Risk vs. reward: how predators and prey respond to aging olfactory cues

Many animals use olfaction to find food and avoid predators, and must negotiate environments containing odors of varying compositions, strengths, and ages to distinguish useful cues from background noise. Temporal variation in odor cues (i.e., “freshness”) seems an obvious way that animals could distinguish cues, yet there is little experimental evidence for this phenomenon. Fresh cues provide a more reliable indicator of donor presence than aged cues, but we hypothesize that the benefits of responding to aged cues depend on whether the cue indicates the proximity of a predator or a potential meal. As prey cannot remain eternally risk averse in response to predator odor, we predict that antipredator responses should diminish as predator cues age. In contrast, animals searching for food should investigate aged prey cues if investigation costs are sufficiently low and the potential benefit (a meal) sufficiently high; thus, we predict that predators will maintain interest in aged prey cues. We tested these ideas using free-ranging rats (Rattus spp.) in two separate experiments; firstly assessing giving-up densities in the presence of predator odor, and secondly examining investigation rates of prey odors. As predicted, giving-up densities dropped once predator odor had aged, but investigation rates remained similar for aged and fresh prey odor. Thus, rats used temporal variation in odor cues to evaluate the cost–benefit relationship of responding to predator and prey odors. We suggest that the ecological significance of variable cue age needs more research and should be considered when interpreting behavioral responses to olfactory information.     

Defensive behavior of ants in a mutualistic relationship with aphids

Mutualistic relationships between ants and aphids are well studied but it is unknown if aphid-attending ants place a greater relative importance on defending aphids from aphid-predators or from competing ant colonies. We tested the hypothesis that aphid-attending ants defend their aphids against aphid-predators more aggressively than against ants from neighboring colonies. We conducted introduction trials by placing an individual non-predatory insect, an aphid-predator, or a foreign conspecific ant on the leaf of a resident ant. We found that ants did not attack non-predatory insects, but did attack competing ants and aphid-predators. When we presented resident ants with both the threats (i.e., predator and competitor) at the same time, residents always attacked potential competitors as opposed to aphid-predators. We suggest this behavior may reduce the likelihood of raids by neighboring colonies. Ants appear to balance both the energetic costs of making an attack and the costs associated with losing aphids to a predator, against the benefits of signaling their defensive ability to rivals and/or preventing rivals from gaining knowledge of a potential food resource.     

Harvest policies and nonmarket valuation in a predator — prey system

Although prey may not have commercial value, their economic value can be ascertained in a predator-prey model if the predator has a harvest value. The economic optimal (recovery) path of the predator and prey are carefully described when growth is quadratic in the predator (prey) and linear in prey (predator). Parameter values, in part, resembling Pacific halibut are used to provide numerical illustrations.     

The costs and benefits of predator inspection behaviour in Thomson's gazelles

Summary When Thomson's gazelles (Gazella thomsoni) detect stalking predators, such as cheetahs (Acinonyx jubatus) and lions (Panthera leo), they often approach and follow the predator for up to 72 min (average 14 min). Coursing predators are rarely approached. Gazelle groups were more likely to approach cheetahs if the groups were larger, if the vegetation was low, or if the cheetahs came closer to the group. Immature gazelles were more likely to approach than adults, and a higher proportion of group members participated in inspection behaviour in small groups than in large ones. Gazelles approached closer in less risky situations: if they were in larger groups or if the vegetation was low. Inspection behaviour caused cheetahs to move further between rests and between hunting attempts. Approaching cheetahs was risky, particularly for younger gazelles (probability of being killed while inspecting a cheetah was 1 in 5000 approaches for adults and 1 in 417 approaches for half-grown/adolescent gazelles), and the risks were higher than monitoring cheetahs from a distance. The time costs of predator inspection were also considerable (less than 4.2% of daylight time budget), suggesting that the benefits must be substantial to offset these costs. The results suggested that inspection behaviour was multifunctional, causing stalking predators to move out of the vicinity, enabling gazelles to monitor the predators' movements, and providing an opportunity, particularly for younger animals, to learn about predators. By approaching, gazelles also inform predators that they have been detected and alert other gazelles to the predators' presence.     

