M. Bory St. Vincent ("Voyage aux Quatre Isles d'Afrique" tome 1 page 222.)
has described some balls of lava from the Isle of Bourbon, which have a
closely similar structure. His explanation, however (if I understand it
rightly), is very different from that which I have given; for he supposes
that they have rolled, like snowballs, down the sides of the crater. M.
Beudant ("Voyage en Hongrie" tome 2 page 214.), also, has described some
singular little balls of obsidian, never more than six or eight inches in
diameter, which he found strewed on the surface of the ground: their form
is always oval; sometimes they are much swollen in the middle, and even
spindle-shaped: their surface is regularly marked with concentric ridges
and furrows, all of which on the same ball are at right angles to one axis:
their interior is compact and glassy. M. Beudant supposes that masses of
lava, when soft, were shot into the air, with a rotatory movement round the
same axis, and that the form and superficial ridges of the bombs were thus
produced. Sir Thomas Mitchell has given me what at first appears to be the
half of a much flattened oval ball of obsidian; it has a singular
artificial-like appearance, which is well represented (of the natural size)
in Figure 4. It was found in its present state, on a great sandy plain
between the rivers Darling and Murray, in Australia, and at the distance of
several hundred miles from any known volcanic region. It seems to have been
embedded in some reddish tufaceous matter; and may have been transported
either by the aborigines or by natural means. The external saucer consists
of compact obsidian, of a bottle-green colour, and is filled with finely
cellular black lava, much less transparent and glassy than the obsidian.
The external surface is marked with four or five not quite perfect ridges,
which are represented rather too distinctly in Figure 4. Here, then, we
have the external structure described by M. Beudant, and the internal
cellular condition of the bombs from Ascension. The lip of the saucer is
slightly concave, exactly like the margin of a soup-plate, and its inner