Changes in predator community structure shifts the efficacy of two warning signals in Arctiid moths

Abstract

Summary1. Polymorphism in warning coloration is puzzling because positive frequency-dependent selection by predators is expected to promote monomorphic warning signals in defended prey. 2. We studied predation on the warning-coloured wood tiger moth (Parasemia plantaginis) by using artificial prey resembling white and yellow male colour morphs in five separate populations with different naturally occurring morph frequencies. 3. We tested whether predation favours one of the colour morphs over the other and whether that is influenced either by local, natural colour morph frequencies or predator community composition. 4. We found that yellow specimens were attacked less than white ones regardless of the local frequency of the morphs indicating frequency-independent selection, but predation did depend on predator community composition: yellows suffered less attacks when Paridae were abundant, whereas whites suffered less attacks when Prunellidae were abundant. 5. Our results suggest that spatial heterogeneity in predator community composition can generate a geographical mosaic of selection facilitating the evolution of polymorphic warning signals. This is the first time this mechanism gains experimental support. Altogether, this study sheds light on the evolution of adaptive coloration in heterogeneous environments.

Experimental evidence suggests that specular reflectance and glossy appearance help amplify warning signals

“…It is therefore possible that variations in predator communities maintain the dull beetles found within certain Oreina populations. Indeed recent work by Valkonen et al 43 and Nokelainen et al 44 both show that variation in predator communities can play an important role maintaining the signal diversity in aposematic species as theorized and modelled by Endler and Mappes 45 .…”

Avoidance of an aposematically coloured butterfly by wild birds in a tropical forest

“…Responses from captive birds might be influenced by their appetite (Sandre et al , 2010), food deprivation, and artificial environments with constrained viewing, whereas natural environments are heterogeneous and offer a wider variety of alternative food, which might alter decision‐making strategies. For example, a study with natural bird populations using artificial models of the wood tiger moth [ Parasemia plantaginis (Linnaeus)] suggested that spatial heterogeneity in a predator community creates a mosaic of selection facilitating polymorphism (Nokelainen et al , 2013). Also, another study with wild birds showed that achromatic (non‐coloured) Heliconius models were attacked significantly more than coloured models of a local pattern, demonstrating the importance of aposematic signals in avoiding predation (Finkbeiner et al , 2014).…”

Does spatial variation in predation pressure modulate selection for aposematism?

“…The current study revealed a tremendous variation in predation pressure (30%–80% attacked) across a large geographic area in South and East Asia, which is also supported by global patterns of predation gradients (Roslin et al., 2017). Local variations in predator community composition and differences in predation pressure have also been shown to be an important driver of diversity in aposematic signals (Mochida, 2011; Nokelainen et al., 2014; Valkonen et al., 2012). Theoretical modeling (Endler & Mappes, 2004) and field experiments with artificial prey (Willink, 2014) have shown that a predator's willingness to attack different colors or conspicuousness may vary leading to variations in warning signal.…”

“…Similar methods have been applied to estimate predation pressure on resting lepidopteran, such as tiger moth where the species is immobile during daytime (Nokelainen et al., 2014). However, two‐thirds of attacks on butterflies occur when they are in flight and the remainder of the attacks occur when they are at rest where the wings are closed most of the time.…”

“…Seasonal variation in predator communities has been shown to impose seasonally dependent selection pressure on aposematic signals of varying conspicuousness (Mappes, Kokko, Ojala, & Lindström, 2014). As aposematic and mimetic signals occur over large geographic areas (Davis Rabosky et al., 2016), spatial variation in the abundance and identity of predators could result in corresponding changes in predation pressure (Nokelainen, Valkonen, Lindstedt, & Mappes, 2014; Valkonen et al., 2012). Hence, we proposed the hypothesis that spatial variation in predation pressure drives conspicuousness‐dependent spatial variation in aposematic and mimetic signal fitness (Figure S1a, b).…”