of our readers may be glad to have a drawing of the Platonic solids,
since they play so large a part in the building up of elements. The regular
solids are five, and five only; in each:

(1) The lines are equal.
(2) The angles are equal.
(3) The surfaces are equal.

[Illustration]

It will be seen that the tetrahedron is the fundamental form, the
three-sided pyramid on a triangular base, _i.e._, a solid figure formed
from four triangles. Two of these generate the cube and the octahedron;
five of these generate the dodecahedron and the icosahedron.

The rhombic dodecahedron is not regular, for though the lines and surfaces
are equal, the angles are not.

NOTES.

Mr. C. Jinarajadasa[1] writes:

The asterisk put before metargon in the list of elements should be omitted,
for metargon had been discovered by Sir William Ramsey and Mr. Travers at
the same time as neon (see _Proceedings of the Royal Society_, vol. lxiii,
p. 411), and therefore before it was observed clairvoyantly. It is not,
however, given in the latest list of elements in the Report of November 13,
1907, of the International Atomic Weights Commission, so it would seem as
though it were not yet fully recognised.

Neon was discovered in 1898 by Ramsey and Travers, and the weight given to
it was 22. This almost corresponds with our weight for meta-neon, 22.33;
the latest weight given to neon is 20, and that corresponds within
one-tenth to our weight, 19.9. From this it would seem that neon was
examined in the later investigations and meta-neon in the earlier.

He says further on a probable _fourth_ Interperiodic Group:

Thinking over the diagrams, it seemed to me likely that a fourth group
exists, coming on the paramagnetic side, directly under iron, cobalt,
nickel, just one complete swing of the pendulum after rhodium, ruthenium,
palladium. This would make four interperiodic groups, and they would come
also _periodically_ in the table too.

I took the diagram for Osmium, and in a bar postulated only three columns
for the first element of the new groups, _i.e._, one column less than in
Osmium. This would make 183 atoms in a bar; the new group then would follow
in a bar, 183, 185, 187. Here I found to my surprise that the third
postulated group would have a remarkable relation to Os, Ir, Pt.

Thus

Os.--245 (in a bar); less 60 = 185
Ir. 247 less 60 = 187
Pt. 249 le