121 Variations of Magnetic Hysteresis with Frequency. 



/* = 2734. 8=11840. Max.B = 10900. 

 I i = 9465. I 3 =-654. l-,= -32. 



1 = 8780. IIW'°i = -00353. 



Though the applied E.M.F/s in these experiments were 

 only approximately sinusoidal, they illustrate the main 

 features of the phenomenon being discussed. Thus we have 

 with diminution of the impedance of the circuit while B] 

 is kept approximately constant, the increase in the reflected 

 upper harmonics of C and the diminution of the corre- 

 sponding harmonics of F, the increase of fi and of I and the 

 increase and subsequent decrease of the reflected energies I 3 

 and I 5 . 



Attention is drawn to the great variation in the ratio of 

 I to 2? 1 ' 57 with change of wave-form of H, and therefore of 

 B, thus showing that the formula given in § 12 for I can 

 only apply to a series in which the wave-form of H is nearly 

 sinusoidal. 



In order to fully investigate the phenomenon drawn 

 attention to in this paragraph, it would be necessary to de- 

 termine the impressed E.M.F. wave as well as that of H and 

 of B. This is being done, and will form the subject of a 

 future communication. 



Finall}*, the matter dealt with in this section seems to me 

 to have an important practical significance in the case of 

 transformers. In calculating their losses all that has hitherto 

 been counted as iron loss is the energy that is dissipated as 

 heat in the iron. But now we see that it is possible for the 

 iron to send back a considerable amount of energy to the 

 primary circuit, where it also is dissipated as heat. This 

 reflected energy should be counted against the iron. 



Thus Experiment 2 above may, for the sake of illustration, 

 be supposed to refer to a transformer on open secondary. 

 The true dissipation of energy associated with its operation 

 per cm. 3 of iron per cycle is Ij or 961)2 ergs, of which 

 1 = 8820 ergs, hitherto called the total iron loss, is dissipatea 

 as heat in the iron, and — (I 3 -fI-,) = 872 ergs is reflected by 

 the iron to the primary circuit, where it also is dissipated as 

 heat. 



I desire to thank Mr. K. S. Cross and Mr. E. Machin for 

 valuable assistance in carrying out the experiments discussed 

 in this paper. 



