Transverse Fields and Allied Phenomena. 



409 



with the interpretation I have ventured to give of Maxwell's 

 conception of viscosity, the phenomena of magnetostriction 

 become more closely akin, following the same course in all 

 metals. 



In discussing experiments like the present it is always desir- 

 able to devise a model of the simplest kind, which shall sug- 

 gestively reproduce the phenomena in question as fully as 

 possible. The model need not be looked upon as an actual 

 occurrence. I will therefore insert a simple torsional mechan- 

 ism around which the present experiments may be conveniently 

 grouped. If we regard the rod to be twisted as a bundle of 

 longitudinal fibers elastically bound together, then the effect of 

 twisting is in the main an elongation of the fibers increasing 

 from the axis to the circumference. Inasmuch as the external 

 fibers are now helical in form, the stretch in question has a 

 horizontal component acting along the circumference of any 

 section, tending to restore the fiber to its original straight 

 form. In a solid rod fibers inclined at 45° to the axis are sub- 

 ject to traction (with the accompanying compression at right 

 angles thereto) only. Other fibers are both stressed and rotated. 

 For the present purposes, however, a description in terms of 

 the principle stresses is unnecessary. 



CD CD 6 G G 



© G G G 



CD G G 



G G G 



G G G 



G G G 



Fig.i 



The case may be illustrated by the annexed model, in which 

 the two parallel discs A and B are kept apart by the rod C, 

 pivoted in a depression in each. Rubber fibers, a, b, c, . . . . 

 symmetrically grouped, hold the two discs together and secure 

 the rod (J in place. If torque be applied to B relative to A 

 the rotation is forcibly resisted. 



The bands a, o, c, . . . correspond to lines of stress which 

 bind the non-magnetic molecules, longitudinally. In the sim- 

 plest case one would assume that in a direction along the lines 

 of stress molecules are closer together than are the successive 

 lines themselves in a direction at right angles thereto ; in other 

 words, the distance apart of lines of stress is large as compared 

 with the distance apart of the molecules which conduct stress. 



