MAGNETIC SHIELDING IN HOLLOW IRON CYLINDERS. 647 



The extent of the changes which occur in the ordinate values of B x during the 

 transition process depends upon the particular point which has been selected on I^-Hi 

 hysteresis loop upon which to superpose the cyclic alternations of H 2 and also upon the 

 maximum value of these alternations. This ordinate change of B x may vary from zero 

 to a maximum. The conditions which prevail when the change is zero have already 

 been stated in § 19, and this occurs at the crossing of the curves there described. Fig. 

 xv. illustrates such a position, the final ordinate values of B x at the extremes of the 

 cyclic alternations of H 2 being almost the same as when H 2 is zero. If H^ had been 

 taken weaker, the final ordinate values would have shown an increase, if stronger 

 a decrease (see figs. xn. and xiii.). The ordinate changes are a maximum (see fig. 

 xvii. ) when B 2 has in the first instance been decreased from a maximum, or in other 

 words when a suitable intermediate point on the hysteresis loop has been selected upon 

 which to superpose the reversals of H 2 . The effect of hysteresis due to the first 

 magnetising process are in this instance almost completely obliterated, and the final 

 values reached by B x at the extremes of the H 2 reversals approximate to the values 

 they would have assumed had repeated reversals of H 2 been superposed upon Bj 

 (increased from zero induction) and supported by H a , the force first acting. 



§ 26. But the ordinate changes of Bj also depend upon the maximum values of the 

 superposed alternating force. Various maximum values limiting the amplitude of H 2 

 were not tried, but it appears from a comparison of the CLL, B c curves of figs. xi. and 

 xii. that if the maximum values of the superposed reversals had been reduced, the 

 ordinate changes would likewise have been less. In other words, the disappearance of 

 the effects of hysteresis due to the force first acting would have been less complete. 



Indeed, Mr Zeno Crook* has shown that an alternating current of 10 amperes per 

 sq. mm. in an iron rod reduces the area of the hysteresis loop due to a cyclic variation 

 of the longitudinal force to a greater extent than an alternating current of 2*5 amperes 

 per sq. mm. The experiments of Messrs Gerosa and Finzi t already referred to are well 

 known. They have shown the almost total collapse of the usual hysteresis loop in an 

 iron wire under the above-mentioned conditions. The effect is ascribed to the violent 

 agitation of the molecules brought about by the rapid reversals of the circular 

 magnetisation, due to the alternating current in the wire, and which corresponds in the 

 experiments described above to the alternations (in steps) of the superposed force H 2 . 



§27. The experiments carry the matter a step further. Figs, xv., xvi. and xvn. 

 show not only the total ordinate change, but the rate of change under the various 

 conditions, as the superposed force H 2 is being repeatedly reversed. This rate of change 

 of the Bi induction component is seen to be least at the extreme values of the superposed 

 reversals, and greatest when the molecules are left freer to obey the constant force H l5 as 

 H 2 approximates to zero. The effect of the first superposition of H 2 in comparison with 

 subsequent reversals in producing change of B 1 is very marked. As has been remarked 



* American Journal of Science, vol. clxi. p. 365, 1901. 

 + " Magnetic Induction in Iron," Ewing, 3rd ed., p. 332, 1900. 

 TRANS. ROY. SOC. EDIN., VOL. XL. PART III. (NO. 26). 5 c 



