240 Proceedings of the Royal Society of Edinburgh. [Sess. 
direction of a quaternary axis and a binary axis ; and Quittner’s results show 
that the opposite occurs. The most probable hypothesis seems to be that 
the term + cos 60 is due to the actual crystalline structure ; and that, in 
low fields, its effect is overbalanced by a similar effect, of opposite sign, due 
to the existence of crystalline flaws. 
Weiss postulated the existence of flaws or incipient fissures parallel to 
the planes of the cubic faces. Seeing that the internal field parallel to the 
cubic edges is negative, these flaws may not be necessary to explain the 
more rectilinear rise of magnetisation under increasing field parallel to the 
cubic edges than of that parallel to the binary or the ternary axes ; but 
(. Proc . R.S.E., 1905, p. 1061) they readily explain the relative order of 
magnitude observed by Weiss in the case of residual magnetisation in these 
directions. They cannot account for the maximum and minimum which 
appear, in weak fields, in the abnormal plate, fig. 5. To explain these we 
must postulate an excess of flaws in planes perpendicular to the 135° diagonal. 
Foliation parallel to the dodecahedral planes or to the octahedral planes 
could produce that effect. Quittner, pointing out the readiness with which 
magnetite cleaves in octahedral planes, postulated incipient splitting parallel 
to them. 
In fig. 8, let AD, AB, AC be Quittner’s axes a, b, c. Foliation parallel 
to any of the octahedral planes at D, B, E, and G (the pair D, G are identical, 
as also are the pair B, E) diminishes the permeability in the direction ab in 
the cubic face DAB ; and this gives the result observed in fig. 5, in the 
plane of the axes a, b, at H = 3T5. Also, if the foliation occurs equally in 
each of these four octahedral planes, the result is to diminish equally the 
