22
Lars Chittka and Adriana Briscoe
the photoreceptors according to the same criteria. At the very least it would mean
that there are other, more important, criteria, or that evolutionary constraints
might have hindered the animal from evolving along the same lines as the model
calculations. The fact that the calculations arrive at similar color vision systems as
nature is tempting, but it does not necessarily imply that one has found the
criterion which has driven the evolution of bee color vision. In fact, sets of color
receptors similar to those of bees occur in animals that occupy entirely different
ecological niches.
Finally, close inspection of the model results reveals that the color receptors of
bees are only nearly optimal. For example, the optimum long wave receptor for
coding flowers is at Amax =550 nm, whereas the visual pigments of bee green
receptors are maximally sensitive at 540 nm. Since the modeled optima are fairly
broad, performance of the real receptors is only 2% below the theoretical
optimum for flower coding. This discrepancy is small when considering the large
range over which the receptors were varied, but what causes it? If bee visual
pigments can be freely tuned, why have they not achieved a perfect match with
floral colors? One possibility is that there are tradeoffs with other visual activities.
For example, it has been assumed that the bee green receptor is optimally matched
to green foliage, and might thus serve as a background detector (Menzel 1979);
but leaves also reflect most strongly at 550 nm (Chittka 1996a), so that, again, the
theoretical optimum is at longer wavelengths than the peak sensitivity of the bee
green receptors.
We are confident that there are other adaptive explanations that might be tried,
and eventually, one might be successful: but we also wish to warn that trying a
large number of adaptive explanations can lead to speculation. Gould and
Lewontin (1979) caricaturized this approach in the following terms: "If one
adaptive argument fails, try another" and "In the absence of a good adaptive
argument... attribute it to imperfect understanding of where an organism lives or
what it does." We are sure that some of our readers will recognize their own
thinking in these words. We do not wish to discourage sensory ecologists to
continue searching for adaptive explanations where at present they seem hard to
find. In what follows, however, we list a number of tools that should make this
search less speculative.
3 Phylogenetic Studies
One possible reason why animals in different ecological contexts have conserved
traits is phylogenetic constraint. For example, there is little reward in searching
for the adaptive significance of why bees have six legs, because leg number is
evolutionarily conservative in insects. Mapping traits on an established phylogenetic tree will reveal if the trait is variable within a given taxon, and whether the
search for adaptations will be worth our time (Brooks and McLennan 1991;
Harvey and Pagel 1991 ). It is this simple evolutionary reasoning that is absent in
many studies of sensory ecology. Often, each species was regarded as an entity
Lars Chittka and Adriana Briscoe
the photoreceptors according to the same criteria. At the very least it would mean
that there are other, more important, criteria, or that evolutionary constraints
might have hindered the animal from evolving along the same lines as the model
calculations. The fact that the calculations arrive at similar color vision systems as
nature is tempting, but it does not necessarily imply that one has found the
criterion which has driven the evolution of bee color vision. In fact, sets of color
receptors similar to those of bees occur in animals that occupy entirely different
ecological niches.
Finally, close inspection of the model results reveals that the color receptors of
bees are only nearly optimal. For example, the optimum long wave receptor for
coding flowers is at Amax =550 nm, whereas the visual pigments of bee green
receptors are maximally sensitive at 540 nm. Since the modeled optima are fairly
broad, performance of the real receptors is only 2% below the theoretical
optimum for flower coding. This discrepancy is small when considering the large
range over which the receptors were varied, but what causes it? If bee visual
pigments can be freely tuned, why have they not achieved a perfect match with
floral colors? One possibility is that there are tradeoffs with other visual activities.
For example, it has been assumed that the bee green receptor is optimally matched
to green foliage, and might thus serve as a background detector (Menzel 1979);
but leaves also reflect most strongly at 550 nm (Chittka 1996a), so that, again, the
theoretical optimum is at longer wavelengths than the peak sensitivity of the bee
green receptors.
We are confident that there are other adaptive explanations that might be tried,
and eventually, one might be successful: but we also wish to warn that trying a
large number of adaptive explanations can lead to speculation. Gould and
Lewontin (1979) caricaturized this approach in the following terms: "If one
adaptive argument fails, try another" and "In the absence of a good adaptive
argument... attribute it to imperfect understanding of where an organism lives or
what it does." We are sure that some of our readers will recognize their own
thinking in these words. We do not wish to discourage sensory ecologists to
continue searching for adaptive explanations where at present they seem hard to
find. In what follows, however, we list a number of tools that should make this
search less speculative.
3 Phylogenetic Studies
One possible reason why animals in different ecological contexts have conserved
traits is phylogenetic constraint. For example, there is little reward in searching
for the adaptive significance of why bees have six legs, because leg number is
evolutionarily conservative in insects. Mapping traits on an established phylogenetic tree will reveal if the trait is variable within a given taxon, and whether the
search for adaptations will be worth our time (Brooks and McLennan 1991;
Harvey and Pagel 1991 ). It is this simple evolutionary reasoning that is absent in
many studies of sensory ecology. Often, each species was regarded as an entity