portant adaptations by means of one heterochronic alteration, as neoteny in
descendant Gryphaea species of the English Jurassic produces shells of both
markedly increased size (by retention of juvenile growth rates over an unchanged
lifetime) and stabilized shape to prevent foundering in muddy environments
(achieved by "bringing forward" the proportions of attached juveniles into the
unattached stage of adult ontogeny); second, an illustration of pervasiveness and
equal (or greater) power than selective forces (to exemplify the strength and high
relative frequency of such positive influences), as geographic variation of the type
species, Cerion uva, on Aruba, Bonaire, and Curasao, a subject of intense
quantitative study and disagreement in the past, becomes resolved in multivariate
terms, with clear distinction between local adaptive differences and the pervasive
general pattern of an extensive suite of automatic sequelae, generated by
nonadaptive variation in the geometry of coiling a continuous tube, under definite
allometric regularities for the genus, around an axis.
- For the second, or conceptually positive, meaning of constraint as a term
for nonstandard causes of evolutionary change, I present a model that compares the
conventional outcomes of direct natural selection, leading to local adaptation, with
two sources that can also yield adaptive results, but for reasons of channeling by
internal constraints rather than by direct construction under external forces of
natural selection. In this triangular model for aptive structures, the functional
vertex represents features conventionally built by natural selection for current
utilities. At the historical vertex, currently aptive features probably originated for
conventionally adaptive reasons in distant ancestors; but these features are now
developmentally channeled as homologies that constrain and positively direct both
patterns of immediate change and the inhomogeneous occupation of morphospace
(especially as indicated by "deep homologies" of retained developmental patterns
among phyla that diverged from common ancestry more than 500 million years
ago). At the structural vertex, two very different reasons underlie the origin of
potentially aptive features for initially nonadaptive reasons: physical principles that
build "good" form by the direct action of physical laws upon plastic material (as in
D'Arcy Thompson's theory of form), and architectural sequelae (spandrels) that
arise as nonadaptive consequences of other features, and then become available for
later cooptation (as exaptations) to aptive ends in descendant taxa. These two
structural reasons differ strongly in the ahis-toricist implications of direct physical
production independent of phyletic context vs. the explicit historical analysis
needed to identify the particular foundation for the origin of spandrels in any
individual lineage. - As a conceptual basis for understanding the importance of recent advances
in evo-devo (the study of the evolution of development), the largely unknown
history of debate about categories of homology, particularly the distinction
between convergence and parallelism, provides our best ordering device—for we
then learn to recognize the key contrast between parallelism as a positive deep
constraint of homology in underlying generators (and therefore as a structuralist
theme in evolution) and convergence as the oppo-