SUPERPOSED MAGNETISMS IN IRON AND NICKEL. 



201 



Experiment IV. Helical Current = 2-405 Amp. 



Linear Current, 



0-893 



1-520 



2323 



2-812 



4-334 



4-680 



Twist, 



368 



819 



1355 



1561 



1865 



1890 



Experiment V. Helical Current = 3*338 Amp. 



Linear Current, 



0-867 



1-494 



1-797 



2-298 



2-702 



3-277 



4-519 



Twist, . 



439 



929 



1226 



1639 



1909 



2168 



2374 



The representative curves are shown on Plate XXIX., nickel groups 

 A and B. The chief points of difference between the behaviour of iron and 

 nickel are these: first, the direction of twist in the nickel is the reverse of that 

 in the iron; and second, there is no maximum in the nickel A group of curves. 

 The free end of the nickel wire twists in the direction opposite to the hands 

 of a watch, as looked at from above, when the wire is traversed by a down 

 current, and is magnetised with north pole downwards. This agrees with 

 Barrett's discovery,* that nickel contracts when magnetised. The jjossibility 

 of a maximum, again, depends upon how the amount of contraction varies 

 with the magnetisation, and also, since the abscissae represent currents and not 

 magnetisations, upon the relation which holds between these last. 



The B curves are very similar in form to the B curves of the iron. It will 

 be noticed that curve IV. of this series lies for the most part higher than curve 

 III., although the steady helical current is smaller in the former; also that I., 

 II., and III. seem to fall together, as belonging to the same set, while IV. and 

 V. form a system by themselves. The reason of this would seem to be that 

 between the dates, June 2nd and 4th, namely, on which these sets were taken, 

 the nickel wire underwent some physical change. Probably this was of the 

 nature of a change in temper, since on the latter date the nickel wire was for 

 1 an instant traversed by a current of sufficient strength to make it glow red hot. 

 Taking this consideration into account, and neglecting curve A, III., which is 

 obviously a bad experiment, we conclude that the twist due to the superposi- 

 tion of circular and longitudinal magnetisations in nickel wire increases with 



* See Nature, vol. xxvi. 1882. 



VOL. XXXII. PART I. 



2K 



