62 H. NAGAOKA AND K. HONDA. 



at first takes place very slowly, but on reaching saturation, the 

 rate of decrease becomes very rapid. So far as the present ex- 

 periment goes, the rate does not diminish except in annealed 

 wolfram steel, in which we notice a slight flattening. 



7. Nickel (Fig. 1). — The behaviour of annealed nickel ovoid 

 as regards the length change is nearly the same as that already 

 observed by one of us. With an increasing field, the contraction 

 reaches an asymptotic value, which in the present case is greater 

 than that obtained by Bidwell from experiments on a nickel wire. 

 The explanation of this discrepancy is to be sought for partly in 

 the difference of shape, and partly in the difterence of treatment, 

 as will be clearly illustrated by experiments on the change of 

 volume. We have also reason to believe that repeated annealing 

 alters the elastic behaviour of ferromagnetics as regards the 

 strain wrought by magnetization. Plotting the curve of length 

 change against the intensity of magnetization, we find a slight 

 bend when the magnetization becomes saturated and the con- 

 traction approaches its asymptotic value. 



Change of Volume. 



8. Experiments by several physicists prove that magnetiza- 

 tion produces change of volume in ferromagnetics, in contradiction 

 to the popular belief which is based on Joule's experiment. The 

 alteration of volume accompanying the magnetization of ferromag- 

 netics is generally very small in weak fields, but as will be seen 

 from the present experiment, the phenomenon becomes more marked 

 as the field is made stronger. As we have already remarked, the 

 change of volume as measured by Cantone^' in an iron ovoid must 



1) Cantone, Mem. delîa K. Accad. dei Lincei 6, 487, 1891. 



