Seasonal shift from bottom-up to top-down impact in phytophagous insect populations

Although many studies now examine how multiple factors influence the dynamics of herbivore populations, few studies explicitly attempt to document where and when each is important and how they vary and interact. In fact, how temporal variation in top-down (natural enemies) and bottom-up (host plant resources) factors affect herbivore dynamics has been suggested as a particularly important yet poorly understood feature of terrestrial food webs. In this study we examined how temporal changes in predator density (wolf spiders, sheet-web builders, and mirid egg predators) and host-plant resources (plant quality and structural complexity) influence the population dynamics of the dominant phytophagous insects on Atlantic-coast salt marshes, namely Prokelisia planthoppers (Homoptera: Delphacidae). We designed a factorial experiment in meadows of Spartina altemiflora to mimic natural variation in vegetation quality and structure by establishing two levels of plant nutrition (leaf nitrogen content) by fertilization, and two levels of habitat complexity by adding leaf litter (thatch). We then assessed seasonal changes in the strength of bottom-up (plant quality) and top-down (predator) impacts on planthopper populations. Planthopper populations responded positively to increased plant quality treatments in late summer. Despite the greater number of planthopper adults colonizing fertilized Spartina plots compared to unfertilized controls, the offspring of these colonists were much less abundant at the end of the season in fertilized plots, particularly those with thatch. The initial colonization effect was later erased because arthropod predators selectively accumulated in fertilized plots where they inflicted significant mortality on all stages of planthoppers. Predators rapidly colonized fertilized plots and reached high densities well in advance of planthopper colonization, a response we attribute to their rapid aggregation in complex-structured habitats with readily available alternative prey. Our results suggest that plant resources not only mediate the strength of predator impacts on herbivore populations, but they also promote the coupling of predator and prey populations and thus influence when enemy impacts are realized.