Introduction

Presence of predators often causes prey to modify traits, including behavior, physiology, morphology, and life history, in order to reduce predation risk (Lima and Dill 1990, Lima 1998a, Agrawal 2001). Such risk-induced trait responses and their ensuing consequences, termed risk effects (also nonlethal effects and fear effects; terminology reviewed in Peacor et al. 2020), have received much attention. Whereas risk-induced trait response in prey generally benefit prey by reducing predation risk, they also may have associated costs to fitness components in the prey. These consequences of the trait response on prey are termed non-consumptive effects (NCEs), and include reduced growth or condition, reduced fecundity, or increased mortality from other predators, and consequent effects on prey population growth rate and prey abundance (Abrams 1984, Brown 1988, Lima 1998b, Peckarsky et al. 2008, Ohgushi et al. 2012, Sheriff et al. 2020, Peacor et al. 2022). The risk-induced trait-responses of the prey can also have indirect effects (termed trait-mediated indirect effects, TMIEs; Peacor et al. 2020), with other species including their resources, competitors, or other predators (Sih et al. 1998, Werner and Peacor 2003). NCEs and TMIEs have received much attention given their potential to profoundly influence ecological communities and management of natural resources (reviewed in Werner and Peacor 2003, Schmitz et al. 2004, Creel and Christianson 2008, Heithaus et al. 2008, Peckarsky et al. 2008, Ohgushi et al. 2012, Say-Sallaz 2019).
Indeed, it is often claimed that risk effects are as important to the net effect of predators on prey abundance as their direct consumptive effects (Sheriff et al. 2020). However, that conclusion has been based largely on experiments done in relatively simple systems with only a few interacting species and little temporal or spatial environmental heterogeneity (Nelson 2007, Sheriff et al. 2020), even when performed in natural settings (Sheriff et al. 2020, Peacor et al. 2022). Thus, if we are to incorporate risk effects into management plans, it is important to know that the effects demonstrated in simplified systems are generalizable to more complex systems incorporating many interacting species and spatial and temporal heterogeneity.
Here we investigate risk effects of fish on the abundance of different zooplankton prey species in complex experimental communities. Here we investigate risk effects of fish on the abundance of different zooplankton prey species in complex experimental communities. The novelty of our approach is that we look at these effects across multiple experiments, each with a very similar experimental conditions, objectives, measurables and implementation. Thus, any observed variability in the effect of predation risk across experiments arises from differences in the experiments that are purely incidental (i.e., differences that were not of consequence to the study design or implementation). These incidental differences in environmental conditions are much smaller than the environmental variation that would be expected in natural settings. Our approach thereby represents a conservative examination of contingency in natural settings, which allows us to ask whether risk effects on prey abundance are robust and generalizable.