POLES OF A MAGNET. 657 



symmetrical masses will be 2\xdx ; if, then, 2 1 is the length of the 

 magnet, we shall have 



3 = 2 ( 



Jo 



n n 



Xdx = \x 



Jo J 



L 3 



This equation will give the constant L, but the distance a from 

 the middle of the magnet to the centre of gravity of each layer 

 is defined, on the contrary, by the condition 



= \xdx; 



and, whatever is the mode of distribution, outside the case of two 

 masses we have always L>. 



For a linear distribution, for instance, in which the density A is 

 proportional to x, the ratio of the values of a and L to the semi- 

 length of the magnet will be 



L s/ 

 7 = V 



^=0-737- 



These few examples will be sufficient to show the difficulties 

 presented by the problem of the distribution of magnetism and 

 the position of the poles. 



1219. BODIES WHICH ARE FEEBLY MAGNETIC OR DIAMAG- 

 NETIC. After iron, nickel, and cobalt, the most powerfully magnetic 

 bodies are the oxides and the salts of iron, but in a far lower 

 degree; the magnetisation is still more feeble in diamagnetic sub- 

 stances. The experiments present then special difficulties arising 

 on the one hand from the smallness of the effect to be measured, 

 and on the other from the sources of error due to the presence of 

 the smallest traces of iron in the bodies in question. 



When diamagnetic phenomena were discovered, Faraday ascribed 

 them to a polarity the inverse of that which a magnetic body would 

 give; he gave up this explanation when he had found that all the facts 



VOL. II. U U 



