From cryptic to colorful: Evolutionary decoupling of larval and adult color in butterflies

Abstract

Many animals undergo complete metamorphosis, where larval forms change abruptly in adulthood. Color change during ontogeny is common, but there is little understanding of evolutionary patterns in these changes. Here, we use data on larval and adult color for 246 butterfly species (61% of all species in Australia) to test whether the evolution of color is coupled between life stages. We show that adults are more variable in color across species than caterpillars and that male adult color has lower phylogenetic signal. These results suggest that sexual selection is driving color diversity in male adult butterflies at a broad scale. Moreover, color similarities between species at the larval stage do not predict color similarities at the adult stage, indicating that color evolution is decoupled between young and adult forms. Most species transition from cryptic coloration as caterpillars to conspicuous coloration as adults, but even species with conspicuous caterpillars change to different conspicuous colors as adults. The use of high-contrast coloration is correlated with body size in caterpillars but not adults. Taken together, our results suggest a change in the relative importance of different selective pressures at different life stages, resulting in the evolutionary decoupling of coloration through ontogeny.

Keywords: Butterflies; caterpillars; color; coupled; ontogenetic.

Figure 1

Phylogenetic signal in coloration in Australian butterflies and caterpillars. (A) Maximum clade credibility (MCC) tree showing the primary color (the brightest if there were more than one) in caterpillars (inner circle) and adults (outer circle) for species sampled with phylogenetic information (N = 98 spp.). (B) Observed differences between the male upperside phylogenetic signal and male underside (blue), caterpillar (red), and female (light and dark gray) for each of the 2500 different trees. For all trees, the difference is higher than zero, suggesting that male upperside color has the lowest phylogenetic signal regardless of the underlying phylogenetic relationships.

Figure 2

(A) Color space for caterpillars and male upperside coloration across Australian butterflies. Each point represents a primary color (e.g., occupying more than 30% of the wing or caterpillar body). Lines connect colors present as primary colors in the same species. (B) Histograms of randomly expected differences in color diversity (D _dif) between caterpillars and male adults (upperside) using three different color distance metrics (A, B, C described in Supporting Material). Arrows indicate the observed D value (all P < 0.05). Values on the left of the random distribution indicate that male adult butterflies are more diverse in coloration than caterpillars across species (ontogenetic divergence). Complete table of results presented in Table S2.

Figure 3

Color strategies in caterpillars and adults. (A) Pie charts showing the percentage of species that employ either of the three anti‐predatory color strategies. (B) Correlation between contrasting colors (the brightest color) used by caterpillars and their adults. If the colors used by the caterpillar and the adult in the same species are the same, then they would follow the diagonal path shown by the dotted line. Only species where both adults and caterpillars have contrasting colorations are shown. (C) Association between caterpillar size and internal contrast in a subset of species (N = 48) for which information on caterpillar body size was available.