408 C. Barus — Torsional Magnetostriction in Strong 



KirchhofPs theory lias recently been examined in an elab- 

 orate experimental research due to Nagaoka and Honda,* with 

 results showing an imperfect agreement with facts both for 

 iron and for nickel. They not only include known reciprocal 

 relations of stress and magnetization, but investigate new data 

 particularly referring to changes of bulk. They find that 

 •KirchhofPs new constants are complicated functions of strain, 

 or that the expressions are applicable only to infinitely small 

 strains. 



As the phenomena, therefore, remain almost equally trouble- 

 some in terms of the constants of the theory, very little assist- 

 ance has been gained from it; and in the entire absence of a 

 better theory it is still permissible to look at magnetostriction 

 from tentative points of view. 



2. I have long been of the opinion that a statistical treat- 

 ment of the subject, such as was suggested for viscosity by 

 Maxwell, might have much in its favor as compared with 

 the purely elastic treatment independent of the mechanism, 

 sketched above. The reciprocal relations of stress and mag- 

 netization are, as it were, incidents in a much more varied 

 phenomenon. Indeed the original explanation of the Wiede- 

 mann effect, as given by Wiedemann himself in terms of 

 Weber's theory of re voluble molecular magnets (a theory 

 which in Ewing's bands has been shown to include hysteresis) 

 seems to have been too consistently ignored. The conception 

 of a magnetic configuration which breaks down under stress 

 but is restored when the stress has sufficiently vanished or 

 been reversed seems to be a reasonable one, supposing the 

 breakdown involves no chemical change, as it does for instance 

 in tempering. In Maxwell's hypothesis any deformation due 

 to molecular instability is a viscous deformation. Now when 

 the breakdown is gradual in character, as it must be when 

 depending on temporary local intensities in the distribution of 

 heat motion, the deformation will be gradual as actually 

 observed in the ordinary phenomena of viscosity. If however 

 breakdown is instantaneous (kaleidoscopic as it were), due for 

 instance to the molecular shake-up accompanying magnetiza- 

 tion or its withdrawal, then viscosity is instantaneously a 

 minimum, and the deformation correspondingly sudden or 

 " static." There seems to be no theoretical gap here whatever, 

 and I hope to show (more directly in the subsequent paper 

 than in the present, which has an introductory character) that 



* Nagaoka and Honda : Journ. of the College of Science, Tokyo, Japan, vol. 

 ix, part iii, pp. 353-391, 1898. 



