ÖJ,(1 



26 H. NAGAOKA. 



vertical component of the terrestrial magnetic field being neutralised 

 by a current in the dmall coil placed within the main coil. The reading 

 of the tor.sion circle was then taken, and the two vertical rods (rr) were 

 so placed, that they just touched the horizontal r(jd (d) of the cross on 

 its opposite sides. The wire was now twisted by turning the torsion 

 circle, and held in the twisted condition for ,S(.)nie time. The reading 

 of the scale was noted. The circle was then turned in the opposite 

 direction so ;is to make the cross free of the vertical rods. When the 

 residual twist was small, the torsi<jn circle was brought back complete- 

 ly to its original position, and the small amount of residual twist was 

 given by the difference in the initi;d and final scale readings. For 

 l-irge residual twists, however, turning the torsion circle back to its 

 origin;d position would have thrown the reflected hcmn off the sc;de. 

 Accordingly the torsion circle was turned back tlu'ongh a convenient 

 and known angle, so that the wire hung free, and the reflected l^eam 

 remained on the scale. 



In the preliminary course of experiments, after the wire was 

 freed from torsional oscillations, the elastic ufter-efFect was measured 

 by simultaneously noting the successive scale readings and the corres- 

 ponding times after the release. Wlien the untwisting due to the after- 

 effect had become very slow, the magnetizing fjrce was applied, and 

 the corresponding scale readings noted. 



Before entering' into the in\ estin'ation of the effect of magnetiza- 

 tion on the permanent twist, it was desirable to have some knowledge 

 of the torsional after-effect of nickel. In every experiment, the wire 

 was twisted and left for some time in this state. On releasing the 

 torsion, torsional oscillations ensued. After its cessation, the wire 

 continued gradually to untwist in virtue of the elastic after-effect. It 

 was then necessary to know when the untwisting due to the after 

 efïect should cease, for otherwise the untwistings due to magnetization 



