ON THE ORIGIN OF MAGNETISM. 413 



involves a heating. Any substance when magnetised becomes hotter, 

 and when demagnetised, cools. But this phenomenon is produced by 

 true variations of magnetisation only, and not as a result of the apparent 

 variations due to the change of orientation of spontaneous magnetisation. 

 It is given by 



At = ^^— Af , 

 2c. d 



where N is the coefficient of molecular field already defined, c the specific 

 lieat, d the density, and I- the square of the intensity of magnetisation. 

 The phenomenon is only of real importance in the neighbourhood of 

 the Curie Point. Below this point magnetisation is mainly of the 

 apparent type, and above it the values of I^ quickly become very small. 

 At the Curie Point the effect is far from being negligible ; in fact, it 

 reaches for nickel a value of about 1° in fields which are readily attained. 

 The extent of temperature variation calculated by the molecular field 

 theory has been shown to agree with that observed. 



The general effect of these results is to show that the molecidar 

 field theory is firmly supported by experiment, and these various pheno- 

 mena enable us to determine the numerical value of the molecular field. 

 For metals at ordinary temperature it is found to be of the order of 

 magnitude 10' gauss. Now it is easy to show that in the most favour- 

 able circumstances the magnetic field produced by the magnetic moments 

 of the molecules of a ferromagnetic body cannot be greater than 

 10' gauss. It is, therefore, impossible for the mutual actions repre- 

 sented by the molecular field to be of a magnetic nature. It is just a 

 notation for forces of a non-magnetic character, with a symbol borrowed 

 from magnetism. I prefer, in place of the primitive definition given 

 earlier, the equivalent definition 



where U is the intrinsic energy and I the intensity of magnetisation, 

 both reckoned per unit volume. This definition is advantageous in 

 that it does not prejudge the nature of the forces. 



In relation to the question to-day under discussion we can thus 

 conclude : 



One of the essential conditions for the production of strong 

 magnetism — or ferromagnetism — is the existence, between the mole- 

 cules possessing magnetic moments, of important mutual actions which 

 are numerically expressed by the molecular field, and are certainly of a 

 non-magnetic nature. 



It does not appear to be impossible that the forces may be electro- 

 static; that, however, is at present a pure supposition. 



I would like now to draw your attention to another condition which 

 governs the display of magnetic phenomena. All the theories of mag- 

 netism due to Weber, Ampere, Lord Kelvin, J. A. Ewing, Langevin, 

 and others have assumed that the molecules or atoms possess magnetic 

 moments. But it is only since we have been in possession of the kinetic 

 theory and the molecular field theory that we have been able easily 



