314 BELL SYSTEM TECHNICAL JOURNAL 



This is what is sometimes called the normal magnetization curve, 

 sometimes the initial curve ; I will adopt the latter term. 



At this point it is well to recall that most of the curves actually found 

 in the literature are B-vs.-II curves, not I-vs.-II. Since in the right- 

 hand member of the equation B — H -\- 4x7 the second term is 

 usually enormously greater than the first, a 5-vs.-// curve usually looks 

 exactly like an I-vs.-II curve plotted on a smaller scale. At very high 

 fieldstrengths, however, a B-vs.-II curve continues climbing upward 

 with a constant slope while the corresponding /-vs. II curve runs 

 parallel to the i7-axis. 



The Initial Curve 



The form of the initial curve is peculiar and distinctive. De- 

 parting from the origin of the (J, //) coordinate-plane, it ascends, 

 bends upward, passes through a point of inflection, bends over but 

 never quite turns downward; it goes off towards a horizontal asymp- 

 tote, toward a maximum or saturation value of magnetization. Nearly 

 all initial curves display these features, the point of inflection and 

 the saturation; but in all other details, in the lengths and curvatures 

 of the arcs before and after the point of inflection, in the scale of the 

 curve and of its parts, they differ very much from one substance to 

 another, and are altered very much by mechanical and thermal 

 treatments. 



Well-annealed substances, iron and nickel and permalloy for 

 instance, display curves which tempt the onlooker to divide them into 

 three segments: a slowly-rising and eventually upward-bending arc 

 starting from the origin, a relatively steep-climbing portion including 

 the point of inflection, a final arc drawing itself close up to the asymp- 

 tote. A good example is shown in Figure 1. The distinction is 

 accentuated by the hysteresis-loops which originate from the various 

 points of the curve. In the prevalent theories of magnetization, as we 

 shall eventually find, these segments are supposed to result from 

 difTerent processes occurring inside the metal. I will therefore adopt 

 this separation of the curve into three parts, warning the reader to 

 remember that at best there is always something arbitrary in subdi- 

 viding a continuous curve, and at worst there are substances in which 

 the division into three segments becomes quite impossible to make. 



The first segment, extending from the origin to what some call the 

 instep of the curve, may be regarded as a parabolic arc so long as 

 the field is rather low — for iron and nickel, inferior to about one 

 gauss; and for these metals it is sensibly a straight line so long as 

 the field is below say a tenth, or to be safe a hundredth of a gauss. 



