MAGNETO-ELASTIC SOURCE OF NOISE 337 



magnetic lield along the axis substantially weakened the effect was 

 additional confirmation, for it is well known " of ferro-magnetic ma- 

 terials that strong magnetization in a given direction reduces their 

 susceptibilit}' in other directions. 



Of course it was inferred at once that the same stresses that brought 

 about changes in the intensities of circular magnetization of the speci- 

 mens also were causing changes in the intensities of longitudinal 

 magnetization. A coil of insulated wire was connected to the input 

 of the amplifier. When one of the ferro-magnetic specimens was 

 placed along the axis of this coil so that the coil winding was approxi- 

 mately midway between its ends, and then was stimulated to vibration, 

 similar sounds were heard from the speaker as with the previous 

 arrangement. As an interesting comparison it was found with a soft 

 steel rod specimen six feet long and one half inch in diameter that 

 a few more than fifty turns of wire on the coil were necessary to produce 

 a sound in the speaker of the same loudness as that obtained when the 

 amplifier input was simply connected to the two ends of the rod. 

 Taking account of the cross-sectional areas available to the circular 

 and to the longitudinal magnetizations, it thus w^as shown that the 

 two classes of efifect were not of different orders of magnitude. 



There now was prepared a specimen planned especially to emphasize 

 the effect of circular magnetization. A soft steel tube six feet long 

 and having an external diameter of three eighths inch and a bore 

 diameter of one eighth inch was obtained, and midway between the 

 two ends a small opening was cut between the outer surface and the 

 bore. Two similar windings of insulated wire were placed, each 

 encircling closely with four turns the longitudinal wall cross-section 

 of the tube between one end and the mid-point. Switching arrange- 

 ment was provided for connecting these two windings in series either 

 so as to encircle the total wall cross-section in one sense, or so as to 

 encircle the two halves in opposite senses. The tube then was placed 

 at the axis of a six-foot long solenoid, and the entire assembly w^as 

 mounted with its axis horizontal and lying in the magnetic East-West 

 direction. Sources of direct current and of sixty cycle alternating 

 current were available. 



Demagnetization of the specimen was accomplished by passing 

 initially strong alternating current through the solenoid and the bore 

 windings, either successively or simultaneously, and then tapering 

 this current off uniformly to zero value. The effectiveness of the 

 treatment could be inferred from the following observations. The 

 specimen was made to acquire strong residual magnetism in the 

 longitudinal direction by passing direct current momentarily through 



