Health-dependent vulnerability to predation affects escape responses of unguarded chinstrap penguin chicks

Behav Ecol Sociobiol (6) 6:778 78 DOI.7/s65-6-- ORIGINAL ARTICLE Health-dependent vulnerability to predation affects escape responses of unguarded chinstrap penguin chicks J. Martín & L. de Neve & V. Polo & J. A. Fargallo & M. Soler Received: 9 December 5 /Revised: 8 March 6 / Accepted: May 6 / Published online: June 6 # Springer-Verlag 6 Abstract Predators may select more often to attack the more vulnerable prey or those with an inferior health status. Thus, prey should be able to assess their own vulnerability to predation and modify their antipredatory behavior accordingly. When approached by predator skuas, unguarded penguin chicks flee short distances, and usually aggregate in dense packs, but there is a clear interindividual variability in their responses under similar conditions. We hypothesized that this variability in escape responses might be related to the perceived vulnerability to predation of each individual chick. We simulated predator attacks to chinstrap penguin chicks and analyzed the sources of variation in their escape response, such as the presence of adults or the density of other chicks, and the sex, age, body condition, and health status of responding chicks. Chicks allowed shorter approach distances when they had a better health condition (i.e., a greater T-cell-mediated immunity, CMI), when they were younger, and when the density of adults around was higher. Sex and density of other chicks were not important. Similarly, chicks fled from the experimenter to longer distances when they had a lower CMI and when the density of adults was lower. Therefore, Communicated by C. Brown J. Martín (*) : L. de Neve : V. Polo : J. A. Fargallo Departamento de Ecología Evolutiva, Museo Nacional de Ciencias Naturales, C.S.I.C., José Gutiérrez Abascal, 86 Madrid, Spain e-mail: Jose.Martin@mncn.csic.es M. Soler Departamento de Biología Animal y Ecología, Facultad de Ciencias, Universidad de Granada, 87 Granada, Spain escape characteristics of chicks depended on the presence of adults that can deter predators and on the healthdependent vulnerability of chicks. Keywords Prey vulnerability. Penguins. Health condition. Immunocompetence. Fear. Antipredatory behavior Introduction When predation risk changes, prey animals should be able to modify the magnitude of their antipredatory behavior to the actual level of threat because antipredatory responses can be costly in terms of loss of time and energy (Lima and Dill 99; Ydenberg and Dill 986). Moreover, recent studies have shown that predators may respond to changing vulnerability of their prey on quite a subtle level, selecting more often to attack the more vulnerable individuals (Cresswell and Quinn ; Quinn and Cresswell ). Thus, prey should be able to assess their own vulnerability to predation and modify their antipredatory behavior accordingly. A prey can theoretically assess vulnerability on the basis of many measures such as microhabitat characteristics, predator behavior, group size, or activity (Smith and Belk ; Stankowich and Blumstein 5). Prey vulnerability should be highest when animals are under energetic stress because they are forced to allocate fewer resources to antipredatory behavior (Lima 998).For example,by foraging in more risky habitats or reducing vigilance rate, prey may increase intake rate, but may also increase their vulnerability (Caraco et al. 98; Quinn and Cresswell ). Vulnerability of individuals within groups may decrease when group size increases because of the dilution effect, where the risk of any one individual being selected is reduced simply by being in a

Behav Ecol Sociobiol (6) 6:778 78 779 larger group (Roberts 996), or because animals may rely on collective detection of a threat via other wary animals in the group (e.g., Creswell et al. ). In addition, predators are usually considered to prey upon individuals of inferior phenotypic quality because these would be easier to subdue and capture, although this hypothesis has only rather rarely been tested (Temple 986; Møller and Erritzøe ; Pole et al. ). These studies found that some raptors differentially preyed upon mammals that were heavily parasitized (Temple 986), and that passerine birds falling prey to feral cats have small spleens, which implies that avian prey often have a poor health status (Møller and Erritzøe ). These data provided indirect evidence that predators preferentially prey on individuals in poor condition that suffer from parasitism and disease and with an inferior health status. This selection might be explained as simply a