First Documentation of Cultural Transmission of Predator Recognition by Larval Amphibians

Predation is a pervasive selective agent shaping a prey's behaviour, morphology and life history. To survive, prey animals have to respond adaptively to predation threats and this can be achieved through learned predator recognition. Cultural transmission of predator recognition is likely a widespread means of learning in social animals, including mammals, birds and fishes. However, no studies have investigated the cultural transmission of predator recognition in amphibians. In our study, we examined whether naïve woodfrog (Rana sylvatica) tadpoles can acquire the recognition of the odour of a predatory tiger salamander (Ambystoma tigrinum) from experienced conspecifics. After conditioning some tutors to recognize salamander odour, we paired naïve observer tadpoles with either a salamander‐naïve or salamander‐experienced tutor and exposed the pairs to either salamander odour or a water control. Observers were subsequently tested alone for a response to salamander odour. We found that when given salamander odour, observer tadpoles that were paired with a salamander‐experienced tutor successfully learned to recognize the salamander odour as a threat, whereas the observers paired with salamander‐naïve tutors did not. Likewise, tadpoles exposed to the water control did not learn to recognize the salamander regardless of whether they were paired with a naïve or experienced tutor. This is the first study demonstrating cultural transmission of predator recognition in an amphibian species.     

Population and behavioural responses of native prey to alien predation

The introduction of invasive alien predators often has catastrophic effects on populations of naïve native prey, but in situations where prey survive the initial impact a predator may act as a strong selective agent for prey that can discriminate and avoid it. Using two common species of Australian small mammals that have persisted in the presence of an alien predator, the European red fox Vulpes vulpes, for over a century, we hypothesised that populations of both would perform better where the activity of the predator was low than where it was high and that prey individuals would avoid signs of the predator’s presence. We found no difference in prey abundance in sites with high and low fox activity, but survival of one species—the bush rat Rattus fuscipes—was almost twofold higher where fox activity was low. Juvenile, but not adult rats, avoided fox odour on traps, as did individuals of the second prey species, the brown antechinus, Antechinus stuartii. Both species also showed reduced activity at foraging trays bearing fox odour in giving-up density (GUD) experiments, although GUDs and avoidance of fox odour declined over time. Young rats avoided fox odour more strongly where fox activity was high than where it was low, but neither adult R. fuscipes nor A. stuartii responded differently to different levels of fox activity. Conservation managers often attempt to eliminate alien predators or to protect predator-naïve prey in protected reserves. Our results suggest that, if predator pressure can be reduced, otherwise susceptible prey may survive the initial impact of an alien predator, and experience selection to discriminate cues to its presence and avoid it over the longer term. Although predator reduction is often feasible, identifying the level of reduction that will conserve prey and allow selection for avoidance remains an important challenge.     

Improving anti‐predator responses of hatchery reared salmonids by social learning

Predation shortly after release is the main source of mortality among hatchery‐reared fish used to restore or enhance endangered salmonid populations. We found, that hatchery‐reared salmonid young originating from endangered stocks have weak innate responses to their natural fish predators. The ability to avoid predation in fish can be improved through social learning from experienced to naïve individuals. Huge benefits would be achieved, if social learning processes could be successfully applied on a large scale to enhance viability of hatchery fish prior to release into the wild. By using model predators together with chemical cues from real predators we tested if social learning could be used to train hatchery‐reared salmonid young to avoid fish predators. As there are clear differences in social behaviour among the salmonid species, we first examined whether these differences affect the probability and efficiency of learning anti‐predator skills from trained demonstrators. We compared anti‐predator responses of observers (fish trained by using experienced fish as demonstrators) with those of control fish, which had been ‘trained’ by untrained naïve conspecifics. We also examined how the efficiency of social learning depends on the ratio of experienced to naïve fish involved in social transmission trials. The results of these experiments will give guidelines how social learning could be utilized in developing hatchery scale training protocols.     

