818 Mr. G. W. Heaps on the Effect of 



III. A and II. B are the corresponding curves when H is 

 parallel to an edge at right angles to the first. In general 

 one can only say that II. B and III.B become approximately 

 equal if the current is perpendicular to the two directions it 

 there has (see I., B & C). 



Several other curves were obtained ; in one, when H was 

 perpendicular to a face of the crystal and the current flowed 

 along an edge, a decrease of resistance as large as 13'0 x 10~ 4 

 per ohm was measured. Collectively the results indicate that 

 even a regular crystal may have an exceedingly complex 

 internal structure. The magnetic anisotropy of magnetite 

 may be better understood if one remembers that Lord Kelvin's 

 theory of magnetism will not be valid in the case of a ferro- 

 magnetic crystal*. A fundamental assumption of this theory 

 is that the magnetic susceptibility, though different in different 

 directions, is independent of the field strength. Clearly this 

 assumption cannot be made in the case of a magnetic sub- 

 stance. If A, B, and C are components of magnetization 

 and X, Y, and Z components of field strength, for a para- 

 or diamagnetic crystal we would have the relations : 



A=K 1 X, 



B = K 2 Y, 



C = K S Z; 



but for a ferromagnetic substance it would be necessary 

 to write : 



A = X[K 1 +/ 1 (X*,Y 2 ,Z 2 )], 



B = Y[K 2 +/ 2 (X 2 ,Y 2 ,Z 2 )], 



C=Z[K,+/,(X% Y\ Z 2 )]. 



For a regular non-magnetic crystal the susceptibility is the 

 same along all the axes, so that K 1 =K 2 =K 3 , and the 

 equations are the same as for an isotropic substance. For a 

 regular magnetic crystal, however, we have the terms 

 /(X 2 , Y 2 , Z 2 ) entering, and these functions cannot be taken as 

 the same along all the axes even if the crystal belongs to the 

 regular system. 



In the case of magnetite Weiss t has shown experimentally 

 that a regular ferromagnetic crystal may have different 

 properties along different axes, and his work, together with 

 that of Y. Quittner J, indicates that magnetite is even more 



* See Voigt, Gott. Nachr. p. 331 (1900). 

 t Journ. d. Phys. v. p. 435 (1896). 

 \ Arm. d. Phys. p. 289 (1909). 



