208 Prof. J. A. Ewing and Mr. W. Low. [Mar. 24, 



II. Swedish. Wrought Iron. 



Outside field. 



% 



^6— outside field 





47T. 



outside field. 



6,690 



27,960 



1700 



4-18 



8,900 



Zv, /oU 



lOOU 



3*34 



9,510 



30,820 



1700 



3-24 



10,000 



31,210 



1690 



3 12 



10,360 



31,630 



1700 



305 



10,810 



31,720 



1670 



2-94 



10,880 



32,060 



1690 



2-95 



11,200 



32,360 



1690 



2-90 



III. Cast Iron. 



Outside field. 





1 -outside field. 



1 



4tt. 



outside field. 



3,900 



19,660 



1250 



5 04 



6,400 



21,930 



1240 



3-42 



7,710 



22,830 



1200 



296 



8,080 



23,520 



1230 



2-91 



9,210 



24,580 



1220 



2-67 



9,700 



24,900 



1210 



2-57 



10,610 



25,600 



1190 



2 '46 



Fig. 4 shows by curves the relation of J to ^/outside field for Low- 

 inoor iron and for cast iron, in the manner introduced by Rowland 

 for showing the relation of §j to The curves have the same kind 

 of inflection that a curve of fi and §j begins to have when the mag- 

 netising force is raised sufficiently high.* The range through which 

 the permeability of iron may vary is well shown by comparing the 

 values reached here (probably in the extreme case less than 3) with 

 the value 20,000, which was found by one of us in the case of a 

 soft wire exposed to a very small magnetising force and kept at the 

 same time in a state of mechanical vibration. f 



The quantity (^ — outside field)/47ris nearly constant in the Swedish 

 iron, but diminishes with increased induction in the Lowmoor iron 

 and in the cast iron. If the outside field were an accurate measure of 



* This feature of the curve of fi and % was not noticed by Rowland himself, who 

 applied to his curve an empirical formula which fails to take account of it. It has, 

 however, been noticed by several later observers (Fromme, ' Wiedemann, Annalen,' 

 vol. 13, p. 695 ; Ewing, loc. cit., p. 574; Bid well, loc. cit., p. 495). 



f Ewing, loc. cit., p. 567. 



