d to reduce the amount of pollen, when
rendered by the closure of the flowers superfluous, yet hardly any of
the above special modifications can have been thus determined, but
must have followed from the laws of growth, including the functional
inactivity of parts, during the progress of the reduction of the pollen
and the closure of the flowers.

It is so necessary to appreciate the important effects of the laws of
growth, that I will give some additional cases of another kind, namely
of differences in the same part or organ, due to differences in relative
position on the same plant. In the Spanish chestnut, and in certain
fir-trees, the angles of divergence of the leaves differ, according to
Schacht, in the nearly horizontal and in the upright branches. In the
common rue and some other plants, one flower, usually the central or
terminal one, opens first, and has five sepals and petals, and five
divisions to the ovarium; while all the other flowers on the plant are
tetramerous. In the British Adoxa the uppermost flower generally has
two calyx-lobes with the other organs tetramerous, while the surrounding
flowers generally have three calyx-lobes with the other organs
pentamerous. In many Compositae and Umbelliferae (and in some other
plants) the circumferential flowers have their corollas much more
developed than those of the centre; and this seems often connected with
the abortion of the reproductive organs. It is a more curious fact,
previously referred to, that the achenes or seeds of the circumference
and centre sometimes differ greatly in form, colour and other
characters. In Carthamus and some other Compositae the central achenes
alone are furnished with a pappus; and in Hyoseris the same head yields
achenes of three different forms. In certain Umbelliferae the exterior
seeds, according to Tausch, are orthospermous, and the central one
coelospermous, and this is a character which was considered by De
Candolle to be in other species of the highest systematic importance.
Professor Braun mentions a Fumariaceous genus, in which the flowers in
the lower part of the spike bear oval, ribbed, one-seeded nutlets; and
in the upper part of the spike, lanceolate, two-valved and two-seeded
siliques. In these several cases, with the exception of that of the
well-developed ray-florets, which are of service in making the flowers
conspicuous to insects, natural selection cannot, as far as we can
judge, have come into play, or only