Insular tammar wallabies (Macropus eugenii) respond to visual but not acoustic cues from predators

We studied the way in which a population of tammar wallabies(Macropus eugenii), which have been isolated from mammalianpredators since the last ice age, responded to the sight andsound of historical and ontogenetically and evolutionarilynovel predators. Tammars were shown a range of visual stimuli,including taxidermic mounts of two evolutionarily novel predators,a red fox (Vulpes vulpes) and a cat (Felis catus), and a modelof an extinct predator, the thylacine (Thylacinus cynocephalus).Controls were a conspecific, the cart on which all mounts werepresented, and blank trials in which spontaneous change in behaviorwas measured. We played back recorded sounds to characterizeresponses to acoustic cues from predators and to a putativeconspecific antipredator signal. Treatments included the howlsof dingoes (Canis lupus dingo), an evolutionarily novel predator;calls of a wedge-tailed eagle (Aquila audax), a historicaland current predator; and wallaby foot thumps. Controls werethe song of an Australian magpie (Gymnorhina tibicen) and ablank trial. After seeing a fox, wallabies thumped their hindfeet in alarm, suppressed foraging, and increased looking.The sight of a cat similarly suppressed foraging and increasedlooking. The sounds of predators did not influence responsiveness,but wallabies foraged less and looked more after thump playbacks.Our results suggest that tammars respond to the sight, butnot the sounds, of predators. In contrast, the response to footthumps demonstrates that this particular sound functions asan antipredator signal. We suggest that responsiveness to visualcues has been preserved under relaxed selection because predatormorphology is convergent, but vocalizations are not.     

Know thy enemy: Behavioural response of a native mammal (Rattus lutreolus velutinus) to predators of different coexistence histories

Abstract Predation is recognized as a major selective pressure influencing population dynamics and evolutionary processes. Prey species have developed a variety of predator avoidance strategies, not least of which is olfactory recognition. However, within Australia, European settlement has brought with it a number of introduced predators, perhaps most notably the red fox (Vulpes vulpes) and domestic cat (Felis catus), which native prey species may be unable to recognize and thus avoid due to a lack of coexistence history. This study examined the response of native Tasmanian swamp rats (Rattus lutreolus velutinus) to predators of different coexistence history (native predator‐ spotted‐tail quoll (Dasyurus maculatus), domestic cats and the recently introduced red fox). We used an aggregate behavioural response of R. l. velutinus to predator integumental odour in order to assess an overall behavioural response to predation risk. Rattus lutreolus velutinus recognized the integumental odour of the native quoll (compared with control odours) but did not respond to either cat or fox scent (compared with control odur). In contrast, analyses of singular behaviours resulted in the conclusion that rats did not respond differentially to either native or introduced predators, as other studies have concluded. Therefore, measuring risk assessment behaviours at the level of overall aggregate response may be more beneficial in understanding and analysing complex behavioural patterns such as predator detection and recognition. These results suggest that fox and cat introductions (and their interactive effects) may have detrimental impacts upon small native Tasmanian mammals due to lack of recognition and thus appropriate responses.     

Phenotypic modifications to conspecific density arising from predation risk assessment

Organisms often perceive predation risk through visual, auditory, or chemical cues that accompany or persist after an attack on other prey individuals. In this paper an argument is developed that suggests that it is adaptive for species that use such indirect cues to include conspecific density in the assessment of predation risk, and to respond to conspecific density by modifying phenotype (e.g. behavior, morphology of life history). A model based on this argument predicts that at equivalent (including negligible) indirect cue levels an organism should adopt less vulnerable phenotypes at lower conspecific density. Further, the phenotypic modifications to differences in conspecific density are predicted to be on the same order of magnitude as phenotypic responses to differences in predator density, to be more pronounced at higher indirect cue levels, and can be extended to responses to differences in the density of heterospecific species that share predators. This "risk assessment" mechanism is qualitatively different from other mechanisms, such as the dilution and "many eyes" effects that predict behavioral responses to conspecific density. If species use conspecific or heterospecific density to assess predation risk as predicted, there may be implications for the role and evolution of traits used to perceive conspecific and heterospecific densities, species aggregation, and population dynamics, and should be considered in the design of experiments of nonlethal effects of predators.     

