124 Mr. H. Nagaoka on the Effects of Torsion and 



are that, although suggestive of the symmetrical form given 

 in fig. 5, the curve is now distinctly one-sided with respect 

 to the line of zero twisting, and indicates much greater 

 magnetization for positive than for negative twists. These 

 peculiarities are still more pronounced when the stress is 

 increased to 209 kilogs., as shown in fig. 7. 



This experiment is of special interest, illustrating as it does, 

 the manner in which the magnetization-cycle changes charac- 

 ter just as the extraordinary phenomenon of change of sign is 

 about to show itself. 



As usual, the first effect of twisting the wire is to increase 

 the intensity of magnetization, while the effect of untwisting 

 it is to decrease it. Then, as the untwisting is continued as 

 negative twisting, the magnetization tends to recover its former 

 value. But for the maximum negative twist of two right 

 angles, the magnetization does not nearly recover its former 

 value. Thus it is evident that in such a strength of field the 

 increase of longitudinal stress tends to make the increase of 

 magnetization during negative twisting gradually less and 

 less, until finally for a certain load the increase does not take 

 place at all for the particular range of twist. This is shown 

 in fig. 8, which is the curve for a tension of 782 kilogs. weight 

 per sq. centim. 



A study of these four curves (figs. 5, 6, 7, 8) shows the 

 character of the changes wrought in the cycles as the load is 

 increased. The symmetry is first lost, the negative loop be- 

 coming smaller and smaller. Then, as shown in fig. 7, it 

 ceases to be a loop, the course of the return curve from greatest 

 negative twist lying above the other, and never cutting it. 

 Then, as in fig. 8, the upward course of the curve on the 

 negative side of zero twist vanishes away altogether ; while 

 at the same time the phenomenon of reversal of magnetic 

 polarity shows itself. Thus the double-looped curve for low 

 tensions passes gradually into a single-looped curve as the 

 tension is increased. And after this single-looped curve is 

 obtained, the phenomenon of reversed polarity begins to 

 appear. The passage from the double-looped to the single- 

 looped curve betokens a peculiar alteration in the lagging- 

 eflFect in nickel — an alteration which has no analogue in the 

 case of iron. 



To study more carefully the law of hysteresis in nickel — to 

 use Professor E wing's word — the experiments were repeated 

 in stronger fields. 



Figs. 9, 10, 11, 12 (PI. III.) show the march of events in a 

 field of 4*94 units, the tensions increasing from 400. For 

 smaller tensions the curves are of the approximately symme- 

 trical form shown already in fig. 5, and do not call for special 