Predator–prey naïveté, antipredator behavior,and the ecology of predator invasions

We present a framework for explaining variation in predator invasion success and predator impacts on native prey that integrates information about predator–prey naïveté, predator and prey behavioral responses to each other, consumptive and non‐consumptive effects of predators on prey, and interacting effects of multiple species interactions. We begin with the ‘naïve prey’ hypothesis that posits that naïve, native prey that lack evolutionary history with non‐native predators suffer heavy predation because they exhibit ineffective antipredator responses to novel predators. Not all naïve prey, however, show ineffective antipredator responses to novel predators. To explain variation in prey response to novel predators, we focus on the interaction between prey use of general versus specific cues and responses, and the functional similarity of non‐native and native predators. Effective antipredator responses reduce predation rates (reduce consumptive effects of predators, CEs), but often also carry costs that result in non‐consumptive effects (NCEs) of predators. We contrast expected CEs versus NCEs for non‐native versus native predators, and discuss how differences in the relative magnitudes of CEs and NCEs might influence invasion dynamics. Going beyond the effects of naïve prey, we discuss how the ‘naïve prey’, ‘enemy release’ and ‘evolution of increased competitive ability’ (EICA) hypotheses are inter‐related, and how the importance of all three might be mediated by prey and predator naïveté. These ideas hinge on the notion that non‐native predators enjoy a ‘novelty advantage’ associated with the naïveté of native prey and top predators. However, non‐native predators could instead suffer from a novelty disadvantage because they are also naïve to their new prey and potential predators. We hypothesize that patterns of community similarity and evolution might explain the variation in novelty advantage that can underlie variation in invasion outcomes. Finally, we discuss management implications of our framework, including suggestions for managing invasive predators, predator reintroductions and biological control.     

Naïveté in novel ecological interactions: lessons from theory and experimental evidence

The invasion of alien species into areas beyond their native ranges is having profound effects on ecosystems around the world. In particular, novel alien predators are causing rapid extinctions or declines in many native prey species, and these impacts are generally attributed to ecological naïveté or the failure to recognise a novel enemy and respond appropriately due to a lack of experience. Despite a large body of research concerning the recognition of alien predation risk by native prey, the literature lacks an extensive review of naïveté theory that specifically asks how naïveté between novel pairings of alien predators and native prey disrupts our classical understanding of predator–prey ecological theory. Here we critically review both classic and current theory relating to predator–prey interactions between both predators and prey with shared evolutionary histories, and those that are ecologically ‘mismatched’ through the outcomes of biological invasions. The review is structured around the multiple levels of naïveté framework of Banks & Dickman (2007), and concepts and examples are discussed as they relate to each stage in the process from failure to recognise a novel predator (Level 1 naïveté), through to appropriate (Level 2) and effective (Level 3) antipredator responses. We discuss the relative contributions of recognition, cue types and the implied risk of cues used by novel alien and familiar native predators, to the probability that prey will recognise a novel predator. We then cover the antipredator response types available to prey and the factors that predict whether these responses will be appropriate or effective against novel alien and familiar native predators. In general, the level of naïveté of native prey can be predicted by the degree of novelty (in terms of appearance, behaviour or habitat use) of the alien predator compared to native predators with which prey are experienced. Appearance in this sense includes cue types, spatial distribution and implied risk of cues, whilst behaviour and habitat use include hunting modes and the habitat domain of the predator. Finally, we discuss whether the antipredator response can occur without recognition per se, for example in the case of morphological defences, and then consider a potential extension of the multiple levels of naïveté framework. The review concludes with recommendations for the design and execution of naïveté experiments incorporating the key concepts and issues covered here. This review aims to critique and combine classic ideas about predator–prey interactions with current naïveté theory, to further develop the multiple levels of naïveté framework, and to suggest the most fruitful avenues for future research.     

Social learning of predators in the dark: understanding the role of visual,chemical and mechanical information

The ability of prey to observe and learn to recognize potential predators from the behaviour of nearby individuals can dramatically increase survival and, not surprisingly, is widespread across animal taxa. A range of sensory modalities are available for this learning, with visual and chemical cues being well-established modes of transmission in aquatic systems. The use of other sensory cues in mediating social learning in fishes, including mechano-sensory cues, remains unexplored. Here, we examine the role of different sensory cues in social learning of predator recognition, using juvenile damselfish (Amphiprion percula). Specifically, we show that a predator-naive observer can socially learn to recognize a novel predator when paired with a predator-experienced conspecific in total darkness. Furthermore, this study demonstrates that when threatened, individuals release chemical cues (known as disturbance cues) into the water. These cues induce an anti-predator response in nearby individuals; however, they do not facilitate learnt recognition of the predator. As such, another sensory modality, probably mechano-sensory in origin, is responsible for information transfer in the dark. This study highlights the diversity of sensory cues used by coral reef fishes in a social learning context.     

