646 MR JAMES RUSSELL ON 



for shield B. Both those comparisons show that the effects of hysteresis due to the 

 force first acting have practically disappeared. 



In both figs. (xvi. and xvn.), we observe, in the second place, that the ordinate 

 changes measuring the induction component in the direction of the force first acting 

 finally become cyclic (as in fig. xv.), the curves returning upon themselves after a few 

 reversals of the superposed force. 



§ 24. But these figs. (xvr. and xvn.) also show points of difference. The ordinate 

 changes in the latter fio-. cover a much wider range than the ordinate change in the 

 former. This is accounted for by the fact, that in fig. xvn. the varying force is super- 

 posed upon an induction due to H c , which is the true magnetising force in the iron. 

 The difference between the particular point selected on the descending limb of the 

 B c -H c cycle, and the maximum negative values of B c reached at the extremes of the 

 final cyclic changes, when a varying force is superposed, is of the order of B c = 6500. 

 When, however, the longitudinal force is that first acting upon the iron, the hysteresis 

 loop is sheared over, and the ordinate distance between its descending limbs and the 

 maximum negative values of B z reached at the extremes of the superposed B c cycles 

 must be relatively small. This is exemplified in fig. xvi. 



The same differences between the maximum and minimum values in the direction 

 of the force first acting also appear when the ordinate changes have become cyclic. In 

 fig. xvil, with the force first acting =2 C.G.S. units, the ordinate change is B c = 1330, 

 while in fig. xvi., with the force first acting = 2*8 C.G.S. units, the ordinate change 

 only amounts to B^ = 230. In fig. xv. the cyclic change of B ? = 950, but in this 

 instance the value of the force first acting is larger, viz. : — Hj = 20'4 C.G.S. units. On 

 the other hand, the points of inflection of the final cyclic curves are much more sharply 

 marked in figs. xv. and xvi. than they are in fig. xvil, where the numerical values 

 of the superposed varying force H z due to the solenoid are greatly in excess of the true 

 magnetising force acting in the iron. The points of inflection are therefore flattened. 

 In figs. xv. and xvi. the superposed alternating force is H c (which when acting alone 

 is the true magnetising force in the iron). The maximum values of clBJdH,. are there- 

 fore rapidly passed in reference to the numerical value on the horizontal axis. The 

 differences discussed in this section are therefore uon-essential, and due to both magnetic 

 circuits not being wholly completed in the iron. They are differences of degree only. 



§ 25. Consequently the conclusions now to be drawn may be expressed as before 

 (§ 19) in terms of the order of field superposition. Of the two magnetising forces at 

 right angles to each other, let H^ be the force first acting, H 2 the force superposed, and 

 let B x and B 2 be the two components of the resultant induction in the direction of H! 

 and H 2 respectively. We have seen that when the superposed magnetising force II 2 , 

 first increased from zero and then repeatedly reversed, carries the B 2 induction 

 component in steps round a series of cycles, the B x component due to Hj kept at a 

 constant value, is found to respond by passing through a series of changes or 

 oscillations in the process of likewise becoming cyclic. 



