238 
Proceedings of the Royal Society of Edinburgh. [Sess. 
Thus the variation of magnetisation exactly follows the theoretical variation 
of the field. As saturation is approached these variations tend to disappear. 
The effect of y I/ 4 , if present, is similar to that of y L r 2 . 
In the irregular plate an abnormality appears, even at low fields, in 
the existence of a maximum along one diagonal and a minimum along the 
other. But, in a truly cubic arrangement, y L' 0 is zero. Therefore the 
effect must either be due to deformation of the crystal by relative elonga- 
tion along the former diagonal (§ 8), or it must be caused by internal flaws 
in planes perpendicular to the latter diagonal, which would reduce the 
permeability in the direction of the diagonal itself. 
] 
I 
S~f6 
— ''■"-A r 
6 96 
' 
298 
H - 
H = 
198 
f2f) 
H 
/26 
63 
H - 
63 
3/6 ' 
t 
3 
e 
i 
a 
0 ? 6 ' 90 TJs 7&o o ? 5 ' ?o / 36 ' no 
Fig. 6. 
Quittner's diagrams of the transverse effect are not reproduced here. 
Zero values occur in its variation where stationary values of the parallel 
effect occur, just as the theory indicates ; and its maxima and minima 
also follow the theoretical indications, both as regards the normal and the 
abnormal conditions. In fig. 5 the values of y I/ 2 , y T' 2 , and y L' 4 are repre- 
sented respectively by r= 2 (sin 2 20 — |) (dashes), r = sin 4 0 (full line), and 
r = 2 (sin 2 2 0 — ^t) (dotted curve). If r and 0 be represented by rectangular 
co-ordinates, the correspondence between the variations of r and the varia- 
tions of the magnetisation in Quittner’s curves becomes at once evident. 
18. The left-hand part of fig. 7 shows Quittner’s observations on the 
parallel component of magnetisation in the case of an octahedral plate. In 
weak fields an effect corresponding in type to L' 0 appears. In stronger 
fields, three maxima and three minima occur. The same takes place in 