Learned Recognition of Intraspecific Predators in Larval Long-Toed Salamanders Ambystoma macrodactylum

The ability of prey to detect predators and respond accordingly is critical to their survival. The use of chemical cues by animals in predator detection has been widely documented. In many cases, predator recognition is facilitated by the release of alarm cues from conspecific victims. Alarm cues elicit anti‐predator behavior in many species, which can reduce their risk of being attacked. It has been previously demonstrated that adult long‐toed salamanders, Ambystoma macrodactylum, exhibit an alarm response to chemical cues from injured conspecifics. However, whether this response exists in the larval stage of this species and whether it is an innate or a learned condition is unknown. In the current study, we examined the alarm response of naïve (i.e. lab‐reared) larval long‐toed salamanders. We conducted a series of behavioral trials during which we quantified the level of activity and spatial avoidance of hungry and satiated focal larvae to water conditioned by an injured conspecific, a cannibal that had recently been fed a conspecific or a non‐cannibal that was recently fed a diet of Tubifex worms. Focal larvae neither reduced their activity nor spatially avoided the area of the stimulus in either treatment when satiated, and exhibited increased activity towards the cannibal stimulus when hungry. We regard this latter behavior as a feeding response. Together these results suggest that an anti‐predator response to injured conspecifics and to cannibalistic conspecifics is absent in naïve larvae. Previous studies have shown that experienced wild captured salamanders do show a response to cannibalistic conspecifics. Therefore, we conducted an additional experiment examining whether larvae can learn to exhibit anti‐predator behavior in response to cues from cannibalized conspecifics. We exposed larvae to visual, chemical and tactile cues of stimulus animals that were actively foraging on conspecifics (experienced) or a diet of Tubifex (naïve treatment). In subsequent behavioral treatments, experienced larvae significantly reduced their activity compared to naive larvae in response to chemical cues of cannibals that had recently consumed conspecifics. We suggest that this behavior is a response to alarm cues released by consumed conspecifics that may have labeled the cannibal. Furthermore, over time, interactions with cannibals may cause potential prey larvae to learn to avoid cannibals regardless of their recent diet.     

Predator experience homogenizes consistent individual differences in predator avoidance

In the presence of predators, many prey species exhibit immediate behavioral responses like the avoidance of risky areas, which imposes opportunity costs, for instance, in the form of reduced foraging. Thus, prey species should be able to discriminate between different predator types and adjust their response to the imminent predation risk. In our current study, we evaluated the relative importance of innate versus learned components of predator recognition and avoidance in the guppy (Poecilia reticulata). We used a feral guppy population occurring in Germany and compared avoidance reactions of each focal individual towards both coevolved piscine predators from their original distribution range and novel, presently co-occurring predator species. Wild-caught, predator-experienced as well as laboratory-reared, predator-naïve individuals showed strong avoidance responses towards all predator animations. Avoidance was stronger in small-bodied than in large-bodied individuals in both cohorts; however, this effect was significant only in predator-naïve fish. Moreover, wild-caught individuals showed a significantly higher within-individual variance (across the six predator species) along with a lower among-individual variance in predator avoidance, which resulted in a lower behavioral repeatability in this cohort. Our results suggest that consistent individual differences in risk-taking behavior (also referred to as the personality trait ‘boldness’) are modified by predator exposure and learning about predators.     

Sex‐Specific Response of Pardosa milvina (Araneae: Lycosidae) to Experience with a Chemotactile Predation Cue

Predators frequently leave behind chemical information (i.e., semiochemicals such as pheromones or kairomones) that can be detected by their prey and used to avoid areas where predators are likely present. Prey that have interacted indirectly with predators via chemical information thus may gain insight into their risk of being consumed that naïve individuals lack. Pardosa milvina (Araneae: Lycosidae) is a chemosensitive wolf spider that shows adaptive responses to chemotactile cues deposited by the larger wolf spider Tigrosa helluo. We raised offspring from P. milvina to examine the effect of experience with a predation cue on activity, foraging, and antipredator behavior. Spiders differed in activity and foraging behavior across ontogeny and between sexes, but there was no effect of experience with a predation cue. However, a sex‐specific effect of experience was found in antipredator behavior. Male spiders, but not females, used experience with a predator cue to increase their survival in the presence of a live predator. Specifically, naïve males were attacked sooner than experienced males, indicating that prior exposure to predator cues can modify Pardosa antipredator behavior. Intersexual differences in how spiders respond to experience with a predation cue likely reflect the risk of predation faced by males and females in nature.     

