Enemy-Risk Effects: A Brief Conceptual Overview
Many organisms exhibit responses to natural enemies (predators and parasitoids; we frequently use “predator/prey” as a catchall that includes parasitoid/host relationships), including within-generation changes in behavior (e.g, reduced activity, increased refuge use, increased group size; Lima 1998), physiology (Hawlena & Schmitz 2010; Clinchy et al. 2013), morphology (Bourdeau & Johansson 2012; Hulthén et al. 2014) and life history (Miner et al. 2005; LaManna & Martin 2016; Relyea et al. 2018). These responses typically have costs in terms of reduced feeding and growth rates and, ultimately, reduced fitness (Kerfoot & Sih 1987; Stamps 2007; Orrock et al. 2013) and population growth rates (Creel & Christianson 2008). Because these responses often involve niche shifts (e.g., in prey diets or habitat use), they also affect prey interactions with other species (Werner & Peacor 2003). For example, anti-predator responses can alter competition among prey (Werner & Anholt 1996), increase exposure to other predators (Sih et al. 1998; Fouzai et al. 2019) or to diseases (Duffy et al. 2011; Shang et al. 2019), and alter impacts on their own resources (Schmitz et al. 2004). Notably, if prey exhibit strong, effective anti-enemy responses, predators might actually kill few prey (i.e., have weak consumptive effects, CEs) but instead have large impacts on prey fitness and prey interactions with the rest of their community (Preisser et al. 2005). These three levels of effect (individual response, impacts on fitness/populations, community effects) are best defined as enemy-induced trait responses, non-consumptive effects, and trait-mediated indirect effects (Peacor et al. 2020). Box 2 discusses this terminology in greater detail.
Experiments to evaluate the relative strength of CEs and NCEs typically contrast the total effect of actual predators (CE + NCEs) with the effect of constrained predators (e.g., predators caged or artificially manipulated to prevent use of mouthparts) or predator cues (NCEs only) on prey. A meta-analysis of these experiments found that the importance of NCEs was highly variable, but on average roughly the same magnitude as CEs (Preisser et al. 2005). For biocontrol, trait-mediated indirect effects cascading to the plant may be even more relevant. Enemies frequently have very strong positive effects on plants due to trait shifts by herbivores, even when CEs are relatively small (Schmitz et al. 2004; Creel & Christianson 2008).
Behavioral ecology theory and experiments suggest that prey typically exhibit stronger trait responses when perceived risk is higher and when the marginal costs of response are lower (Lima 1998). When perceived risk reflects actual risk, predators that are more dangerous in the absence of prey defenses can induce such strong anti-predator responses that they kill fewer prey (but cause stronger NCEs) than less dangerous predators. Thus, predation rate is often not a good measure of predation risk, and therefore not always a good indicator of total effect on prey (CEs + NCEs). Perceived risk, however, is not always proportional to actual risk. Perceived risk depends on not just the type of predator and its attack success, but also on the type and strength of predator cues or prey alarm cues (Kats & Dill 1998; Stankowich & Blumstein 2005; Ferrari et al. 2010), on the habitat per se (Verdolin 2006; Thaker et al. 2011), and on prey sensory/cognitive capacities (Kats & Dill 1998; Ferrari et al. 2010; Bedoya-Perez et al. 2019). Predators that are not very dangerous, but difficult to locate and assess (e.g., ambush predators) can induce strong anti-predator responses and thus strong NCEs (Sih 1992; Preisser et al. 2007). Prey may even respond to an organism that is incapable of killing them if the cues are sufficiently close to those of a dangerous enemy (Fill et al. 2012). Box 3 discusses how prey perceive risk in more detail and implications for enemy-risk effects and biocontrol.
Marginal costs of enemy-induced trait responses are higher (and prey exhibit weaker trait responses) if prey are energy stressed (hungry), resources or mates are abundant but more accessible only if prey show little anti-predator response, or if prey have high reproductive value (more to lose; Clark 1994; Houston et al. 1999). For herbivores, the strength of the enemy-risk effect depends on, among other things, plant abundance and quality, herbivore condition, and life history stage (McArthur et al. 2012; Stephan et al. 2017).
The role of enemy-risk effects in community dynamics becomes more complex when we consider multiple enemies and intraguild predation (IGP), common occurrences in biocontrol systems. With multiple agents of mortality, enemy-risk effects can often blend into CEs where a trait response to an enemy (e.g., a shift in microhabitat use) increases mortality from another enemy (Sih et al. 1998), environmental stressors (Schmitz et al. 1997), or even pesticides (Janssens & Stoks 2013). With IGP, predators are also potentially prey, and thus also exhibit enemy-induced trait responses and NCEs. The mix of CEs and NCEs involving multiple species then influences community outcomes including biocontrol efficacy. We discuss this in more detail in a later section.
Many of these predictions about enemy-risk effects assume that prey exhibit adaptive responses to enemies that they have co-evolved with. Prey lacking evolutionary (or developmental) history with enemies (or specific enemies) often exhibit much weaker anti-enemy responses and thus suffer heavy mortality (strong CEs) when novel enemies appear (e.g., island prey or prey in fishless ponds; Cox & Lima 2006; Carthey & Blumstein 2018). This depends on the cue or functional similarity of new enemies to the prey’s familiar enemies (Sih et al. 2010; Carthey & Banks 2014; Saul & Jeschke 2015). Given that biocontrol often involves introducing enemies that have a coevolutionary history with the target pest, but not with non-target organisms, the effect of evolutionary history on CEs versus NCEs is clearly a salient issue that we discuss in more detail below.