seems more
closely to resemble the state of things in the _Baeckea_, and which
occurred in a double hyacinth, wherein both anthers and ovules were
borne on the same placenta. Probably, though the fact is not stated, the
ovary of the hyacinth was open; and we are told that the flower was
double--that it was, in fact, modified and changed in more organs than
one; while in the _Baeckea_ nothing at all unusual was observed till the
ovary was cut open. The style was present even in those flowers where
there was no axile placenta; hence in these cases it could not be, as
Lindley stated it to be in the closely allied _Babingtonia_, a
prolongation of the placenta.[184]

=Formation of pollen within the ovules.=--This has now been recorded in
two instances by Mr. S. J. A. Salter in _Passiflora caerulea_ and in _P.
palmata_,[185] and by the author in _Rosa arvensis_.[186]

[Illustration: FIG. 99.--Pollen within the ovule of _Passiflora_ (after
Salter).]

In the case of the passion-flower there were various malformations in
the ovaries, which were all more or less split open at the distal end,
indicating a tendency towards dialysis. The pollen-bearing ovules were
borne on the edges of these ovaries, and presented various intermediate
conditions between anthers and ovules, commencing at the distal
extremity of the carpel with a bi-lobed anther, and passing in series to
the base of the ovary, an antheroid body of ovule-like form, a modified
ovule containing pollen, an ovule departing from a perfectly natural
condition only in the development of a few grains of pollen in its
nucleus, and, finally, a perfect, normal ovule.

In the flowers of the Rose the stamens exhibited almost every
conceivable gradation between their ordinary form and that of the
carpels, while some of the ovules contained pollen in greater or less
abundance. Speaking generally, the most common state of things in these
flowers was the occurrence on the throat of the calyx, in the position
ordinarily occupied by the stamens, and sometimes mingled with those
organs, of twisted, ribbon-like filaments, which bore about the centre
one or more pendulous, anatropous ovules on their margins. Immediately
above the latter organs were the anther-lobes, more or less perfectly
developed, and surmounting these a long style, terminating in a fringed,
funnel-shaped stigma. Sometimes the ovules were perfect, at other times
the nucleus protruded through the foramen, while in a th