254 On the Magnetic Torsion of Iron and Nickel Wires. 



The behaviour of nickel then, though in agreement with 

 Maxwell's hypothesis, is not accepted by Wiedemann as 

 affording confirmation of it. 



But we have further means of testing the hypothesis. It 

 has been said that an iron rod when very strongly magnetized 

 is contracted instead of being elongated. Moreover, when the 

 rod is stretched by a weight contraction occurs with smaller 

 magnetizing forces than when it is unstretched. In accordance 

 with Maxwell's explanation, therefore, an iron wire spirally 

 magnetized by very strong currents should twist (just as if 

 it were a nickel wire) in the opposite direction to that indi- 

 cated by Wiedemann for iron ; and this reversal of the twist 

 should take place at an earlier stage of the magnetization if 

 the wire were stretched. Now this is exactly what I have 

 found to be the case. The free end of an iron wire, through 

 which a constant current is passing, can be made to twist in 

 either direction by varying the current through the surrounding 

 helix, while with a certain medium strength of current there 

 is no movement at all. And again, when the wire is loaded 

 with a weight, the current which produces no torsion is con- 

 siderably diminished, the reversal of the torsion also, of course, 

 occurring with smaller currents. 



It is not easy to see how Wiedemann's theory could fairly 

 be made to explain these phenomena. He would be compelled 

 to assume that when the iron is more intensely magnetized, 

 that is, when the molecules are more completely turned in the 

 same direction, the excess of longitudinal over transverse mo- 

 lecular friction, instead of becoming more marked, as might 

 naturally be expected, would be decreased, and, with a suffi- 

 cient degree of magnetization, even converted into a defi- 

 ciency. It would also be necessary to assume that when the 

 molecules of a wire are drawn apart by stretching the friction 

 between their ends is nevertheless greater. Such assumptions 

 would, I think, be highly unscientific. 



On the other hand, Maxwell's explanation, which does not 

 seek to go behind the magnetic elongations and retractions 

 detected by Joule and others, fits the newly observed facts 

 easily and naturally. 



Note on the Experiments. 



It seems desirable to give a few details of the experiments above 

 referred to, though too much importance must not be attached to 

 the quantitative results : — 



The iron wire used was *7 mm. in diameter and 20 cm. long. It 

 was suspended in an upright position from a fixed clamp, and 

 passed through a helix consisting of 876 turns of copper wire in 



