ABSTRACT
Judging by the volume of writings about evolutionary constraints, they are an important topic in evolutionary biology. However, their involvement in shaping patterns of evolutionary change from morphological stasis to adaptive radiation remains contentious. This is at least in part because of the paucity of robust analyses of potential examples of constraints, whether of a more absolute or a relative nature. Here, we argue that what is needed to explore the type of constraints and bias on evolutionary change that may emerge from the way in which phenotypic variation is generated is an integrative approach applied to systems that can be tackled at different levels of biological organization. This is illustrated using research on the evolution of patterns in butterfly wing eyespots that has applied a combination of evolutionary genetics and evo-devo to an emerging model species with the beginnings of a comparative approach to describe patterns of variability among the extant taxa of two species-rich genera.
Subject(s)
Biological Evolution , Butterflies/anatomy & histology , Animals , Butterflies/genetics , Butterflies/growth & development , Models, Biological , Phenotype , Selection, Genetic , Species Specificity , Wings, Animal/anatomy & histology , Wings, Animal/growth & developmentABSTRACT
Temperature-induced variation and norms of reaction have been analyzed for wing pattern elements of six species belonging to the African butterfly genus Bicyclus (Lepidoptera, Nymphalidae, Satyrinae). Five of these species are sympatric in Malawi and exhibit seasonal polyphenism in the savannah-rainforest ecotone. The sixth species originated from Cameroonian equatorial rainforest. The organisms were laboratory reared under four different temperature conditions ranging from 17-28°C. The variation in response to temperature is described by principal component analysis (PCA). Discrimination on the basis of plastic wing pattern characters was performed by discriminant function analysis (DFA) and unweighed pair-group method algorithm (UPGMA) clustering. A phylogenetic reconstruction based on adaptive plastic wing characters was compared with a cladogram built on "nonadaptive" characters. Results demonstrate that: (1) Phenotypic plasticity of wing pattern characters in response to temperature in laboratory-reared organisms is reminiscent of variation induced by seasonal change in the field. (2) Different wing pattern characters are under different control: "exposed" characters of butterflies at rest position are highly sensitive to temperature variation, whereas "hidden" characters, only visible during active behavior, are dominated by species differences. In general the sensitivity of the former can be attributed to their proposed function in deflecting predators. (3) The sexes differ especially in the size of those eyespots that are displayed during active behavior. (4) Species from seasonal and aseasonal environments react in a broadly similar manner to temperature variation. However, savannah species and species of aseasonal rainforest exhibit relatively shallow reaction norms, whereas reaction norms are steeper in species from the savannah-rainforest ecotone. Such a strong response was also apparent in so-called correlation networks between principal components for these species. (5) Phylogenetic distances are to some extent reflected in ordination in both PCA-space and DFA-space: closely related species of the safitza group remain close in both ordinations. The more distantly related species differ in ordination from a pattern as suggested by a phylogenetic reconstruction. It is argued that the wing pattern variation of these species reflects both adaptive processes and historical relationships.