Orientation behavior of Propylaea japonica toward visual and olfactory cues from its prey–host plant combination

The lady beetle Propylaea japonica (Thunberg) (Coleoptera: Coccinellidae) is an important predator of aphids in agroecosystems. The inundative release of coccinellid beetles can be an effective biological control strategy. An understanding of how biological control agents perceive and use stimuli from host plants is the key to successfully implement commercially produced predators. Here, we studied the relative role of visual and volatile cues. Dual‐choice assays using foraging‐naïve and foraging‐experienced P. japonica adults were conducted using cotton plants [Gossypium hirsutum L. (Malvaceae)] with or without infestation by the cotton aphid, Aphis gossypii (Glover) (Hemiptera: Aphididae). Overall, experienced beetles were more attracted than naïve beetles toward cues associated with aphid‐infested plants. Experienced beetles were also more responsive to olfactory cues compared with naïve beetles. Both foraging‐naïve and ‐experienced lady beetles integrate olfactory and visual cues from plants infested with aphids, with an apparently greater reliance on olfactory cues. The results suggest that foraging experience may increase prey location in P. japonica.     

Predator‐Naïve Brown Trout (Salmo trutta) Show Antipredator Behaviours to Scent from an Introduced Piscivorous Mammalian Predator Fed Conspecifics

Introduced mammalian predators may pose a high risk for native and naïve prey populations, but little is known about how native fish species may recognize and respond to scents from introduced mammalian predators. We investigated the role of diet‐released chemical cues in facilitating predator recognition, hypothesizing that native brown trout (Salmo trutta) would exhibit antipredator behaviours to faeces scents from the introduced American mink (Neovision vison) fed conspecifics, but not to non‐trout diets. In treatments‐control and replicate stream tank experiments, brown trout showed significant antipredator responses to faeces scent from mink fed conspecifics, but not to faeces scent from mink fed a non‐trout diet (chicken), or the non‐predator food control, Eurasian beaver (Castor fiber). We conclude that native and naïve brown trout show relevant antipredator behaviours to an introduced mammalian predator, presumably based on diet‐released conspecific alarm cues and thereby estimate the predation risk.     

Determining sensitive stages for learning to detect predators in larval bronzed frogs: Importance of alarm cues in learning

Successful survival and reproduction of prey organisms depend on their ability to detect their potential predators accurately and respond effectively with suitable defences. Predator detection can be innate or can be acquired through learning. We studied prey–predator interactions in the larval bronzed frogs (Sylvirana temporalis), which have the innate ability to detect certain predators. We conducted a series of experiments to determine if the larval S. temporalis rely solely on innate predator detection mechanisms or can also learn to use more specific cues such as conspecific alarm cues for the purpose. The results of our study clearly indicate that larval S. temporalis use both innate and learned mechanisms for predator detection. Predator-naïve tadpoles could detect kairomones alone as a potential threat and responded by reducing activity, suggesting an innate predator detection mechanism. Surprisingly, predator-naïve tadpoles failed to detect conspecific alarm cues as a potential threat, but learned to do so through experience. After acquiring the ability to detect conspecific alarm cues, they could associate novel predator cues with conspecific alarm cues. Further, post feeding stages of larval S. temporalis are sensitive for learning to detect conspecific alarm cues to label novel predators.     

Response of the zoophytophagous predators Macrolophus pygmaeus and Nesidiocoris tenuis to volatiles of uninfested plants and to plants infested by prey or conspecifics

Knowledge about the orientation mechanisms used by two important predaceous mirids (Macrolophus pygmaeus Rambour and Nesidiocoris tenuis (Reuter)) in finding their prey (whitefly Bemisia tabaci (Gennadius) and the tomato borer Tuta absoluta (Meyrick)) is limited. In a Y-tube olfactometer, we tested the behavioral responses of naïve and experienced predators to uninfested plants, herbivore-induced plant volatiles (HIPVs) from plants infested with T. absoluta and/or B. tabaci, the sex pheromone of T. absoluta, and volatiles produced by plants injured by the predators. Nesidiocoris tenuis responds to volatiles produced by uninfested plants only after experience with the plant, whereas naïve and experienced M. pygmaeus show positive chemotaxis. Both predators are attracted to volatiles from prey-infested plants, and we provide the first evidence that experience affects this response in M. pygmaeus. Infestation of the same plant by both prey species elicited similar responses by the two predators as plants infested by either herbivore singly. Neither predator responded to sex pheromones of T. absoluta. Macrolophus pygmaeus avoided plants injured by conspecifics, while N. tenuis females were attracted by such plants. The implications of these results for augmentative biological control are discussed.     

