ix
sloping upwards and six downwards, the funnels radiating outwards from a
central globe, and these two parts being united by a connecting rod (see,
again, sodium, Plate I).

The funnel (shown flat as an isosceles triangle, standing on its apex) is a
somewhat complicated structure, of the same type as that in sodium (Plate
VI, 2), the difference consisting in the addition of one more globe,
containing nine additional atoms. The central globe is the same as in
sodium, but the connecting rod differs. We have here a regular arrangement
of five globes, containing three, four, five, four, three atoms
respectively, whereas sodium has only three bodies, containing four, six,
four. But copper and silver, its congeners, have their connecting rods of
exactly the same pattern as the chlorine rod, and the chlorine rod
reappears in both bromine and iodine. These close similarities point to
some real relation between these groups of elements, which are placed, in
the lemniscates, equi-distant from the central line, though one is on the
swing which is going towards that line and the other is on the swing away
from it.

CHLORINE: Upper part {12 funnels of 25 atoms 300
{Central globe 10
Lower part same 310
Connecting rod 19
----
Total 639
----
Atomic weight 35.473
Number weight 639/18 35.50
(The Atomic Weights are mostly from Erdmann, and the Number Weights are
those ascertained by us by counting the atoms as described on p. 349,
January, and dividing by 18. Prof. T.W. Richards, in _Nature_, July 18,
1907, gives 35.473.)

BROMINE (Plate V, 3).--In bromine, each funnel has three additional bodies,
ovoid in shape, an addition of 33 atoms being thus made without any
disturbance of form; two pairs of atoms are added to the central globe, and
a rearrangement of the atoms is effected by drawing together and lessening
the swing of the pair of triplets, thus making symmetrical room for the
newcomers. The connecting rod remains unchanged. The total number of atoms
is thus raised from the 639 of chlorine to 1439. Over and over again, in
these investigations, were we reminded of Tyndall's fascinating description
of crystal building, and his fancy of the tiny, ingenious