252 Mr. 8. Bid well on the Magnetic Torsion 



preliminary nature, and I had hoped to be able to carry them 

 further ; but so far as they go they appear to be conclusive, and 

 a short account of them may be of interest at the present time 

 in connection with the recent publication of Prof. Wiedmann. 



In two papers communicated to the Royal Society *, I 

 have pointed out that the effect of magnetization upon the 

 dimensions of an iron rod is not so simple as it had previously 

 been believed to be. Joule enuciated the law that the 

 elongation in a given bar is at first proportional to the 

 magnetic intensity, and that it ceases to increase after the 

 iron is fully " saturated." My own experiments show that 

 if the magnetization is carried beyond the point at which the 

 elongation reaches a maximum, the length of the rod, instead 

 of remaining unchanged, steadily diminishes, until, when the 

 magnetizing current has attained a certain strength, the 

 original length of the rod is unaffected, and if this strength 

 be exceeded actual retraction is produced. From some 

 further experiments, details of which are not yet published, it 

 appears that effects of the same character occur when rings 

 are used instead of straight rods. The diameter of an iron 

 ring surrounded by a coil of wire was found to be increased 

 when a comparatively small current was passed through the 

 coil and diminished when the current was strong. 



I have also carried further Joule's experiments regarding 

 the effects of magnetization upon the length of iron wires under 

 tension and ascertained, among other things, that a wire when 

 stretched by a weight attains its maximum elongation with a 

 smaller external magnetizing force than when it is free, 

 retraction apparently beginning at an earlier stage of magneti- 

 zation . 



Lastly, I have repeated and confirmed the experiments of 

 Barrett f upon nickel. Barrett discovered that the length of 

 a nickel bar when longitudinally magnetized was diminished. 

 So far as appears from his published papers on the subject, 

 which are very brief, he worked only with comparatively 

 strong magnetizing currents, and it seemed to me possible 

 that, as in the case of iron, weaker currents might cause 

 elongation. But on trying the experiment I found that this 

 was not so. Magnetizing forces which were so small as to 

 produce no sensible effect whatever upon iron caused con- 

 siderable diminution in the length of the nickel bar, and the 

 curve of retraction given in my paper (p. 131) clearly passes 

 through the intersection of the axes, and shows quite con- 



* Proc. Roy. Soc. xl. pp. 109, 257. 



t ' Nature,' xxvi. 585, and Brit. Assoc. Rep. 1882. 



