360 BELL SYSTEM TECHNICAL JOURNAL 



fying to make a connection between the theories of Ewing and Weiss. 

 Perhaps they are the units from which arise the separate clicks which 

 constitute the Barkhausen effect. As for the initial curve, attempts 

 must be made to explain it either by supposing that the increasing 

 field wheels the magnetic moments of the several zones gradually 

 into parallelism with itself, or — what is more probable — that the 

 field abruptly reverses, one after the other, all the magnetic moments 

 which initially are inclined to it at angles superior to 90°. By suitably 

 combining these two images, one may copy almost any possible form 

 of initial curve. I cannot enter into these questions, except to 

 answer as far as possible what I designated as the second question to 

 be asked in testing the theory: what observable quantity is to be 

 compared with the ' ' residual magnetization ' ' predicted from the theory 

 of Weiss? 



A piece of iron brought to saturation by a large applied field is 

 supposed to consist of these magnetized zones, their moments all 

 directed either parallel or at least at inclinations of less than 90° to 

 the field. The applied fieldstrength should elevate the magnetization 

 of each to a value somewhat greater (corresponding to an intersection- 

 point somewhat farther along the "curve" of Fig. 13) than the 

 predicted "residual magnetization"; but the values of n and I and 

 hence their product are so enormous that the addition is only slight. 

 The saturation intensity of magnetization of the iron, /max., should 

 then be very nearly equal to the predicted residual magnetism, 

 if all the magnetic moments are parallel ; or to one half of the predicted 

 residual magnetization, if the magnetic moments are distributed at 

 random over the directions inclined at less than 90° to the applied 

 field. In the former case, the variation of /max. with T should follow 

 the curve of Fig. 9; in the latter case, a curve of the same form. 

 The actual observations upon iron, nickel, cobalt and magnetite are 

 shown in that figure, and the reader may judge of the agreement 

 for himself. 



Comparison of Ewing' s Theory with that of Langevin and Weiss 

 At first glance the Ewing model and the Langevin-Weiss conception 

 of a ferromagnetic substance seem extremely different; contradictory, 

 in fact. In Ewing's view, the perpetual effort of the applied field to 

 align the elementary magnets is hindered by the forces which these 

 exert on one another. In Langevin 's theory, the antaonigst of the 

 applied field is the thermal agitation. Now Langevin 's theory is 

 competent to deal with paramagnetic substances which are difficult to 

 magnetize, but not with iron and the like which are strongly mag- 