result of a differential hunting success (Fitzgibbon 99; Funston et al. ), but predators could also be able to identify and choose to attack these most vulnerable prey (Cresswell and Quinn ; Quinn and Cresswell ). In any case, health-dependent vulnerability of prey should also affect their antipredator responses. In the penguin breeding colonies, predation by skuas (Catharacta spp.) is the major cause of chick mortality (Hunter 99; Young 99; Emslie et al. 995). One of the parents uses to provide continuous protection to young chicks (Viñuela et al. 995), but relatively older chicks are often left unattended by both parents while they forage in the sea, and chicks often form more or less dense aggregations until emancipation (Davis 98; Williams 995; Penteriani et al. ). Chicks are still vulnerable to predation by skuas during this chick aggregation stage (Davis 98),and grouping may function to provide protection from predators and/or from inclement weather (Davis 98; Lawless et al. ; Besnard et al. ; Le Bohec et al. 5). However, older chicks are very difficult to subdue by skuas because these older chicks are much heavier than skuas and can defend themselves very aggressively (Young 99). Skuas, then, behave more like scavengers and tend to feed on carcasses of penguins (Norman et al. 99). Nevertheless, skuas are often observed harassing chicks within the aggregations and can still kill some chicks (Davis 98; Young 99). When approached by skuas, unguarded chicks of the chinstrap penguin, Pygoscelis antarctica, use to flee short distances and usually aggregate in dense packs of chicks, which are supposed to decrease predation risk (Cresswell 99a; Young 99). However, there is an apparently clear variability in the escape responses of different individual chicks under similar conditions. Some individuals flee sooner than others, and even some individuals do not flee at all but standstill and face the approaching skua, threatening it with open bill and loud shouts (personal observation). We hypothesized that this variability might be related to the perceived vulnerability to predation of each individual chick, which might depend not only on the probability of being defended by adults from an attack, and on the dilution effect of being in a group with other chicks, but also on the statedependent ability of each individual chick to defend itself. In this paper, we simulated predatory attacks and examined the escape response of chinstrap penguin chicks during the chick aggregation phase in an Antarctic breeding colony. We analyzed the sources of variation in the escape response of chicks, such as the presence of some adult penguins around, or the density of other chicks, and examined whether the characteristics of these responses could be state-dependent, i. e., related to the chicks health status and body condition, independently of age, sex, and the presence of conspecifics. Materials and methods Study area and general methods We performed the study at the Vapour Col chinstrap penguin colony (, breeding pairs) on Deception Island, South Shetlands (6 S, 6 W) during the austral summer. In this colony, penguin eggs and chicks are commonly predated by Subantarctic skua (C. antarctica). During the breeding season, we followed up a total of penguin nests ( nests in each of subcolonies) from incubation until the chick aggregation phase. The chicks of these nests were individually marked with insulation tape and later with numbered metallic rings on one flipper. Flipper-bands were removed at the end of the study to avoid their potential detrimental costs to chicks (Jackson and Wilson ). Because we had visited daily the nest to control whether the eggs had hatched, we knew the exact age of each of these chicks at the moment of carrying out this study. Sex was determined by molecular procedures using DNA extracted from blood taken from the external foot web from young chicks (Ellegren 996). Escape behavior The experiment was done shortly after the guard phase, when adults left chicks unguarded and these started to form aggregations (Viñuela et al. 996; Penteriani et al. ), and before independence. Unguarded penguin chicks found in aggregations were between 8 and 8 days of age (mean± SE=.±. days, n=68). We searched the area between 9 and 6 h (GMT) by walking slowly between the groups of chicks until a marked penguin chick was sighted using binoculars. We noted without disturbing it the number of other chicks and adult penguins in a -m circle around the focal chick. This ratio was chosen after some trials because it allowed us to easily count

