PROFESSOR J. A. EWINO ON THE MAGNETIC QUALITIES OP NICKEL. 335 



by step, to about 150 c.g.s. units. Then the bar was demagnetised, a load was 

 applied to the lever producing a stress of compression, and under this the bar was 

 again magnetised. The process was repeated under one and another of a series of 

 loads, the greatest of which produced a compressive stress of 20 kilos, per sq. mm. 



The results of this group of experiments are shown in fig. 12, Plate 17, in the form 

 of curves connecting 3 (the intensity of magnetism) with $ (the magnetising force), for 

 each of the following states of stress : 0, T9, 3'5, G'8, 10, 13'3, and 19'8 kilogrammes 

 per square millimetre. It will be noticed that the effect of compressive stress in 

 augmenting the magnetic susceptibility of nickel is no less remarkable than the effect 

 of tensile stress was shown (in the former paper) to be in reducing the susceptibility. 

 The influence of stress is especially noticeable in the neighbourhood of the bend, or 

 what WIEDEMANN calls the " Wendepunct " of the curves. This is well shown by 

 the following Table, which gives the maximum value of the magnetic susceptibility *c 

 for each state of stress : 



Concurrently with these observations another group was taken to determine the 

 influence which the presence of these stresses of compression during magnetisation 

 had on the amount of residual magnetism held by the metal when the magnetising 

 current was broken at each stage in the process, the stress being maintained constant 

 while the magnetising current was made and broken. Curves of the residual magne- 

 tism (3 r ) in its relation to < are given in fig. 13, for the same set of loads as the 

 curves of induced magnetism in fig. 12 refer to. They show that a state of compres- 

 sion during the application and removal of magnetising force augments the residual 

 magnetism even more than it augments the induced magnetism. In other words, the 

 ratio of residual to induced magnetism is increased by the presence of compressive 

 stress. It was found, in the former paper, that tensile stress reduced this ratio, so 

 much, indeed, that under a strong pull there was scarcely any retentiveness left. 

 Here, under compression, we have the opposite effect : there is enormous retentiveness 

 when the stress is considerable. With no load the maximum value in the ratio of 

 residual to induced magnetism is 0'56 ; with a compressive stress of 10 kilos, per sq. 

 mm. it is 0'91 ; with one of 19'8 kilos, per sq. mm. it reaches the astonishing value 

 of 0-96. 



