Dennis Paulson wrote:
>Oh boy, Burt, you've got this particular adaptationist excited now. Tell
>me that Ruby-crowned Kinglets evolved an eye ring and Golden-crowned
>Kinglets an eye line just by genetic accident, or that half the vireos
>and warblers have wing bars and half don't just by genetic accident.


The observation that half the warblers and half the vireos have wingbars
whereas the remainders of those groups do not can conceivably be
explained by just a couple of evolutionary events. If an ancestral
species has wing bars, its descendants will have wing bars unless mutants
that lack wing bars appear in a descendant population and selection or
drift take the mutant alleles to fixation (ignoring epigenetic influences
for the sake of simplicity). Suppose the common ancestor of warblers and
vireos had wing bars. If one species of vireo and one species of warbler
lost wingbars, all the descendant species of those ancestral populations
would lack wingbars. Explaining two losses of wingbars as resulting from
drift sounds a lot more plausible than trying to explain independent
losses in dozens of species, which is the assumption implicit in raw
comparisons of how many have them vs. how many do not. And those that do
have wingbars? That may require only one event -- the acquisition of
wingbars in the common ancestor.


To the extent that species share evolutionary histories, they are not
independent of one another. Treating them as though they are is a kind of
"evolutionary pseudoreplication" when comparing how many species have a
trait and how many lack it. There are straightforward methods for
assessing the history of traits, including degree of evolutionary
independence, using hierarchical phylogenies.


DP >Or that some sparrows have white outer tail feathers and some don't,
>again for no particular reason. I don't even have a hypothesis for some
>of the possible functions, but those little markings, recurring again and
>again on birds of different origins but similar life styles--they're
>"good for" something! It costs--biochemically--to make the longer
>feathers of a crest. My working hypothesis is that the crest would
>indeed be selected against if it were of no significance to the bird.

Experiments modifying ornaments such as long tails in various bird species
support the idea that they are selected for by mate choice. An
interesting implication from sexual selection theory is that selection for
such a trait in males alone will nevertheless usually produce the same
trait in females (correlated response to selection on homologous traits).
In such cases, the trait can actually be *non*adaptive for females.
Selection against such traits in females can eventually lead to sexual
dimorphism, but it requires accumulation of mutations that prevent
expression of the trait in females but not in males. (For the sake of
fairness, the existence of nipples in male mammals is probably an example
of a trait selected for in *females* that is simply expressed in males by
default [correlated response])


So crests in birds such as jays may well be the product of sexual
selection on males, even though they are present in both sexes. (Along
the lines of burt Guttman's suggestion, if these crests incur little cost
in females, there will be little pressure for females to "lose" them, and
that slight negative selection will be swamped by sexual selection for the
crest in males, and they could be retained by females indefinitely despite
having no function in females) Finally, groups such as orioles in which
some species have bright plumages in both sexes, some are strongly
dichromatic, and some are weakly dichromatic, may represent cases where
multiple "steps" in the evolution of a sexually selected character have
been preserved because different lineages are evolving (females losing
male trait) at different rates. It's fun stuff, but you really need
phylogenetic evidence to play the game.


David Wright
dwright at u.washington.edu