94 K. HONDA. 



We tluis obtain incredibly large values for the change of volume. 

 In nickel and soft iron, there is at first decrease of volume, 

 and then follows an increase; in wolfram steel, the diminution 

 of volume reaches a maximum and then gradually decreases. 

 The above result for soft iron agrees feirly well with that of 

 BidwelF* for un annealed iron ring. But in the experiment with 

 ovoids"^ made of the same specimens, there was always small in- 

 crease of volume for nickel, steel and soft iron. The amount of the 

 change at the field of 100 C. G. S. units was 0.7 x lO"", 3.1 x 10"^ 

 and 2.8 X 10"' for these metals respectively. Hence the question 

 now arises whether the change of volume is so influenced by the 

 shape of these metals. To settle this point, fresh experiments on 

 the change of volume were undertaken with a dilatometer. The 

 answer was in the negative, the result being in rough agree- 

 ment with that for the ovoids. The initial decrease of volume 

 was never observed, but the volume always increased with the 

 increase of the magnetizing field. The discrepancy between the 

 calculated and the experimental result is perhaps due to the 

 œolotropy of the materials. For, if it were not isotropic, the 

 lateral dilatation by longitudinal magnetization would not coin- 

 cide with the change of length by circular magnetization. It will 

 also be explained by the a3olotropy of the specimens that in weak 

 fields, Bidwell's calculation resulted in the large diminution of 

 volume of iron rings in contradiction to the experimentally 

 established fact. 



1) loc. cit. 



2) Nagaoka and Honda, loc. cit. 



