Phenotypic divergence between closely related species has long interested biologists. Taxa that inhabit a range of environments and have diverse natural histories can help understand how selection drives phenotypic divergence. In butterflies, wing color patterns have been extensively studied but diversity in wing shape and size is less well understood. Here, we assess the relative importance of phylogenetic relatedness, natural history, and habitat on shaping wing morphology in a large dataset of over 3500 individuals, representing 13 Heliconius species from across the Neotropics. We find that both larval and adult behavioral ecology correlate with patterns of wing sexual dimorphism and adult size. Species with solitary larvae have larger adult males, in contrast to gregarious Heliconius species, and indeed most Lepidoptera, where females are larger. Species in the pupal-mating clade are smaller than those in the adult-mating clade. Interestingly, we find that high-altitude species tend to have rounder wings and, in one of the two major Heliconius clades, are also bigger than their lowland relatives. Furthermore, within two widespread species, we find that high-altitude populations also have rounder wings. Thus, we reveal novel adaptive wing morphological divergence among Heliconius species beyond that imposed by natural selection on aposematic wing coloration.