Abstracts (first author)
Batesian mimicry, morphospace occupancy, and the shaping of warning signal diversity in butterfly communities
Mimicry as a defensive strategy is one of the most compelling example of adaptation. Mimicry communities often involve numerous species, and both mutualistic (Müllerian) and deceptive (Batesian) mimics coexist. In deceptive mimicry, palatable prey mimic unprofitable species (e.g. chemically defended prey which predators avoid) with a negative impact on avoidance learning by predators. In mutualistic mimicry, appearances of defended prey converge on a similar warning signal, thereby reinforcing it and decreasing the per-capita cost of training predators. Theoretical and experimental studies on mimicry phenotypes abound, but studies empirically testing their predictions in real mimicry communities remain scarce. Indeed, the quantitative distribution of phenotypes within and among mimicry complexes is largely unknown, and how phenotypic variations are influenced by the type of mimicry, selection intensity, and/or phylogeny remains unaddressed. We first developed a novel framework that enables for the automatic and precise quantification and comparison of colour pattern. We then used this tool on over 2000 specimens, consisting of 130 butterfly species, collected from distinct Neotropical butterfly communities in the Peruvian Amazon. We quantified the distribution of phenotypes and their structure into a number of separate mimicry optima, using a morphological space encompassing the variation and frequencies of all coexisting colours patterns. We analysed this structure to extract the ecological and phylogenetic patterns underlying the coexistence of multiple mimicry groups within a given locality. We then demonstrated the influence of deceptive vs. mutualistic mimics on phenotypic variability around a mimicry optimum in order to address the effective impact of deception on mimicry.