MAGNETIC SHIELDING IN HOLLOW IRON CYLINDERS. 645 



of two distinct sets of readings for the same values of H c , the galvanometer connections 

 being reversed at the beginning of the second set. 



In fig. xv. the values of B, thus obtained are plotted against the values of the 

 superposed cyclic force H c . On this force being increased by steps to a maximum, the 

 B, component at first rises above, but finally falls somewhat below its original value 

 when H c = 0. This curve is approximately the same as the B, curve given in fig. xil. 

 (for a slightly higher value of B,) under the LC conditions, the ordinates being for 

 present purposes expanded five times. H c now carries the B c component round a 

 complete cycle, and it is at once seen that the B, component responds. After a few 

 reversals the curve closes upon itself and becomes cyclic. Fig. xv. thus shows (1) the 

 B t curve when H c is increased from zero to a maximum (LC conditions) ; (2) the process 

 by which the higher values of the B 2 component are reached under the LCC conditions 

 at the extremes of the H c reversals ; and (3) the changes which occur in the B, component 

 of the resultant induction after the process has become cyclic, under the influence of the 

 superposed cyclic force H c . 



§ 23. Obviously the pre-existing induction upon which a force at right angles is 

 superposed may be at any stage of the hysteresis loop. The (rase where the induction is 

 at one of the extremes of this loop, sustained by a magnetising force from zero, has just 

 been given. In figs. xvi. and xvn. the varying cyclic force is superposed upon points 

 of this B-H cycle, where B is changing most rapidly with respect to H, and where 

 consequently it might be expected a superposed force would have a maximum effect 

 upon the pre-existing induction. In the former fig. H; is still the force first acting, 

 the induction being left at a small negative value of B { = -296. In the latter fig. H c is 

 the force first acting, the induction being in this case left at a small positive value of 

 B C =1760. The experimental data for figs. xvi. and xvn. will be found in Tables XII. 

 and X. respectively. At the top of each table the negative values of the force first 

 acting are given in C.G.S. units, and also the positive maximum values from which they 

 were reduced. The various columns have the" same signification as in Table XL with 

 this exception, that in Table XII. both sets of B.G. readings have been tabulated, from 

 which the average readings have been derived. 



Both figs, may be discussed together. As in fig. xv. the ordinates measure the 

 component of the induction in the direction of the force first acting ; the abscissae, the 

 values of the alternating force which is superposed in steps. In both cases, we observe 

 that, in the first -place, the superposition of the cyclic force causes the ordinate values 

 rapidly to decrease, or rather to increase negatively, and to approximate to the values 

 which would have held had the induction due to the force first acting been a negative 

 increment superposed upon zero induction. Compare the B; ordinate value attained at 

 the extremes of the H c cycle (fig. xvi.) with the Bj curve of fig. xi. under the LCC 

 conditions at a value of H c = 12. Also the ordinate value attained at the extremes of 

 the H, reversals (fig. xvn.), viz., B c = -4900, may be transferred to fig. vn. at a 

 position on the horizontal axis of H = -2 C.G.S. units, and compared with the B-H cycle 



