664 MR JAMES RUSSELL ON 



nearly the latter in value. It is of special interest to note that as H t is increased the 

 maximum of each curve occurs nearer the vertical axis, thus approximating more 

 nearly to the CTT curves, which have their maxima in the vertical axis. This 

 effect is marked in the case of both shields. The superposition of the unidirectional 

 transverse field is evidently the first step in the vibrational process (§ 46), which only 

 becomes final after repeated reversals of H« — CTT conditions. The stronger the first 

 molecular tap (increased values of H f ) the nearer is the approach to the final conditions 

 for the same values of the transverse field. 



We may therefore conclude that the results arrived at in the immediately pre- 

 ceding paragraph in reference to the CTT conditions are to a certain extent applicable 

 here, but that the effects usually associated with hysteresis, and with molecular 

 vibration, are not so marked, where the transverse field is merely superposed, as to wipe 

 out all apparent connection between the shielding ratios and the value of dB/dH due to 

 the circular field. 



Conclusions under TCC conditions. 



§ 51. The TCC curves have not been plotted for the various values of H t . 

 Reference to Tables XV. and XVI. shows that when repeated reversals of the circular 

 field are superposed upon a pre-existing induction due to the transverse field, the 

 shielding ratio curves (TCC) given in figs. xx. and xxi., when H ( =20'9, are also 

 typical for higher values of H { . When it was found that repeated reversals of H t increased 

 the circular magnetisation and diminished the shielding ratio, it was obvious that 

 repeated reversals of H c would probably increase the permeability of the shield to the 

 magnetising force which remained constant, i.e., the transverse field, and that the 

 shielding ratio would also increase. This deduction was confirmed by the above 

 experimental results. In figs. xx. and xxi. the shielding ratio curves are, at the early 

 stages of induction and when the values of dB c /dH c are a maximum, seen to be greatly 

 increased above their normal and, as we have seen, their theoretical values under the TC 

 conditions (§ 48). Repeated reversals of H c therefore increase the permeability of the 

 iron to the constant transverse force, and concurrently with this, the wholly circular 

 induction was found to be slightly less that under the TC conditions (see figs. xxxn. 

 and xxxiil). On H c being further increased the shielding ratio curves rapidly descend 

 and coincide with the curves under the TC conditions, until they reach their minimum 

 asymptotic values. There appears to be a slight tendency for these curves to cross, as 

 is seen in fig. xx. , but this difference is within the limits of experimental error. 



(When Ht=4"37 the increase of the shielding ratio was not obtained (see Table 

 XV.) for shield A, and only slightly obtained (see Table XVI.) for shield B. Owing to 

 the low values of the shielded field measured in comparison with the higher value of 

 the leakage field due to H c , as also the smallness of the galvanometer deflections, it 

 does not necessarily follow that the same increase in the shielding ratio would not have 

 been obtained, if more perfectly homogeneous shields had been available.) 



