338 BELL SYSTEM TECHNICAL JOURNAL 



the solenoid winding, and its magnetized condition was verified by 

 exploration with a magnetic compass. Then upon applying the 

 demagnetizing cycle either to the solenoid or to the bore winding this 

 evidence of the magnetized state would disappear. 



With the halves of the bore winding in series aiding and connected 

 to the amplifier input, tapping on the end of the demagnetized tube 

 produced a low but distinct sound in the speaker. Upon reversing 

 one half of the bore winding the loudness of this sound usually was 

 slightly reduced, but occasionally was slightly increased. Following 

 the momentary passage of direct current through the bore winding 

 with its halves connected in either sense, the loudness of the sound 

 from the speaker was tremendously and permanently increased. Now 

 upon reversing one half of the bore winding the loudness of this sound 

 always was reduced markedly — although never to so low a level as 

 that produced by the demagnetized specimen. Also, it was noted 

 that with the bore windings in series aiding and connected to the 

 amplifier input and starting with the tube demagnetized, the moment- 

 ary' passage of direct current through the solenoid winding ("thereby 

 imposing a state of residual longitudinal magnetization upon the 

 specimen) was followed always by a moderate but marked increase 

 in the response to tapping. 



The foregoing results established quite firmly that the effect under 

 investigation was due largely if not wholly to variations in the intensity 

 of circular magnetization accompanying the applications of stresses. 

 The presence of the effect to a slight degree with a specimen which 

 presumably was in a demagnetized state remained unexplained, since 

 testing equipment was not available for extending the inquiry further. 

 Several plausible explanations suggested themselves. Perhaps a 

 specimen of ferro-magnetic material could not be demagnetized 

 completely, or — what amounted to the same thing — perhaps the state 

 of complete demagnetization was unstable, and was followed im- 

 mediately and spontaneously by the appearance of some magneti- 

 zation. Again, it might have been that the state of complete de- 

 magnetization was reasonably stable of itself, but was readily dis- 

 turbed by the initial application of the mechanical stresses. It 

 seemed reasonable to expect that any such self-magnetization would 

 have arisen most pronouncedly along the paths of least reluctance. 

 For the time being, it was necessary to leave this point to conjecture. 

 Certain of the results described in the paragraph immediately pre- 

 ceding clearly were attributable to lack of homogeneity of the specimen. 



That a simple length of iron rod should be capable of functioning 

 as a complete alternating current generator appealed to the writers 



