the Magnetic Hysteresis of Cobalt. 273 



The standardization of the galvanometer was effected in 

 the usual manner with a standard secondary coil placed in 

 the interior of a long primary solenoid of known dimensions. 

 From the observations the magnetizing force (H) employed 

 and the corresponding flux-density or induction (B) were 

 calculated and the results set out in the accompanying tables. 

 The first set of experiments consisted in taking a careful set 

 of complete magnetic (B.H.) cycles (see fig. 1, Plate VII.), 

 and from these the hysteresis losses in ergs per cub. centim. 

 per cycle corresponding to known maximum flux densities 

 were calculated by measuring the delineated areas. A 

 Steinmetz curve was then constructed showing the relation 

 of the hysteresis loss in each cycle to the maximum induction 

 for that cycle (see fig. 2, Plate VIII.). This was converted 

 by taking logarithms (ordinary) of both variables and plotting 

 a logarithm curve (see fig. 3). This last curve proved to be 

 nearly a straight line. 



If W = magnetic hysteresis energy loss in ergs per cub. 

 centim. per cycle, and 

 B = maximum flux-density (induction) during the cycle, 

 I =maximum magnetization during the cycle; 



then the logarithmic curves show that over a wide range 



W = O01 B 16 , or 

 W = 0-527 I 162 . 



These exponential expressions W = ?7B M and W = 7? / l" for W 

 in terms of B and 1 show that the relation found by Mr. C. P. 

 Steinmetz * for iron and steels of various composition, and by 

 Dr. A. E. Kennelly | for nickel, approximately holds good 

 for this sample of cobalt, viz. that the hysteresis loss varies as 

 the l'6th power of the maximum flux-density during the cycle. 



It is not a little curious that for materials differing so much 

 as the above cast cobalt and soft annealed transformer iron 

 the hysteretic exponent should in both cases be so near to 1"6, 

 x\t low inductions the hysteretic exponent increases in value, 

 however, in all cases. In the case of iron it is well known 

 that magnetic hysteresis becomes zero before the maximum 

 induction becomes zero. There is, in fact, a non-hysferetic 

 range of cyclical magnetization. Hence it follows that for some 

 small value of the maximum flux^density there must be a range 

 of flux-density during which hysteresis increases from zero to 

 a finite value; in other words, its rate of change is then very 

 large. Hence log 10 W will have a very large negative value 



* ' Electrician,' vol. xxviii. pp. 384, 408, 425 ; also vol. xxxii. p. 672. 

 f ' Electrician/ vol. xxviii. p. 6e6 (1892). 



Phil. Mag. 8. 5. Vol. 48. No. 292. Sept. 1899. X 



