EFFECTS OF TOP PREDATOR SPECIES ON DIRECT AND INDIRECT INTERACTIONS IN A FOOD WEB

Abstract

Current theory on trophic interactions in food webs assumes that ecologically similar species can be treated collectively as a single functional unit such as a guild or trophic level. This theory implies that all species within that unit transmit identical direct and indirect effects throughout the community. We evaluated this assumption by conducting experiments to compare the direct and indirect effects of three top-predator species, be- longing to the same hunting spider guild, on the same species of grasshopper and on old- field grasses and herbs. Observations under field conditions revealed that each spider species exhibited different hunting behavior (i.e., sit-and-wait, sit-and-pursue, and active hunting) and occupied different locations within the vegetation canopy. These differences resulted in different direct effects on grasshopper prey. Grasshoppers demonstrated significant be- havioral (diet) shifts in the presence of sit-and-wait and sit-and-pursue species but not when faced with actively hunting species. Grasshopper density was significantly reduced by spider species that occupied lower parts of the vegetation canopy (sit-and-pursue and actively hunting species), but it was not significantly reduced by the sit-and-wait spider species that occupied the upper parts of the canopy. These direct effects manifested themselves differ- ently in the plant trophic level. The sit-and-wait spider caused indirect effects on plants by changing grasshopper foraging behavior (a trait-mediated effect). The sit-and-pursue spider caused indirect effects by reducing grasshopper density (density-mediated effects); the effects of changes in grasshopper behavior were thus not reflected in the plant trophic level. The actively hunting spiders had strictly density-mediated indirect effects on plants. The study offers mechanistic insight into how predator species within the same guild can have very different trophic effects in food webs. Thus classical modeling approaches that treat all predator species as a single functional unit may not adequately capture biologically relevant details that influence community dynamics.