62 Prof. G. Wiedemann's Magnetic Researches. 



last lie further away from the extreme torsions than the per- 

 manent torsions do. 



If we consider the changes in temporary magnetism in 

 the first torsion, we observe in every case an increase and 

 then a decrease, evidently because the molecular magnets, 

 directed more or less axially by the magnetizing-force, obey 

 at first that force still further in consequence of vibration, 

 and are then carried from their axial position into more 

 oblique positions. This is confirmed by the fact that with 

 very strong temporary magnetism the first rise through 

 vibration is less than with less powerful magnetization, since 

 in the first case the molecules are already more nearly in the 

 axial position. 



With permanent longitudinal magnetization the directing 

 force of the external magnetizing-power is absent ; the mole- 

 cules, disregarding mutual action, are affected only by the 

 displacements produced by torsion, which consequently mani- 

 fest themselves more distinctly than in the previous experi- 

 ments. The moment at the limiting-points is different, and 

 thus also the position of the molecules. But, again, the most 

 rapid rotation of the molecules at the beginning of each defor- 

 mation is seen,, within the given limits. 



A similar behaviour is observed in the case of temporary 

 transverse magnetization, in which the magnetic directing- 

 force which acts upon the molecules is one-sided; and so 

 the moments at the limiting-points cannot be equal. Never- 

 theless, the rotations of the molecules with rising and falling 

 torsion here also preserve the above-mentioned character. 



§5. 



In exact analogy with the mechanical and magnetic beha- 

 viour described in § 4, it may be shown that if a wire which 

 has been subjected to a temporary deformation (as a special 

 case, this deformation may be zero), then suffers another defor- 

 mation, the molecules again retain the position corresponding 

 to the first deformation up to a certain degree. 



In order to make these experiments, a wire was temporarily 

 twisted by forces 0, a, b, gradually increasing, and its tempo- 

 rary torsion measured, as also with decreasing forces b, a, 0. 

 For this purpose the same apparatus was used, as in my former 

 experiments. A flat weight 7 y x provided with a hook was 

 attached to the thread q which acts upon the torsion-circle ; to 

 this a second weight /3 /?, with a hole through the centre, was 

 attached by lateral threads a (3 7, u x ft 7^ These threads were 

 connected to a horizontal wire 8 e, which was soldered to the 

 frame of a pulley rj suspended by the thread f 6. The end 6 



