312 H. XAOAOKA 



Reviewing the results of these experiments we see that twisting 

 produces many singular effects on the magnetic properties of nickel. 

 In the first place, maximum susceptibility occurs, for the twisted wire, 

 at lower values of the magnetizing force than for the untwisted wire. 

 The amount of twist which must he applied to olitain the maximinn 

 susceptibility in the weakest field is nbout 3° per cm. F(M" higher 

 twists, the " Wendepunkt " occurs at the higher values of the field. 

 The critical value of twist is nearly constant for all values of longi- 

 tudinal stress. Also, the wire which, twisted to this critical amount, 

 gives the greatest maximum susceptibility, gives at the same twist the 

 greatest maximum diftei'ential susceptibility — in other words the curve 

 rises most abruptly to its turning jioint. I'he field for maximum 

 differential susceptibility is very slightly smaller than the field for 

 maximum susceptibility. As the latter is passed, the differential sus- 

 ceptibility diminishes markedly in value and remains pretty constant 

 in the higher fields. Ultinaately the magnetic susceptibility of the 

 twisted nickel becomes less thim that of the normal wire, the curves 

 crossing each other in high magnetic fields. Again the field for 

 maximum susceptibility inci-eases with load ; but the maximum 

 susceptibility itself diminishes. Also the susceptibility in fields of 

 moderate strenffth is more sensitive to twisting for the greater loads. 

 This effect of load is more marked with regard to the residual magnet- 

 ism. Thus witli a load of 10 kgs. in a field of i7, a wire with a twist 

 of 3° per cm. shows 50 times as much residual magnetism as it did 

 in the untwisted conditiou. Finnllv the wire released from torsion 

 behaves in a difterent way from the normal one, the m:ignetic effect 

 of stress onl living the removal of it. 



il. The effecl of combined longitudinal and torsional 

 stresses on the retenliveness of nickel wire. 



After the preceding experiments w^re finished, I determined to 



