Data from: Indirect effects of predators control herbivore richness and abundance in a benthic eelgrass (Zostera marina) mesograzer community

1. Herbivore communities can be sensitive to changes in predator pressure (top-down effects) and resource availability (bottom-up effects) in a wide range of systems. However, it remains unclear whether such top-down and bottom-up effects reflect direct impacts of predators and/or resources on herbivores, or are indirect, reflecting altered interactions among herbivore species. 2. We quantified direct and indirect effects of bottom-up and top-down processes on an eelgrass (Zostera marina) herbivore assemblage. In a field experiment, we factorially manipulated water column nutrients (with Osmocote™ slow-release fertilizer) and predation pressure (with predator exclusion cages) and measured the effects on herbivore abundance, richness and beta diversity. We examined likely mechanisms of community responses by statistically exploring the response of individual herbivore species to trophic manipulations. 3. Predators increased herbivore richness and total abundance, in both cases through indirect shifts in community composition. Increases in richness occurred through predator suppression of common gammarid amphipod species (Monocorophium acherusicum and Photis brevipes), permitting the inclusion of rarer gammarid species (Aoroides columbiae and Pontogeneia rostrata). Increased total herbivore abundance reflected increased abundance of a caprellid amphipod species (Caprella sp.), concurrent with declines in the abundance of other common species. Furthermore, predators decreased beta diversity by decreasing variability in Caprella sp. abundance among habitat patches. 4. Osmocote™ fertilization increased nutrient concentrations locally, but nutrients dissipated to background levels within 3 m of the fertilizer. Nutrient addition weakly affected the herbivore assemblage, not affecting richness and increasing total abundance by increasing one herbivore species (Caprella sp.). Nutrient addition did not affect beta diversity. 5. We demonstrated that assemblage-level effects of trophic manipulations on community structure are the result of distinct and often indirect responses of herbivore species. These results underscore the importance of understanding herbivore–herbivore interactions in a system commonly subjected to both eutrophication and overfishing