Socially acquired predator recognition in complex ecosystems

Social animals acquire information on predator identities through social learning, where individuals with no prior experience learn from experienced members of the group. However, a large amount of uncertainty is often associated with socially acquired information especially in cases of cross-species learning. Theory predicts that socially acquired information from heterospecifics should take more repetitions to develop in complex ecosystems where the number of participants is greater. Our work focuses on coral reef fish as their social and communal lifestyles, along with their complex life histories, make them an ideal model to test for socially acquired predator recognition. Specifically, we tested if Pomacentrus wardi were capable of transmitting the recognition of an unknown predator, Pseudochromis fuscus, to closely related Pomacentrus moluccensis and phylogenetically distant Apogon trimaculatus. Individuals of both species were able to learn the predator's identity from experienced P. wardi based on a single conditioning event. It is somewhat surprising how fast social learning occurred particularly for the distantly related cardinalfish. This study demonstrates the widespread nature of social learning as a method of predator recognition in biologically complex ecosystems, and highlights that the benefits of responding to uncertain information may override the costs associated with lost foraging opportunities.     

Movement of foraging Tundra Swans explained by spatial pattern in cryptic food densities

We tested whether Tundra Swans use information on the spatial distribution of cryptic food items (below ground Sago pondweed tubers) to shape their movement paths. In a continuous environment, swans create their own food patches by digging craters, which they exploit in several feeding bouts. Series of short (<1 m) intra-patch movements alternate with longer inter-patch movements (>1 m). Tuber biomass densities showed a positive spatial auto-correlation at a short distance (<3 m), but not at a larger distance (3-8 m). Based on the spatial pattern of the food distribution (which is assumed to be pre-harvest information for the swan) and the energy costs and benefits for different food densities at various distances, we calculated the optimal length of an inter-patch movement. A swan that moves to the patch with the highest gain rate was predicted to move to the adjacent patch (at 1 m) if the food density in the current patch had been high (>25 g/m2) and to a more distant patch (at 7-8 m) if the food density in the current patch had been low (<25 g/m2). This prediction was tested by measuring the response of swans to manipulated tuber densities. In accordance with our predictions, swans moved a long distance (>3 m) from a low-density patch and a short distance (<3 m) from a high-density patch. The quantitative agreement between prediction and observation was greater for swans feeding in pairs than for solitary swans. The result of this movement strategy is that swans visit high-density patches at a higher frequency than on offer and, consequently, achieve a 38% higher long-term gain rate. Swans also take advantage of spatial variance in food abundance by regulating the time in patches, staying longer and consuming more food from rich than from poor patches. We can conclude that the shape of the foraging path is a reflection of the spatial pattern in the distribution of tuber densities and can be understood from an optimal foraging perspective.     

Specific targeting of host individuals by a kleptoparasitic bird

Kleptoparasitism is a tactic used to acquire food opportunistically and has been shown to provide several benefits, including greater food intake rate and the acquisition of items not normally available during self-foraging. Host individuals may differ in their ability to defend themselves against kleptoparasitic attacks and therefore identifying those host individuals that are particularly vulnerable to attack could both provide energetic benefits and increase the efficiency of kleptoparasitism as a foraging strategy. Here, we show that the kleptoparasitic fork-tailed drongo (Dicrurus adsimilis) specifically targets juveniles when following groups of cooperatively breeding pied babblers (Turdoides bicolor). Drongos give alarm calls upon sighting a predator, thus providing extra predator vigilance to foraging pied babblers. However, drongos also use alarm calls to steal food items. During kleptoparasitic attacks, drongos give false alarm calls and then swoop down to steal food items dropped by alarmed babblers. Juvenile pied babblers are particularly vulnerable to attack because they (a) spend a longer period handling prey items prior to consumption and (b) respond to alarm calls primarily by immediately moving to cover, in contrast to adults who respond by looking up and visually scanning the surrounding area. Drongos attack juvenile babblers significantly more often than adults, with attacks on juveniles more likely to result in the successful procurement of a food item. This patterns of attack suggests that drongos are able to differentiate between individuals of different age when targeting pied babblers, thus increasing the efficiency of kleptoparasitism as a foraging strategy.     