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.     

Prey Responses to Predator's Sounds: A Review and Empirical Study

Many animals assess their risk of predation by listening to and evaluating predators' vocalizations. We reviewed the literature to draw generalizations about predator discrimination abilities, the retention of these abilities over evolutionary time, and the potential underlying proximate mechanisms responsible for discrimination. Broadly, we found that some prey possess an ability to respond to a predator after having been evolutionarily isolated from a specific predator (i.e., predators are allopatric) and that some prey are predisposed to respond to certain types of predators that they coevolved with but without having ecological experience. However, these types of studies are lacking, and relatively, few studies have examined predator discrimination abilities in ungulates. To begin addressing these knowledge gaps, we performed field experiments on Mule deer (Odocoileus hemionus) in which we investigated the ability of deer to discriminate among familiar predators [coyotes (Canis latrans) and mountain lions (Puma concolor)] and an evolutionary relevant predator with which deer have had no recent exposure [locally extinct wolves (Canis lupus)]. We found that Mule deer respond to and discriminate among predators based on predator vocalizations and have retained an ability to respond to wolves that have been extinct from the study area since the early 20th century. Previous playback studies have shown that responses vary among human‐habituated and non‐habituated populations and differ according to human proximity. Deer greater than 0.5 km from human residences allocated more time to heightened responses both before and after stimulus playback. Our findings may help predict how prey–predator interactions may change as a result of the recovering wolf population with a basis in ecological and evolutionary experience in predator discrimination and desensitization.     

Conspecific alarm cues, but not predator cues alone, determine antipredator behavior of larval southern marbled newts, Triturus pygmaeus

Predation imposes selection on the ability of prey to recognize and respond to potential threats. Many prey species detect predators via chemoreception, particularly in aquatic environments. Also, chemical cues from injured prey are often perceived as an indication of predation risk. However, because antipredatory behavior can be costly, prey responses should depend on the current level of risk that each predator poses, which may depend on the type of chemical cues detected. We exposed larval newts, Triturus pygmaeus, to chemical cues from predator larval beetles or to alarm cues from conspecific larval newts and examined the behavioral changes of larval newts. Results showed that larval newts reduced activity levels when conspecific alarm cues were present but not when the predator cues alone were present. These results might suggest that larval newts are unable to recognize predator chemicals. To avoid costs of unnecessary antipredatory behaviors, larval newts may benefit by avoiding only predators that represent a current high level of threat, showing only antipredatory responses when they detect conspecific alarm cues indicating that an actual predatory attack has occurred.     

Selective Learning in a Marsupial

Behavioural plasticity allows animals to adjust rapidly to local environmental conditions, but at the risk of erroneously changing behaviour in response to irrelevant events. Adaptive biases or predispositions constrain learning and reduce such potential costs. Preferential learning about complex biologically‐meaningful stimuli, such as predators, has been investigated in only a few systems and there have been no experimental tests for the presence of adaptive biases in a marsupial. We have previously shown that tammar wallabies (Macropus eugenii) became fearful of a model fox (Vulpes vulpes) after it was repeatedly paired with an aversive event. Tammars generalized their acquired response to a cat (Felis catus), but not to a non‐predator (juvenile goat, Capra hircus), suggesting that they might have a bias to associate predators with frightening events. The present study tested this idea directly. We used an experimental design identical to that of earlier predator‐training experiments, but substituted a model goat for the fox as the stimulus predicting a capture attempt. A control group had the same total experience of the goat and of a human with a net, but without any predictive relationship between these two events. We detected no change in behaviour towards the goat, or to any of an array of control stimuli, as a consequence of training. This finding contrasts strongly with the effects of the same pairing procedure using a fox model. Taken together, these studies provide the first evidence for an adaptive predisposition to acquire a fear of predators in marsupials. Learning processes in this group are thus evolutionarily convergent with those previously described in eutherian mammals.     

