MAGNETIC SHIELDING IN HOLLOW IKON CYLINDERS. 655 



first four sequences for shields A and B, plotted as ordinates against values of H c as 

 abscissae. WhenH c =0, the curves have their origin at values corresponding to those 

 plotted in fig. VI. when the transverse field is the only force acting upon the iron, which 

 in this case is equal to 20'9 C.Gr.S. units. As these curves leave their origin in the 

 vertical axis, a general comparison of figs. xx. and xxi. with the theoretical shielding ratios 

 ((g)) given in fig. iv. shows that the differences which exist in the magnetic qualities of 

 the two shields A and B are well brought out in the shielding ratios. For shield A they 

 rise higher, attain their maximum values earlier, and ultimately descend to lower values 

 than for shield B. Leaving out of consideration meanwhile the curves obtained under 

 the CTT conditions given above, these remarks are undoubtedly true for those obtained 

 under the TC and TCC conditions, and also, although to a less extent, for those obtained 

 under the CT conditions. Indeed, more careful examination shows that, when incre- 

 ments of the circular field are superimposed upon a pre-existing induction due 

 to the transverse field (TC condition), the shielding ratio curves coincide (except 

 near their origin) with the theoretical ((g)) curves, within the limits of experimental 

 error. 



S 40. As the conclusions arrived at in this investigation, when both transverse and 

 circular fields act upon the iron, and indicated in the last sentence of § 39, are not in 

 harmony with the inductions drawn from the experiments of Stefan and Du Bois, I 

 quote from the hitter's recent contributions on the subject. He writes as follows* : — 

 " One of the thick wrought-iron cylinders was uniformly wound with wire, by 

 means of which a circular field of several C.G.S. units could be excited. 

 Besides the corresponding circular magnetisation this produced considerable 

 polarity, owing to the welded joint and general want of homogeneity. Not- 

 withstanding this and the high permeability corresponding to the circular 

 field, the shielding ratio proved practically unaffected by it. This result lends 

 support to the assumption, made throughout, that shielding against moderate 

 disturbing fields depends, within due limits, upon the constant permeability 

 for small increments or decrements of force superimposed upon any condition 

 of magnetisation." 

 In a footnote he adds : — 



" This result is given by the late Professor Stefan without proof, and is probably 

 empirical." 

 This quotation states the problem clearly. The distinction here drawn or implied 

 between the permeability of the iron to the circular field and the permeability of the 

 iron to the transverse field is a valid one. It can hardly be maintained that the above 

 conclusion, even if correct, could be arrived at otherwise than empirically. The problem 

 is therefore one for experimental evidence. Stefan's results are stated to be given with- 

 out proof ; and whether Du Bois' data are sufficient or insufficient to establish his con- 

 clusions will become evident as we proceed. 



* Electrician, "Magnetic Shielding," vol. xl. p. 654 (1898). 

 TRANS. EOY. SOC. EDIN., VOL. XL. PART III. (NO. 26). 5 d 



