Although many organisms experience relatively predictable environmental conditions throughout development, others, as a result of ontogenetic changes, inhabit very different environments across life stages, requiring them to navigate into habitats or microhabitats that they have not previously encountered. For holometabolous insects, late instars must often locate a safe site before undergoing pupation. In Lepidoptera, many pre-pupae cease feeding and wander in search of an appropriate pupation site; their preferences display a shift in behavioral responses to habitat-specific environmental cues, relative to those of feeding larvae. Having previously determined that pre-pupae of Epargyreus clarus (Cramer) (Lepidoptera: Hesperiidae) move from their Fabaceae host plant down into the leaf litter, where they construct a pupal refuge out of dried leaves and silk, we investigated the sensory cues used by E. clarus in selecting a pupation location. We further explored how responses to these cues vary between actively feeding larvae and non-feeding pre-pupae, both for individual insects and for stage-based cohorts. Using Y-tube assays, we determined that pre-pupae used visual (light) and gravitational cues to move from host plants into the leaf litter, whereas the odor cues we tested did not appear to influence their decision-making. Directional responses to both light and gravity reversed over ontogeny, allowing the two life stages to make use of the same cues to produce stage-appropriate adaptive behaviors.

All holometabolous insects undergo a pupal life stage, a transformative period during which the insects are immobile and thus particularly vulnerable to both natural enemies and harmful abiotic conditions. For multivoltine species like the silver-spotted skipper [Epargyreus clarus (Cramer) (Lepidoptera: Hesperiidae)], which produces both diapausing and nondiapausing generations throughout much of its range, both the duration of the pupal stage and the ecological challenges faced by pupae can differ among generations. We conducted a set of field experiments to investigate the seasonal and annual variation in pupal mortality for E. clarus pupae experiencing different biotic and abiotic conditions. We also examined the behavioral and ecological factors influencing the construction and persistence of pupal shelters by prepupal larvae. Surprisingly, measures of both cumulative and daily pupal predation were significantly higher during the relatively short (10–14 d) nondiapausing (summer) generations, compared with the diapausing (winter) generations, despite a nearly 20-fold longer pupal duration recorded for the latter. Indirect evidence from field censuses suggested that this intergenerational difference in mortality was due to seasonal variation in consumption of pupae by generalist vertebrate predators. The presence of a shelter increased survival in summer, though not in winter, perhaps because winter pupae were likely to be buried under autumnal leaf litter, regardless of initial shelter status. When constructing their shelters, prepupal E. clarus larvae did not prefer host leaves over nonhost leaves, suggesting that induced preferences are unlikely to play an important role in this process. Despite finding marked differences in the decomposition rates of shelter leaves derived from host vs. nonhost plants, several lines of evidence suggest that these differences are unlikely to impact E. clarus pupal mortality during either the summer or winter generations.