Patterns of extinction risk in the butterfly fauna of eastern North America and Britain and Ireland

Date of Completion

January 2007


Biology, Ecology|Biology, Entomology




The fossil record shows definitive patterns resulting from the mass extinctions of the geologic past: species with particular traits went extinct, while other species, with different traits, survived. It is generally accepted that the current human domination of ecosystems is causing an extinction rate comparable to past mass extinctions, but is the current pattern of biodiversity loss similar to that seen in the fossil record? Over the past few decades, studies of the correlates of rarity and extinction risk among a wide variety of taxa have begun to answer this question. The study of extinction risk is an important contribution to the developing science of conservation biology, enabling a greater understanding of the causes of species loss and allowing prediction of which species are the most vulnerable. ^ This study used multivariate logistic regression to identify correlates of extinction risk for the butterfly faunas of eastern North America and Great Britain and Ireland, and to statistically control for intercorrelations among these traits. The models identify seven statistically independent correlates of extinction risk among the butterflies of eastern North America: winter diapause, particularly in the larval stage; belonging to a large genus; feeding on herbaceous larval hosts; a small geographic range; high larval host plant specificity; association with prairie, barrens, coniferous forest, tropical forest, or wetland habitat (and not anthropogenic, tundra, or desert habitat); and few broods per year. In Britain and Ireland, geographic range size and habitat specificity were the primary correlates of extinction risk, although several other traits had some predictive power. Relationships between species traits and extinction risk are discussed, and causal mechanisms for the patterns observed are hypothesized. The patterns of extinction risk described in this study show similarities to previous studies of butterflies in Europe and southeast Asia, as well as to patterns of extinction risk among other extant and fossil taxa. The models constructed can be used to predict the conservation status of individual species based on their actual traits, or based on hypothetical changes in these traits. ^