arrest of
vegetative growth, due to the antagonism of reproduction. If this were
the whole explanation of the phenomenon, we should expect the quantity
of seed to increase as this vegetative growth diminished, since the seed
is the product of the reproductive energy of the plant, and its quantity
the best measure of that energy. But is this the case? The ranunculus
has comparatively few seeds, and the flowers are not numerous; while in
the same order the larkspur and the columbine have far more seeds as
well as more flowers, but there is no shortening of the raceme or
diminution of the foliage, although the flowers are large and complex.
So, the extremely shortened and compressed flower-heads of the
compositae produce comparatively few seeds--one only to each flower;
while the foxglove, with its long spike of showy flowers, produces an
enormous number.

Again, if the shortening of the central axis in the successive stages of
hypogynous, perigynous, and epigynous flowers were an indication of
preponderant reproduction and diminished vegetation, we should find
everywhere some clear indications of this fact. The plants with
hypogynous flowers should, as a rule, have less seed and more vigorous
and abundant foliage than those at the other extreme with epigynous
flowers. But the hypogynous poppies, pinks, and St. John's worts have
abundance of seed and rather scanty foliage; while the epigynous
dogwoods and honeysuckles have few seeds and abundant foliage. If,
instead of the number of the seeds, we take the size of the fruit as an
indication of reproductive energy, we find this at a maximum in the
gourd family, yet their rapid and luxuriant growth shows no diminution
of vegetative power. So that the statement that plant modifications
proceed "along an absolute groove of progressive change" is contradicted
by innumerable facts indicating advance and regression, improvement or
degradation, according as the ever-changing environment renders one form
more advantageous than the other. As one instance I may mention the
Anonaceae or custard-apple tribe, which are certainly an advance from
the Ranunculaceae; yet in the genus Polyalthea the fruit consists of a
number of separate carpels, each borne on a long stalk, as if reverting
to the primitive stalked carpellary leaves.

_On the Origin of Spines._

But perhaps the most extraordinary application of the theory is that
which considers spines to be an indication of the "ebbing