a carpel.[325]

Agardh has observed a similar thing in a hyacinth, one half of the fruit
of which contained seeds, and the other half, anthers. B. Clarke
mentions an instance in _Mathiola incana_ in which the carpels were
disunited, and antheriferous at the margin.[326]

The passage of pistils to stamens in willows has been frequently
remarked, as in _Salix babylonica_, _silesiaca_, _cinerea_, _Caprea_ and
_nigricans_. One of the most curious illustrations of this
transformation in this genus is given by Henry and Macquart (Erst.
Jahrb. des bot. Vereines am m. et n. Rhein., 1837). In the flowers in
question the series of changes were as follows:--first, the ovary opened
by a slit, and then expanded into a cup; next, anther-cells were
developed on the margin of the cup, with stigmas alternating with them,
the ovules at the same time disappearing; lastly, the margin became
divided, and bore three perfect anthers, which in the more perfect
states were raised on three filaments.

_Campanula persicifolia_, _C. rapunculoides_, and _C. glomerata_ have
been observed to present an anther surmounting the pistil.[327] Double
tulips often present this change, and a like appearance has been
observed in _Galanthus nivalis_, and _Narcissus Tazetta_.

Moquin mentions the existence of this condition in a female plant of
maize, some of the pistils of which were wholly or partially converted
into anther-like organs. Mohl has recorded an analogous malformation in
_Chamaerops humilis_, and in which the three carpels were normally
formed, and only differed from natural ovaries in this, that along the
two edges of the ventral suture there was a yellow thickening, which a
cross section of the ovary showed to be an anther-lobe filled with
pollen.[328]

In _Tofieldia calyculata_ a similar substitution of a stamen for a
carpel has been observed by Klotsch,[329] and Weber[330] gives other
instances in _Prunus_ and _Paeonia_. Corresponding alterations may be met
with in cultivated tulips, in the cowslip and other plants. In most of
the above cases the transmutation has been perfect, but in quite an
equal number of cases a portion only of the carpel is thus changed,
generally the style or the stigma; thus Baillon describes the stigmas of
_Ricinus communis_ as having been in one instance antheriferous.[331]
Moggridge figures a flower of _Ophrys insectifera_ in which the
rostellate process was replaced by an anther.[332]

Mohl remarks that the