Prey responses to fine‐scale variation in predation risk from combined predators

While it is well documented that organisms can express phenotypic plasticity in response to single gradients of environmental variation, our understanding of how organisms integrate information along multiple environmental gradients is limited in many systems. Using the freshwater snail Helisoma trivolvis and two common predators (water bugs Belostoma flumineum and crayfish Orconectes rusticus), we explored how prey integrate information along multiple predation risk gradients (i.e. caged predators fed increasing amounts of prey biomass) that induce opposing phenotypes. When exposed to single predators fed increasing amounts of prey biomass, we detected threshold responses; intermediate amounts of consumed biomass induced phenotypic responses, but higher amounts induced little additional induction. This suggests that additional increases in predator‐induced traits with greater predator risk offer minimal increases in fitness or that a limit in the response magnitude was reached. Additionally, the response thresholds were contingent on the predator and focal trait. For shell width, responses were generally detected at a lower amount of consumed biomass by water bugs compared to crayfish. Within the crayfish treatments, we found that the shell thickness response threshold was lower than the shell width response threshold. When we combined gradients of consumed biomass from both predators, we found that the magnitude of response to one predator was often reduced when the other predator was present. Interestingly, these effects were often detected at consumed biomass levels that were lower than the threshold concentration necessary to elicit a response in the single‐predator treatments. Moreover, our combined predator treatments revealed that snails shifted from discrete responses to more continuous (i.e. graded) responses. Together, our results reveal that organisms experiencing multiple environmental gradients can integrate this information to make phenotypic decisions and demonstrate the novel result that an exposure to multiple species of predators can lower the response threshold of prey.     

Prey Responses to Predator Chemical Cues: Disentangling the Importance of the Number and Biomass of Prey Consumed

To effectively balance investment in predator defenses versus other traits, organisms must accurately assess predation risk. Chemical cues caused by predation events are indicators of risk for prey in a wide variety of systems, but the relationship between how prey perceive risk in relation to the amount of prey consumed by predators is poorly understood. While per capita predation rate is often used as the metric of relative risk, studies aimed at quantifying predator-induced defenses commonly control biomass of prey consumed as the metric of risk. However, biomass consumed can change by altering either the number or size of prey consumed. In this study we determine whether phenotypic plasticity to predator chemical cues depends upon prey biomass consumed, prey number consumed, or both. We examine the growth response of red-eyed treefrog tadpoles (Agalychnis callidryas) to cues from a larval dragonfly (Anax amazili). Biomass consumed was manipulated by either increasing the number of prey while holding individual prey size constant, or by holding the number of prey constant and varying individual prey size. We address two questions. (i) Do prey reduce growth rate in response to chemical cues in a dose dependent manner? (ii) Does the magnitude of the response depend on whether prey consumption increases via number or size of prey? We find that the phenotypic response of prey is an asymptotic function of prey biomass consumed. However, the asymptotic response is higher when more prey are consumed. Our findings have important implications for evaluating past studies and how future experiments should be designed. A stronger response to predation cues generated by more individual prey deaths is consistent with models that predict prey sensitivity to per capita risk, providing a more direct link between empirical and theoretical studies which are often focused on changes in population sizes not individual biomass.     

Chemical defense of toad tadpoles under risk by four predator species

Many organisms use inducible defenses as protection against predators. In animals, inducible defenses may manifest as changes in behavior, morphology, physiology, or life history, and prey species can adjust their defensive responses based on the dangerousness of predators. Analogously, prey may also change the composition and quantity of defensive chemicals when they coexist with different predators, but such predator‐induced plasticity in chemical defenses remains elusive in vertebrates. In this study, we investigated whether tadpoles of the common toad (Bufo bufo) adjust their chemical defenses to predation risk in general and specifically to the presence of different predator species; furthermore, we assessed the adaptive value of the induced defense. We reared tadpoles in the presence or absence of one of four caged predator species in a mesocosm experiment, analyzed the composition and quantity of their bufadienolide toxins, and exposed them to free‐ranging predators. We found that toad tadpoles did not respond to predation risk by upregulating their bufadienolide synthesis. Fishes and newts consumed only a small percentage of toad tadpoles, suggesting that bufadienolides provided protection against vertebrate predators, irrespective of the rearing environment. Backswimmers consumed toad tadpoles regardless of treatment. Dragonfly larvae were the most voracious predators and consumed more predator‐naïve toad tadpoles than tadpoles raised in the presence of dragonfly cues. These results suggest that tadpoles in our experiment had high enough toxin levels for an effective defense against vertebrate predators even in the absence of predator cues. The lack of predator‐induced phenotypic plasticity in bufadienolide synthesis may be due to local adaptation for constantly high chemical defense against fishes in the study population and/or due to the high density of conspecifics.     

