336 PROFESSOR J. A. EWING ON THE MAGNETIC QUALITIES OF NICKEL. 



By way of showing more fully the influence of stress in facilitating the magnetisa- 

 tion of nickel we may draw, in ROWLAND'S manner, curves connecting the permeability 

 yx with the induction 93. This is done in fig. 14 for three states of stress : (l) no load, 

 (2) a compression of 10 kilos, per sq. mm., (3) a compression of 19'8 kilos, per sq. 

 mm. In the first the maximum permeability is only 71, in the second it is 212, and 

 in the third it is 357. The points directly found from the observed values of 93 and 

 are marked by dots upon these curves. 



The nickel bar was then softened by heating it to redness in a charcoal fire, and 

 allowing it to cool slowly ; and experiments similar to the foregoing were made with 

 it in the annealed state. The relation of induced and residual magnetism to magne- 

 tising force was examined while the annealed bar was in three states of stress : (1) 

 under no IcTad ; (2) under a compressive stress of 3 '5 kilos, per sq. mm.; (3) under 

 a compressive stress of 6 '8 kilos, per sq. mm. The stress was 'not increased beyond 

 this for fear of hardening the bar by producing permanent set. Figs. 15 and 16 

 show the results of this group of tests. In fig. 15 the induced values of 3 are shown 

 by full lines and the residual values by broken lines, in relation to $. In fig. 16 the 

 permeability p. is shown in relation to the induction 93. It will be noticed by com- 

 paring figs. 15 and 16 with figs. 12 and 14 that the effect of annealing this bar is (as 

 with the nickel wire used in former experiments) to increase the permeability at early 

 stages of the magnetising process, but to reduce it at later stages, and to reduce the 

 highest value to which the magnetism of the metal was raised. Fig. 1 6 shows that 

 the curve of p. and 93 for nickel suffers the same kind of inflection as the corre- 

 sponding curve for iron* when the magnetisation is pushed to high values. 



The ratio of residual to induced magnetism in this annealed nickel bar (as in all 

 former samples both of iron and of nickel) passes a maximum in the neighbourhood of 

 the " Wendepunct." The values of this maximum are 0'84, 0*88, and 0'91 for the 

 three states of stress examined, namely, no load, 3 '5 kilos, per sq. mm. and 6 '8 kilos, 

 per sq. mm. respectively. 



A further experiment was made on the annealed bar to determine the initial value 

 of the magnetic susceptibility under very feeble magnetising forces. For this purpose 

 a new induction coil was wound on the bar, with many more turns than the former 

 coil, and the early part of the curve of 3 and was examined by the ballistic method 

 as before. Fig. 1 7 shows the results for the two conditions of no load and a com- 

 pressive stress of about 5 kilos, per sq. mm. With this, as with the nickel wire 

 tested in earlier experiments, the curve of magnetisation is at first a sensibly straight 

 line with a definite inclination. The initial value of the susceptibility is from 2 to 

 2 '5, and the initial permeability is therefore 25 or 30. The initial permeability 

 determined by this experiment has been utilised in plotting the point where the 

 curve p. and 5B in fig. 16 cuts the axis of /*. 



It is interesting to notice that the residual magnetism after a very weak field has 

 * 'Phil. Trans.,' 1885, p. 574; 'Roy. Soc. Proc.,' vol. 42, p. 208. 