78 Behav Ecol Sociobiol (6) 6:778 78 individuals just before the attack and because it was considered to include individuals actually affected by the attack (i.e., a larger circle would include individuals too far from the focal chick that very often did not respond to the approach). Then, we approached chicks by simulating an approaching skua by walking directly and fast (ca. m/ min) toward the focal chick. Previous studies have suggested that penguins and other animals do not actually respond to the presence of a human but to the speed of approach of a large object (Burger and Gochfeld 98; Van Heezik and Seddon 99; Yorio and Boersma 99), which would be similar to the approach of a predator. Thus, presumably, a rapid direct approach should be considered by penguins as a higher risk of predation because it should indicate a high probability of a serious attack (Martín et al. ). We are confident that with this procedure, we simulated a predator attack because observations of skua approaches to groups of chicks indicated that these reacted with a similar antipredatory behavior to natural skua predators than to the human experimenter (Martín et al. ; unpublished data). Similarly, many other animals react to humans as if they were potential predators (Frid and Dill ). To avoid confounding effects that may affect risk perception (e.g., Burger and Gochfeld 98), the same person wearing the same clothing performed all approaches in a similar way, walking at the same medium speed, while another person recorded the chicks behavior. Also, because approach distance in birds may be dependent on intruder starting distance (Blumstein ), we started approaches at the same distance in all cases (approximately 5 m). This starting distance was established after some preliminary trials. Penguins typically made a short flight at relatively fast speed, and then stopped, or later walked slowly. We noted the approach distance as the distance between the penguin and the observer when the penguin first moved (straight-line measured to the nearest.5 m), and the distance fled as the distance across which a penguin fled away from the observer until first stopping (Ydenberg and Dill 986; Martín and López ; Martín et al. ). Body condition and cell-mediated immunity Immediately after the simulated attack, we captured the focal chick with a net and measured the flipper length (mean±se=9.±. cm; range=7.6.5 cm) with a rule to the nearest mm, bill length (9.±. mm; range=..8 mm) with a caliper to the nearest. mm, and body mass (.±. kg; range=.. kg) with a Pesola spring balance to the nearest 5 g. After removing variation with age, male chicks were significantly heavier than females (age: F,65 =., P=.7; sex: F,65 =.69, P<.), had significantly longer flippers (age: F,65 =.85, P=.5; sex: F,65 =., P=.5), and significantly longer bills than females (age: F,65 =9.7, P<.; sex: F,65 =5.6, P=.). Body condition was estimated as the residuals from an ANCOVA with log-body mass as the dependent variable, sex as a factor, and log-flipper length and log-bill length as covariates (R =.7, F,6 =.9, P<.) (Tella et al. ). To assess the health condition of chicks, we used a delayed-type hypersensitivity test; we measured the cellmediated immune (CMI) response by challenging the immune system through the subcutaneous injection of a novel mitogen (phytohemagglutinin, PHA-P, Sigma) (Smits et al. 999). PHA is a mitogen that produces a local swelling due to a prominent perivascular accumulation of T-lymphocytes followed by macrophage infiltration (Smits et al. 999). This method is considered to be a reliable method of measuring CMI (reviewed by Norris and Evans ). We injected.5 mg PHA dissolved in. ml of physiological saline solution (Bausch and Lomb) in the right external foot web (Moreno et al. 998; Tella et al. ). The thickness of the foot web was measured three times at the injection site with a digital caliper (. mm) just before injection. The repeatability of the three measurements calculated as the intraclass correlation coefficient (Lessells and Boag 987) was high (r=.97, F 6,9 =86.6, P<.). The chicks were liberated at the exact capture point and recaptured after h (± h) after injection to measure again three times the thickness of the foot web with the same procedure (repeatability: r=.99, F 6,9 =57., P<.). We used the mean difference in thickness between the day of injection and the following day as an estimate of CMI (Moreno et al. 998; Smits et al. 999). Data analysis To investigate the relationships between escape characteristics (approach distance and distance fled) and chick characteristics (sex, age, body condition, and CMI) and the initial density of other chicks and adults around, we used general linear modeling (GLM) (Crawley 99) with normal error distribution. Because some of the explanatory variables may covary, we used a forward stepwise procedure. We tested the variation explained by each variable in separate models, the variable explaining more variance was added to the model, and the significance of the remaining variables was tested again until no additional variable reached significance. The dependent variables were log-transformed to ensure normality. Residuals from the final model were normally distributed. Results Measures of CMI ranged between. and. mm (mean± SE=.±.6 mm). A GLM model (R =.6, F,57 =.7,