Food and predators affect egg production in song sparrows

Although the possibility that food and predators may interact in limiting avian populations has long been recognized, there have been few attempts to test this experimentally in the field. We conducted a manipulative food addition experiment on the demography of Song Sparrows (Melospiza melodia) across sites that varied in predator abundance, near Victoria, British Columbia, Canada, over three consecutive breeding seasons. We previously showed that food and predators had interactive effects on annual reproductive success (young fledged per female). Here, we report the effects on egg production. Our results show that food limits the total number of eggs laid over the breeding season ("total egg production") and that interactive food and predator effects, including food effects on nest predation, determine how those eggs are "parceled out" into different nests. Food addition alone significantly affected total egg production, and there was no significant interannual variability in this result. At the same time, both food and predators affected the two determinants of total egg production: "clutch number" (total number of clutches laid) and average clutch size. Both clutch number and size were affected by a food x predator x year interaction. Clutch number was lower at low-predator locations because there was less nest predation and thus less renesting. Food addition also significantly reduced nest predation, but there was significant interannual variation in this effect. This interannual variation was responsible for the food x predator x year interactions because the larger the effect of food on nest predation in a given year, the smaller was the effect of food on clutch number; and the smaller the effect of food on clutch number, the larger was the effect of food on clutch size. Potential predator and year effects on total egg production were thus cancelled out by an inverse relationship between clutch number and clutch size. We suggest that combined food and predator effects on demography could be the norm in both birds and mammals.     

Risk taking during parental care: a test of three hypotheses applied to the pied flycatcher

According to life-history theory, there will often be a conflict between investment in current versus future reproduction. If a predator appears during breeding, parents must make a compromise between ensuring the growth and survival of offspring (nest defence, feeding and brooding of young), and reducing the risk of predation to ensure their own survival. We model three hypotheses for the outcome of this conflict which are particularly relevant for altricial birds. They are not mutually exclusive, but focus on different costs and benefits. (1) Parental investment is determined by the parents’ own risk of predation. This hypothesis predicts that a lone parent should take smaller risks than a parent that has a mate. (2) Parental investment is related to the reproductive value of the offspring: Parents are predicted to take greater risks for larger broods, larger-sized or older offspring. (3) Finally, we present the new hypothesis that parental investment is related to the harm that offspring would suffer during a period of no parental care (incubation, brooding, feeding). This hypothesis predicts that parents should take greater risks for younger offspring, or for offspring in poorer condition, because the marginal benefit of parental care is largest in such cases. Hence, one may also expect that lone parents should take greater risks than two parents because their offspring are more in need of care. We tested these hypotheses on the pied flycatcher (Ficedula hypoleuca) by presenting a stuffed predator of the parents (a sparrowhawk, Accipiter nisus) close to the nest when parents were feeding the young. Risk taking was measured as the time that elapsed until the first visit to the nest. Most support was found for the ‘‘harm to offspring’’ hypothesis. Previous studies have usually measured the intensity of nest defence against typical nest predators, and have found evidence for the ‘‘reproductive value of offspring’’ hypothesis. However, our model predicts that the importance of the reproductive value of the offspring should decrease relative to the harm that offspring would suffer if they were not cared for when the predator type changes from a nest predator to a predator of adults, and when conditions for breeding turn from good to bad. Received: 13 April 1995/Accepted after revision: 11 March 1996     

Position preferences within groups: do whirligigs select positions which balance feeding opportunities with predator avoidance?

The benefits and costs of group living are likely to be asymmetric within a group. Animals at the edge of a group are more at risk from predators, according to the selfish herd hypothesis, but are also more likely to obtain scattered food resources. Does an animal's choice between these two conflicting positions depend on its body reserves? The hunger level of marked whirligig beetles (Coleoptera: Gyrinidae) was manipulated and the positions of individuals relative to the rest of the group on the surface of the water were determined with image analysis software. In 12 out of 13 groups, of approximately 18 beetles each, hungry beetles were closer to the edge of a group and had a higher distance to their nearest neighbor than well-fed beetles. Hungry beetles at the edge obtained nearly all of the food particles dropped onto the surface of the water. These results show that position preferences within groups may involve a dynamic feedback between foraging, predator avoidance, and shortterm hunger levels.     

Predator biomass,prey density,and species composition effects on group size in recruit coral reef fishes

Group incidence and size are described for recruit parrotfishes, wrasses, and damselfishes on Hawaiian reefs over 3 years (2006–2008) at sites spanning the archipelago (20–28°N, 155–177°W). Coral-poor and coral-rich areas were surveyed at sites with both low (Hawaii Island) and high (Midway Atoll) predator densities, facilitating examination of relations among predator and recruit densities, habitat, and group metrics. Predator and recruit densities varied spatially and temporally, with a sixfold range in total recruit densities among years. Group (≥2 recruits) metrics varied with time and tracked predator and recruit densities and the proportion of schooling species. Groups often included heterospecifics whose proportion increased with group size. A non-saturating relationship between group size and recruit density suggests that the anti-predator benefits of aggregation exceeded competitive costs. Grouping behavior may have overarching importance for recruit survival–even at high recruit densities–and merits further study on Hawaiian reefs and elsewhere.     

