330 BELL SYSTEM TECHNICAL JOURNAL 



The relations between the influence of magnetization on strain, 

 and the influence of strain on magnetization, have been derived from 

 the laws of thermodynamics. It appears that each of the several 

 effects agrees with the theory insofar as the sign is concerned (for 

 instance, tension applied to a wire which shortens when magnetized 

 should diminish its susceptibility, and does) but not always in magni- 

 tude. I have not heard of anyone renouncing the laws of thermody- 

 namics on this account. 



Hysteresis plays a great part in the effect of tension on magnetiza- 

 tion; if a constant magnetizing field is applied to a wire while the 

 tension is being cyclically varied, the magnetization when plotted as 

 function of extension follows a hysteresis-loop. Also the first appli- 

 cation of a load to a wire is likely to make a sudden and violent change 

 in the value of /. Some avoid these troubles, or try to, by shaking 

 the wire continually or by continually applying an alternating magnetic 

 field during the measurements. These introduce further complica- 

 tions. In fact, if all the data that could be assembled concerning 

 the effect of strains upon magnetization were to be sought out, I 

 suspect that "the world itself could not contain the books that 

 should be written." 



The Barkhausen Effect 

 Imagine once more a piece of some ferromagnetic substance, en- 

 circled by a magnetizing coil, through which the current is being 

 steadily increased; encircled also by a loop, which is connected to 

 the voltage terminals of an oscillograph, or to some other device which 

 moment by moment records the electromotive force impressed upon 

 the loop by the changing magnetization. This electromotive force, 

 as I have said, is proportional to the rate-of -change dBjdt of the 

 induction, for which the changing of the magnetizing field is respon- 

 sible. It is a measure of the rate of magnetization of the sample 

 girdled by the loop. The magnetizing field is being increased con- 

 tinuously; were the magnetization also to rise continuously towards 

 its saturation-value, as we should probably expect, the voltage-curve 

 would be smooth. However, when a sensitive oscillograph is used, 

 the curve is a succession of sharp teeth. The magnetization of the 

 sample evidently proceeds by small but sudden jumps. These can 

 be shown — in the most literal sense of the word "to show" — by 

 connecting a telephone-receiver through an audion-amplifier to the 

 loop. Listening at the receiver, one hears a rustling or a crackling 

 sound ; it has been compared with rain beating upon a tin roof, also 

 with coal rattling dpwn a chute. Barkhausen discovered the effect 

 in this way. 



