The recent discussion by Ben Jones of the value of so-called
less than perfect phenotypes uses the evolution of flight as
an example where incremental changes towards a flight mechanism
can be incrementally more fit.
While this may be true for big, furry animals, its probably not
true for insects. The work of Joel Kingsolver and Mimi Cole has
shown in an elegant fashion that the incremental addition of
length to pronotal lobes in proto-flying insects actually has
no beneficial areodynamic effects. Thus, the insect would have
no incremental gain in fitness via flight by incremental lengthening of
its proto-wings until they had become large enough to actually
affect its gliding ability.
The flying-squirrel hypothesis for the evolution of insect flight
does not appear possible.
What could have happened, is selection for larger
and larger surfaces for thermodynamic control of body temperature
(large surfaces for basking) which were then coopted for flight
once they were large enough.
The lesson from this is that small, incremental changes in phenotype
toward an optimum may be the most obvious path, but might
not be biomechanically or physiologicaly possible and may
have arisen through various
mechanisms at various times in its evolutionary development.
(Sorry, I can't lay my hands of the references to Cole and Kingsolver's
work at the moment. I'll keep looking.)
thomasb at spider.ento.csiro.au