Behav Ecol Sociobiol (6) 6:778 78 78 P=.6) showed that the CMI of chicks was significantly related to age (F,57 =7., P<.) and body condition (F,57 =., P=.) but was independent on sex (F,57 =., P=.7). Thus, both older male and female chicks and with a higher body condition had a greater CMI. Approach distances of penguin chicks varied between and m (mean±se=.9±. m). Variability in approach distances was explained by the chick CMI, age, and initial density of adults m around (GLM model, R =., F,6 =6., P<.) (Table ). Chicks had shorter approach distances when they had a greater CMI (Fig. a), when they were younger (Fig. b), and when the density of adults around was higher (Fig. c). However, neither sex, body condition, nor initial density of other chicks m around was retained in the final model. Distances fled by penguin chicks varied between and 6 m (mean±se=.5±. m). Variability in distances fled was related to the chick CMI and the initial density of adults m around (GLM model, R =., F,6 =., P=.9) (Table ). Chicks fled to longer distances when they had a lower CMI (Fig. a) and when the density of adults was lower (Fig. b). Neither age, sex, body condition, nor initial density of other chicks was retained in the final model. Discussions Our results indicated that penguin chicks differed in the characteristics of their escape response when approached by a simulated predator. Variability in escape behavior was partly explained by the social environment (i.e., the presence of some adults around). Adult penguins are effective in deterring skua attacks, as they normally respond to the skua presence with aggressive defensive behavior until the skua is expelled (Davis 98; Young 99; personal observation). This does not imply that adults close to chicks aggregations are defending all chicks, but chicks might assess that the adult presence lowers their vulnerability to predation, and, then, chicks may decrease the intensity of their escape behavior accordingly. This would reduce the costs of an unnecessarily high antipredatory response. In contrast, the density of other chicks around did not modify the escape response of chicks. This suggests that the possible dilution effect (Cresswell 99a; Roberts 996) of a larger number of chicks per se in aggregations might not be considered by penguin chicks to increase safety. Otherwise, chicks within a larger or denser group should have tolerated closer predator approaches, as individual risk would be reduced (Fernández-Juricic et al. ). It remains possible that the dilution effect of a large number of available prey would be important to prey species where predators attack by surprise. If a predator simply attacks any individual prey, or those prey that exceed a minimum vulnerability level, a prey animal s predation risk will be determined by its absolute vulnerability, which might depend on group size. However, in prey species where predators can examine available prey before choosing to attack the most vulnerable individual, dilution of risk for an individual prey might depend on the relative availability of other more vulnerable prey in the group (Cresswell and Quinn ) and not just on the total number of individuals in the group. This might be the case for penguin chicks and skuas, where skuas often approach to the groups of chicks by walking and examine chicks before launching an actual attack (personal observation). Chick age was another determinant of their approach distance. If the ability to defend from an attack would depend on body size (Young 99), older (=larger) chicks would be less vulnerable to predation and might delay their escape response. However, in contrast to what is expected from age-related differences in body size, younger chicks reacted significantly later to the simulated predator. This Table Results of GLM analyses on how a penguin chick escape behavior (approach and flight distances) was affected by the chick s CMI response and age, and the density of adults around Variable source Adj. SS df F P Coeff. SE Coeff. Approach distance Intercept..6.8.86.7 CMI response.69.69 <...7 Age.88.8..5. Adult density.6 8..6.8.8 Error.66 6 Flight distance Intercept 89.8 6.66 <..5. CMI response.6.6...9 Adult density.6.5.6.98.5 Error 8.77 6

