CONTEMPORARY ADVANCES IN PHYSICS 341 



C. Theories of Ferromagnetism 



To devise a theory of ferromagnetism is not necessarily the same 

 task as to make a theory of magnetism. In studying the properties 

 of paramagnetic and those of diamagnetic bodies, one finds many 

 indications that the ultimate atoms of the elements are magnets of 

 definite and seldom-changing moments, or at least may profitably be 

 so regarded. The theory of line-spectra reinforces this opinion, and 

 it is confirmed by the observations of Gerlach and Stern upon the 

 deflections undergone by free-flying streams of atoms traversing a 

 strong magnetic field with a strong field-gradient.* Now, to say that 

 atoms are magnets is scarcely tantamount to giving an explanation of 

 magnetism. On the contrary, the problem is merely pushed a step 

 further away, and must eventually be faced again and either be 

 solved by explaining why atoms are magnets, or else be given up by 

 conceding that magnetism is one of the fundamental properties of 

 matter. Yet it is quite logical and sensible to aspire to construct a 

 theory of magnetization — of the gradual magnetizing of a substance 

 by an increasing applied field, of the shapes of the /-vs.-// curves, of 

 hysteresis-loops — out of the assumption that the ultimate atoms are 

 permanent magnets. To explain the gradual rise of an /-vs.-// curve 

 by postulating atoms which are already magnetized to saturation, to 

 explain hysteresis by postulating atoms which individually have no 

 hysteresis — these would be triumphs not open to the objection made 

 against many "explanations," that they are achieved by ascribing to 

 the atoms the very properties to be explained. 



We shall presently make the acquaintance of "elementary magnets" 

 — hypothetical beings, of which each magnetizable substance is 

 supposed to consist. To these we shall assign, for the time at least, 

 definite and unchangeable magnetic moments. A magnetic field 

 applied to an assemblage of such magnets could not change the 

 moment of any. Yet it could change the net magnetic moment of 

 the assemblage, which is the resultant of the moments of all the 

 individuals; for it could, directly or indirectly, cause the elementary 

 magnets to align themselves along its own direction. The assemblage, 

 the substance, would be magnetized not through magnetization of the 

 individuals but through orientation of the individuals which make it up. 



That idea is an old one; but by itself it is nearly useless. We must 

 think of some agency which could combat the tendency of the ele- 

 mentary magnets to align themselves along the field; for there must 

 be such a one, as otherwise the weakest possible field would magnetize 

 each substance to saturation; which is not the case. The most 



* I refer for these to my Introduction to Contemporary Physics, pp. 48-50, 383-393. 



