56 Prof. Gr. Wiedemann's Magnetic Researches. 



that there could be no possible doubt of the facts. The expe- 

 riments may even be made with an ordinary magnetic needle 

 in place of the reflecting magnet. The effect is less marked 

 with thinner wires. 



If the torsion acted only like an extension of the wires, it 

 would not be easy to understand why an equatorially mag- 

 netized wire should become longitudinally magnetized by it. 

 If we further assume that the fibres of the wires take up an 

 inclined position, but that the molecules composing them 

 retain their position relative to the axes of the fibres, then in 

 all cases, if the current flows from the movable to the fixed 

 clamp, and the clamp is rotated opposite to the direction of 

 the hands of a watch, the wire would acquire a south pole at 

 the fixed clamp, equally whether it consisted of iron or nickel. 

 The expansion of the fibres might produce under the critical 

 magnetization an increase of polarization in the case of iron 

 and a decrease with nickel, but not a reversal of the polari- 

 zation. 



A simple displacement of the sections, without a consequent 

 rotation of the molecules, would communicate no longitudinal 

 magnetism to the wire at all, and the expansion in length 

 could produce no change. 



Hence under all conditions torsion must produce a rotation 

 of the molecules of iron opposite to that which takes place in 

 nickel. I have already referred these rotations to the two 

 simultaneous phenomena just mentioned, namely, to the rela- 

 tive displacement of the longitudinal fibres, and that of the 

 sections of the twisted wires, whereby contiguous mole- 

 cules suffer rotation in consequence of mutual friction, and in 

 opposite directions in consequence of the two displacements. 

 In iron the friction of the longitudinal fibres is the greatest ; 

 in nickel that of the sections predominates. 



We must evidently employ these various rotations of the 

 molecules, as I have endeavoured to do, to explain reciprocal 

 phenomena. If the molecules of a wire are placed with their 

 axes oblique by means of two currents passing through and 

 round the wire, its fibres and sections must follow their rota- 

 tions, as the molecules follow the displacements of the fibres 

 and sections when torsion takes place, and consequently here 

 also nickel must behave oppositely to iron. The change in the 

 length of the fibres can only exert a secondary quantitative 

 change upon the phenomenon. 



Hence I believe that the phenomena I have observed cannot 

 be taken simply as secondary phenomena to those observed by 

 Mr. Joule and Sir W. Thomson. 



