f)34 PROFESSOR KNOTT ON SOME RELATIONS BETWEEN 



introduced. For example, in one case the wire, subject only to the earth's field, was 

 twisted to and fro through a small angle of ± 14' per centimetre length, and the changes 

 in magnetic moment noted. The range was 23 in scale units. Then the compensating 

 current was applied, the wire being vigorously tapped, and the current gradually 

 increased in value until the neutral condition was attained. The same small to-and-fro 

 twisting was again applied ; and it was found that the range of magnetic change was 36, 

 or half as great again as in the original simply magnetised condition, with a tendency to 

 a negative accumulation of polarity. On the twisting being increased to ± 28' per 

 centimetre length, the range of the magnetic change rose to 92, and the wire during the 

 first cycle acquired a considerable negative polarity. 



In another case the wire was annealed as it hung in a vertical position. On cooling, 

 it became magnetised under the influence of the earth's vertical force, which in Tokyo is 

 somewhat greater than the horizontal force. In this position it was vigorously tapped 

 and then introduced into the coil in such a manner that its south pole pointed north. 

 The magnetometer indicated a large negative polarity, which vigorous tapping only 

 reduced to about two-thirds of its original value. 



To get the wire into a neutral condition it was necessary to apply a positively 

 magnetising current, that is, one whose field would be co-directional with the earth's 

 horizontal field. On the balance being effected, the twisting was proceeded with as in 

 the former experiments. The result was an accumulation of positive polarity, so that 

 here also it is the more recently applied magnetisation which is more powerfully 

 influenced by the twist. 



As already stated, these results seem to point to the existence of a neutral layer 

 separating two regions in the wire, which differ in the sign of their longitudinal polarity. 

 The inner core of the wire is magnetised in accordance with the natural effect of the first 

 magnetising force. The outer shell surrounding this core is magnetised in accordance 

 with the natural effect of the second magnetising force, which has to be applied to bring 

 the wire to a neutral condition. The first effect of a twist being to offer facilities for the 

 molecular groupings to break up and assume configurations of less potential energy, is 

 clearly to tend to increase the longitudinal polarity induced by a constant magnetising 

 force. Evidently the outer shell of molecular groupings will be more sensitive to the 

 disturbing effects of the twist than the inner core ; the magnetic change will therefore be 

 determined by the sign of the average polarity of this outer layer. 



In the hint here given of what might be called a periodic magnetic distribution we 

 may perhaps find an explanation of the curious discovery made by Mr Nagaoka that a 

 twisted nickel wire can sustain a distinct negative polarity in a tolerably strong positive 

 field. 



29. Conclusion. — To summarise concisely the results of Part III. is not easy. I 

 shall content myself with drawing attention to what seem to be the phenomena which 

 have most importance from a theoretical point of view. 



There is, first, the contrast between iron and nickel as regards the sign of the 