Social capital as a key determinant of perceived benefits of community‐based marine protected areas

Globally, marine protected areas (MPAs) have been relatively unsuccessful in meeting biodiversity objectives. To be effective, they require some alteration of people's use and access to marine resources, which they will resist if they do not perceive associated benefits. Stakeholders’ support is crucial to ecological success of MPAs, and their support is likely to depend on their capacity to adapt to and benefit from MPAs. We examined the influence of social adaptive capacity (SAC) on perceived benefits of MPAs in Siquijor, Philippines, in the Coral Triangle. This region has substantial biodiversity and a population of over 120 million people, many of them dependent on marine resources for food and income. The region has many MPAs, most of which are managed under decentralized governance systems. We collected survey data from 540 households in 19 villages with associated MPAs. We evaluated the influence of multiple SAC variables (e.g., occupational multiplicity and social capital) on perceived benefits with decision trees (CHAID) and qualitatively analyzed this relationship with respect to types and recipients of benefits. Our models revealed the key role of social capital, particularly trust in leadership, in influencing perceptions of benefits (χ² = 14.762, p = 0.000). A path analysis revealed that perceptions of distributional equity were a key mechanism through which social capital affected perceived MPA benefits (root mean‐square error of approximation = 0.050). Building social capital and equity within communities could lead to more effective management of MPAs and thus to expenditure of fewer resources relative to, for example, regulation enforcement.     

The effects of opportunistic and intentional predators on the herding behavior of prey

In this article, we study how predator behavior influences the aggregation of prey into herds. Game-theoretic models of herd formation are developed based on different survival probabilities of solitary prey and prey that join the herd and on the predator's preference of what type of prey to search for. For an intentional predator that will only pursue its preferred type of prey, a single herd with no solitaries cannot emerge unless the herd acts as a prey refuge. If neither prey choice provides a refuge, it is shown that an equilibrium always exists where there are both types of prey and the predator does not always search for the same type of prey (i.e., a mixed equilibrium exists). On the other hand, if the predator is opportunistic in that it sometimes shifts to pursue the type of prey that is observed first, there may be a single herd equilibrium that does not act as a prey refuge when there is a high level of opportunistic behavior. For low opportunistic levels, a mixed equilibrium is again the only outcome. The evolutionary stability of each equilibrium is tested to see if it predicts the eventual herding behavior of prey in its corresponding model. Our analysis confirms that both predator and prey preferences (for herd or solitary) have strong effects on why prey aggregate. In particular, in our models, only the opportunistic predator can maintain all prey in a single herd that is under predation risk.     

Benefits and costs of earwig (Forficula auricularia) family life

The evolution of parental care and family group formation critically depends on offspring survival benefits and parental fecundity costs of care under given ecological conditions. Investigations of the functional significance of care in insect species that exhibit facultative parental care have been relatively rare but may be of particular interest for better understanding of benefit and cost schedules at an early evolutionary stage. In this study, aspects of benefits and costs of care were addressed in the sub-social European earwig (Forficula auricularia; Dermaptera: Forficulidae) by manipulating the presence of tending mothers and brood size in a fully crossed experimental design. Larvae growing in broods tended by their mother or of reduced size showed a higher survival probability than larvae growing in untended or large broods, as predicted if maternal care is beneficial and shaped by a trade-off between number and quality of offspring. Analysis of patterns of food consumption and developmental time further suggested that the benefit of maternal attendance is mediated by the maternal provisioning of food, while the quality–quantity trade-off seemed to be driven by sibling rivalry. Further, tending mothers delayed the production of a second clutch, indicating a potential cost of care in terms of lifetime fecundity. This study experimentally shows benefits and potential costs of maternal care and family group formation in the European earwig. More detailed behavioural experiments will be required to fully understand how behavioural interactions among family members mediate these reproductive outcomes.     