A Test of the Multi-Predator Hypothesis: Rapid Loss of Antipredator Behavior after 130 years of Isolation

Many species find themselves isolated from the predators with which they evolved. Isolation often leads to the loss of costly antipredator behavior, which may have adverse consequences if the population should later come into contact with predators. An understanding of both the mechanism (i.e. the degree to which antipredator behavior depends on experience), and of the time course of loss is important to be able to predict how a population will respond to future contact. We studied ‘group‐size effects’– the way in which animals change the time they allocate to antipredator vigilance as a function of group size – and visual and acoustic predator recognition in a population of tammar wallabies (Macropus eugenii), a cat‐sized (6–10 kg) macropodid marsupial. To study group size effects we observed wallabies foraging in four populations – three with some sort of predator and a New Zealand population that was isolated from all predators for about 130 yr. To study predator recognition, we observed the response of New Zealand wallabies to the presentation of a model or taxidermic mount of mammalian predators, and to the broadcast sounds of mammalian and avian predators. We compare these predator recognition experiments with results from a previous study of Kangaroo Island (South Australia) tammars. Complete isolation from all predators for as few as 130 yr led to the loss of group size effects and a rapid breakdown in visual predator recognition abilities. Our results are consistent with a key prediction of the multi‐predator hypothesis – namely, that the isolation from all predators may lead to a rapid loss of antipredator behavior.     

There must be something in the water: assessing the behavioral responses of rusty crayfish (<Emphasis Type="Italic">Orconectes rusticus</Emphasis>) to fish and amphibian predator kairomones

Animals use chemical cues to find food, locate mates, and detect potential predators. Detecting cues in a risky environment can induce behavioral changes to increase survival. Rusty crayfish (Orconectes rusticus) reduce activity, increase refuge use, and make defensive displays after detecting fish predator cues. However, no studies have introduced amphibian cues. We investigated crayfish responses to hellbender salamander (Cryptobranchus alleganiensis, a dominant predator of crayfish) cues and compared these to responses to largemouth bass (Micropterus salmoides) cues. Largemouth bass occur sympatrically with hellbenders and are known to induce distinct responses in rusty crayfish. We randomly assigned crayfish to predator cue and conspecific alarm cue combinations and recorded frozen behavior, appendage movement, locomotion, and refuge use. We found crayfish increased their proportion of time spent frozen and reduced their proportion of time spent active in the tank when exposed to either predator cue. Moreover, these responses were magnified when crayfish were exposed to predator cues in combination with conspecific alarm cues. Our experiment demonstrates evidence in support of the crayfish’s ability to detect and appropriately respond to predator cues alone and in combination with conspecific alarm cues. Future work should investigate the effects of these behavioral changes on trophic dynamics in a natural system.     

Antipredator Responses to a Previously Neutral Sound by Free-living Adult Golden-mantled Ground Squirrels, Spermophilus lateralis (Sciuridae)

I investigated whether free-living adult golden-mantled ground squirrels ( Spermophilus lateralis ) could learn to associate a novel sound with the appearance of a predator. Using a classical conditioning protocol, I presented adult squirrels with a novel tone followed by the appearance of a model hawk. After repeated trials, I again measured the squirrels' behavioral responses to the tone and compared them with their pretraining responses. Individuals that experienced the tone paired with the hawk responded with antipredator behavior of longer duration after pairing than before and these responses were indistinguishable from responses to natural, conspecific alarm calls. By examining the process, rather than simply the pattern, of call recognition ontogeny, this study provides new evidence for a mechanism by which animals can develop the ability to respond to both conspecific and heterospecific calls.     

Effects of ocean acidification on learning in coral reef fishes

Ocean acidification has the potential to cause dramatic changes in marine ecosystems. Larval damselfish exposed to concentrations of CO(2) predicted to occur in the mid- to late-century show maladaptive responses to predator cues. However, there is considerable variation both within and between species in CO(2) effects, whereby some individuals are unaffected at particular CO(2) concentrations while others show maladaptive responses to predator odour. Our goal was to test whether learning via chemical or visual information would be impaired by ocean acidification and ultimately, whether learning can mitigate the effects of ocean acidification by restoring the appropriate responses of prey to predators. Using two highly efficient and widespread mechanisms for predator learning, we compared the behaviour of pre-settlement damselfish Pomacentrus amboinensis that were exposed to 440 μatm CO(2) (current day levels) or 850 μatm CO(2), a concentration predicted to occur in the ocean before the end of this century. We found that, regardless of the method of learning, damselfish exposed to elevated CO(2) failed to learn to respond appropriately to a common predator, the dottyback, Pseudochromis fuscus. To determine whether the lack of response was due to a failure in learning or rather a short-term shift in trade-offs preventing the fish from displaying overt antipredator responses, we conditioned 440 or 700 μatm-CO(2) fish to learn to recognize a dottyback as a predator using injured conspecific cues, as in Experiment 1. When tested one day post-conditioning, CO(2) exposed fish failed to respond to predator odour. When tested 5 days post-conditioning, CO(2) exposed fish still failed to show an antipredator response to the dottyback odour, despite the fact that both control and CO(2)-treated fish responded to a general risk cue (injured conspecific cues). These results indicate that exposure to CO(2) may alter the cognitive ability of juvenile fish and render learning ineffective.     

