MAGNETISM AND TWIST IN IRON iND NICKEL. 487 



same peculiarity ; but lie does not seem to have tried to co-ordinate these observations, 

 and has made no attempt to follow up the inquiry. 



In what follows, the recorded twists are produced by reversal of one of the currents, 

 the other being kept steady. A very few reversals of the one current suffices to bring the 

 wire into a cyclic condition as regards the twist produced. After the twist, correspond- 

 ing (say) to steady line current and reversed coil current, has been observed, the coil 

 current, or longitudinal field, is in its turn kept steady, and the twist taken that 

 corresponds to the reversal of the line current. In general these two twists are different. 

 We shall distinguish them hereafter as the field-reversal twist and the current-reversal 

 twist, the unqualified word " current " being always used in the sense of the current 

 along the wire. The current that produces the longitudinal magnetic field will be 

 distinguished as the coil current. 



If no magnetic asymmetry exists in the wire, the magnitude of the field-reversal 

 twist should be the same whatever be the direction of the steady current, and the 

 magnitude of the current-reversal twist should be the same whatever be the direction of 

 the steady field. This was always found to be the case when the current (whether coil 

 current or line current) that was being reversed was tolerably strong. But when it was 

 weak, the magnitude of the twist sometimes depended on the direction of the steady 

 current. When this happened, the mean of the twists obtained for the two directions of 

 steady current or field was taken as the value of the twist for the particular combination. 



4. Numerical Results for Iron. — The iron wire used in the first series of experiments 

 had a diameter of 0*8 mm. It was threaded through the coils in the manner described 

 in section 2, and the twists under various combinations of tension, current, and field 

 were measured. These are tabulated in Table I. The first column gives the tension in 

 kilogramme-weight per square centimetre of section ; the second, the current (C) along 

 the wire in amperes ; and the third, the longitudinal field (H) acting on the wire in 

 C.G.S. electromagnetic units. The fourth and fifth columns contain the resulting twists 

 for field reversal and current reversal respectively. The twists are given in thousandths 

 of radians, and are the total twists on the wire. To reduce to twists per unit centimetre, 

 we should divide by half the length of the wire, if that half length were wholly included 

 in the magnetising coil. But from the form of the arrangement it is clear that the half 

 length is not subject along its whole extent to the same field. If, however, we divide the 

 total twist by 35, we shall get a result that cannot differ by more than 10 per cent, from 

 the twist per centimetre in the heart of the coil. The object of the present inquiry not 

 being the accurate determination of the amounts of twist corresponding to a given 

 combination of currents, it is not essential to make the final reduction to twists per unit 

 length. 



