Abstract
Müllerian mimicry theory states that frequency dependent selection should favour geographic convergence of harmful species onto a shared colour pattern. As such, mimetic patterns are commonly circumscribed into discrete mimicry complexes each containing a predominant phenotype. Outside a few examples in butterflies, the location of transition zones between mimicry complexes and the factors driving them has rarely been examined. To infer the patterns and processes of Müllerian mimicry, we integrate large-scale data on the geographic distribution of colour patterns of all social bumble bees across the contiguous United States and use these to quantify colour pattern mimicry using an innovative machine learning approach based on computer vision and image recognition. Our data suggests that bumble bees exhibit a manifold of similar, but imperfect colour patterns, that continuously transition across the United States, supporting the idea that mimicry is not discrete. We propose that bumble bees are mimicking a perceptual colour pattern average that is evolutionarily transient. We examine three comimicking polymorphic species, Bombus flavifrons, B. melanopygus, and B. bifarius, where active selection is driving colour pattern frequencies and determine that their colour pattern transition zones differ in location and breadth within a broad region of poor mimicry. Furthermore, we explore factors driving these differences such as mimicry selection dynamics and climate.
