Rapid environmental change may cause mismatches in species’ evolved responses to informative cues required for developmental decisions, with negative fitness consequences resulting in evolutionary traps. For insects relying on stationary photoperiod cues to initiate winter diapause, warmer summer temperatures can accelerate development rates, phenology, and exposure to cues, leading to direct development into another generation before the onset of winter. It is hypothesized that this facultative generation may experience high mortality if individuals do not have sufficient time to complete development to the appropriate life stage for overwintering, creating a lost generation with severe consequences for population growth. We test this lost generation hypothesis in 20 butterfly species that show plasticity in their voltinism, or number of generations per year, in response to climatic variability across space and time.
Results/Conclusions
We used 18 years of citizen science monitoring of butterflies in Ohio to estimate the phenology and abundance for each species’ generation, which were then used to model voltinism and population growth rates. We demonstrated that the relative size of the last generation, a proxy for choosing direct development over diapause, varies with exposure to environmental cues during the previous generation through locally-adapted responses to photoperiod and temperature. Annual population growth rates were primarily driven by negative density-dependence. Across species, larger relative sizes of the last generation were associated with higher population growth rates and weakened density-dependence. We observed wide interspecific variation in the impacts of winter conditions, such as arrival of frost and mean temperature, on population growth rates. This study suggests that evolutionary traps for insects with flexible voltinism are rare and that increased voltinism with longer growing seasons will likely benefit insect population growth in temperate climates. We show that the lost generation hypothesis does not explain population dynamics in Ohio butterflies, although it may explain population crashes in species with mismatched responses to cues in parts of their range.