PARAMAGNETIC & DIAMAGNETIC SUBSTANCES 299 



greater than the first and quite masks it. If there were no 

 interrnolecular and interatomic agitation that is, if the temperature 

 of the body were absolute zero the molecules might set under 

 sufficient force all with their axes in the direction of H, like the 

 magnets in E wing's model. The intensity of magnetisation would 

 then be a maximum and would be independent of the field when this 

 had reached the value needed to secure parallelism. But molecular 

 and atomic agitation that is, heat with the resulting collisions, 

 produces disturbance of the axes. 



Langevin examined specially the case of a paramagnetic gas such 

 as oxygen. Let us imagine that we are dealing with a constant mass 

 kept at constant unit volume. Let its moment in field H be M. 

 Let the moment be increased to M + dM.. The work done from 

 outside on the gas is therefore HdM. But work done on a 

 gas at constant volume goes to increase its temperature. If 

 we keep the temperature constant we must take away heat dQ = 



II<7M. But if T is the temperature, -^ is a perfect differential, 



TT 



for it is equal to the entropy removed, d(f>. Then 7^. dM. is 



aUo a perfect differential. Now the condition of the gas, in- 

 cluding the condition M, is a function of H and T only. 



It follows easily that M must be a function of ^ 



-/l> 



But as far as experiment has yet gone, when T is constant 



MocH. 



M 

 a r ^ 



Therefore 



or 



M 

 H 



C 

 T' 



But M/H is proportional to the susceptibility, so that this 

 result gives Curie's law. This cannot hold, however, down to 

 the absolute zero. There we shall have a moment M due 

 to parallel alignment of all the molecules. But of course the gas 

 law assumed ceases to hold before we get to zero temperature. 



By a treatment analogous to that by which the mean square of 

 the velocity is found, Langevin found that the magnetic suscepti- 



M 2 

 bility should be znr,, where R is the gas constant. 



The magneton.* P. Weiss, in conjunction with Kammerlingh- 

 * Weiss, Journal de Physique, 5 ser. I. (1911), pp. 900 and 965. 



