60 



Mr. R H. M. Bosauquet on 



Ifr=B, 



L = KBD. 



If C=-D, 





As r=B is the only solution possible, the method can only 

 be used with a telephone under the conditions given by Max- 

 well, and is impracticable on account of the trouble met with 

 in obtaining the double adjustment. 



V. Note on Magnetization. — On Sequences of Reversals. 

 By R H. M. Bosanquet*. 



THE present paper arose out of a recent discussion, at a 

 meeting of the Physical Society, on Prof. Ayrton's paper 

 on the magnetic resistance of a broken ring. I alluded on that 

 occasion to the determinations of the magnetism of rings and 

 bars which I had made. In answer to a question as to the 

 course I had pursued as to ascending or descending values of 

 magnetism, I stated that in all cases I had taken the values in 

 the ascending order. But I omitted to make it clear that the 

 observations were made by reversal, so that the direct effects 

 of residual magnetism were necessarily eliminated. The 

 arrangement I have always employed may be described as an 

 ascending sequence of reversals. 



There can be no doubt, however, that, in sequences of re- 

 versals, the history of previous magnetizations produces some 

 effect. I still consider that the best way of attacking the 

 problem is to take the first ascending sequence, in which the 

 history is a minimum, and then as a subsequent study to deal 

 with the effects of the history. 



Immediately after the discussion above alluded to I made a 

 few experiments on the first bar I ever completely determined, 

 with the view of ascertaining the general character presented 

 by series of successive reversals. The original determinations 

 will be found at Phil. Mag. (1884) xvii. pp. 531-6, Soft-Iron 

 Bar I. 



I shall confine myself at present to the values of magnetic 

 resistance. The reason for preferring this datum is that, when 

 plotted with the induction as abscissa, the form of the curve 

 of magnetic resistance is approximately independent of the 

 arrangement of the metal, the only material differences occur- 

 ring in the region of saturation. Thus values for bars and 

 rings differ only by a constant, which may be regarded as due 

 to the shape. Bars with pole-pieces have similar curves lying 

 between those for bars and rings. See Phil. Mag. xxii. p. 303, 

 and the paper above referred to. 



* Conmiunicated by the Physical Society : read April 23, 1887. 



