.112 Prof. T. R. Lyle on the Variations of 



determined by the wave-tracer. The difference I— (U + E), 

 which Fleming has called the kinetic hysteresis, was deter- 

 mined for each experiment, and is given in the tables that are 

 to follow; and its values for different induction densities and 

 frequencies for both rings are plotted in fig. 3 (p. Ill) against 

 B the maximum induction in each case. The results expressed 

 by these curves seem to thoroughly verify the observations of 

 Steinmetz and of Siemens both as to the existence of such a 

 quantity as kinetic hysteresis and as to the general character 

 of its variation with induction density and period. As this 

 question has been much disputed the verification, by a new 

 method, of the results obtained by previous investigators is 

 not without significance. 



11. In Table I. (p. 113) are given the analytical results 

 deduced as indicated in the preceding paragraphs from a series 

 of experiments with Ring I., in which the period was approxi- 

 mately *019 sec. The wave-forms of the magnetizing currents 

 in experiments 1 to 13 were approximately sinusoidal, and 

 the E.M.F. impressed on the circuit was practically of the 

 sine form in all experiments with the exception of No. 15, 

 Table I. 



In fig. 4 (p. 114) the more important characteristics of the 

 induction-waves given in Table I. Nos. 1 to 13, i. e. produced 

 by q.p. sine currents, are plotted against the amplitudes B x 

 of the first harmonic of these waves. These curves are 

 typical of any series of induction-waves of constant period 

 produced by currents of similar wave-forms. No such 

 regularity, however, would be obtained if the wave-forms of 

 the magnetizing currents were allowed to vary, as will be 

 seen by marking on fig. 4 the points for the characteristics 

 of the induction- wave of Experiment 15, in which the H 

 wave was greatly distorted (made saddle- shaped) by artificially 

 distorting the applied E.M.F. wave. 



The characteristics /^ and 0, which are the connecting 

 links between the H wave and the B wave it produces, both 

 fall to small values for small values of B 1? 6 probably 

 vanishing with B^ Fig. 4 shows clearly how both rise to 

 maxima and then diminish as B : increases. 



The curve for b 3 the ratio of the third (B 3 ) to the first 

 harmonic Bj of B is striking, apparently issuing from the 

 origin (see series 2, Table II. p. 115, & fig. 5, p. 118), it rises 

 quickly between Bi = and B 1 =1000 (q.p.), from which it 

 continues for larger values of Bj as a straight line. Hence 

 for all values of B x greater than 1000 the amplitude B 3 of 

 the third harmonic of B is of the form 



aBi-f/SBr. 



