340 BELL SYSTEM TECHNICAL JOURNAL 



cable range; while with nearly all other materials / is apparently 

 proportional to // as far as the field can be carried. Here again there 

 is a contrast so great that it can serve as the basis of a useful distinc- 

 tion. Yet all the intermediate stages between the two extremes are 

 exhibited by iron at the various temperatures between 700° and 

 800° C. Furthermore, there is reason from theory (as we shall see) 

 for supposing that the magnetization of any substance would cease 

 to be proportional to // and would approach a limit, if we could force 

 the field to high enough or the temperatures to low enough values. 

 In fact, there is at least one of the substances conventionally called 

 "paramagnetic" (it is gadolinium sulphate) for which / was found 

 to approach a limit, when the applied fieldstrength was increased 

 while the substance was maintained at the unprecedentedly low tem- 

 perature of l.°3 K. It is therefore evidently something of an accident 

 that in practice we nearly always meet either with substances for 

 which the ratio //// is constant within the accuracy of measurement 

 throughout the feasible range of the fieldstrengths, or else with sub- 

 stances for which that ratio varies greatly and unmistakably with 

 the field. 



Presence or absence of hysteresis is the third and last of the usual 

 criteria. Iron and cobalt and nickel and some of their alloys and 

 the Heusler alloys exhibit hysteresis-loops, and residual magnetism, 

 and coercive force; and the normal magnetization curve must be 

 distinguished from curves obtained by other procedures for varying 

 the applied field, and one must bother with demagnetizings or else 

 take account of the prior magnetization of whatever sample he is 

 working with. Other substances are free from these complications. 

 Here also it is probable that in iron all the measurable features of 

 hysteresis dwindle off continuously to zero as the metal is heated. 

 On the other hand, it appears that gadolinium sulphate, in spite of 

 acquiring a curvature in its /-vs.-// curve at extremely low temper- 

 atures, does not acquire hysteresis and residual magnetism. Perhaps, 

 then, it is better to take the presence of hysteresis rather than the 

 inconstancy of the ratio //// as the sign of that curious quality, 

 whatever it may be, which makes iron notable among metals. 



The general conclusion seems to be the same, as for many other 

 classifications — that is to say: It is possible to draw distinctions 

 between "ferromagnetic" and "paramagnetic" substances, valid for 

 extreme cases of the two types, not sharply marked for intermediate 

 cases; but it happens that for the time being the intermediate cases 

 are in practice not conspicuous; and consequently the distinctions — 

 any one of the three which I mentioned — are useful and worth the 

 making. 



