r parts of the flower likewise undergo modification. The changes
most commonly met with are such as may be classed under Goethe's theory
of retrograde metamorphosis; for instance, if a supplementary bud be
developed in the axil of a sepal, that sepal is likely to be more than
ordinarily leaf-like in appearance. The dislocation of the affected
sepal from its fellows is a very frequent occurrence; in cases of this
kind the detached sepal is placed below the others, thus approximating,
in position as well as in function, to the bracts. In some of the
instances of proliferous pears, on which I shall have occasion to
comment, the sepals are described as sharing in the succulent character
of the fruit.

The petals, under such circumstances, often exist in the guise of sepals
or of small leaves; and instances are recorded wherein the place of the
calyx and corolla was supplied by a succession of overlapping green
scales, from the axils of which the new buds arose. M. Germain de Saint
Pierre records such a case in _Trifolium repens_, wherein the calyx and
corolla were replaced by overlapping scales, in the axils of each one of
which arose a flower; above there was a row of stamens, and in the
centre a pistil in the guise of a trifoliate leaf.[139] Such instances
seem to afford an extreme degree of a more common change, viz., the
diminished size and contracted appearance of the sepals and petals when
affected with axillary prolification. They have also a close
relationship to such developments as we see in the wheat-ear carnation,
in certain species of the genus _Maesa_ and others, wherein the calyx is
repeated over and again, to the partial or complete suppression of the
other parts of the flower. All these cases may be in part explained by
the operation of the principle of compensation.

So far as the androecium is concerned, the stamens either remain
unaltered, or they are present in a more or less petal-like condition;
but it far more frequently happens that the stamens are entirely
suppressed, the adventitious bud supplying their place; thus was it in
the _Dianthus_ represented in the adjoining woodcut, fig. 66, where the
stamens were entirely absent, and their places supplied by
flower-bearing branches. This _Dianthus_ has the more interest from its
similarity to the one described by Goethe, Metam. der Pflanzen, cap.
16, sect. 105; but in that instance median prolification also existed.
For my specimens I am indebted to M