Host plant use by insect herbivores is predicted to be the target of strong selection due to the prominent role host plant identity plays in determining herbivore fitness. Because both bottom-up factors (e.g., plant foliar availability, nutrition, and physical and chemical defenses) and top-down factors (i.e., risk of attack from natural enemies) tend to be temporally variable, field experiments quantifying preference-performance relationships for polyphagous and oligophagous insect herbivores are needed to explicitly examine how the fitness landscape changes over time. The Silver-spotted skipper (E. clarus; Hesperiidae), a widespread North American legume-feeder, has incorporated a variety of non-native species into its diet, including soybean as well as the invasive species kudzu and Chinese wisteria. We experimentally assessed the preference and performance of E. clarus on six commonly used hosts, quantifying the temporal variation (seasonal and interannual) in both plant quality and natural enemy-mediated mortality, to determine the relative importance of ecological factors in mediating patterns of host use. Our field protocol involved placing neonate E. clarus larvae on each of the six host plants growing in a common garden and monitoring establishment, survival, and growth throughout larval development. Over the three years (8 larval generations), we individually monitored the fates of over 3500 larvae.
Results/Conclusions
All six plant species examined (3 native, 3 non-native) were used as hosts in both the field and in oviposition experiments; both adult and larval preference varied significantly among hosts, with the most preferred hosts including one native and one non-native species. Larval performance in both the presence and absence of natural enemies demonstrated consistent effects of host plant species but variable top-down effects, with a strong seasonal component for both factors. Across all hosts, foliar quality (water content, %N, C:N ratio) declined between early summer and mid-summer and remained low into the fall. These predicatable declines in foliar quality precipitated a sharp (~25%) increase in larval development time, coincident with the warmest field temperatures. Field estimates of predation tended to increase from early to mid/late summer, but were highly variable with respect to host plant. Notably, the strong effect of rearing host on development time was a poor predictor of mortality risk from natual enemies, failing to support predictions of the slow-growth-high-mortality hypothesis. Notably, the plant with the lowest nutritional quality (and longest larval development time) appeared to provide enemy-free space to E. clarus larvae, selecting for inclusion of this host in the diet of E. clarus.