MAGNETISM AND TWIST IN IRON AND NICKEL. 503 



residual twist seems to vary very little in value, so long as it is derived from the cessation 

 of a field higher than the field that corresponds to the maximum twist. This is shown in 

 the accompanying Table, giving the limiting twist ranges for different field cycles and 

 the residual ranges corresponding. In a good symmetrical cycle the residual range is 

 double either of the residual twists measured from the mean position. 



It is possible that with high enough fields nickel might be made to show the same 

 peculiarity ; but up to the highest fields which I have been able to use the residual 

 twist in nickel is always smaller than the limiting twist. 



The cyclic graphs for nickel (figs. 2 and 3) bring out very clearly the differences in 

 the field and current cycles. In one particular they agree, namely, in the fact that the 

 rate of untwisting is always less than the rate of twisting near the limiting range of 

 twist. This, of course, is the usual law of the after-effect, and holds for all the four 

 types of curves depicted. In the details of its form, however, the current-reversal graph 

 differs distinctly from the field-reversal graph. This difference must correspond to some 

 fundamental difference in the characters of the longitudinal and circular magnetic fields. 

 Such a difference is not far to seek, inasmuch as the distribution throughout the wire of 

 the magnetic field due to a current passing along it is known to be quite different from 

 the distribution of the field due to a current flowing in a coil surrounding it. 



It may safely be concluded, then, that the different features presented by the field- 

 reversal and current-reversal graphs are the result of the magnetic after-effect. We 

 have already pointed out that, in the field-reversal cycle, the reversed field needed to 

 reduce the residual twist to zero was almost identical in value with what Dr J. Hopkin- 

 son has called the coercive force. 



11. Effect of Simultaneous Reversal of both Magnetising Forces. — It has 

 generally been stated, or at any rate left to be inferred, that the reversal of both 

 magnetising forces which produce the Wiedemann effect is unaccompanied by any 

 twisting. This is not so, however. A small twist is generally produced. The simplest 

 way of effecting the simultaneous reversal of both is to make them depend on the same 

 current. Thus, let the current be led through the iron or nickel wire and the magnetis- 

 ing coil set in series in the same circuit. Then, if the current is reversed, a twist is 

 produced, although the reversal of the current implies the simultaneous reversal of the 

 field. This is shown in the following table, which gives the result of an experiment 

 made with a nickel wire under the same conditions of tension and length as in Nos. 7 and 

 8 of Table II. The first column gives the line current, the second the longitudinal field, 

 and the third the twist due to the simultaneous reversal of both current and field. 



A comparison of the first combination with experiments Nos. 7 and 8 of Table II. will 

 show that nearly all the fields are, for a current of 176 amperes, above the field corre- 

 sponding to the point of equal twists. For smaller currents this will also be the case ; 

 and we may plausibly assume it to hold for the last and smallest field and current given. 

 In fact, the twist '04 for this last combination is in the same direction as the twist 2"5 for 



