310 BELL SYSTEM TECHNICAL JOURNAL 



that for every magnetizable substance there is a distinctive I-vs.-H 

 relation — it is evident that this relation, if it exists, must be the 

 fundamental fact of magnetism. The first object of research in 

 ferromagnetism is to discover it for all of the ferromagnetic materials; 

 the second, to devise for each of these materials a model, accounting 

 for the particular form of /-vs.-// relation which it displays. 



On setting about to collate the recorded samples of /-vs.-// curves, 

 one promptly encounters the last and greatest of the troubles of 

 ferromagnetism. There are infinitely many such curves to be col- 

 lected, for there is a limitless variety of ferromagnetic substances! 



This is not always realized, because of the unfortunate practice of 

 referring to "the three ferromagnetic metals, iron, nickel and cobalt," 

 as though there were but three I-vs.-H relations to be determined. 

 But in addition, there are ferromagnetic alloys: binary alloys of iron 

 with nickel, of nickel with cobalt, of cobalt with iron; ternary and 

 yet more complex alloys comprising these and other elements, or 

 consisting entirely of elements none of which by itself is ferromagnetic. 

 Anyone acquainted with the diversities of alloys would be prepared 

 to find a truly vast variety of qualities exhibited by these; and he 

 would not be disappointed. Indeed, an alloy may contain one of its 

 elements in so small a proportion as to appear quite negligible — so 

 small, as to be considered a mere casual impurity — so slight, as to be 

 difficult to detect and difficult to expel — and yet so great, as to influence 

 the magnetization in the most drastic fashion. Iron containing a 

 fraction of a per cent of carbon differs as much from pure iron, in 

 regard to its magnetic properties, as either differs from nickel. (Per- 

 haps even what is now called "pure" iron contains a minimal amount 

 of some undetected yet potent impurity, the ultimate removal of 

 which will reveal a whole new set of phenomena!) So there is not a 

 triad, but a multitude of ferromagnetic substances, each of which 

 may be expected to have a distinctive /-vs.-// relation of its own. 



But for each of these substances there is, as it turns out, not one 

 but a legion of I-vs.-H relations. The curve depends very much on 

 the temperature of the sample — to such an extent, indeed, that as 

 the temperature is raised, the ferromagnetism varies rapidly, dimin- 

 ishes, and finally vanishes. The curve depends also upon the mechan- 

 ical stresses prevailing in the material, compression and tension and 

 torsion and the complicated combinations of these. It is also liable 

 to be altered by an electric current flowing in the material. 



Degree of crystallization likewise matters a great deal. Most of 

 the samples of metal used in the past have consisted of very great 

 numbers of very small crystals, millions of them to a cubic centimeter. 



