Biology Reference
In-Depth Information
computing residuals from a regression. What the phylogenetic comparative methods
correct is the type I error rate
the probability of incorrectly favoring the alternative
hypothesis over a true null hypothesis. That correction insures that when we claim the
traits we hypothesize to be integrated or otherwise constrained have evolved in concert,
they actually exhibit greater covariation than expected by chance. Further assurance can
be obtained by computing the K statistic proposed by
Blomberg et al. (2003)
, or the per-
mutation test presented by
Klingenberg and Gidaszewski (2010)
, to test explicitly that the
trait distribution is more congruent with the phylogeny than would be expected of a
randomly evolving trait. The phylogenetic comparative methods may not be able to correct
our inferences regarding the slope of that relationship, but they do provide firmer grounds
for claiming that such relationships exist. In this very specific sense, they can and do give
us a better basis for testing hypotheses of historical or phylogenetic constraint, but only if
we keep in mind that there is much more to those hypotheses than the similarity of sister
taxa. Congruence with phylogeny that is consistent with a model of random evolution is
no more evidence of constraint than it is evidence of adaptation.
EVOLUTIONARY ALLOMETRY
Gould (1966)
characterized allometry as “the study of size and its consequences”. This
description may seem rather extravagant, but does capture the importance of size for
many aspects of biology. Many studies of allometry have investigated the influence of size
on ecological role or functional performance, sometimes finding complex relationships
that produce unexpected results. For example, a bigger snake may be able to eat absolutely
bigger fish, but changes in jaw proportions that allow snakes to catch bigger and faster
fish may also force them to choose relatively smaller ones. Conversely, a different jaw
allometry that allows snakes to eat relatively larger prey might alter their strike mechanics
and force them to switch to prey that are also relatively slower. Because geometric
morphometrics is able to partition morphology into independent size and shape com-
ponents, it is able to provide more direct answers to questions about relationships of
size and shape than could be extracted from older methods.
Questions about the evolutionary role of morphological allometry concern many
more topics than its influence on mechanical functions or ecological interactions. One of
these questions is what proportion of shape variation is correlated with size? This may
seem very specific and narrow, but the answer may determine the ability of the allometry
to constrain evolutionary change. The long-term stability of allometric patterns may also
be a factor in their role as constraints. A closely related question is whether allometric
patterns provide an evolutionary line of least resistance, a notion predicated on their
potential to predict or constrain responses to selection (
Schluter, 1996; Marroig and
Cheverud, 2005
). The converse question is whether selection can change allometric pat-
terns to direct evolution along a new axis. Dynamic evolution of allometric patterns
could play an important role in the rapid morphological diversification of an adaptive
radiation.
Gould's expansive representation of the importance of allometry should also serve as a
reminder that not all allometric studies are concerned with the relationship of size and
