1890.] upon Magnetic Changes of Length in Wires. 479 



Professor J. J. Thomson has shown* that the Villari effect is dyna- 

 mically connected with the changes of length undergone by an iron rod 

 when magnetised, an iron rod being lengthened in a weak magnetic 

 field and shortened in a strong one. Now cobalt behaves oppositely 

 to iron in this respect, a rod of cobalt becoming shorter in a weak 

 field, longer in a strong one.f Professor Thomson^ therefore, pre- 

 dicted that a Villari reversal would be found to occur in cobalt and 

 that it would be of the opposite character to that in iron. Some 

 experiments made by Mr. ChreeJ with a cobalt rod under pressure 

 gave results in accordance with Professor Thomson's expectation. 



But no reversal of the effect of magnetisation in diminishing the 

 length of an unstretched nickel rod has ever been- observed. Such 

 a rod always becomes shorter in a magnetic field, whether strong or 

 weak. It appears to attain its shortest length in a field of about 

 750 units, but it does not pass a minimum » and become longer again 

 in stronger fields, like cobalt. Applying. Professor Thomson's 

 reasoning, therefore, to the case of nickel, we should expect that it 

 would be found not to exhibit any Villari reversal, either of the 

 nature of that in iron, or of that in cobalt. Both Sir William 

 Thomson and Professor Ewing have in fact looked for one and failed 

 to find it. But in a paper read at the meeting of the Physical 

 Society, on 21st March, 1890,§ Mr. Herbert Tomlinson gave an 

 account of some experiments which, as he believed," showed that a 

 Villari critical point really existed in nickel, though it was only to be 

 reached by the application of comparatively great magnetising forces. 

 If Mr. Tomlinson is right, 1 venture to suggest that his results may 

 possibly be brought into harmony with Professor J. J. Thomson's 

 mathematics by consideration of the experiments described in the 

 present paper : for, as I understand Professor Thomson's argument, 

 he has hitherto taken no account of the effects of mechanical stress 

 upon magnetic changes of length. 



Cobalt. — The results for cobalt show that the changes of length 

 which this metal undergoes when magnetised are almost, if not en- 

 tirely, unaffected by tensional stress^ at least within the limits of the 

 experiments. Having regard to the verv marked influence of tension 

 upon iron and nickel, this cannot but be regarded as a most re- 

 markable fact. 



(depending upon the magnitude of the stretching force) for which stretching pro- 

 duces no effect. 



* ' Applications of Dynamics to Physics and Chemistry,' p. 54. 



f < Phil. Trans.,' A, 1888, p. 227. 



X Supra, p. 41. 



§ 2fot yet published. 



