We compiled a large database of 58 059 point locality records for 70 species and 434 subspecies of heliconiine butterflies and used these data to test evolutionary hypotheses for their diversification. To study geographical patterns of diversity and contact zones, we mapped: (1) species richness; (2) mean molecular phylogenetic terminal branch length; (3) subspecies richness and the proportion of specimens that were subspecific hybrids, and (4) museum sampling effort. Heliconiine species richness is high throughout the Amazon region and peaks near the equator in the foothills and middle elevations of the eastern Andes. Mean phylogenetic terminal branch length is lowest in the eastern Andes and tends to be low in species-rich areas. By contrast, areas of high subspecies richness, where subspecies overlap in range and/or hybridize, are concentrated along the course of the Amazon River, with the eastern Andes slopes and foothills relatively depauperate in terms of local intraspecific phenotypic diversity. Spatial gradients in heliconiine species richness in the Neotropics are consistent with the hypothesis that species richness gradients are driven at least in part by variation in speciation and/or extinction rates, resulting in observed gradients in mean phylogenetic branch length, rather than via evolutionary age or niche conservatism alone. The data obtained in the present study, coupled with individual case studies of recently evolved Heliconius species, suggest that the radiation of heliconiine butterflies occurred predominantly on the eastern slopes of the Andes in Colombia, Ecuador, and Peru, as well as in the upper/middle Amazon basin.

No abstract

Traditional anatomical analyses captured only a fraction of real phenomic information. Here, we apply deep learning to quantify total phenotypic similarity across 2468 butterfly photographs, covering 38 subspecies from the polymorphic mimicry complex of Heliconius erato and Heliconius melpomene. Euclidean phenotypic distances, calculated using a deep convolutional triplet network, demonstrate significant convergence between interspecies co-mimics. This quantitatively validates a key prediction of Müllerian mimicry theory, evolutionary biology’s oldest mathematical model. Phenotypic neighbor-joining trees are significantly correlated with wing pattern gene phylogenies, demonstrating objective, phylogenetically informative phenome capture. Comparative analyses indicate frequency-dependent mutual convergence with coevolutionary exchange of wing pattern features. Therefore, phenotypic analysis supports reciprocal coevolution, predicted by classical mimicry theory but since disputed, and reveals mutual convergence as an intrinsic generator for the unexpected diversity of Müllerian mimicry. This demonstrates that deep learning can generate phenomic spatial embeddings, which enable quantitative tests of evolutionary hypotheses previously only testable subjectively.

Climatic gradients frequently predict large‐scale ecogeographical patterns in animal coloration, but the underlying causes are often difficult to disentangle. We examined ecogeographical patterns of reflectance among 343 European butterfly species and isolated the role of selection for thermal benefits by comparing animal‐visible and near‐infrared (NIR) wavebands. NIR light accounts for ~50% of solar energy but cannot be seen by animals so functions primarily in thermal control. We found that reflectance of both dorsal and ventral surfaces shows thermally adaptive correlations with climatic factors including temperature and precipitation. This adaptive variation was more prominent in NIR than animal‐visible wavebands and for body regions (thorax‐abdomen and basal wings) that are most important for thermoregulation. Thermal environments also predicted the reflectance difference between dorsal and ventral surfaces, which may be due to modulation between requirements for heating and cooling. These results highlight the importance of climatic gradients in shaping the reflectance properties of butterflies at a continent‐wide scale.