Predator selection on phenotypic variability of cryptic and aposematic moths

Abstract

Natural selection generally favours phenotypic variability in camouflaged organisms, whereas aposematic organisms are expected to evolve a more uniform warning coloration. However, no comprehensive analysis of the phenotypic consequences of predator selection in aposematic and cryptic species exists. Using state-of-the-art image analysis, we examine 2800 wing images of 82 moth species accessed via three online museum databases. We test whether anti-predator strategy (i.e., camouflage or aposematism) explains intraspecific variation in wing colour and pattern across northern hemisphere moths. In addition, we test two mutually non-exclusive, ecological hypotheses to explain variation in colour pattern: diel-activity or dietary-niche. In this work, taking into account phylogenetic relationships, moth phenotypic variability is best explained by anti-predator strategy with camouflaged moths being more variable in wing patterning than aposematic species.

Fig. 1. The species assembly of this work illustrating between-species differences in moth appearance.

The moth images exemplify interspecific phenotypic variation across moth anti-predator strategies, diel-activity and dietary preference as flagged by the legend. Images are not normalised nor in actual scale. The first half on the left side shows species that are considered aposematic in literature and the second half on the right shows species that are broadly considered camouflaged. Two species are not shown for artistic reasons.

Fig. 2. The moth assembly illustrating some examples of moth interspecific (rows) versus intraspecific (columns) phenotypic variation between aposematic and camouflage anti-predator strategies.

Here, six species are shown. Notice how the marking size of the pattern characteristically varies more in camouflaged than aposematic species, albeit there is phenotypic variation across all moths. The figure highlights the care required to make helpful interpretations of phenotypic differences using between-species (i.e., interspecific) versus within-species (i.e., intraspecific) data, which characterise different biological hierarchies.

Fig. 3. An ultrametric tree showing the phylogenetic relationships of the moth species used in this study.

The legends annotate the alternative ecological hypotheses predicting their appearance as: anti-predator strategy (aposematism, camouflage), diel-activity (cathemeral, diurnal, nocturnal) and dietary preference (monophagous, mycophagous, oligophagous, polyphagous).

Fig. 4. A heatmap plot that summarises significant associations of the moth colour and pattern as regards the alternative ecological hypotheses predicting their appearance.

On the x axis values stands for forewing and hindwing, x for mean (i.e., interspecific variation) and cv for the coefficient of variation (i.e., intraspecific variation). On the y axis, different pattern and colour metrics are compared against their alternative hypotheses for moth phenotypic variability: anti-predator strategy, diel-activity and dietary-niche. The panel shows Eta Squared statistics for effect sizes and flags significant P values obtained from the phyloANOVA analysis (i.e., the higher η2-values and lower P values indicate higher statistical significance). All associated test statistics are based on two-tailed significance levels.

Fig. 5. The primary variables of interest highlighting within-species phenotypic variability through variation of coefficients (cv).

A shows case examples of different camouflaged and aposematic moth species and their characterisations. Boxplots are organised in descending order with respect to statistical significance and separate anti-predator strategy (B) and moth diel-activity (C). The boxplot shows minimum and maximum (whiskers), the median line and the interquartile range. Wing pattern values are shown for fore- and hindwings. B, C show the data of n = 2800 wing images of 82 species collated under respective anti-predator strategy (B) and diel-activity (C).