Prey risk assessment depends on conspecific density

In many systems, the number of prey killed by predators increases with prey density. This in turn generates higher levels of the indirect signals that prey use to assess predation risk. A model developed by Peacor (2003) showed that prey that respond to predator cues without accounting for conspecific density will consistently over‐ or under‐estimate risk and therefore invest improperly in anti‐predator defense. We tested this model using Rana temporaria tadpoles as prey and Aeshna cyanea dragonfly larvae as predators. As assumed by the model, prey reduced risky activity with increasing concentrations of predator kairomones and increased activity at high prey density. However, prey did not react to changes in cue or density if the ratio of cue‐to‐density remained constant. Prey therefore monitored their per capita risk, strongly supporting Peacor's model.     

Temporal variation in behavioral responses to dietary cues from a gape‐limited predator in tadpole prey: A test of the phylogenetic relatedness hypothesis

The ability of prey to recognize and adequately respond to predators determines their survival. Predator‐borne, post‐digestion dietary cues represent essential information for prey about the identity and the level of risk posed by predators. The phylogenetic relatedness hypothesis posits that prey should respond strongly to dietary cues from closely related heterospecifics but respond weakly to such cues from distantly related prey, following a hierarchical pattern. While such responses have mostly been observed in prey at their first encounter with predators, whether prey maintain such hierarchical levels of investment through time remains unclear. We investigated this question by exposing Rhacophorus arboreus tadpoles to the non‐consumptive effect of gape‐limited newt predators Cynops pyrrhogaster that were fed one of five prey diets across a gradient of phylogenetic relatedness: frog tadpoles (Rhacophorus arboreus, Rhacophorus schlegelii, Pelophylax nigromaculatus, and Hyla japonica) and medaka fish (Oryzias latipes). Predators’ diet, time, and their interaction significantly influenced tadpole activity level. We found support for the phylogenetic relatedness hypothesis: Investments in defense were stronger to cues from tadpole diets than to cues from fish diet. However, such a hierarchical response was recorded only in the first four days following predator exposure, then gradually disappear by day 8 on which the tadpoles exhibited similar activity level across all predator treatments. The findings suggest that, at least under the threat of gape‐limited predators, prey use phylogenetic information to evaluate risk and appropriately invest in defense during early encounters with predators; however, energy requirements may prevent prey from maintaining a high level of defense over long exposure to predation risk.     

Behavioural responses of free‐ranging western grey kangaroos (Macropus fuliginosus) to olfactory cues of historical and recently introduced predators

Predation risk influences foraging decisions and time allocation of prey species, and may result in habitat shifts from potentially dangerous to safer areas. We examined a wild population of western grey kangaroos (Macropus fuliginosus) to test the efficacy of predator faecal odour in influencing time allocated to different behaviours and inducing changes in habitat use. Kangaroos were exposed to fresh faeces of a historical predator, the dingo (Canis lupus dingo), a recently introduced predator, the red fox (Vulpes vulpes), a herbivore (horse, Equus caballus) and an unscented control simultaneously. Kangaroos did not increase vigilance in predator‐scented areas. However, they investigated odour sources by approaching and sniffing; more time was spent investigating fox odour than control odours. Kangaroos then exhibited a clear anti‐predator response to predator odours, modifying their space use by rapidly escaping, then avoiding fox and dingo odour sources. Our results demonstrate that wild western grey kangaroos show behavioural responses to predator faeces, investigating then avoiding these olfactory cues of potential predation risk, rather than increasing general vigilance. This study contributes to our understanding of the impact of introduced mammalian predators on marsupial prey and demonstrates that a native Australian marsupial can recognize and respond to the odour of potential predators, including one that has been recently introduced.     