Social learning enhances search image acquisition in foraging brook trout

Brook trout feed primarily on prey items for which they have developed a search image. Periodically, fish are required to update their search images to match temporal variability in prey abundance. It is currently unknown how brook trout develop search images, but there is evidence to suggest that it could be influenced by social learning, wherein a fish learns a novel behavior through observation alone. Here we demonstrate that adult brook trout use social learning to quickly develop search images for novel prey. We trained a set of demonstrators to develop a search image for canned mealworms and subsequently moved demonstrators to treatment pools that contained naïve bystanders. We also had control pools that contained naïve bystanders and sham demonstrators that had not been trained on mealworms. Over an 8-day period, bystanders in treatment pools consumed 68 % of mealworms, compared to only 36 % in control pools. Moreover, social learning of search images was rapid as bystanders in treatment pools began feeding on mealworms in less than 1 day, whereas it took 7 days for bystanders in control pools to feed heavily on mealworms. Social learning of search images confers an adaptive advantage by reducing energetic costs of foraging.     

Within‐patch oviposition site shifts by spider mites in response to prior predation risks decrease predator patch exploitation

In egg‐laying animals with no post‐oviposition parental care, between‐ or within‐patch oviposition site selection can determine offspring survival. However, despite the accumulation of evidence supporting the substantial impact predators have on oviposition site selection, few studies have examined whether oviposition site shift within patches (“micro‐oviposition shift”) reduces predation risk to offspring. The benefits of prey micro‐oviposition shift are underestimated in environments where predators cannot disperse from prey patches. In this study, we examined micro‐oviposition shift by the herbivorous mite Tetranychus kanzawai in response to the predatory mite, Neoseiulus womersleyi, by testing its effects on predator patch exploitation in situations where predatory mites were free to disperse from prey patches. Adult T. kanzawai females construct three‐dimensional webs on leaf surfaces and usually lay eggs under the webs; however, females that have experienced predation risks, shift oviposition sites onto the webs even in the absence of current predation risks. We compared the predation of eggs on webs deposited by predator‐experienced females with those on leaf surfaces. Predatory mites left prey patches with more eggs unpredated when higher proportions of prey eggs were located on webs, and egg survival on webs was much higher than that on leaf surfaces. These results indicate that a micro‐oviposition shift by predator‐experienced T. kanzawai protects offspring from predation, suggesting adaptive learning and subsociality in this species. Conversely, fecundity and longevity of predator‐experienced T. kanzawai females were not reduced compared to those of predator‐naïve females; we could not detect any costs associated with the learned micro‐oviposition shift. Moreover, the previously experienced predation risks did not promote between‐patch dispersal of T. kanzawai females against subsequently encountered predators. Based on these results, the relationships of between‐patch oviposition site selection and micro‐oviposition shift are discussed.     

Selection of aphid prey by a generalist predator: do prey chemical defences matter?

1. For predators, prey selection should maximise nutrition and minimise fitness costs. In the present study, it was investigated whether a generalist predator [Chrysoperla carnea (Stephens) lacewing larvae] rejected harmful, chemically‐defended prey [Brevicoryne brassicae (Linnaeus) aphids] when non‐defended prey [Myzus persicae (Sulzer) aphids] were available. 2. It was tested: (i) whether consuming different prey species affects predator mortality; (ii) whether naïve predators reject chemically‐defended prey while foraging when non‐defended prey are available; (iii) whether the relative abundance of each prey affects the predator's prey choice; and (iv) whether predators learn to avoid consuming chemically‐defended prey after exposure to both prey species. 3. Consumption of B. brassicae yielded greater C. carnea mortality than M. persicae consumption, but naïve C. carnea did not reject B. brassicae in favour of M. persicae during foraging. When presented at unequal abundances, naïve predators generally consumed each aphid species according to their initial relative abundance, although, predation of non‐defended prey was less than expected when defended prey were initially more abundant, indicating a high consumption of B. brassicae impeded M. persicae consumption. With experience, C. carnea maintained predation of both aphid species but consumed more M. persicae than B. brassicae, indicating a change in behaviour. 4. Although prey choice by C. carnea may change with experience of available prey, prey chemical defences do not appear to influence prey choice by naïve predators. This inability to avoid harmful prey could facilitate wider, indirect interactions. Myzus persicae may benefit where high consumption of B. brassicae hinders predators in the short term, and in the long term, increases predator mortality.     

