The Fruitful Facets of Galton's Polyhedron 363
then new features will first appear in males, since novelty must be added
terminally under Haeckel's law.
The specification of such channels as constraints on adaptation does not, of
itself, push natural selection to a periphery of unimportance; for, as Darwin argued
in another context (see pp. 251-260), existing channels must have evolved for
some adaptive reason in ancestors—and if selection constructs adaptation, then
Darwinian forces can reclaim their creative power, whatever limits these inherited
channels may place upon current adaptation. In fact, Eimer did toy with the idea
that selection might have built the primary channels—doing so in a burst of
fanciful speculation, much in the mode that he himself would later castigate as the
primary weakness of adaptationist thinking.
For the primary law of longitudinal stripes to spots to transverse stripes in
coloration, for example, Eimer suggested that the initial state might represent an
adaptation of ancestors (under the false assumption that monocots precede dicots in
the geological record of angiosperms): "The fact of the original prevalence of
longitudinal striping might be connected with the original predominance of the
monocotyledonous plants, whose linear organs and linear shadows would have
corresponded with the linear stripes of the animals" (1890, p. 57). Eimer then
extended this speculation by guessing that conversion to spots "might be connected
with the development of a vegetation which cast spotted shadows"; and the final
transition to transverse stripes "with the shadows, for example, of the branches of
woody plants—thus the marking of the wild cat escapes notice among the branches
of trees" (p. 57).
From this phyletic fancy, Eimer moved to more conventional Darwinian
reasons for other channels. Male preponderance "might possibly be explained by
the fact that the males fight the battle of existence more than the females, and
therefore must always be first to respond to new demands" (p. 58). And ornamental
waves moving from posterior to anterior might also gain a selectionist basis "by
the fact that the part of the body farthest from the head is most in need of mimicry,
because it is least protected in other ways by the sense organs, and because it is at a
special disadvantage; that it is the last part to be withdrawn from the pursuit of an
enemy" (p. 58).
But Eimer could not, ultimately, grant even this much power to selection and
adaptation for two major reasons. First, he decided that several of his channels
expressed no evident utility and, even if adaptive in final expression, could not
have possessed any selective value in incipient states (1890, p. 59):
But all this does not explain the first occurrence of the new characters, nor
the undeviating course of the evolution in a particular direction. For when a
number of varying individuals are compared it is seen that the variations of
all tend to a definite end, and that the majority of the intermediate forms
show stages in the development of the characters, which are absolutely
without use to them. This cannot be explained except by