e compound fruit with the two kinds either blended
together, both externally and internally, or segregated in various
ways. This tree can be propagated by cuttings, and retains its
diversified character. The so-called trifacial orange of Alexandria and
Smyrna[911] resembles in its general nature the bizzarria, but differs
from it in the _sweet_ orange and citron being blended together in the
same fruit, and separately produced on the same tree: nothing is known
of its origin. In regard to the bizzarria, many authors believe that it
is a graft-hybrid; Gallesio on the other hand thinks that it is an
ordinary hybrid, with the habit of partially reverting {392} by buds to
the two parent-forms; and we have seen in the last chapter that the
species in this genus often cross spontaneously.

Here is another analogous, but doubtful case. A writer in the
'Gardener's Chronicle'[912] states that an _AEsculus rubicunda_ in his
garden yearly produced on one of its branches "spikes of pale yellow
flowers, smaller in size and somewhat similar in colour to those of _AE.
flava_." If as the editor believes _AEsculus rubicunda_ is a hybrid
descended on one side from _AE. flava_, we have a case of partial
reversion to one of the parent-forms. If, as some botanists maintain,
_AE. rubicunda_ is not a hybrid, but a natural species, the case is one
of simple bud-variation.

The following facts show that hybrids produced from seed in the
ordinary way, certainly sometimes revert by buds to their parent-forms.
Hybrids between _Tropaeolum minus_ and _majus_[913] at first produced
flowers intermediate in size, colour, and structure between their two
parents; but later in the season some of these plants produced flowers
in all respects like those of the mother-form, mingled with flowers
still retaining the usual intermediate condition. A hybrid Cereus
between _C. speciosissimus_ and _phyllanthus_,[914] plants which are
widely different in appearance, produced for the first three years
angular, five-sided stems, and then some flat stems like those of _C.
phyllanthus_. Koelreuter also gives cases of hybrid Lobelias and
Verbascums, which at first produced flowers of one colour, and later in
the season flowers of a different colour.[915] Naudin[916] raised forty
hybrids from _Datura laevis_ fertilised by _D. stramonium_;