78 Behav Ecol Sociobiol (6) 6:778 78 Approach distance (m),5,5 Immune response (CMI) 6 8 6 8 5 a b c Age (days) Adults m around Fig. Relationships between the approach distance of penguin chicks in response to a simulated predator attack and a the T-cell-mediated immune response (CMI), b the age of chicks, and c the initial density of adults m around the focal chick might suggest that the ability, or the need, to recognize and react to predator approaches develops with age. When chicks are young, at least one of the parents is always in the nest guarding the chicks from approaching predators (Viñuela et al. 995); thus, young chicks might not need to show escape responses, whereas when older chicks are left unguarded, they need to defend by themselves and should escape when a predator approaches. Our results further suggested that penguin chicks with a worse health state, as indicated by their T-cell immune response, were more frightened by our approach, which affects their escape response characteristics (i.e., they fled sooner and ran farther). Some studies have shown a link Distance fled (m) Distance fled (m) 6 5,5,5 Immune response (CMI) b 6 5 a Adults m around Fig. Relationships between the distance fled by penguin chicks in response to a simulated predator attack and a the T-cell-mediated immune response (CMI), and b the initial density of adults m around the focal chick between risk-taking behavior under the threat of predation and personality traits like fear (e.g., Quinn and Cresswell 5), but the interaction between these behaviors and health was almost unknown. Our results may be explained if, as occurs in other bird species, antipredator behavior performance was affected by their nutritional condition and presence of disease (Lindström et al. ). Similarly, lizards with a poor health state (i.e., lower CMI response) were shier when confronting with a simulated predator (López et al. 5). Other indirect evidences of this relationship between health state and shyness have been found in experiments on fear responses under chronic stress. Stress may increase the levels of corticosteroids, with the concomitant immunosuppressive effects of these hormones (Cohn 997). Thus, an index of fear in rats was positively correlated with plasma levels of corticosterone (Gamallo et al. 986). In hens, tonic immobility, a specific fear-related response, was prolonged after the basal plasma levels of corticosterone were artificially increased (Jones et al. 988). Penguin chicks with a worse health state might have suffered some previous stress situations such as a high exposure to predation risk or social harassment by adults, or sibling competition for food. For example, experimental

Behav Ecol Sociobiol (6) 6:778 78 78 feeding experiments have shown that northern bobwhite chicks in poor nutritional condition show a weaker CMI (Lochmiller et al. 99). Also, our results indicated a positive relationship between CMI and body condition in chinstrap penguin chicks, as it also occurs in other penguin species (Tella et al. ). Hatching asynchrony is common in penguins (Moreno et al. 99), larger penguin chicks of double broods get most of the food provided by the parents (Blanco et al. 996), and CMI differs between larger and smaller fledglings, with smaller siblings tending to have a lower CMI (Tella et al. ). Also, it remains possible that, if the immune response were heritable (e.g., Brinkhof et al. 999), chicks with a lower CMI might be the sons of late breeders, which had poorer health and a lower CMI than early breeders (Moreno et al. 998). Predators should prey upon individuals of inferior phenotypic quality (Temple 986; Møller and Erritzøe ). Thus, several species of passerine birds falling prey to cats had smaller spleens, and, thus, a lower immunocompetence as well than other conspecifics (Møller and Erritzøe, but see Smith and Hunt for a review). Recent studies have suggested that, when predators are able to monitor a large number of potential prey, they might use subtle variations in behavior of different prey to select them according to their vulnerability (Cresswell and Quinn ; Quinn and Cresswell ). Thus, predators would avoid predictably unsuccessful chases. It is likely that if health state affects general antipredatory behavior performance (Lindström et al. ), the ability of poor condition penguin chicks to defend themselves may be lower, forcing them to flee sooner and for longer. Only penguin chicks in