186 D. M. Walsh
mapping of genotype space into phenotype space. Some changes in genotype space
may correspond to no changes in phenotype space. Small changes in genotype
space may issue in large changes of phenotype space. This phenomenon goes by
the name of ‘developmental constraint’. Developmental constraint is defined as:
... a bias in the production of variant phenotypes or a limit on phe-
notypic variability caused by the structure, character, composition, or
dynamics of the developmental system. [Maynard Smithet al., 1985,
266]
C. H. Waddington [1942; 1957] was perhaps the first to make clear the signif-
icance of developmental constraint for evolution. Waddington drew attention to
three salient features of development. The first is that an organism is made up
of relatively few discrete, stable tissue types, all of which differentiate from a sin-
gle, homogeneous precursor. The second is that the development of phenotype is
heavily buffered. In a wild population, there is little phenotypic variation, despite
an enormous amount of genetic variation and environmental perturbation.^4 The
third is that given sufficient environmental stress, development can be induced
to produce new stable phenotypes without a corresponding change in underlying
genotype.
Quite what the implications of this might be for evolutionary biology has been
an issue of some genuine concern and confusion ([Amundson, 1994]; [Maynard
Smithet al., 1985]; [Gould, 2002]). The most common approach is to suppose that
while selection is the only adaptation-promoting force in evolution, it is to some
degree at the mercy of ontogeny [Goodwin, 1982]. Sometimes development makes
unavailable those phenotypes that, otherwise, selection would promote. Those who
think of development playing primarily this role cast it as a constraint against the
adaptation-promoting powers of selection.
For instance developmental constraints frustrate selection by restrict-
ing the phenotypic variation selection has to act upon. Adaptations
would be able to evolve only to optima within the constrained space
of variability. [Wagner and Altenberg, 1996, 973]The nature of the existing developmental system somehow constrains
or channels acceptable change, so that selection is limited in what it
can achieve given some starting anatomy. [Raff, 1996, 294–295]It doesn’t follow from this that developmental constraint should play only a
minor explanatory role. But it does follow that development and selection com-
pete for explanatory relevance [Walsh, 2003]. Development and selection are both
candidate explanantia of the distribution of biological form. Whereas selection
distributes form in phenotype space by promotingadaptedness, development dis-
tributes form in a way that isindependentof its adaptedness. Adaptedness is the
(^4) The extent of the genetic variation was only realized with the advent of electrophoresis
[Lewontin, 1974].