||Plastic phenotypes are expected to be favoured in heterogeneous environments
compared with stable environments. Sensory systems are interesting
to test this theory because they are costly to produce and support, and
strong fitness costs are expected if they are not tuned to the local environment.
Consistently, the visual system of several species changes with the
conditions experienced during early development. However, there is little
information on whether the amplitude of the change, that is the reaction
norm, differs between visual environments. Given the rapid change of many
ecosystems, especially eutrophication for aquatic habitats, it is crucial to
determine down to which spatial scale, change in the reaction norm occurs.
We addressed this issue by quantifying the between-habitat variation in the
expression of a UV-sensitive opsin in a newt. In western France, this species
breeds in ponds of small forest patches, where water filters out UV, and in
agricultural ponds where UV transmission is variable. We raised larvae from
both habitats with or without exposure to UV. Opsin expression was
reduced in larvae from agricultural habitats when raised without UV,
whereas it was low in larvae from forest ponds under all lighting conditions.
Thus, the variation in the reaction norm of opsin expression was lower in
stable filtering environments and higher in heterogeneous environments. Its
variation occurred between habitats across a small spatial scale. We discuss
the hypotheses for this pattern and for the maintenance of residual opsin
expression in forest populations.