660 MR JAMES RUSSELL ON 



magnetisation was nota.bly increased ; the magnetisation curve was found in that case 

 to lie above the normal curve everywhere except in the region of strongest magnetis- 

 ation." 



§ 47. The difference between the dash line curves of figs. xxxn. and xxxiil 

 speaks to a decrease of the magnetic lines converging upon the iron (which as we have 

 seen accompanies the increase of the circular induction) unber the CTT conditions, 

 as compared with the somewhat greater number of magnetic lines converging upon the 

 iron (which accompanies the decrease of the circular induction) under the TC 

 conditions. 



But we know that when a very long cylindrical rod of circular section is placed 

 in a uniform transverse field, the induction B is given by 



H ( /x + 1 



so that when //■ is large the induction in the rod approximates to twice the value of the 

 undisturbed transverse field. Further, that if the long rod be a hollow cylinder, the 

 number of lines of the transverse field converging upon the iron will be considerably 

 less than the above ratio as the wall of the cylinder is reduced in thickness. Reference 

 may be made to the lines of force diagram given by Du Bois in the articles already 

 referred to. # But Table IV. (§ 13) shows that for both shields the lines of force 

 converging upon the iron is practically the same as if the shields had been solid. This 

 discrepancy between theory and experiment is fully accounted for by the fact that the 

 conditions of great length relative to thickness are not fulfilled — the ratios of length to 

 diameter for both shields being about 4*5 to 1. 



In figs, xxxii. and xxxiii. the dotted horizontal lines measure the transverse field 

 undisturbed by the presence of the iron shields, or H ( = 20'9 C.G.S. units. Comparison 

 with the dash line curves shows that under the TC conditions the ratio which at its 

 origin is approximately equal to two, shows a slight increase at low values of H c , but 

 that as the circular field is further increased the ratio tends to fall below two. Under 

 the CTT conditions, on the other hand, there is no such initial rise, but the ratio falls 

 slowly but continuously as H c is increased. 



These facts point to the conclusion that, although the dominating factor in determining 

 the number of magnetic lines of the transverse field converging upon the iron is for 

 large values of /x the geometrical factor, still the permeability of the iron to the trans- 

 verse field is to a certain extent at least distinctly seen to respond to the permeability 

 impressed upon the iron by the superposed circular field. On the other hand, the effect 

 of repeated reversals of the transverse field superposed upon the circular is to 

 abnormally increase the induction due to the latter (full line curves CTT), and 

 concurrently with this to decrease somewhat the permeability of the shields to the trans- 

 verse field (dash line curves CTT). 



* Electrician, vol. xl. p. 513. 



