Magnetic Permeability of various Alloys of Iron. 135 



Hopkinson, and probably than that used by Ewing. In a specimen of very pure 

 iron from the Elswick works, the maximum induction in the same field as ours, 

 45 C.Gr.S., is given as 17450 by Messrs. Lydall and Pocklington in the Proceedings 

 of the Royal Society^ vol. 52, p. 228. From tests made on a sample of Hadfield's 

 dynamo " steel," Prof. Ewing gives the maximum induction at 45 as 16900, more 

 complete tests at Finsbury College, with a ring of the same steel, gave the follow- 

 ing values : — 



Maximum induction, . . 16480, for H = 45 '5. 



Permeability, . . . 1600, ,, „ = 8-0. 



Coercive force, . . . 1*75 ,, ,, = 65 '0. 



Loss in ergs per cycle, . . 12080, ,, ,, = 65*0. 



The correction for the demagnetizing action of the ends raises the residual 

 induction and ^iermeability in our specimen of iron, as will be seen in Note A 

 here appended. 



In conclusion, we desire again to express our thanks to Mr. Allen and 

 Mr. Wills for frequent assistance in the numerous observations I'equired in the 

 second part of this paper. 



Note A. 

 As we have already pointed out, a correction is necessary for the demagnetizing 

 reaction of the ends in our rods, especially in those of high permeability. Prof. 

 J. A. Ewing, F.R.S., has shown {Phil. Trans., 1885, Part ii., p. 536) that "even 

 when dealing with the softest iron, we may take a rod, whose length is not less 

 than 300 diameters, as giving results scarcely different from those given by a ring 

 or longer rod": and, " in hard iron and in steel a smaller ratio of length to dia- 

 meter would, no doubt, give an equally good approximation to the condition of 

 endlessness." In his work on Magnetic Induction in Iron and other Metals, p. 31, 

 Ewing gives a table of correcting factors for rods of various lengths. The rods 

 we have tested are 200 diameters long, and, according to Ewing, the actual mag- 

 netizing force is the original force in the solenoid, diminished by 0-000125 times 

 the induction for any given point. Thus, if we take a point on the B— H curve 

 of our annealed iron (this curve is repeated in each plate), where B = 16000, we 

 find H = 16-5 ; then the actual value of H for this induction will be 16-5 - 2 = 14-5, 

 since 0-000125 x 16000 = 2; and so on for any other point on the curve. It there- 

 fore follows that the values of H must not be measured from the vertical axis OB, 

 but from a line inclined to the right, and drawn through O to a point whose 

 co-ordinates are B = 16000 and H = 2 : this line will make an angle of two degrees 

 with OB. It is easy to draw such a line on our plates, by placing a straight-edge 



