Why Sensory Ecology Needs to Become More Evolutionary
33
For such experiments, it is critical that animals be tested in environments which
are realistic in terms of food distribution, predator and mate density, and abiotic
parameters (Endler 1980). If we test the fitness of bearers of different color vision
systems, e.g., bees with different color receptors, in an environment where
detectability of flowers is not a limiting factor to search time, or where
identification of the most rewarding flowers is not critical, then it is likely that we
will find no fitness differences.
In this sense, pollinators such as bumble bees and solitary bees are ideal subjects
because we need not worry about "building" an environment: they can forage in
natural arrays of flowers, but because they raise all their offspring in one place,
we can readily evaluate their fitness. The second prerequisite for fitness tests is
that we fmd variance between individuals of a species in sensory traits. If there is
such variation, we can either use naturally occurring phenotypes for fitness tests,
or use strongly deviant phenotypes created by selection experiments.
If we do not fmd differences between individuals, we might consider creating
phenotypes by experimental manipulation. This may be difficult in color vision
systems (we cannot selectively paint over only one type of color receptor, or alter
its spectral sensitivity), but in other sensory systems such manipulations appear
feasible. An elegant alternative may be to test organisms in which mutants for
sensory traits are available, such as Drosophila melanogaster. It is also feasible to
create transgenic Drosophila with a modified repertoire of visual pigments (Britt
et al. 1993), and to measure their fitness. Using these techniques, Drosophila has
become a model to study the receptor-neuronal control of insect behavior; but
these tools have not yet been used to study the role of such mechanisms for
survival in nature. The difficulty, in this case, is to create an environment for
Drosophila flies that is natural enough to allow for realistic tests of fitness.
Unfortunately, however, we have not yet succeeded in creating transgenic animals
more amenable to realistic fitness tests, such as bees.
10 Conclusion
We do not wish to abandon the notion that sensory systems are adaptive. Instead,
we want sensory ecologists to consider constraint, evolutionary inertia, and
random processes as possible alternatives to adaptive explanations, not to replace
an adaptive scenario entirely. We are also emphasizing that if we want to
demonstrate adaptiveness, it is not sufficient to show that sensory traits appear
well (or even optimally) matched to the environment. Instead, we must show that
animals carrying the sensory characteristics in question are fitter than those that
do not. To do this, we must exploit heritable variation in natural populations,
create new phenotypes by manipulation of existing traits or selection experiments,
or use transgenic animals. Using such methodology, we may even eventually
understand why bees have the color receptors they do, why most bee species lack
red receptors, and why some other insects with entirely different life-styles have
color receptors similar to bees.
33
For such experiments, it is critical that animals be tested in environments which
are realistic in terms of food distribution, predator and mate density, and abiotic
parameters (Endler 1980). If we test the fitness of bearers of different color vision
systems, e.g., bees with different color receptors, in an environment where
detectability of flowers is not a limiting factor to search time, or where
identification of the most rewarding flowers is not critical, then it is likely that we
will find no fitness differences.
In this sense, pollinators such as bumble bees and solitary bees are ideal subjects
because we need not worry about "building" an environment: they can forage in
natural arrays of flowers, but because they raise all their offspring in one place,
we can readily evaluate their fitness. The second prerequisite for fitness tests is
that we fmd variance between individuals of a species in sensory traits. If there is
such variation, we can either use naturally occurring phenotypes for fitness tests,
or use strongly deviant phenotypes created by selection experiments.
If we do not fmd differences between individuals, we might consider creating
phenotypes by experimental manipulation. This may be difficult in color vision
systems (we cannot selectively paint over only one type of color receptor, or alter
its spectral sensitivity), but in other sensory systems such manipulations appear
feasible. An elegant alternative may be to test organisms in which mutants for
sensory traits are available, such as Drosophila melanogaster. It is also feasible to
create transgenic Drosophila with a modified repertoire of visual pigments (Britt
et al. 1993), and to measure their fitness. Using these techniques, Drosophila has
become a model to study the receptor-neuronal control of insect behavior; but
these tools have not yet been used to study the role of such mechanisms for
survival in nature. The difficulty, in this case, is to create an environment for
Drosophila flies that is natural enough to allow for realistic tests of fitness.
Unfortunately, however, we have not yet succeeded in creating transgenic animals
more amenable to realistic fitness tests, such as bees.
10 Conclusion
We do not wish to abandon the notion that sensory systems are adaptive. Instead,
we want sensory ecologists to consider constraint, evolutionary inertia, and
random processes as possible alternatives to adaptive explanations, not to replace
an adaptive scenario entirely. We are also emphasizing that if we want to
demonstrate adaptiveness, it is not sufficient to show that sensory traits appear
well (or even optimally) matched to the environment. Instead, we must show that
animals carrying the sensory characteristics in question are fitter than those that
do not. To do this, we must exploit heritable variation in natural populations,
create new phenotypes by manipulation of existing traits or selection experiments,
or use transgenic animals. Using such methodology, we may even eventually
understand why bees have the color receptors they do, why most bee species lack
red receptors, and why some other insects with entirely different life-styles have
color receptors similar to bees.