Conspecific density determines the magnitude and character of predator-induced phenotype

The benefits in survival gained from predator-induced phenotypes often come at a cost to other components of fitness. Therefore, the level of expression of an induced phenotype should mirror the level of risk in the environment. When a predator exhibits a saturating functional response the risk of mortality to a given prey decreases as prey density increases. Therefore, for a given predator threat, investment in defense should be lower in prey at high density relative to those at low density. In this study, I test whether the magnitude of predator-induced morphological plasticity decreases with increasing conspecific density by exposing pine woods tree frog (Hyla femoralis) tadpoles at three different densities to predators (present or absent) in a factorial experiment. Tadpole morphology was not affected by changes in density in the absence of predators. However, predators had a significant, density-dependent effect on tadpole morphology. Specifically, the magnitude of morphological response was graded and larger for animals in the low density (high risk) environment. This study demonstrates that tadpoles can modulate phenotypic plasticity in response to mortality risk as a function of both the density of conspecifics and chemical cues from predators, which suggests that they are able to detect and respond to fine-scale changes in the threat environment. In addition, this study highlights the need for analytical approaches that allow morphological plasticity studies to elucidate allometric relationships in addition to simply quantifying size-corrected traits. Electronic supplementary material The online version of this article (doi:) contains supplementary material, which is available to authorized users.     

Two common invertebrate predators show varying predation responses to different types of sentinel prey

Sentinel prey (an artificially manipulated patch of prey) are widely used to assess the level of predation provided by natural enemies in agricultural systems. Whilst a number of different methodologies are currently in use, little is known about how arthropod predators respond to artificially manipulated sentinel prey in comparison with predation on free‐living prey populations. We assessed how attack rates on immobilized (aphids stuck to cards) and artificial (plasticine lepidopteran larvae mimics) sentinel prey differed to predation on free‐moving live prey (aphids). Predation was assessed in response to density of the common invertebrate predators, a foliar‐active ladybird Harmonia axyridis (Coleoptera: Coccinellidae), and a ground‐active beetle Pterostichus madidus (Coleoptera: Carabidae). Significant increases in attack rates were found for the immobilized and artificial prey between the low and high predator density treatments. However, an increased predator density did not significantly reduce numbers of free‐living live aphids included in the mesocosms in addition to the alternate prey. We also found no signs of predation on the artificial prey by the predator H. axyridis. These findings suggest that if our assessment of predation had been based solely on the foliar artificial prey, then no increase in predation would have been found in response to increased predator density. Our results demonstrate that predators differentially respond to sentinel prey items which could affect the level of predation recorded where target pest species are not being used.     

Behavioural response of bullfrog tadpoles to chemical cues of predation risk are affected by cue age and water source

When confronted by signals of predators presence, many aquatic organisms modify their phenotype (e.g., behaviour or morphology) to reduce their risk of predation. A principal means by which organisms assess predation risk is through chemical cues produced by the predators and/or prey during predation events. Such responses to predation risk can directly affect prey fitness and indirectly affect the fitness of species with which the prey interacts. Accurate assessment of the cue will affect the adaptive nature, and hence evolution, of the phenotypic response. It is therefore, important to understand factors affecting the assessment of chemical cues. Here I examined the effect of the age of chemical cues arising from an invertebrate predator, a larval dragonfly (Anax junius), which was fed bullfrog tadpoles, on the behavioural response (activity level and position) of bullfrog tadpoles. The bullfrog response to chemical cues declined as a function of chemical cue age, indicating the degradation of the chemical cue was on the order of 2–4 days. Further, the decay occurred more rapidly when the chemical cue was placed in pond water rather than well water. These results indicate a limitation of the tadpoles to interpret factors that affect the magnitude of the chemical cue and hence accurately assess predation risk. These findings also have implications for experimental design and the adaptation of phenotypic responses to chemical cues of predation risk.     

Having the guts to compete: how intestinal plasticity explains costs of inducible defences

Predators commonly induce phenotypic changes that make prey better at surviving predation at the cost of reduced growth. While we have a good understanding of how trait changes affect predation risk, we lack a mechanistic understanding of why predator‐induced phenotypes differ in growth. Using two mesocosm experiments, we combined phenotypic plasticity theory with predictions from optimal digestion theory to demonstrate that intra‐ and interspecific competition induced relatively long guts while predators induced relatively short guts. The longer guts induced by competition appear to be an adaptive response that allows more efficient digestion and more rapid growth whereas the shorter guts induced by predators appear to result from a tradeoff of building larger tails in predator environments at the cost of smaller bodies. By combining these two bodies of theory, we now have a much better understanding of the mechanisms that cause the phenotypic trade‐offs that select for inducible defences.     

