under the Operation of Feeble Magnetic Forces. 239 



reversal of current 50 is thus 8 divisions, and being propor- 

 tional to the current can be deduced for any other case. At 

 the bottom of the table, where the current is 1493, the dis- 

 placement rendered latent by the coil is thus about 240 divi- 

 sions ; and since at this point the uncompensated displacement 

 is nearly of the same amount, we see that the value of k (as 

 above defined) is here doubled. Thus, if «£) denote the mag- 

 netizing force in C.Gr.S. measure, we have 



*=6-4(l+-8£). 



The form of the relations of k to *£) for small forces is pretty 

 accurately demonstrated by the observations. On the other 

 hand, the reduction to absolute measure is rather rough* — a 

 point of less consequence, inasmuch as the constants may be 

 expected to vary according to the sample and condition of the 

 iron. 



The observations in Table I. give a good deal more than the 

 extreme range of magnetization due to the reversal of a force. 

 In all cases the two residual magnetizations (when the force 

 is zero) are recorded ; while in the two latter, where the range 

 is greatest, further intermediate points are included. The 

 results are plotted in Plate II. fig. 2, where it will be seen 

 that the curves start backwards in a horizontal direction after 

 a maximum or minimum of force. Special observations (not 

 recorded in the table) were directed to this point. Neither at 

 the maxima nor at the zeros of force was there any evidence 

 of failure of compensation when a small backward movement 

 was made. 



The curves do not differ much from parabolas ; and in 

 other cases, where the applied magnetic forces were all of 

 one sign, I have found that after a large movement in one 

 direction, the curve representing a backward movement 

 coincides somewhat closely with a parabola whose magnitude 

 is nearly the same under different circumstances, and which 

 is placed so that its axis is vertical and vertex coincident with 

 the point where the backward movement commences. The 

 reader will not forget that to obtain the real curves fully ex- 

 pressing the relation between magnetization and force, we must 

 add the effect, proportional to the force, rendered latent by the 

 compensating-coil. 



On the basis of this parabolic law we may calculate the 

 influence of hysteresis in the magnetization of iron upon the 

 apparent self-induction and resistance of the magnetizing- 

 coil, when periodic currents of moderate power are allowed 



* In all probability the number 6*4, applicable wben § = 0, is too 

 small. 



