244 Proceedings of the Poyal Society of Edinburgh. [Sess. 
as to indicate the general nature of the correspondence between the form 
of these curves and the form of Weiss’s curve. It must, of course, be 
remembered that the general equation just given represents the value of 
that part of an internal field which is parallel to the direction of magnetisa- 
tion supposed to have attained its constant saturation value, while Weiss’s 
curves represent the component of magnetisation in the direction of the 
external field supposed to be constant. But it is the variation of the 
internal field which causes the variation of the magnetisation, so the general 
nature of the two sets of curves should be very similar. 
The chief difference between the two sets lies in the breadth of the 
minimum in the neighbourhood of 90°. Suppose that the sin 20 term were 
Fig. 10. 
absent. In this case, with the same maximum and minimum values, the 
curve would be 1 — 0*8 sin 2 (9, which is represented by the dotted line. The 
marked flatness at the apex of the lower full curve is therefore due to the 
L' 4 effect. Hence if, at the lower fields, the sin 2 20 term were reversed 
in sign, the apex of the curves would be very sharp and the sides hollow. 
This might occur if the sign of the true structural term were negative, 
while a similar term of positive sign preponderated at the lower fields until 
increasing magneto-striction caused its relative obliteration. But the law 
connecting magnetisation with field intensity is very determinative. If 
the lower left-hand curve represented the total field (internal plus external) 
with the line of zero field lying not much below its vertex, and if the 
magnetisation were proportional to the square of the field, the curve of 