Behavioral responses of native prey to disparate predators: naiveté and predator recognition

It is widely accepted that predator recognition and avoidance are important behaviors in allowing prey to mitigate the impacts of their predators. However, while prey species generally develop anti-predator behaviors through coevolution with predators, they sometimes show accelerated adoption of these behaviors under strong selection pressure from novel species. We used a field manipulation experiment to gauge the ability of the common ringtail possum (Pseudocheirus peregrinus), a semi-arboreal Australian marsupial, to recognize and respond to olfactory cues of different predator archetypes. We predicted that ringtails would display stronger anti-predator behaviors to cues of the invasive European red fox (Vulpes vulpes) in areas where fox impacts had been greatest, and to cues of the native lace monitor (Varanus varius) in areas of sympatry compared with allopatry. We found that ringtails fled quickly and were more alert when exposed to the fecal odors of both predators compared to neutral and pungent control odors, confirming that predator odors are recognized and avoided. However, these aversive responses were similar irrespective of predator presence or level of impact. These results suggest that selection pressure from the fox has been sufficient for ringtails to develop anti-predator behaviors over the few generations since foxes have become established. In contrast, we speculate that aversive responses by ringtails to the lace monitor in areas where this predator is absent reflect recent coexistence of the two species. We conclude that rapid evolution of anti-predator behaviors may occur when selection is strong. The maintenance of these behaviors should allow re-establishment of predator–prey relationships if the interactants regain sympatry via range shifts or management actions to reintroduce them to their former ranges.     

Red fox trap success is correlated with piping plover chick survival

Predation management is an important component of managing species of conservation concern. The piping plover (Charadrius melodus; plover), a disturbance-dependent and conservation-reliant shorebird that nests on sandy beaches and barrier islands on the Atlantic Coast, was listed under the United States Endangered Species Act in 1986, with habitat loss and predation stated as key causes of its decline. We evaluated the relationship between a suite of predators and plover chick survival from 2015–2018. We used a camera grid to establish indices of the abundance of 3 known chick predators: red fox (Vulpes vulpes), raccoon (Procyon lotor), and domestic cat. We used camera detections in a survival model to assess potential relationships between predator species detection and plover chick survival. Plover chick survival was negatively related with red fox detection, but not with detection of the other 2 predators. In addition to the correlation with red fox detections, chick survival was negatively related to high plover nesting density. Our results suggest that red foxes were predators of piping plover chicks during our study, likely augmented by other density-dependent sources of mortality. Targeted predator management could aid in conservation of piping plovers in this system as a short-term solution, but long-term recovery plans must also address habitat limitation.     

Bottom-up and top-down processes interact to modify intraguild interactions in resource-pulse environments

Top predators are declining globally, in turn allowing populations of smaller predators, or mesopredators, to increase and potentially have negative effects on biodiversity. However, detection of interactions among sympatric predators can be complicated by fluctuations in the background availability of resources in the environment, which may modify both the numbers of predators and the strengths of their interactions. Here, we first present a conceptual framework that predicts how top-down and bottom-up interactions may regulate sympatric predator populations in environments that experience resource pulses. We then test it using 2 years of remote-camera trapping data to uncover spatial and temporal interactions between a top predator, the dingo Canis dingo, and the mesopredatory European red fox Vulpes vulpes and feral cat Felis catus, during population booms, declines and busts in numbers of their prey in a model desert system. We found that dingoes predictably suppress abundances of the mesopredators and that the effects are strongest during declines and busts in prey numbers. Given that resource pulses are usually driven by large yet infrequent rains, we conclude that top predators like the dingo provide net benefits to prey populations by suppressing mesopredators during prolonged bust periods when prey populations are low and potentially vulnerable.     

