Introduction:
Our contemporary understanding of the evolution of bright and conspicuous color patterns is rooted in the work of the early pioneers of evolutionary themes such as natural selection. Charles Darwin developed the theory of sexual selection to explain the presence of conspicuous ornamentation, but realised it could not account for the presence of bright colors in non-reproductive Lepidopteran larvae (Darwin 1871). Alfred Russel Wallace, on the other hand, was skeptical of sexual selection and instead built on the work of John Jenner Weir and Henry Walter Bates, to outline a theory of aposematic warning signals, that was later developed further by Edward Bagnall Poulton (Caro 2017; Caro and Ruxton 2019; Marchant 1916; Poulton 1890).
Aposematic and sexually selected color patterns are highly diverse, but such signals are often characterised by high visual contrast both between pattern components within an organism, and to the background against which the organism is viewed (Andersson 1994; Ruxton et al. 2019; Stevens and Ruxton 2012). Brighter and more conspicuous signals are commonly associated with more potent defences and greater reproductive fitness, such that predators are more easily deterred (Aronsson and Gamberale-Stille 2008; Forsman and Herrström 2004; Forsman and Merilaita 1999; Halpin et al. 2020; Prudic et al. 2006; Stevens et al. 2010), rivals are more wary, and potential mates more interested when signals are highly contrasting (Andersson 1994; Endler 1983; Ryan and Keddy-Hector 1992; Svensson and Wong 2011).
High signal contrast can be achieved via two interconnected visual pathways: achromatic contrast (luminance/brightness) and chromatic contrast (hue/saturation). In vertebrates, achromatic contrast is measured as a single intensity value received by longwave sensitive photoreceptors, whereas hue is perceived through opponent processing by two or more photoreceptors that differ in their peak wavelength sensitivity (Kelber and Osorio 2010; Vorobyev and Osorio 1998). Consequently, different colors contribute to phenotypic contrast in different ways: longwave colors (e.g., red, orange, and yellow) contribute to both achromatic and chromatic contrast, whereas shortwave colors (e.g., blue and ultraviolet (UV)) only significantly affect chromatic contrast (Stevens and Ruxton 2012; Umbers 2013).
For this reason, conspicuous signals frequently generate high visual contrast by combining bright long wavelength colors with low luminance black (Stevens and Ruxton 2012). Short wavelength colors, including UV, can also create high contrast and are occasionally incorporated into seemingly conspicuous signals (Umbers 2013). However, evidence for the efficacy of UV in aposematic signals has been mixed, with no compelling confirmation that naturally occurring UV signals are effective at deterring predators despite some evidence that UV signals can be learned (Lyytinen et al. 2001; Werner et al. 2014a; Werner et al. 2014b; Werner et al. 2012). Moreover, rather than preventing attacks UV containing (UV+) signals can instead deflect attacks to more expendable body parts (Olofsson et al. 2010), or they may actually increase predation risk (Lyytinen et al. 2004). Despite the discovery of UV reflectance attracting much attention, perhaps due to our own inability to perceive such signals, we currently lack a complete understanding of if, or to what extent, UV reflectance contributes to aposematic signaling.
In the Neotropics two independent, and completely unrelated, radiations of bright conspicuous colors have drawn much scientific attention: the heliconiian butterflies (Heliconiinae; Nymphalidae) and the poison frogs (Dendrobatidae: Anura). Both groups are found in similar rainforest habitats, are highly toxic, are at risk from similar predatory taxa, and have become renowned for their high diversity of species and bright colors (Merrill et al. 2015; Stynoski et al. 2015).
Despite many similarities in color diversity, chemical defense, the visual environment, and the predator community, ultraviolet reflective colors are relatively common in heliconiian butterflies but seemingly rare in poison frogs (Briscoe et al. 2010; Bybee et al. 2012; Yeager and Barnett 2020). Indeed, we recently described the first example of UV-reflectance in poison frogs, from an Ecuadorian population ofOophaga sylvatica . We found that although UV shows up brightly in photographs, it adds little to internal color pattern contrast (Yeager and Barnett 2020). As coloring in both groups has been selected under the influence of similar UV-sensitive predators for the purpose of aposematism, the lack of UV reflection in frogs is perplexing. Here we expand these previous findings to describe UV reflectance in two more species of poison frog (Ameerega bilinguis and Epipedobates tricolor ). We compare the contribution of UV to signal contrast between these two dendrobatid frogs and five species of heliconiian butterflies, and then discuss the importance of predator versus conspecific vision to point to potential explanations for the evolution of UV reflectance in these groups.