CONTEMPORARY ADVANCES IN PHYSICS 339 



sorely tempted to have recourse after the first few efforts to devise a 

 better definition. Let us, nevertheless, at least take notice of a few 

 of the alternative proposals. 



Materials are classified into diamagnetic and paramagnetic and 

 ferromagnetic. To distinguish those of the first sort is relatively 

 easy, since in any of them the magnetization / called forth by an 

 applied field H is antiparallel to H (in isotropic materials, at least; 

 in crystals the angle between / and H lies between 90° and 180°). 

 In materials of the second or of the third sort, the vectors / and H 

 are parallel and point in the same sense, or at least are inclined to 

 one another at angles smaller than 90° — provided, that is to say, 

 that the material was demagnetized before H was applied. To 

 distinguish between paramagnetic and ferromagnetic bodies, therefore, 

 we must seek some other criterion. 



The magnetization of iron, nickel, cobalt, certain of their alloys 

 with one another and with other metals, and the Heusler alloys, may 

 attain values enormously greater than those which can be impressed 

 upon other substances with the highest possible fieldstrengths. One 

 might therefore select some intermediate value for /, and say that all 

 substances for which / may surpass this critical value are ferro- 

 magnetic, all others paramagnetic (or diamagnetic). In practice this 

 is usually convenient, because of the great contrast between the 

 substances which I just listed and practically all others. Among 

 the elementary metals apart from the iron-cobalt-nickel triad, one of 

 the most magnetizable is platinum, which shares a column of the 

 periodic table with that triad; yet its susceptibility is only 2 -10"^, 

 and an applied fieldstrength of 20000 gauss would impart to it a 

 magnetization of less than one unit, which is utterly negligible com- 

 pared to those which are easily imprinted even upon the less magnet- 

 izable of the substances which I named. The contrast is therefore 

 great enough to be the basis for a useful definition. Yet it must be 

 regarded as accidental, that in practice we are nearly always con- 

 fronted with extreme cases of one sort or the other. If we travel along 

 the iron-manganese or the nickel-chromium series of alloys (to take 

 but two instances), or if we follow pure iron through a sufficient range 

 of rising temperatures, we find a continuous series of intermediate 

 stages between one extreme and the other; and in principle it is 

 necessary to take account of these. 



The magnetizations of iron, cobalt, nickel, certain of their alloys 

 and the Heusler alloys increase, when the applied field is continuously 

 increased, in the curious ways which I described above, attaining 

 maximum limiting-values at fieldstrengths well within the practi- 



