We examined density dependence in population attributes and community impact of a generalist predator by experimentally mimicking natural variation in initial cohort densities produced by synchronous egg hatch in Mantis religiosa (Mantodea: Mantidae). Mantid cohorts within the normal range of emergence from a single egg mass were established in a replicated, well-controlled open field experiment. On the scale of the progeny from a single female, density-dependent food limitation caused mortality and ontogenetic asynchrony to increase with increasing density. All cohorts converged to a common level of abundance and biomass because both development rate and population size declined with increasing initial density. Numbers and biomass of other arthropods generally declined with increasing initial density of mantids, although there were both positive and negative effects on different taxa. The abundance of hemipterans (almost exclusively herbivorous mirids) increased in the presence of mantids; this was an indirect effect as large in magnitude as any of the direct reductions in abundance of other taxa. Per capita interaction strengths of mantids on most taxa generally were weak except for the strong positive interaction with hemipterans. In spite of different mantid development rates among treatments, predator load (proportion of arthropod biomass present as predators) for all three treatments, attributable mainly to mantid biomass, converged to approximately five times control level by the end of the experiment. The differences in predator loads between control and treatment plots thus may represent different levels of predator saturation: one for control plots, where predator load was constant over time and in which generalists contributed relatively little to predator biomass, and a higher one for treatment plots, in which generalists comprised the bulk of predator biomass. Predator load may therefore be an indicator of the relative importance of generalist vs. specialist predators in terrestrial arthropod assemblages.