Predators,fire or resources: What drives the distribution of herbivores in fragmented mesic forests?

Trophic interactions and disturbance events can shape the structure and function of ecosystems. However, the effects of drivers such as predation, fire and climatic variables on species distributions are rarely considered concurrently. We used a replicated landscape‐scale predator management experiment to compare the effects of red fox Vulpes vulpes control, time‐since‐fire, vegetation type and other environmental variables on native herbivore distributions. Occurrence data for four native herbivores and an invasive predator – the red fox – were collected from 240 sites across three baited (for lethal fox control) and three unbaited forest blocks (4659–9750 ha) in south‐western Victoria, Australia, and used to build species distribution models. The herbivore taxa were as follows: red‐necked wallaby Macropus rufogriseus, black wallaby Wallabia bicolour, grey kangaroo Macropus fuligenosus and Macropus giganteus and common brushtail possum Trichosurus vulpecula. Fox control and fire had little effect on herbivore occurrence, despite the literature suggesting it can influence abundance, while climate, proximity to farmland and topography were more influential. This may be because the region’s high productivity and agricultural pastures subsidise food resources for both predators and prey within the forest blocks and so dampen trophic interactions. Alternatively, these drivers may affect herbivore abundance, but not herbivore occurrence. Understanding the drivers of herbivore distributions is an important step in predicting the effects of herbivory on other species, particularly after management interventions such as predator control and prescribed burns.     

Predation-Related Costs and Benefits of Conspecific Attraction in Songbirds—An Agent-Based Approach

Songbirds that follow a conspecific attraction strategy in the habitat selection process prefer to settle in habitat patches already occupied by other individuals. This largely affects the patterns of their spatio-temporal distribution and leads to clustered breeding. Although making informed settlement decisions is expected to be beneficial for individuals, such territory clusters may potentially provide additional fitness benefits (e.g., through the dilution effect) or costs (e.g., possibly facilitating nest localization if predators respond functionally to prey distribution). Thus, we hypothesized that the fitness consequences of following a conspecific attraction strategy may largely depend on the composition of the predator community. We developed an agent-based model in which we simulated the settling behavior of birds that use a conspecific attraction strategy and breed in a multi-predator landscape with predators that exhibited different foraging strategies. Moreover, we investigated whether Bayesian updating of prior settlement decisions according to the perceived predation risk may improve the fitness of birds that rely on conspecific cues. Our results provide evidence that the fitness consequences of conspecific attraction are predation-related. We found that in landscapes dominated by predators able to respond functionally to prey distribution, clustered breeding led to fitness costs. However, this cost could be reduced if birds performed Bayesian updating of prior settlement decisions and perceived nesting with too many neighbors as a threat. Our results did not support the hypothesis that in landscapes dominated by incidental predators, clustered breeding as a byproduct of conspecific attraction provides fitness benefits through the dilution effect. We suggest that this may be due to the spatial scale of songbirds’ aggregative behavior. In general, we provide evidence that when considering the fitness consequences of conspecific attraction for songbirds, one should expect a trade-off between the benefits of making informed decisions and the costs of clustering.     

Overcoming prey naiveté: Free‐living marsupials develop recognition and effective behavioral responses to alien predators in Australia

Naiveté in prey arises from novel ecological mismatches in cue recognition systems and antipredator responses following the arrival of alien predators. The multilevel naiveté framework suggests that animals can progress through levels of naiveté toward predator awareness. Alternatively, native prey may be preadapted to recognize novel predators via common constituents in predator odors or familiar predator archetypes. We tested predictions of these competing hypotheses on the mechanisms driving behavioral responses of native species to alien predators by measuring responses of native free‐living northern brown bandicoots (Isoodon macrourus) to alien red fox (Vulpes vulpes) odor. We compared multiple bandicoot populations either sympatric or allopatric with foxes. Bandicoots sympatric with foxes showed recognition and appropriate antipredator behavior toward fox odor via avoidance. On the few occasions bandicoots did visit, their vigilance significantly increased, and their foraging decreased. In contrast, bandicoots allopatric with foxes showed no recognition of this predator cue. Our results suggest that vulnerable Australian mammals were likely naïve to foxes when they first arrived, which explains why so many native mammals declined soon after fox arrival. Our results also suggest such naiveté can be overcome within a relatively short time frame, driven by experience with predators, thus supporting the multilevel naiveté framework.     

