214 A. Tana^adate on the Thermal Effect due 



curves (fig. 23) show the integral amount of temperature corre- 

 sponding to different currents after given time has elapsed, the 

 ordinate being the reading of the Siemens dynamometer, which 

 is proportional to c 2 . The curves show that the difference of 

 temperature due to Joule's effect is sufficiently expressed by 

 the form ac 2 , where a is constant. From these curves the 

 correction was found for each experiment, and uc 2 is drawn 

 below the time-ordinate in each curve by dashes (figs. 1-19) ; 

 and the true rate of the generation of heat arising from the 

 reversals of magnetization is taken to be the algebraic differ- 

 ence of the ordinate of the tangent at origin, and this line. 

 It is seen that this term «c 2 is so small, when «£) is less than 10, 

 that the line in the diagram is scarcely distinguishable from 

 the time axis. 



When the field *£) was increased above 30 C.G.S. units, the 

 heat due to the current alone was so great that a large part of the 

 heat due to magnetic reversals was neutralized by it. Special 

 experiments were made in which a continuous current was 

 sent through both the rings for more than half-an-hour, when 

 the time-variation of temperature became very small, and 

 then the current was alternated. Fig. 20 shows three experi- 

 ments of this kind. It is seen that in all the cases the effect 

 of the continuous current was to heat the wooden ring ; and 

 when the current was alternated, the temperature of the iron 

 ring rose suddenly almost in straight line. It must be re- 

 marked here that when the current was made to alternate its 

 strength diminished, partly owing to induction and partly 

 to the increase of contact resistance. In the curve corre- 

 sponding to «jp= 32*8 the heating-effect due to current was made 

 to remain the same by suddenly taking off some resistance in 

 the circuit at the instant of commencing the alternations. In 

 other two this was not done: the difference can be seen from 

 the values of <£) given alongside the curves. 



The rate of heating thus obtained was divided by the 

 number of complete reversals per second to find the heat due 

 to one complete reversal of magnetization. The temperatures 

 so found are plotted against <£) in fig. 27, the number near the 

 points gives the rate of alternations per second. 



Ballistic Experiment. 

 In order to compare the heating-effect thus far obtained 

 with the hysteresis in the same ring, the amount of magneti- 

 zation, was now investigated ballistically. For this purpose a 

 secondary coil was wound uniformly over each of the rings, 

 and the current induced by changes of magnetization in the 

 iron ring due to small sudden changes of magnetizing field 

 was measured by the throw of a ballistic mirror-galvanometer 

 in the circuit. The secondary coil over the wooden ring was 