Phenotypically plastic responses of green frog embryos to conflicting predation risk

Predators have been shown to alter the timing of switch points between life history stages, but few studies have addressed switch point plasticity in prey exposed simultaneously to conflicting predation pressure. We tested hatching responses of green frog (Rana clamitans) embryos subject to perceived predation risk from chemical cues released by two stage-specific predators, predicting that these predators would elicit: (1) directional hatching responses when presented independently, and (2) intermediate phenotypic responses when presented simultaneously. R. clamitans embryos in outdoor exclosures were exposed to cues from an egg predator (freshwater leeches; Nephelopsis obscura), a larval predator (dragonfly nymphs, Aeschna canadensis), and both predators in a 2 × 2 factorial experiment, and changes in hatchling size, hatchling developmental stage, and hatching time were compared to those for control embryos. Leeches alone induced embryos to hatch at a smaller size and an earlier developmental stage than controls, while dragonfly nymphs elicited a delay in egg hatching time that was associated with larger size and later developmental stage at hatching. Embryos failed to respond to simultaneous exposure to both predators, implying that responses to each occurred concurrently and were therefore dampened. Our results indicate that prey under threat from conflicting predators may manifest intermediate defensive phenotypes. Such intermediate responses may result in elevated rates of prey mortality with possible consequences at the population level.     

Do prey select for vacant hunting domains to minimize a multi‐predator threat?

Many ecosystems contain sympatric predator species that hunt in different places and times. We tested whether this provides vacant hunting domains, places and times where and when predators are least active, that prey use to minimize threats from multiple predators simultaneously. We measured how northern Yellowstone elk (Cervus elaphus) responded to wolves (Canis lupus) and cougars (Puma concolor), and found that elk selected for areas outside the high‐risk domains of both predators consistent with the vacant domain hypothesis. This enabled elk to avoid one predator without necessarily increasing its exposure to the other. Our results demonstrate how the diel cycle can serve as a key axis of the predator hunting domain that prey exploit to manage predation risk from multiple sources. We argue that a multi‐predator, spatiotemporal framework is vital to understand the causes and consequences of prey spatial response to predation risk in environments with more than one predator.     

Fleeing towards death – leech‐induced behavioural defences increase freshwater snail susceptibility to predatory fish

Prey species are often exposed to multiple predators, which presents several difficulties to prey species. This is especially true when the response to one predator influences the prey’s susceptibility to other predators. Predator‐induced defences have evolved in a wide range of prey species, and experiments involving predators with different hunting strategies allow researchers to evaluate how prey respond to multiple threats. Freshwater snails are known to respond to a variety of predators with both morphological and behavioural defences. Here we studied how freshwater snails Radix balthica responded behaviourally to fish and leech predators, both separately and together. Our aim was to explore whether conflicting predator‐induced responses existed and, if so, what effect they had on snail survival when both predatory fish and leeches were present. We found that although R. balthica increased refuge use when exposed to predatory fish, they decreased refuge use when exposed to predatory leeches. When both predators were present, snails showed a stronger response towards leech than fish and responded by leaving the refuge. This response made the snails more susceptible to fish predation, which increased snail mortality when exposed to both fish and leech compared to fish only. We show that predators that have a relatively low predation rate can substantially increase mortality rates by indirect effects. By forcing snails out of refuges such as rock and macrophyte habitats, leeches can indirectly increase predation from molluscivorous fish and may thus affect snail densities.     

When directional selection reduces geographic variation in traits mediating species interactions

Although we often focus on the causes of geographic variation, understanding processes that act to reduce geographic variation is also important. Here, we consider a process whereby adaptive foraging across the landscape and directional selection exerted by a conifer seed predator, the common crossbill (Loxia curvirostra), potentially act to homogenize geographic variation in the defensive traits of its prey. We measured seed predation and phenotypic selection exerted by crossbills on black pine (Pinus nigra) at two sites in the Pindos Mountains, Greece. Seed predation by crossbills was over an order of magnitude higher at the site where cone scale thickness was significantly thinner, which was also the cone trait that was the target of selection at the high predation site. Additional comparisons of selection differentials demonstrate that crossbills exert selection on black pine that is consistent in form across space and time, and increases in strength with increasing seed predation. If predators distribute themselves in relation to the defensive traits of their prey and the strength of selection predators exert is proportional to the amount of predation, then predators may act to homogenize trait variation among populations of their prey in a process analogous to coevolutionary alternation with escalation.     

