ds an arrested development of the perianth is habitual,
as in _Oncidium abortivum_ (fig. 217), where, on a large branching
panicle, numerous abortive, but few perfect, flowers are produced.
In a similar way the petals and labellum of _Odontoglossum
Uro-Skinneri_ have been found reduced to filamentous processes.

[Illustration: FIG. 217.--Flower of _Oncidium abortivum_, magnified.]

=Abortion of the stamens.=--Atrophy of one or more stamens is of very
common occurrence, as a general rule, in many genera of plants, _e.g._
_Scrophularia_, _Erodium_, many _Restiaceae_, &c. &c. As a strictly
teratological condition atrophy of the stamens is more rare than
complete suppression. It has been noticed in _Arabis alpina_, _Cerastium
glomeratum_, _C. tetrandrum_, _Rhamnus catharticus_, _Anemone_,
_Hepatica_, &c. It happens frequently among Orchids both wild and
cultivated. In the _Hymenocallis_ flowers described by the elder Morren,
four out of five stamens were atrophied. In other flowers, otherwise
perfectly formed, one abortive stamen was found bearing a spherical
indehiscent anther. All these atrophied anthers of _Hymenocallis_ were
found to contain pollen, differing at first sight but little from what
is usual, but presenting this important peculiarity, that while the
normal pollen does not burst until it comes into contact with the
stigma, in the abnormal flowers the outer coat of the pollen-grains
split while still within the anther, from which latter, indeed, they
could not escape, owing to the indehiscent nature of the latter. Again,
the pollen-tube of the abnormal grains cracked, in its turn, on mere
exposure to the air, and liberated the fovilla, so that the pollen of
these atrophied anthers was necessarily impotent, because it opened
before it could be applied to the stigma, even had that been rendered
possible by the opening of the anther.

An abortive condition of the stamens and of the pollen, is of very
common occurrence among hybridised plants. Gaertner and other writers
have spoken of this defective condition as contabescence.[535] It forms
one reason for the sterility so frequently observed in the case of true
hybrids. In some hybrid passion-flowers, while all other parts of the
flower were apparently perfect, even to the ovules, the stamens were
atrophied, and distorted, and contained little or no pollen; the few
grains of the latter being smaller than usual. (See under Heterogamy,
pp. 193-196, and p. 398.)