Behavioral games involving a clever prey avoiding a clever predator: An individual-based model of dusky dolphins and killer whales

Faced with an intermittent but potent threat, animals exhibit behavior that allows them to balance foraging needs and avoid predators and over time, these behaviors can become hard-wired adaptations with both species trying to maximize their own fitness. In systems where both predator and prey share similar sensory modalities and cognitive abilities, such as with marine mammals, the dynamic nature of predator-prey interactions is poorly understood. The costs and benefits of these anti-predator adaptations need to be evaluated and quantified based on the dynamic engagement of predator and prey. Many theoretic models have addressed the complexity of predator-prey relationships, but few have translated into testable mechanistic models. In this study, we developed a spatially-explicit, geo-referenced, individual-based model of a prototypical adult dusky dolphin off Kaikoura, New Zealand facing a more powerful, yet infrequent predator, the killer whale. We were interested in two primary objectives, (1) to capture the varying behavioral game between a clever prey and clever predator based on our current understanding of the Kaikoura system, (2) to compare evolutionary costs vs. benefits (foraging time and number of predator encounters) for an adult non-maternal dusky dolphin at various levels of killer whale-avoidance behaviors and no avoidance rules. We conducted Monte Carlo simulations to address model performance and parametric uncertainty. Mantel tests revealed an 88% correlation (426 × 426 distance matrix, km²) between observed field sightings of dusky dolphins with model generated sightings for non-maternal adult dusky dolphin groups. Simulation results indicated that dusky dolphins incur a 2.7% loss in feeding time by evolving the anti-predator behavior of moving to and from the feeding grounds. Further, each evolutionary strategy we explored resulted in dolphins incurring an additional loss of foraging time. At low killer whale densities (appearing less than once every 3 days), each evolutionary strategy simulated converged towards the evolutionary cost of foraging, that is, the loss in foraging time approached the 2.7% loss experienced by evolving near shore-offshore movement behavior. However, the highest level of killer whale presence resulted in 38% decreases in foraging time. The biological significance of these losses potentially incurred by a dusky dolphin is dependent on various factors from dolphin group foraging behavior and individual energy needs to dolphin prey availability and behavior.     

Effects of combining an intraguild predator with a cannibalistic intermediate predator on a species-level trophic cascade

A greater diversity of natural enemies can in some cases disrupt prey suppression, particularly when natural enemies engage in intraguild predation, where natural enemies compete with and prey upon each other. However, empirical studies have often demonstrated enhanced prey suppression despite intraguild predation. A recent theoretical study proposed the hypothesis that, when the intermediate predator is cannibalistic, intraguild predation can reduce cannibalism within the intermediate predator population, leading to little change in intermediate predator mortality and thus enhanced prey suppression. The goal of this study was to examine this hypothesis empirically. Two summer-long field enclosure experiments were conducted in cotton fields. We investigated the effects of adding an intraguild predator, Zelus renardii, on (1) the abundance of a cannibalistic intermediate predator, Geocoris pallens, (2) the abundance of a herbivore, Lygus hesperus, and (3) cotton plant performance. G. pallens adult abundance did not increase, even when food availability was high and natural enemies were absent, suggesting that density-dependent cannibalism imposes an upper limit on its densities. Furthermore, although Z. renardii is an intraguild predator of G. pallens, G. pallens long-term densities were unaffected by Z. renardii. In the presence of the intermediate predator, the addition of the intraguild predator Z. renardii enhanced suppression of L. hesperus, and there were suggestions that Z. renardii and G. pallens partitioned the L. hesperus population. Effects of herbivore suppression cascaded to the plant level, improving plant performance. In conclusion, we provide empirical support for the hypothesis that the addition of an intraguild predator may enhance prey suppression if the intermediate predator expresses density-dependent cannibalism. Intraguild predation and cannibalism co-occur in many communities; thus their joint effects may be broadly important in shaping predator effects on herbivores and plant performance.     

Predator inspection behaviour in three-spined sticklebacks (<Emphasis Type="Italic">Gasterosteus aculeatus</Emphasis>): body size,local predation pressure and cooperation