Bulgy tadpoles: inducible defense morph

Predator induced morphological defenses are marked morphological shifts induced directly by cues associated with a predator. Generally, remote cues, i.e., chemical substances emitted from predators or injured conspecifics, are considered to be ideal signals to induce morphological change in aquatic environments rather than close cues, i.e., close chemical or tactile cues, since chemical substances that can propagate over relatively long distances and persist for a long period may allow organisms to keep safe and to deliberately change their morph. In fact, most organisms adopting an inducible morphological defense utilize remote chemical cues to detect predation risk and to produce morphological defenses. In this paper, we report a unique and functionally well designed inducible morphological defense strategy where the induction process requires close cues from a predator. The tadpoles of Rana pirica exhibited a bulgy bodied morphology when threatened with predation by larval salamanders, Hynobius retardatus, in close proximity. Predation trials and a function experiment showed that the induced bulgy morph is an adaptive defense phenotype against the gape-limited predator larval H. retardatus. Furthermore, R. pirica tadpoles use two adaptive strategies in terms of cost saving, i.e., adjustment of the extent of bulginess according to predation risk and reversibility by actual shrink of bulgy body after removing the predation threat. In general, R. pirica hatch earlier than H. retardatus. In natural ponds, during the early developmental stage R. pirica tadpoles live in close proximity to young H. retardatus larvae. As they grow, the salamanders gradually become serious predators and the predator–prey interaction becomes intimate. After a while, predation, cannibalism and metamorphosis decrease the number of salamanders in the ponds, and the predator–prey interaction weakens. Such a phenology in the predator–prey interaction allows the evolution of a close-cue detection system and adaptive cost-saving strategies. Our results highlight that the characteristics of the inducible defense depend on the intensity and specificity of the predator–prey system.     

Influence of Body Size on the Responses of Fathead Minnows, Pimephales promelas, to Damselfly Alarm Cues

A wide diversity of aquatic organisms release chemical alarm cues upon encountering or being attacked by a predator. These alarm cues can be used by nearby individuals to assess local predation risk. Receivers warned by chemical alarm cues gain a survival benefit when encountering predators. Animals that are in the same prey guild (i.e. that co‐occur and share the same predators) may learn to recognize each others’ chemical alarm cues. This ability may confer an adaptive advantage if the prey animals are vulnerable to the same predators. However, if the prey grow to different sizes and as a consequence are no longer vulnerable to the same suite of predators, then there should no longer be an advantage for the prey to respond to each others’ alarm cues. In this study, we exposed small and large fathead minnows (Pimephales promelas) to cues from syntopic injured damselfly larvae (Enallagma boreale), cues from injured mealworm larvae (Tenebrio molitor) and to distilled water. Small minnows exhibited antipredatory behaviour and increased shelter use in response to injured damselfly cues but not to the controls of injured mealworm or distilled water. On the contrary, large minnows exhibited no significant change in shelter use in response to any of the injured cues. These data demonstrate that fathead minnows exhibit an antipredator response to damselfly alarm cues, but only when minnows are small and members of the same prey guild as damselfly larvae. These results demonstrate the considerable flexibility in the responses to heterospecific alarm cues.     

Constrained plasticity in switching across life stages: pre- and post-switch predators elicit early hatching

The timing of many life history events shows phenotypic plasticity in response to the risk of predation. Theory predicts that increased risk of mortality in an early stage should select for switching earlier, while a higher risk after the transition should select for switching later. Here we examined the effects of stage-specific predation risk on the timing of hatching of Rana temporaria. Embryos were exposed to chemical cues from either an egg predator (Haemopis sanguisuda) or a tadpole predator (Aeshna cyanea) to evaluate three specific hypotheses: (1) a fixed intermediate response, (2) a ‘fixed predator’ response (i.e., either anticipation or delay), and (3) a specific predator response (both anticipation and delay). Rana temporaria embryos did not discern between pre- and post-hatching specific predators, and they hatched prematurely regardless predator type. These results suggest that R. temporaria embryos respond to predation risk in a fixed way by hatching early, and that they use cues stemming from injured conspecifics, which provides a simple, conservative mechanism of risk assessment. In conclusion, our data are not anticipated by the theoretical consideration that organisms should spend less time in more dangerous environments, but they confirm an invariable adjustment of hatching time in response to an inscrutable predation risk (response to a fixed-predator) in connection with a consistent mechanism mediating the perception of predation risk.     

