Prof. G.Wiedemann's Magnetic Researches. 61 



The results of these observations for the last torsion and 

 detorsion after attaining a permanent condition, are repre- 

 sented in Plate I. figs. 2-9, by curves whose abscissa represent 

 torsions and ordinates the corresponding magnetic moments. 

 The curves marked I., II., III. represent the results obtained 

 with the first, second, and third torsion and detorsion of the 

 soft-iron bar. 



The curves show that after the wires have been accommodated 

 by repeated deformation, the curves representing the mag- 

 netism are nearly symmetrical for rising and falling torsion, 

 beginning from the minimal and maximal torsions of 0° and 

 210°. With the temporarily longitudinally magnetic iron 

 wires — the moments are nearly equal for the later torsions — 

 they rise similarly in consequence of torsion and detorsion, 

 and then fall again to the limiting-point. The curves, how- 

 ever, rise more rapidly than they fall, so that the maximum of 

 the temporary movements is obtained before the half-torsion 

 or detorsion. 



With the iron wires permanently magnetized longitudinally 

 and temporarily magnetized transversely, the behaviour is 

 similar, only the moments at the two limiting-points of 0° 

 and 210° are different. Starting, however, from these points, 

 the curves run similarly. The moments again change more 

 rapidly at the beginning of torsion or detorsion than at the 

 end. With nickel the moments change in the opposite direc- 

 tion ; but in other respects according to the same rules as 

 with iron. 



We may hence conclude that, within certain limits, the 

 molecules, in consequence of repeated twisting to and fro, 

 become movable ; so that they perform nearly equal motions 

 for equal torsions from the two limiting values. With tem- 

 porary longitudinal magnetization the tendency of the mag- 

 netizing- force is to place the molecules with their magnetic 

 axes longitudinal, in consequence of which a final condition 

 is attained in which they deviate at the limiting-points equally 

 from their most nearly longitudinal position. A priori, we 

 should expect that they would be most nearly longitudinal in 

 the mean position, i. e. at half torsion and detorsion ; and there 

 the wire would show its maximum of longitudinal magnetism. 

 The deviation from this result shows that at the commencement 

 of torsion or detorsion the molecules are more quickly influ- 

 encedby the acting forces. 



It is also to be observed that the permanent torsions into 

 which the bar, left to itself, slowly passes from the temporary 

 torsions 0° and 210° do not at all agree with the torsions at 

 which it shows the maximum of temporary magnetism ; these 



