188 D. M. Walsh
caterpillars imprint on their hosts; they lay their eggs on the type of plant on
which they hatch. Suppose an egg on the host plant gets blown onto another type
of plant, one that has until now never served as a host. This plant, as it turns
out, is a beneficial host and the caterpillar that hatches on it has a considerable
increase in fitness on account of this alone. There is now variation in fitness in this
population, and it is heritable, as our new ‘lucky’ butterfly will lay her eggs on
the newly encountered host. Because of its intergenerational stability, this variant
phenotype is selectable. Adaptive evolution can occur in which this trait increases
in prevalence because of its contribution to survival and reproduction. Yetex
hypothesithere is no change or variation in genetic material and no replication of
the inherited trait. This form of inheritance is not inconsistent with the Mendelian
mechanism — but it doesn’t instantiate it — nor does it violate the Weissman
doctrine of sequestration. Nevertheless, we have a change in a developmental
process that is inheritable, selectable and adaptive. Mendelism and Weissmanism
are sort of an irrelevance.
This suggests that the fragmentation of evolution is in part mistaken. There
may well be discernible Mendelian, Darwinian and Weissmanian (developmental)
processes, but it doesn’t follow that inheritance and the generation of selectable
variants is the domain of the Mendelian process alone. Development can generate
and preserve selectable variants too. This insight motivates the second grade of
ontogenetic involvement.
4.2 Grade II: Developmental Processes as Selecta
The lucky butterfly scenario brings into relief the importance of a biological tru-
ism: phenotypes are not the consequence of replicators alone. Genes may exert
some control over phenotype, but they do not do so in glorious isolation; it takes
genes and environments to make a phenotype. The environment constitutes a
resource for the developing organism. This isn’t news, of course, and biologists
have long grown accustomed to thinking that despite the truism there is a still a
privileged role for replicators to play in the production of phenotypes. Replica-
tors are crucial difference makers. They are presumed to play a more significant,
or perhaps more reliable, role than environmental factors do in determining that
an organism will have one phenotype rather than some other. But replicators
deserve no such deference. Advocates of ecological developmental biology stress
that ecological conditions are frequently just as important in the determination
of specific phenotypes [Gilbert, 2001]. Sex-determination in reptiles is often trig-
gered environmentally. Polyphenisms are induced by specific environmental cues.
Even heritable traits show this dependence on ecological factors. The pattern of
deposition of bone in vertebrates is both highly heritable and highly sensitive to
the gravitational attraction of the earth. Even heritable traits that differ system-
atically between lineages may be under the control of ecological conditions (e.g.
the lucky butterfly). This is not to say that replicators aren’t crucial difference
makers for inheritable traits; it is simply to insist that in principle, any develop-