The visual ecology of a color polymorphic reef fish: the role of aggressive mimicry

Abstract

24Since all forms of mimicry are based on perceptual deception, the sensory ecology of the 25 intended receiver is of paramount importance to test the necessary precondition for mimicry 26 to occur, i.e. model-mimic misidentification, and to gain insight in the origin and evolutionary 27 trajectory of the signals. Here we test the potential for aggressive mimicry by a group of coral 28 reef fishes, the color polymorphic Hypoplectrus hamlets, from the point of view of their most 29 common prey, small epibenthic gobies and mysid shrimp. We build visual models based on 30 the visual pigments and spatial resolution of the prey, the underwater light spectrum and 31 color reflectances of putative models and their hamlet mimics. Our results are consistent with 32 one mimic-model relationship between the butter hamlet H. unicolor and its model the 33 butterflyfish Chaetodon capistratus but do not support a second proposed mimic-model pair 34 between the black hamlet H. nigricans and the dusky damselfish Stegastes adustus. We discuss 35 our results in the context of color morphs divergence in the Hypoplectrus species radiation 36 and suggest that aggressive mimicry in H. unicolor might have originated in the context of 37 protective (Batesian) mimicry by the hamlet from its fish predators. 38 39 40 41 42 43 44 45 46 48 unequivocal identification of new cases of mimicry, their evolutionary dynamics and the very 49 definition and boundaries of the concept of mimicry are still challenging and hotly debated 50 issues among evolutionary biologists (e.g. Vane-Wright 1980; Moynihan 1981; Ruxton et al. 51 2004; Rainey and Grether 2007; Wickler 2013; Dalziell and Welbergen 2016). A key 52realization that has emerged from the ongoing debate was that the perceptual system of the 53 signal receiver must be put at the center of any analysis on the origins and maintenance of a 54 mimicry system (Cuthill and Bennett 1993; Dittrich et al. 1993). Indeed, testing hypotheses 55 of mimic-model relationships can be misleading without an appropriate eye-of-the-beholder 56 approach (Dittrich et al. 1993). This is because the evolution of a mimic signal is shaped not 57 by similarity with the model but by the receiver's percepts of both the signals from model 58 and mimic. Reducing the difference between those percepts below the receiver's threshold 59 for detecting a just noticeable difference (sensu Fechner 1860) will ensure perfect mimicry. 60 It follows that high fidelity might not be the most important requirement for efficient 61 mimicry. Cognitive processes such as generalization (Darst and Cummings 2006; Ham et al. 62 2006), categorization (Chittka and Osorio 2007) and overshadowing (Mackintosh 1976), by 63 which a conflict in the perception of multiple cues leads to only a subset of characteristics of 64 the signal being considered at the expense of others, are likely to affect mimetic accuracy. By 65acting on the receiver percept of the mimic phenotype and not on the phenotype itself, 66 selection will frequently affect only a limited ...