WIREs Water

Research that conceptually transcends boundaries between aquatic and terrestrial ecosystems has a long history of increasing insight into ecology and evolution. To stimulate further cross-fertilization between studies that focus on different ecosystems, we highlight several insights on risk effects-the costs of antipredator behavior-that have emerged in part because of combined advances in aquatic and terrestrial systems. Namely, risk effects (1) are not restricted to structured landscapes where antipredator behavior is easily measurable, (2) can be substantial even when prey experience very low predation rates, (3) are contingent on a three-way interaction between the hunting mode of the predator, escape tactic of the prey, and features of the landscape/physical environment, and (4) can interact with direct predation (consumption) and resource availability (through its effects on prey energy state) to control consumer population size. We conclude by highlighting the value of exploring differences between aquatic and terrestrial risk effects and offering a prospectus for future studies of antipredator behavior and its ecological importance in both eco-domains. © 2014 Wiley Periodicals, Inc. How to cite this article: WIREs Water 2014. doi: 10.1002/wat2.1039 G. C. Williams, originator of the theory of fish schooling that I am here supporting (Williams, 1964(Williams, , 1966 points out that schooling is particularly evident in the fish that inhabit open waters. This fits with the view that schooling is similar to cover-seeking in its motivation. His experiments showed that fish species whose normal environment afforded cover in the form of weeds and rocks had generally less marked schooling tendencies. Among mammals, similarly, the most gregarious species are inhabitants of open grassy plains rather than of forest (Hesse, Allee & Schmidt, 1937 the throng (Springer, 1957) and a parallel to this is commonly seen in the behaviour of the hindmost sheep that a sheepdog has driven into an enclosure: such sheep try to butt or to jump their way into the close packed ranks in front. Behaviour of this kind certainly cannot be regarded as showing an unselfish concern for the welfare of the whole group. -W. D. 8 Behavioral ecologists have also benefited greatly from cross-ecosystem comparisons. For instance, comparative studies between odontocete cetaceans and primates have yielded general insights into the function of coalitions and other social behaviors (e.g., Ref 9). Despite these examples, however, ecologists often underutilize the potential benefits of thinking outside the confines of a particular eco-domain. THE SCOPE OF ANTIPREDATOR BEHAVIOR There is a growing effort to elucidate the consequences of risk effects for community properties. This effort has primarily entailed the use of mesocosm experiments (e.g., see Ref 20 for a review) or fieldwork in aquatic and terrestrial systems where, from the researcher's perspective, well-defined landscape structures (i.e., patchiness) facilitate measures of variability in antipredator behavior (e.g., habitat shifts and use of refugia, 24 and giving-up densities 15 ). Moreover, it has focused largely on invertebrate or small-bodied herbivores that are amenable to experimentation 22 In patchy landscapes, which characterize many terrestrial ecosystems and littoral and benthic zones within aquatic ecosystems, it is well documented that prey individuals often respond to predation risk by shifting to safer habitats or by increasing vigilance. These responses carry the cost of lower foraging rates or reduced access to resources 28,29 Similar patterns of predator-induced DVM have been documented in freshwater planktivores. 1 Large schools of fish, for instance, effectively represent patches offering heightened safety (e.g., via predator detection, confusion, and/or risk dilution) and within which danger can vary as a function of group position. This idea appears to apply in riverine ecosystems as well. For example, Orpwood et al. WIREs Water Cross-fertilizing aquatic and terrestrial research channels formed larger shoals in response to predation risk, but only in structurally simple habitats. Second, risk-induced shifts already are well documented in many systems where researchers can define landscape structure more easily. Awareness of that evidence should remove conceptual barriers for testing the plausibility of similar behaviors in the more subtle patchiness of open-water systems. For example, experiments in old-field meadows found that grasshoppers at risk of predation from spiders avoided the bottom strata of vegetation-where dangerous spiders and preferred food were most abundant-and increased use of the safer but food-poorer canopy. 33 These experiments inspired behavioral models for pinnipeds (i.e., large marine mesopredators) under risk of predation by deep-water sharks. Similar to the spider-grasshopper system, these models predicted that pinnipeds would optimize risk-energy trade-offs by avoiding more dangerous and food-richer deep strata and instead foraging more safely (at a feeding cost) in shallower strata 39 In the mid-1990s, there was little appreciation for the possibility that large-bodied marine herbivores (sea cows and sea turtles) and mesopredators (e.g., cetaceans and pinnipeds), which have low predation rates in many areas, respond to predation risk by shifting foraging habitats. THE CONTINGENCY OF RISK EFFECTS The threat of predation often induces behavioral modification (e.g., avoidance and increased vigilance) that reduces prey foraging investment. 24,53 Consequently, predator presence is generally assumed to benefit species serving as food for prey. Growing evidence suggests, however, that this assumption may not always be valid and that prey responses to predators-and, therefore, the strength of risk effects-instead vary with context. 25 Ambush spiders leave persistent, point-source cues of predation risk, to which herbivorous grasshoppers can respond with chronic habitat shifts, whereas roving spiders do not leave such cues. Consequently, ambush spiders indirectly benefit nutritious grasses by inducing grasshoppers to shift their foraging activity to less preferred plant species, whereas roving spiders do not trigger habitat shifts. Instead, roving spiders have stronger consumptive effects on grasshoppers and thus differ in the kinds of indirect effects that they exert on plants