and the Maximum of Magnetism of Iron , Steel, and- Nickel. 141 



magnets; but it was soon found that little could be done with- 

 out new experiments on the magnetic permeability of substances. 

 Few observations have been made as yet for determining the 

 magnetic permeability of iron, and none, I believe, of nickel and 

 cobalt, in absolute measure. The subject is important, because 

 in all theories of induced magnetism a quantity is introduced 

 depending upon the magnetic properties of the substance, and 

 without a knowledge of which the problem is of little but theo- 

 retical interest ; this quantity has always been treated as a con- 

 stant, although the experiments on the maximum of magnetism 

 show that it is a variable. However, the form of the function 

 has never been determined, except so far as we may deduce it 

 from the equation of Muller, 



I=22(W* tan ^^ 



which, as will be shown, leads to wrong results. The quantities 

 used by different persons are as follows : — . 



Kj Neumann's coefficient, or magnetic susceptibility. (Thorn- 

 son). 



k, Poisson's coefficient. 



fi, coefficient of magnetization (Maxwell), or magnetic per- 

 meability (Thomson). 



X, introduced for convenience in the following paper. 



The relations of these quantities are given by the following 

 equations ;— - 



, 477 K _/A — 1_X~-47T 



4tt/c + 3 - fi + 2 ~~ X + 8tt' 

 _/£—l__ 3k _X— 47r 

 K ~ ~krT ~ 4tt(1— k) ~ : ~IQ^ 3 



The first determination of the value of any of these quantities 

 was made by Thalen. But more important experiments have 

 been made by Weber, Von Quintus Icilius, and more recently 

 by M. Reicke and Dr. A. Stoletow*. The first three of these 

 in their experiments used long cylindrical rods^ or ellipsoids of 

 great length; the last, who has made by far the most important 

 experiments on this subject, has used an iron ring. The 

 method of the ring was first used by Dr. Stoletow in September 

 1871 ; but more than eight months before that, in January 

 1871, 1 had used the same method, but with different apparatus, 

 to measure the magnetism. He plots a curve showing the va- 



* Phil. Mag. January 18/3. 