Is the hibiscus harlequin bug aposematic? The importance of testing multiple predators

Aposematism involves predators learning conspicuous signals of defended prey. However, prey species utilize a wide range of chemical (or physical) defenses, which are not likely to be equally aversive to all predators. Aposematism may therefore only be effective against a physiologically sensitive subset of potential predators, and this can only be identified through behavioral testing. We studied the emerging model organism Tectocoris diophthalmus (Heteroptera: Scutelleridae), an aposematically colored but weakly defended shieldback stinkbug, to test the efficacy of its defenses against a suite of predator types. We predicted the bugs' defenses would be ineffectual against both experienced and naïve birds but aversive to predaceous insects. Surprisingly, the opposite pattern was found. Both habituated wild passerines and naïve chickens avoided the bugs, the chickens after only one or two encounters. To avian predators, T. diophthalmus is aposematic. However, praying mantids showed no repellency, aversion, or toxicity associated with adult or juvenile bugs after multiple trials. Comparison with prior studies on mantids using bugs with chemically similar but more concentrated defenses underscores the importance of dose in addition to chemical identity in the efficacy of chemical defenses. Our results also emphasize the importance of behavioral testing with multiple ecologically relevant predators to understand selective pressures shaping aposematic signals and chemical defenses.     

Responses of naïve and experienced European rabbits to predator odour

The response of prey species to predator scent has been investigated in many mammalian species; however, there is little information about the responses of European wild rabbits at the population level. Therefore, we conducted a simple experiment to investigate the behavioural response of a rabbit population to native predator cues in the wild. We compared the response to the scent of a predator (red fox) in a wild rabbit population bred in semi-natural conditions and naïve to terrestrial predators with the response of a population in a similar environment where terrestrial predators were present. The response to predators was based on rabbit abundance, inferred from pellet counts and measured by the defecation rate per day (DRD). Our results indicate that rabbits responded to the odour of fox faeces in the treatment warrens, resulting in a lower DRD. The main anti-predator behaviour observed was spatial avoidance (warren abandonment), which seemed to be more accentuated for rabbits who had not previously had contact with foxes in the plot where terrestrial predators were excluded. In both the fenced and the unfenced plot, the differences in the effect of the predator odour between the control and treatment warrens disappeared after cessation of treatment, suggesting a flexible and adaptive behaviour of rabbits to predator cues.

     

Chemically mediated learning in juvenile rainbow trout. Does predator odour pH influence intensity and retention of acquired predator recognition?

The prediction that variability in ambient pH will influence the intensity and retention of learned predator recognition in juvenile rainbow trout Oncorhynchus mykiss was tested under laboratory conditions. Juvenile rainbow trout were conditioned to recognize the odour of a novel predator at pH 6·0 or 7·0 and then tested for learned recognition of the predator odour at pH 6·0 or 7·0 at 2 or 7 days post-conditioning. When tested 2 days post-conditioning, rainbow trout exhibited a significant learned antipredator response regardless of predator odour pH. The response was stronger, however, when the test pH matched the conditioning pH. When tested 7 days post-conditioning, rainbow trout only exhibited a learned response when conditioning and testing pH were the same. These results demonstrate that episodic acidification may impair the strength and retention of acquired predator recognition learning. Given the demonstrated survival benefits associated with learned predator recognition in prey fishes, such impairment will probably have considerable negative impacts at both individual and population levels.     

Generalization of learned predator recognition: an experimental test and framework for future studies

While some prey species possess an innate recognition of their predators, others require learning to recognize their predators. The specific characteristics of the predators that prey learn and whether prey can generalize this learning to similar predatory threats have been virtually ignored. Here, we investigated whether fathead minnows that learned to chemically recognize a specific predator species as a threat has the ability to generalize their recognition to closely related predators. We found that minnows trained to recognize the odour of a lake trout as a threat (the reference predator) generalized their responses to brook trout (same genus as lake trout) and rainbow trout (same family), but did not generalize to a distantly related predatory pike or non-predatory suckers. We also found that the intensity of antipredator responses to the other species was correlated with the phylogenetic distance to the reference predator; minnows responded with a higher intensity response to brook trout than rainbow trout. This is the first study showing that prey have the ability to exhibit generalization of predator odour recognition. We discuss these results and provide a theoretical framework for future studies of generalization of predator recognition.