o08 K. HONDA : ON THE MUTUAL INFLUENCE 



Thus when / is less than tlie critical field, the curve is at 

 first convex towards the axis of t ; but it soon becomes concave 

 towards the same axis. When t is further increased, it again 

 becomes convex. When / is greater than the critical field, the 

 curve is at first concave and then becomes convex towards the 

 axis of t in strong fields. 



In the above considerations, we have supposed our ferro- 

 magnetics to be magnotically isotropic and to have no residual 

 magnetism ; but these suppositions are not admissible for iron 

 or nickel tube. Moreover we have not considered the variation 

 of the effective part of the longitudinal field caused by the change 

 of demagnetizing force with increasing transverse field, fîence 

 we can not, strictly speaking, compare the above results with 

 those obtained by experiments. But qualitatively all these points 

 just mentioned are verified in the present experiment, as shown 

 by Figs. 1, 3, 6 & 10. 



In order to get a clearer insight of the phenomenon, 1 com- 

 pared the result calculated from the curve of longitudinal or 

 circular magnetization by the above considerations with that ob- 

 tained by experiments. The comparison was made only in strong 

 longitudinal fields, in which the suppositions introduced in the 

 theoretical consideration were comparatively small. In general, 

 the change of magnetization in weak transverse fields fairly agrees 

 with theory, but in the strong, the deviation becomes more 

 pronounced. By taking the theoretical results calculated from 

 the circular magnetization instead of the longitudinal, the 

 coincidence between theory and experiment becomes closer, 

 especially in strong fields. These results are given in the follow- 

 ing table : 



