The females and males of species often look and act different from each other. These differences are called sexual dimorphism, and when the sexes differ in color we call them dichromatic. Dichromatism in animals has long fascinated naturalists and ecologists alike, and Charles Darwin spend considerable time trying to understand how and why it evolves. Sex differences such as the striking dichromatism in butterflies and birds provided a major challenge for Darwin’s earlier work on natural selection, and prompted him to extend his evolutionary theory with the force of sexual selection. This addition argued that females often prefer to mate with more conspicuous males causing the color of males to evolve away from females.
However, the contemporary co-discoverer of natural selection, Alfred Russel Wallace, disagreed. In his book “Darwinism”, Wallace argued that dichromatism can be the result of natural selection alone, without the need to invoke sexual selection. According to Wallace, females are often more at risk of predation, as they spend more time brooding or carrying heavy eggs. Natural selection should therefore cause females to become more cryptic, and thus their color should evolve away from conspicuous males.
Wallace and Darwin discussed these ideas with explicit examples in mind. In particular, dichromatic butterflies and birds fueled their debate about the evolutionary origins of dichromatism. However, only recently have their competing ideas been put to formal testing in a range of studies investigating color evolution in birds. Surprisingly, even though sexual selection is now a widely accepted part of evolutionary theory, dichromatism in birds commonly follows the Wallacean model. However, outside of a handful of butterfly species, we still did not know the relative importance of these two models in for butterflies in general.
Chris Wheat thought that this question could be answered using a dataset of scanned pictures of all European butterflies from the field guide by Tolman & Lewington. Dirk Zeuss from the University of Marburg had previously used these drawings to analyze the relationship between climate temperature and the lightness of insect body coloration. Realizing that by combining this data with a state-of-the-art phylogeny from the group of Niklas Wahlberg at Lund University, Wheat thought the stage could be set for a detailed look at the evolutionary history of European butterfly colors, and the origins of dichromatism.
Dorsal wing color by sex of European butterflies. (a) depicts the phylogenetic relationships between species. (b) & (h) illustrate the color profiles of each sex, respectively, for each species as the fraction of pixels in each of 50 color clusters. (c) & (g) give the female and male color centroids (average color). (e) plots the distance between the female and male color centroids, used as the metric for dichromatism. (d) & (f) show exemplar wings for males and females of the species that are connected by lines to the other panels. From top to bottom: Erebia rhodopensis, Erebia pharte, Pseudochazara anthelia, Hipparchia semele, Coenonympha glycerion, Melitaea diamina, Polygonia c‐album, Boloria eunomia, Polyommatus damon, Freyeria trochylus, Glaucopsyche alexis, Lycaena hippothoe, Anthocharis cardamines, Colias hyale, and Zerynthia polyxena. Photo credit: Kalle Tunström.
Chris asked me to join the project, and during the summer of 2018 I worked on analyses to contrast the predictions from Wallace and Darwin. I quickly discovered that the color of males has evolved faster than that of females. This is expected under the Darwinian model, and such a pattern is quite often used to argue that dichromatism is the result of sexual selection leading to the divergence of males from females. However, we think that such a conclusion is premature as other processes may generate such results. For example, it could be the case that once species are dichromatic there is a release of evolutionary constraint (due to sexual conflict), which would then allow male color to evolve faster. We therefore needed to test the hypotheses much more directly.
Earlier that year, the group of Pascale Raia published a new method to study the evolution of traits along a phylogeny based on ridge regression. This new method made it easy to not only calculate evolutionary rates, but also to estimate the direction the trait was changing in. I then realized that we could use these evolutionary vectors to give us a much better view of what was exactly happening to male and female color when dichromatism evolved. By looking at the speed and direction of color evolution at those parts of the phylogeny where dichromatism is changing most, we could assess whether changes in males were more important. I found the contribution of male color evolution to dichromatism outweighs the female contribution around 2:1.
We conclude that directional selection on males is the most common cause of butterfly dichromatism, and the Darwinian model is much better supported than the Wallacean model in European butterflies. This likely means that sexual selection is playing an important role, although our data did not allow us to test that directly. We hope that the methods we used here will prove useful for the study of dimorphism in other taxa and different traits.