322 BELL SYSTEM TECHNICAL JOURNAL 



We do not know how the transformation from electrical energy to 

 thermal energy was effected ; but we do know that so much electrical 

 energy vanishes, and that so much thermal energy appears. A broad 

 hysteresis-loop therefore signifies that there will be much dissipation 

 of energy if the sample is exposed to cyclic magnetizing forces, as 

 usually happens in electric machinery in which magnets play an 

 important role; and the heat developed is not merely a sign of energy 

 gone to waste, it is often detrimental to the material, and a bad 

 contribution to that unforgettable history which the magnet is forever 

 piling up. For these reasons the discovery of a new ferromagnetic 

 material of low hysteresis is always welcome. 



As a rule, narrow hysteresis-loops go with high values of initial 

 permeability, and with initial curves easily divided into three stages, 

 and with early saturation; and wide loops go with initial curves 

 which rise slowly and bend upward slowly and display no sharply- 

 marked second segment and approach very tardily to the saturation 

 value. Magnetic hardening, in the sense which I earlier defined, 

 accentuates hysteresis; and the agencies which bring it about — 

 tempering and mechanical hardening and the admixture of certain 

 elements in small quantities, such as carbon into iron — widen the 

 loops and augment the generation of heat. These effects are fre- 

 quently described by giving measurements of the heat W generated 

 in a single cycle of magnetization in which B is carried back and forth 

 between standard values -f- Bo and — 5o of the induction-measure- 

 ments which in their turn are cited by giving the corresponding 

 values of WlBo^-^, the quantity 77 of the formula of Steinmetz.* The 

 value of this quantity is only .00032 for very pure well-annealed iron, 

 leaps to 0.015 when one per cent of carbon is added, leaps again to 

 0.034 when the so-constituted steel is tempered; while the addition 

 of silicon to iron, the very process which raises Mmax. to values excelled 

 only by permalloys, brings the value of rj down to .00011. The 

 permalloys themselves are still more eminent in this regard, some 

 of them having hysteresis-loops only a sixteenth as great in area as 

 those of pure annealed iron. 



To give the area of a loop is not always sufficient; its shape and 

 orientation are very important for theory and for practice. The 

 agencies which harden a material not only widen its hysteresis-loops, 

 but rotate them clockwise around the origin, as the figures show. 

 This rotation tends to decrease the intercept of each loop upon the 



* I should emphasize that for many materials the "law of Steinmetz" is not 

 accurate, so that strictly one should plot the actual curve of hysteresis-loss-w-^o, 

 instead of making a single measurement and using it to determine r; by the assumption 

 that the "law" is valid. 



