral lines and the Zeeman effect. Hence it is very probable that the shift of an absorption 

 band on crystallization would be determined by the Zeeman effect of H c . 



t This results from both the theory of HAVELOCK, 'Roy. Soc. Proc.,' A, vol. 77, p. 170, 1906, and that 

 of LANGEVIN, 'Le Radium,' vol. 7, p. 251, 1910. The law is found to hold experimentally, COTTON and 

 MOUTON, 'Ann. de Chim. et de Phys.,' ser. VIII., vol. 19, p. 155, 1910. 



| There are good reasons for supposing that this law is not accurate for such large fields, for a saturation 

 effect must come in. But the result of assuming that it holds is suggestive, and probably indicates the 

 true order of the effects. 



Vide note, p. 99. 