Hunting‐mediated predator facilitation and superadditive mortality in a European ungulate

Predator‐prey theory predicts that in the presence of multiple types of predators using a common prey, predator facilitation may result as a consequence of contrasting prey defense mechanisms, where reducing the risk from one predator increases the risk from the other. While predator facilitation is well established in natural predator‐prey systems, little attention has been paid to situations where human hunters compete with natural predators for the same prey. Here, we investigate hunting‐mediated predator facilitation in a hunter‐predator‐prey system. We found that hunter avoidance by roe deer (Capreolus capreolus) exposed them to increase predation risk by Eurasian lynx (Lynx lynx). Lynx responded by increasing their activity and predation on deer, providing evidence that superadditive hunting mortality may be occurring through predator facilitation. Our results reveal a new pathway through which human hunters, in their role as top predators, may affect species interactions at lower trophic levels and thus drive ecosystem processes.     

Response of a complex foraging phenotype to artificial selection on its component traits

In nature, where predators must often track dynamic and dispersed prey populations, predator consumption rate, conversion efficiency, dispersal, and prey finding are likely to be important links between foraging and predator–prey population dynamics. Small differences in predator foraging caused by variation in any of the abovementioned traits might lead to significant differences in predator success as well as population dynamics. We used artificial selection to create lines of the predatory mite, Phytoseiulus persimilis in order to determine the potential for or constraints on the evolution of predator foraging behaviors. All four foraging traits demonstrated considerable phenotypic variation. They also exhibited significant realized heritabilities after artificial selection, except that prey finding did not respond to downward selection. Lines that responded to selection did so rapidly, and high-consumption, high-conversion efficiency, and high- and low-dispersal were stable for at least four generations after artificial selection was relaxed. There were some indirect responses to selection among the foraging traits. For example, there was positive correlation between consumption and dispersal. However, none of the correlated responses were of the magnitude of the direct responses we measured on the same trait. We also observed some correlations between foraging traits and life-history traits such as low-consumption and development time (negative), high-consumption and fecundity (positive), and high-conversion efficiency and fecundity (positive), but these were more likely to represent non-genetic constraints. Intrinsic rates of increase in low-consumption and low-conversion efficiency lines were lower than in their respective high lines and the unselected control, whereas rates of increase in dispersal and olfactory response lines did not differ from the unselected control. Thus, traits that make up foraging share partially overlapping genetic architectures with highly heritable phenotypic components, suggesting that each foraging trait will be able to respond rapidly to changes in the density and distribution of resources.     

Reversibility of predator‐induced plasticity and its effect at a life‐history switch point

In natural systems, organisms are frequently exposed to spatial and temporal variation in predation risk. Prey organisms are known to develop a wide array of plastic defences to avoid being eaten. If inducible plastic defences are costly, prey living under fluctuating predation risk should be strongly selected to develop reversible plastic traits and adjust their defences to the current predation risk. Here, we studied the induction and reversibility of antipredator defences in common frog Rana temporaria tadpoles when confronted with a temporal switch in predation risk by dragonfly larvae. We examined the behaviour and morphology of tadpoles in experimental treatments where predators were added or withdrawn at mid larval development, and compared these to treatments with constant absence or presence of predators. As previous studies have overlooked the effects that developing reversible anti‐predator responses could have later in life (e.g. at life history switch points), we also estimated the impact that changes in antipredator responses had on the timing of and size at metamorphosis. In the presence of predators, tadpoles reduced their activity and developed wider bodies, and shorter and wider tails. When predators were removed tadpoles switched their behaviour within one hour to match that found in the constant environments. The morphology matched that in the constant environments in one week after treatment reversal. All these responses were highly symmetrical. Short time lags and symmetrical responses for the induction/reversal of defences suggest that a strategy with fast switches between phenotypes could be favoured in order to maximise growth opportunities even at the potential cost of phenotypic mismatches. We found no costs of developing reversible responses to predators in terms of life‐history traits, but a general cost of the induction of the defences for all the individuals experiencing predation risk during some part of the larval development (delayed metamorphosis). More studies examining the reversibility of plastic defences, including other type of costs (e.g. physiological), are needed to better understand the adaptive value of these flexible strategies.