my exist, that, paradoxical as the assertion may
seem, I consider the exception rather as confirming than lessening the
importance of the character.

(*Footnote. Linnean Society Transactions 12 page 136.)

It may perhaps be unnecessary to remark, that the raphe, or vascular cord
of the outer coat, almost universally belongs to that side of the ovulum
which is next the placenta. But it is at least deserving of notice, that
the very few apparent exceptions to this rule evidently tend to confirm
it. The most remarkable of these exceptions occur in those species of
Euonymus, which, contrary to the usual structure of the genus and family
they belong to, have pendulous ovula; and, as I have long since noticed,
in the perfect ovula only of Abelia.* In these, and in the other cases in
which the raphe is on the outer side, or that most remote from the
placenta, the ovula are in reality resupinate; an economy apparently
essential to their development.

(*Footnote. Abel's China page 377.)

The distinct origins and different directions of the nourishing vessels
and channel through which fecundation took place in the ovulum, may still
be seen in many of those ripe seeds that are winged, and either present
their margins to the placenta, as in Proteaceae, or have the plane of the
wing at right angles to it, as in several Liliaceae. These organs are
visible also in some of those seeds that have their testa produced at
both ends beyond the inner membrane, as Nepenthes; a structure which
proves the outer coat of scobiform seeds, as they are called, to be
really testa, and not arillus, as it has often been termed.

The importance of distinguishing between the membranes of the
unimpregnated ovulum and those of the ripe seed, must be sufficiently
evident from what has been already stated. But this distinction has been
necessarily neglected by two classes of observers. The first consisting
of those, among whom are several of the most eminent carpologists, who
have regarded the coats of the seed as products of fecundation. The
second of those authors who, professing to give an account of the ovulum
itself, have made their observations chiefly, or entirely, on the ripe
seed, the coats of which they must consequently have supposed to be
formed before impregnation.

The consideration of the arillus, which is of rare occurrence, is never
complete, and whose development takes place chiefly after fecundation,
might here, perhaps, be entirely