The apparently maladaptive tendency of fish to approach and inspect potential predators has been explained in terms of useful information gathering or as a signal to the predator that it has been seen. We examined this behaviour in 16 populations of wild-caught stickleback (Gasterosteus aculeatus) from ponds with and without predatory perch (Perca fluviatilis). Three large and three small individuals per population were each exposed to three model predators differing in realism. A final cooperative treatment entailed pairing subjects with a second individual from the same population, but of the alternative size class, during predator presentation. As might be expected, predator inspection behaviour was much greater in the predator-sympatric populations, and only these fish increased their level of inspection as the models became incrementally more realistic. This suggests that reductions occur in the level of costly inspection behaviour in populations without predators. Subject body size had no effect on inspection effort, which suggests a limited role for experience (we assumed larger fish to be older than smaller fish), at least over the relative age differences utilized. However, small predator-sympatric fish were the only subjects to increase inspection significantly when in a cooperative context, perhaps reflecting the inherent value of a relatively larger partner in this context. These results confirm that levels of predator inspection are both population- and situation-dependent, suggesting a trade-off in the potential costs and benefits of this behaviour.Communicated by C. St. Mary     

Antipredator responses of koomal (Trichosurus vulpecula hypoleucus) against introduced and native predators

Antipredator behavior studies generally assess prey responses to single predator species although most real systems contain multiple species. In multi-predator environments prey ideally use antipredator responses that are effective against all predator species, although responses may only be effective against one predator and counterproductive for another. Multi-predator systems may also include introduced predators that the prey did not co-evolve with, so the prey may either fail to recognize their threat (level 1 naiveté), use ineffective responses (level 2 naiveté) or succumb to their superior hunting ability (level 3 naiveté). We analyzed microhabitat selection of an Australian marsupial (koomal, Trichosurus vulpecula hypoleucus) when faced with spatiotemporal differences in the activity/density levels of one native (chuditch, Dasyurus geoffroii) and two introduced predators (red fox, Vulpes vulpes; feral cat, Felis catus). From this, we inferred whether koomal recognized introduced predators as a threat, and whether they minimized predation risk by either staying close to trees and/or using open or dense microhabitats. Koomal remained close to escape trees regardless of the predator species present, or activity/density levels, suggesting koomal employ this behavior as a first line of defense. Koomal shifted to dense cover only under high risk scenarios (i.e., with multiple predator species present at high densities). When predation risk was low, koomal used open microhabitats, which likely provided benefits not associated with predator avoidance. Koomal did not exhibit level 1 naiveté, although further studies are required to determine if they exhibit higher levels of naiveté (2–3) against foxes and cats.     

Boldness by habituation and social interactions: a model

Most studies of animal personality attribute personality to genetic traits. But a recent study by Magnhagen and Staffan (Behav Ecol Sociobiol 57:295–303, 2005) on young perch in small groups showed that boldness, a central personality trait, is also shaped by social interactions and by previous experience. The authors measured boldness by recording the duration that an individual spent near a predator and the speed with which it fed there. They found that duration near the predator increased over time and was higher the higher the average boldness of other group members. In addition, the feeding rate of shy individuals was reduced if other members of the same group were bold. The authors supposed that these behavioral dynamics were caused by genetic differences, social interactions, and habituation to the predator. However, they did not quantify exactly how this could happen. In the present study, we therefore use an agent-based model to investigate whether these three factors may explain the empirical findings. We choose an agent-based model because this type of model is especially suited to study the relation between behavior at an individual level and behavioral dynamics at a group level. In our model, individuals were either hiding in vegetation or feeding near a predator, whereby their behavior was affected by habituation and by two social mechanisms: social facilitation to approach the predator and competition over food. We show that even if we start the model with identical individuals, these three mechanisms were sufficient to reproduce the behavioral dynamics of the empirical study, including the consistent differences among individuals. Moreover, if we start the model with individuals that already differ in boldness, the behavioral dynamics produced remained the same. Our results indicate the importance of previous experience and social interactions when studying animal personality empirically.     

Seasonal changes in the diet and feeding behaviour of a top predator indicate a flexible response to deteriorating oceanographic conditions

Shifts in the diet of top predators can be linked to changes in environmental conditions. In this study, we tested relationships between environmental variation and seasonal changes in diet of a top predator, the grey-headed albatross Thalassarche chrysostoma, breeding at Bird Island, South Georgia in an austral summer of 1999/2000. Oceanographic conditions in that year around South Georgia were abnormal (i.e. anomalously high sea surface temperature to a relative 19-year long-term mean). The diet of grey-headed albatrosses showed high seasonal variation, shifting from cephalopods (42.9 % by mass) in late February to Antarctic krill Euphausia superba (58.3 %) in late April, and grey-headed albatrosses breeding performance was low (16.8 %). This study shows these albatrosses did not manage to find sufficient alternative prey and highlight the risk to top predators if there is an increase in the frequency or severity of food shortages in Antarctic waters.