308 H. XAGAOKA 



wire. The maximum value of tlae dittereiitial susceptibility is \'ei'y 

 great for twisted wire, and is greatest for the twisted wire which has 

 the greatest maximum susceptibility. After passiu»' tlie " Wende- 

 punkt," tlie magnetization curve becomes almost straight, the ditfer- 

 ential susceptibility becoming very suiull and of such a nature as to 

 cause the nickel always to cut the curve for untwisted nickel. This 

 point of intersection occurs sooner for higher twists. And finally 

 the amount of residual magnetism is enormously increased by twist- 

 ino" the wire even throuj^h a small an£>le. 



D CD CD 



In the experiments now to be described, the wire was subjected 

 to extra loads. It was necessary, however, to take anew wire cut 

 of course from the same specimen, and treated preliminarily as the 

 first piece w as. 



The first extra load used was nearly 1 kg.; and with this load 

 exactly the same series of experiments was gone through as in the 

 previous case. In Fig. 2, the curves of magnetization are shown 

 graphically. They are very similar to the curves of Fig. 1 ; only in 

 this case it was not possible with the strength of field used to get the 

 curves for the twisted nickels to intersect the curve for the imtwisted 

 nickel. In all probability, however, the crossing will occur fur high 

 enough fields. 



It was natural to enquire, here, as to the after etfect of the twist. 

 It is known that, because of the elastic after-strain, the wire does not 

 return exactly to its former position when the twisting stress is 

 relieved. It seemed likely that something analogous woidd occur 

 in reference to the magnetic characteristics of the wire. Accordingly 

 the wire, after it had Ijcen twisted, was allowed to hang freely under 

 the action of the load ; and under these conditions, the magnetization 

 curve was obtainctl. Comparing it with the first cur\e, we see that 

 there is distinct evidence of a change in the magnetic projjcrties of 



