560 Messrs. K. Honda and S. Shi 



mizii on 



Here H, B, and W denote respectively the internal field, 

 the magnetic induction, and the hysteresis-loss, all expressed 

 in c.G.s. units. In Swedish iron, the hysteresis-loss is 

 decreased in weak inductions and increased in the strong, b}' 

 cooling it in liquid air. Fleming and Dewar * found no 

 effect of cooling on the hysteresis-loss of iron. In tungsten- 

 steel, nickel, cast and annealed cobalts, the hysteresis-loss is 

 always increased by cooling. These changes are briefly 

 expressed by saying that the cooling in liquid air magnetically 

 hardens the specimens. 



In comparing the hysteresis-loss of different metals, it is 

 to be observed that the hysteresis-loss in nickel is the smallest 

 among them, and that those for tungsten-steel and cast cobalt 

 are greater by, about three times than the hysteresis-loss of 

 Swedish iron, and the same relation also holds between 

 annealed cobalt and tungsten-steel. 



From the courses of the curves in fig. 1, it is evident that 

 Steinmetz's formula giving the relation between the hysteresis- 

 loss and the maximum induction holds for nickel and an- 

 nealed cobalt up to an induction of 3000 C.G.S. ; it holds 

 for cast cobalt and tungsten -steel up to 8000 C.G.s. ; and 

 lastly for Swedish iron, it fails beyond an induction of 18,000 

 C.G.S. If, however, the specimens are cooled in liquid air, 

 the applicable range of the law for induction is notably 

 extended, as may be seen from fig. 1. 



As regards the residual magnetism, the cooling always 

 increases it in a marked degree. 



(d) Hysteresis-loss in Nickel-Steels. 



The hysteresis of nickel-steels was studied at ordinary 

 temperature, and that of seven of them at liquid air tempe- 

 rature, The results are given in Table IV. and in PL IX. 

 figs. 2 a, b. 



The hysteresis-loss of nickel-steels at ordinary temperature 

 is generally small compared with that of other ferromagnetic 

 metals. The values for reversible alloys are, how r ever, com- 

 parable with those of nickel ; but for irreversible alloys they 

 are all very small. Especially nickel-steel of 28*32 per cent. 

 does not almost enclose any area, giving only 16 ergs for an 

 induction of 1200 c.G.s. If the alloy has a greater value 

 for induction, it will be very useful for the construction of 

 transformers. Thus the magnetic state of the irreversible 

 nickel-steels corresponds to that of ferromagnetic metals at 

 high temperatures. As seen from fig. 2 a, the Steinmetz's 

 formula does not apply, except for very weak inductions. 

 * Fleming and Dewar, loc. cit. 