good condition may delay fleeing. Therefore, the characteristics of escape behavior of chicks might be considered as an honest signal of condition (Grafen 99; Vega-Redondo and Hasson 99), Skuas and other avian predators might assess that individual penguin chicks that escape relatively later or face the predator will be more difficult to capture. Similarly, distress calls of some larks, which are used to discourage predators (Cresswell 99b), reflect their health condition (CMI response and body condition) and, thus, may signal their ability to escape (Laiolo et al. ). Further experiments are clearly needed to analyze whether skuas are able to assess prey vulnerability when attacking penguin chicks. Acknowledgements We thank JL Tella and JL Quinn for the helpful comments on the manuscript. This study was based on the Spanish Army Base Gabriel de Castilla ; transport to and from Deception Island was provided by the Spanish Navy ship Las Palmas. We thank the cooperation offered by all participants in the Spanish Antarctic campaign / and especially to JC, the man with long sideburns, for his persistent companionship. Financial support was provided by the Spanish Antarctic Program, MCYT project BOS -7. The manipulations and experimental procedures were approved by the Scientific Committee for Antarctic Research (SCAR). References Besnard A, Gimenez O, Lebreton JD () A model for the evolution of crèching behaviour in gulls. Evol Ecol 6: 89 5 Blanco DE, Yorio P, Boersma PD (996) Feeding behaviour, size asymmetry, and food distribution in Magellanic penguin (Spheniscus magellanicus) chicks. Auk :96 98 Blumstein DT () Flight initiation distance in birds is dependent on intruder starting distance. J Wildl Manage 67:85 857 Brinkhof MWG, Heeb P, Kolliker M, Richner H (999) Immunocompetence of nestling great tits in relation to rearing environment and parentage. Proc R Soc Lond B 66:5 Burger J, Gochfeld M (98) Discrimination of the threat of direct versus tangential approach to the nest by incubating herring and great black-backed gulls. J Comp Physiol Psychol 95: 676 68 Caraco T, Martindale S, Pulliam HR (98) Avian time budgets and distance to cover. Auk 97:87 875 Cohn LA (997) Glucocorticoids as immunosuppressive agents. Sem Vet Med Surg Small Anim :5 56 Crawley MJ (99) GLIM for ecologists. Blackwell, Oxford, UK Cresswell W (99a) Flocking is an effective anti-predation strategy in redshanks, Tringa totanus. Anim Behav 7: Cresswell W (99b) Song as a pursuit-deterrent signal, and its occurrence relative to other anti-predation behaviours of skylark (Alauda arvensis) on attack by merlins (Falco columbarius). Behav Ecol Sociobiol :7 Cresswell W, Quinn JL () Faced with a choice, sparrowhawks more often attack the more vulnerable prey group. Oikos :7 76 Creswell W, Hilton GM, Ruxton GD () Evidence for a rule governing the avoidance of superfluous escape flights. Proc R Soc Lond B 67:7 77 Davis LS (98) Crèching behaviour of Adèlie penguin chicks (Pygoscelis adeliae). NZ J Zool 9:79 86 Ellegren H (996) First gene on the avian W chromosome (CHD) provides a tag for universal sexing of non-ratite birds. Proc R Soc Lond B 6:65 6 Emslie SD, Karnovsky N, Trivelpiece W (995) Avian predation at penguin colonies on King George Island, Antarctica. Wilson Bull 7:7 7 Fernández-Juricic E, Jiménez MD, Lucas E () Factors affecting intra- and inter-specific variations in the difference between alert distances and flight distances for birds in forested habitats. Can J Zool 8: Fitzgibbon CD (99) Why do hunting cheetahs prefer male gazelles? Anim Behav :87 85 Frid A, Dill LM () Human-caused disturbance stimuli as a form of predation risk. Conserv Ecol 6: Funston PJ, Mills MGL, Biggs HC () Factors affecting the hunting success of male and female lions in the Kruger National Park. J Zool 5:9 Gamallo A, Villanua A, Trancho G, Fraile A (986) Stress adaptation and adrenal activity in isolated and crowded rats. Physiol Behav 6:7 Grafen A (99) Biological signals as handicaps. J Theor Biol :57 56 Hunter S (99) The impact of avian predator scavengers on king penguins Aptenodytes patagonicus chicks at Marion Island. Ibis : 5 Jackson S, Wilson RP () The potential costs of flipper-bands to penguins. Funct Ecol 6: 8 Jones RB, Beuving G, Blokhuis HJ (988) Tonic immobility and heterophil/lymphocyte responses of the domestic fowl to corticosterone infusion. Physiol Behav :9 5

