flower-bud, consisting of calyx, corolla, and stamens, but with no trace
of pistil. I have been unable to find recorded any instance of
malformation among Labiates or Borages at all similar to this. It
differed from most other examples of prolification in that the axis was
not prolonged, the adventitious bud occupying precisely the position of
the three lobes of the ovary that were absent. The sole remaining carpel
had a style and a stigma as perfect in appearance as though the pistil
had been complete.

[Illustration: FIG. 60.--Flower of _Dianthus_ affected with median
floral prolification.]

In a flower of _Conostephium_ (_Epacridaceae_) forwarded to me by Mr.
Bentham, there was a similar adventitious bud placed by the side of the
pistil, but as the latter contained the usual number of cells it is
probable that the supernumerary bud in this case originated rather from
the side than the end of the axis.

Certain families of plants present this deviation from their ordinary
structure with greater frequency than others: the following orders seem
to be the most frequently affected by it: _Ranunculaceae_,
_Caryophyllaceae_, _Rosaceae_; while it is commonly met with in
_Scrophulariaceae_, _Primulaceae_ and _Umbelliferae_. Of genera which seem
peculiarly liable to it may be mentioned the following: _Anemone_,
_Ranunculus_, _Cheiranthus_, _Dianthus_, _Dictamnus_, _Daucus_, _Rosa_,
_Geum_, _Pyrus_, _Trifolium_, _Antirrhinum_, _Digitalis_, _Primula_.

A reference to the subjoined list of genera affected by this
malformation, and the knowledge of its comparatively greater frequency
in some than in others of them, will show that it is more often met with
in plants having an indefinite form of inflorescence than in those
having a definite one. The change may affect some only, or the whole of
the flowers constituting an inflorescence; and though it is by no means
a constant occurrence, it very frequently happens that the central or
terminal flower in a definite inflorescence is alone affected, the
others remaining in their ordinary condition, as in pinks (_Dianthus_);
and in the indefinite forms of inflorescence, it is equally common that
the uppermost flower or flowers are the most liable to be thus affected.

In those plants which present this deviation from the ordinary condition
with the greatest frequency, it often happens that the axis is normally
more or less prolonged, either between the various whorls of the flower,