Responses to terrestrial nest predators by endemic and introduced Hawaiian birds

Birds free from nest predators for long periods may either lose the ability to recognize and respond to predators or retain antipredator responses if they are not too costly. How these alternate scenarios play out has rarely been investigated in an avian community whose members have different evolutionary histories. We presented models of two nest predators (rat and snake) and a negative control (tree branch) to birds on Hawai?i Island. Endemic Hawaiian birds evolved in the absence of terrestrial predators until rats were introduced approximately 1,000 years ago. Introduced birds evolved with diverse predator communities including mammals and snakes, but since their introduction onto the island approximately one century ago have been free from snake predation. We found that (a) endemic and introduced birds had higher agitation scores toward the rat model compared with the branch, and (b) none of the endemic birds reacted to the snake model, while one introduced bird, the Red‐billed Leiothrix (Leiothrix lutea), reacted as strongly to the snake as to the rat. Overall, endemic and introduced birds differ in their response to predators, but some endemic birds have the capacity to recognize and respond to introduced rats, and one introduced bird species retained recognition of snake predators from which they had been free for nearly a century, while another apparently lost that ability. Our results indicate that the retention or loss of predator recognition by introduced and endemic island birds is variable, shaped by each species' unique history, ecology, and the potential interplay of genetic drift, and that endemic Hawaiian birds could be especially vulnerable to introduced snake predators.     

Responses of larval dragonflies to conspecific and heterospecific predator cues

Abstract.  1. In cannibalistic populations, smaller individuals are subject to predation by larger conspecifics, and small individuals commonly alter their behaviour in response to cannibals. Little is known, however, about the underlying cues that trigger such responses and how the behavioural responses to conspecific cannibals differ from heterospecific predators.
2. This study tests which cues are used for the detection of conspecific predators in the larva of the dragonfly Plathemis lydia and how the behavioural response to cannibals differed from the response to heterospecific predators.
3. Individuals were exposed to chemical cues, visual cues, and a combination of both cues from conspecifics as well as no predator and heterospecific predator controls during which their activity and feeding rates were observed.
4. Individuals increased their activity, spatial movement and feeding behaviour in response to either visual or chemical cues from conspecific predators, which was opposite to responses displayed with cues from heterospecific predators. Interestingly, the responses to visual and chemical cues from conspecifics combined were weaker than to either cue in isolation and similar to the no cue control.
5. The results clearly indicate that individuals are able to use chemical and visual cues to detect even very subtle differences in phenotype of conspecific predators.
6. The opposite response in behaviour when exposed to conspecific cannibals vs. heterospecific predators suggests that the presence of cannibals will increase the mortality risk of small individuals due to heterospecific predation. This risk-enhancement is likely to have important consequences for the dynamics of predator–prey interactions.     

An island-wide predator manipulation reveals immediate and long-lasting matching of risk by prey

Anti-predator behaviour affects prey population dynamics, mediates cascading effects in food webs and influences the likelihood of rapid extinctions. Predator manipulations in natural settings provide a rare opportunity to understand how prey anti-predator behaviour is affected by large-scale changes in predators. Here, we couple a long-term, island-wide manipulation of an important rodent predator, the island fox (Urocyon littoralis), with nearly 6 years of measurements on foraging by deer mice (Peromyscus maniculatus) to provide unequivocal evidence that prey closely match their foraging behaviour to the number of fox predators present on the island. Peromyscus maniculatus foraging among exposed and sheltered microhabitats (a measure of aversion to predation risk) closely tracked fox density, but the nature of this effect depended upon nightly environmental conditions known to affect rodent susceptibility to predators. These effects could not be explained by changes in density of deer mice over time. Our work reveals that prey in natural settings are cognizant of the dynamic nature of their predators over timescales that span many years, and that predator removals spanning many generations of prey do not result in a loss of anti-predator behaviour.