78 Behav Ecol Sociobiol (6) 6:778 78 Laiolo P, Tella JL, Carrete M, Serrano D, López G () Distress calls may honestly signal bird quality to predators. Proc R Soc Lond B 7:S5 S55 (Suppl) Lawless RM, Buttemer WA, Astheimer LB, Kerry KR () The influence of thermoregulatory demand on contact crèching behaviour in Adèlie Penguin chicks. J Therm Biol 6:555 56 Le Bohec C, Gauthier-Clerc M, Le Maho Y (5) The adaptive significance of crèches in the king penguin. Anim Behav 7: 57 58 Lessells CM, Boag PT (987) Unrepeteable repeatabilities: a common mistake. Auk :6 Lima SL (998) Stress and decision making under the risk of predation: recent developments from behavioral, reproductive and ecological perspectives. Adv Stud Behav 7: 5 9 Lima SL, Dill LM (99) Behavioral decisions made under the risk of predation: a review and prospectus. Can J Zool 68:69 6 Lindström KM, Van de Veen IT, Lagault BA, Lundström JO () Activity and predator escape performance of common greenfinches Carduelis chloris infected with Sindbis virus. Ardea 9: Lochmiller RL, Vestey MR, Boren JC (99) Relationship between protein nutritional status and immunocompetence in northern bobwhite chicks. Auk :5 5 López P, Hawlena D, Polo V, Amo L, Martín J (5) Sources of individual shy-bold variations in antipredatory behaviour of male Iberian rock-lizards. Anim Behav 69: 9 Martín J, López P () Fleeing to unsafe refuges: effects of conspicuousness and refuge safety on the escape decisions of the lizard Psammodromus algirus. Can J Zool 78:65 7 Martín J, de Neve L, Fargallo JA, Polo V, Soler M () Factors affecting the escape behavior of juvenile chinstrap penguins, Pygoscelis antarctica, in response to human disturbance. Polar Biol 7:775 78 Møller AP, Erritzøe J () Predation against birds with low immunocompetence Oecologia :5 5 Moreno J, Carrascal LM, Sanz JJ, Amat JA, Cuervo JJ (99) Hatching asynchrony, sibling hierarchies and brood reduction in the chinstrap penguin Pygoscelis antarctica. Polar Biol : Moreno J, DeLeon A, Fargallo JA, Moreno E (998) Breeding time, health and immune response in the chinstrap penguin Pygoscelis antarctica. Oecologia 5: 9 Norman FI, McFarlane RA, Ward SJ (99) Carcasses of Adèlie penguins as a food source for south polar skuas: some preliminary observations. Wilson Bull 6:6 Norris K, Evans MR () Ecological immunology: life history trade-offs and immune defense in birds. Behav Ecol :9 6 Penteriani V, Viñuela J, Belliure J, Bustamante J, Ferrer M () Causal and functional correlates of brood amalgamation in the chinstrap penguin Pygoscelis antarctica: parental decision and adult aggressiveness. Polar Biol 6:58 5 Pole A, Gordon IJ, Gorman ML () African wild dogs test the survival of the fittest paradigm. Proc R Soc Lond B 7:57 (Suppl) Quinn JL, Cresswell W () Predator hunting behaviour and prey vulnerability. J Anim Ecol 7: 5 Quinn JL, Cresswell W (5) Personality, anti-predation behaviour and behavioural plasticity in the chaffinch Fringilla coelebs. Behaviour :77 Roberts G (996) Why individual vigilance declines as group size increases? Anim Behav 5:77 86 Smith ME, Belk MC () Risk assessment in western mosquitofish (Gambusia affinis): do multiple cues have additive effects? Behav Ecol Sociobiol 5: 7 Smith KG, Hunt JL () On the use of spleen mass as a measure of avian immune system strength. Oecologia 8:8 Smits JE, Bortolotti GR, Tella JL (999) Simplifying the phytohaemagglutinin skin-testing technique in studies of avian immunocompetence. Funct Ecol :567 57 Stankowich T, Blumstein DT (5) Fear in animals: a metaanalysis and review of risk assessment. Proc R Soc Lond B 7: 67 6 Tella JL, Forero MG, Bertellotti M, Donázar JA, Blanco G, Ceballos O () Offspring body condition and immunocompetence are negatively affected by high breeding densities in a colonial sea bird: a multiscale approach. Proc R Soc Lond B 68: 7 Temple SA (986) Do predators always capture substandard individuals disproportionately from prey populations? Ecology 68: 669 67 Van Heezik Y, Seddon PJ (99) Effect of human disturbance on beach groups of jackass penguins. S Afr J Wildl Res :89 9 Vega-Redondo F, Hasson O (99) A game-theoretic model of predator prey signaling. J Theor Biol 6:9 9 Viñuela J, Amat JA, Ferrer M (995) Nest defence of nesting chinstrap penguins (Pygoscelis antarctica) against intruders. Ethology 99: Viñuela J, Moreno J, Carrascal LM, Sanz JJ, Amat JA, Ferrer M, Belliure J, Cuervo JJ (996) The effect of hatching date on parental care, chick growth, and chick mortality in the chinstrap penguin Pygoscelis antarctica. J Zool :5 58 Williams TD (995) The penguins. Oxford University Press, Oxford, UK Ydenberg RC, Dill LM (986) The economics of fleeing from predators. Adv Stud Behav 6:9 9 Yorio P, Boersma PD (99) The effects of human disturbance on Magellanic penguin Spheniscus magellanicus behaviour and breeding success. Bird Conserv Int :6 7 Young E (99) Skua and penguin. Cambridge University Press, Cambridge, UK