118 Proceedings of the Eoyal Society of Edinburgh. [Sess. 
the model as constructed the octet of poles which represents the Weber 
element is pivoted on a vertical needle, and therefore turns about one 
axis only, but of course it has to be thought of as having three degrees of 
rotational freedom. To allow the construction to be more clearly seen, 
fig. 15a (PI. II) shows the Weber element removed from its place and 
standing alongside the frame of fixed magnets. 
In another model, shown in fig. 16 (PI. II), the same outer fixed group is 
used, but the Weber element consists of a quartet of poles occupying the 
corners of a regular tetrahedron — two north poles and two south poles. 
This form is readily derived by removing four of the eight spokes of the 
Weber element which is shown in figs. 15 and I 5a. 
A. W. Hull has suggested that the twenty-six electrons of the iron 
atom are probably grouped so that two of them form a doublet with the 
nucleus of the atom in the middle, and the remaining twenty-four are 
spaced further out along the diagonals of a cube, forming three octets. 
The members of each octet occupy the corners of a cube.'^ On the basis 
of this suggestion we may think of the outer octets as supplying the 
four pairs of fixed magnetic elements shown in the models. The Weber 
element may be constituted simply by the inner doublet ; on the other 
hand, it may include the innermost of the three octets, leaving the other 
two octets to form the fixed magnets. In the latter case it may be 
conjectured that the inner doublet is also included in the part that turns. 
According to either hypothesis the Weber element — the thing that turns 
during magnetisation -need not be more than a comparatively small part 
of the atom. That it is only a very small part has been inferred by 
Messrs A. H. Compton and O. Eognley from the fact that no change in 
the X-ray spectrum reflected from the face of an iron crystal could be 
detected when the crystal was magnetised. j* 
On this point, however, it should be observed that if in the atoms of 
a crystal each Weber element is an octet of electrons (or any cubically 
symmetrical system), its being oriented, by magnetising the substance, 
into the most favourable of its positions of stability cannot be expected 
to alter the X-ray spectrum, no matter what proportion its dimensions 
bear to those of the complete atom. In the model atom there is no 
difference to the eye between one and another of the positions of stability. 
It is therefore of doubtful validity to conclude that the dimensions ( 
of the Weber element are minute in relation to those of the atom, and 
it is open to suppose that the central group of electrons, which in my view 
constitutes the Weber element, extends far enough from the nucleus to 
t Ihid.^ xvi, p. 464, 1920. 
^ Phys. Rev.^ ix, p. 85, 1917. 
