338 BELL SYSTEM TECHNICAL JOURNAL 



The temperature at which I-vs.-T curves intersect the axis of T, 

 or would intersect that axis were it not that they turn aside shortly 

 before reaching it, is known as the Curie-point. For iron, values 

 of Curie-point ranging from 768° to 790° are given; the differences 

 seem to be due partly to uncertainties in deciding just where 

 the I-vs.-T curves "would intersect the T-axis if they continued 

 on downward without turning," partly to the indubitable fact that 

 these intersection-points are not the same for I-vs.-T curves for 

 different values of H, and partly to the use of other definitions of the 

 Curie-point. For nickel, cobalt, and magnetite the Curie-points are 

 in the neighborhood of 360°, 1130° and 550° respectively; and values 

 are recorded for a considerable number of alloys. 



The Curie-point is not the sign of what is properly designated as a 

 "change of phase." Iron suffers changes of phase at temperatures 

 near 900° and near 1400°, changes in which the atom-lattice goes over 

 into an entirely different type, and a number of physical properties 

 are sharply altered ; but the Curie-point is not one of these, it is the 

 locality of merely a rapid (though not absolutely sudden) change in 

 magnetic properties and an evidently-correlated anomaly in specific 

 heat.* As for the real changes of phase, they normally occur at 

 temperatures so high that they do not influence the magnetization of 

 iron below the Curie-point. Yet it is possible to bring one of the high- 

 temperature modifications suddenly down into the low-temperature 

 range, and then its magnetic properties are quite different from those 

 of "ordinary" iron. In certain alloys this possibility is easy to 

 realize; I will quote only the notorious case of a "nickel-steel" 

 discovered by J. Hopkinson, which at 580° C. is merely one of the 

 many non-ferromagnetic metals, remains so as it is cooled all the way 

 down to zero, then turns suddenly into a'modification which is strongly 

 magnetizable and retains this state as it is being heated all the way 

 back to 580° C. But indeed ferromagnetism of alloys is entangled 

 with all the infinite complexities of the behavior and the internal 

 changes of these complicated substances, and varies with all the 

 variations of the more or less durable equilibria between their com- 

 ponents. 



Definition of Ferromagnetism 



Ferromagnetism has sometimes been defined as "the kind of 

 magnetism which iron exhibits" — an easy evasion, to which one is 



* Contrary statements about iron are to be found in the early literature; but they 

 are due partly to inaccurate experiments, and partly to the fact that the change-of- 

 phase which in pure iron lies well above the Curie-point descends when carbon is 

 progressively added to the iron, and before long comes into coincidence with the 

 Curie-point; and if still more carbon is added, the "vanishing of ferromagnetism" 

 takes place at the transition temperature. 



