Do penguins dare to walk at night? Visual cues influence king penguin colony arrivals and departures

Behav Ecol Sociobiol (2010) 64:1145 1156 DOI 10.1007/s00265-010-0930-3 ORIGINAL PAPER Do penguins dare to walk at night? Visual cues influence king penguin colony arrivals and departures Anna P. Nesterova & Céline Le Bohec & David Beaune & Emeline Pettex & Yvon Le Maho & Francesco Bonadonna Received: 15 June 2009 /Revised: 8 February 2010 /Accepted: 10 February 2010 /Published online: 9 March 2010 # Springer-Verlag 2010 Abstract Orientation based on visual cues can be extremely difficult in crowded bird colonies due to the presence of many individuals. We studied king penguins (Aptenodytes patagonicus) that live in dense colonies and are constantly faced with such problems. Our aims were to describe adult penguin homing paths on land and to test whether visual cues are important for their orientation in the colony. We also tested the hypothesis that older penguins should be better able to cope with limited visual cues due to their greater experience. We collected and examined GPS paths of homing penguins. In addition, we analyzed 8 months of penguin arrivals to and departures from the colony using data from an automatic identification system. We found that birds rearing chicks did not minimize their traveling time on land and did not proceed to their young (located in crèches) along straight paths. Moreover, breeding birds' arrivals and departures were affected by the time of day and luminosity levels. Our data suggest that king penguins prefer to move in and out of the colony when visual cues Communicated by C. Brown A. P. Nesterova (*) : E. Pettex : F. Bonadonna Behavioural Ecology Group, CEFE CNRS, 1919 route de Mende, 34293 Montpellier, Cedex 5, France e-mail: apnesterova@gmail.com C. Le Bohec Centre for Ecological and Evolutionary Synthesis, Department of Biology, University of Oslo, PO Box 1066, Blindern, 0316 Oslo, Norway C. Le Bohec : D. Beaune : Y. Le Maho Département d Ecologie, Physiologie, et Ethologie, IPHC CNRS, 23 rue Becquerel, 67087 Strasbourg, Cedex 2, France are available. Still, they are capable of navigating even in complete darkness, and this ability seems to develop over the years, with older breeding birds more likely to move through the colony at nighttime luminosity levels. This study is the first step in unveiling the mysteries of king penguin orientation on land. Keywords Short-range navigation. King penguins. Seabirds. Visual landmarks. Nocturnal movements. Aptenodytes patagonicus Introduction Different environments provide a variety of challenges for animal navigation. Ants (Cataglyphis) foraging in the Saharan desert need to locate their nest in areas that may be limited in visual landmarks features associated with their goal (Wehner 2003). On the other hand, many humancreated environments are saturated with different types of landmarks, leading to the question of which types of landmarks humans (Homo sapiens) should use for navigation (Caduff and Timpf 2008). For other species, seasonal changes can completely modify their habitat. For example, occasional spring snowstorms can result in extensive snow cover on alpine meadows (>1 m); nevertheless, Columbian ground squirrels (Urocitellus columbianus) are able to locate their burrows (Vlasak 2006). Navigational challenges stem not only from the environment but also from the individuals present there. For instance, navigation in crowded environments can be extremely challenging. This is especially obvious in the case of animals that form gregarious colonies. In such colonies, the presence of hundreds or even thousands of conspecifics can obstruct locally available cues. This

1146 Behav Ecol Sociobiol (2010) 64:1145 1156 problem is even more pronounced in species that do not fly. Aggression can also be a problem in gregarious colonies, making a search for a particular location (or goal ) very difficult. Thus, it is to the advantage of an individual to minimize traveling within such crowded colonies. Several studies have investigated how individual animals identify their partners, offspring, or nest locations within crowded colonies. In some species, individuals rely on the vocal cues to identify partners or offspring (king penguins (Aptenodytes patagonicus): Lengagne et al. 1999; emperor penguins (Aptenodytes forsteri): Aubin et al. 2000; macaroni penguins (Eudyptes chrysolophus): Searby et al. 2004; rockhopper penguins (Eudyptes chrysocome): Searby and Jouventin 2005; Magellanic penguins (Spheniscus magellanicus): Clark et al. 2006; kittiwakes(rissa tridactyla): Aubin et al. 2007; Galapagos fur seals (Arctocephalus galapagoensis) and Galapagos sea lions (Zalophus californianus wollebaeki): Trillmich 1981; harbor seals(phoca vitulina): Renouf 1985; northern fur seals (Callorhinus ursinus): Insley 2000; South American fur seals (Arctocephalus australis): Phillips and Stirling 2000). In other species, individuals use olfactory cues to find their nests (blue petrels (Halobaena caerulea): Bonadonna et al. 2001; Mardon and Bonadonna 2009; commondivingpetrels(pelecanoides urinatrix) and South-Georgian diving petrels (Petecanoides georgicus): Bonadonna et al. 2003). Yet other species use a combination of acoustic and olfactory cues for recognition purposes (Antarctic fur seals (Arctocephalus gazella): Dobson and Jouventin 2003a). In the case of Mexican free-tailed bats (Tadarida brasiliensis mexicana), such dual recognition systems allow mothers to find their offspring in the large and dense groups with more than 5,000 young/m 2 (Balcombe 1990; Loughry and McCracken 1991; Balcombe and McCracken 1992). However, acoustic and olfactory signals can only be used over very short distances given the limitations imposed by colonial lifestyles. In king penguins, for example, a vocal recognition range between individuals within a colony is on average 8.8 m (Lengagne et al. 1999; Dobson and Jouventin 2003b). Therefore, animals must use some other cues to navigate to the general area of their goal, and less is known regarding this scale of navigation in the colony. The importance of visual cues for short-range navigation has been documented in a variety of species. For example, rats (Rattus norvegicus) are known to use visual landmarks around the goal (e.g., Suzuki et al. 1980; Benhamou and Poucet 1998), as are homing pigeons (Columbia livia) (reviewed in Cheng et al. 2006), Clark's nutcrackers (Nucifraga columbiana) (e.g., Gould-Beierle and Kamil 1996; e.g., Gibson and Kamil 2001), ants (Cataglyphis fortis) (e.g., Wehner et al. 1996), wasps (Cerceris rybyensis) (e.g., Zeil et al. 1996), honeybees (Apis mellifera) (reviewed in Cheng 2000), cephalopod spp. (reviewed in Alves et al. 2008), and many other species (reviewed in Healy 1998; Shettleworth 1998). Potentially, visual cues can be used at larger scales than acoustic or olfactory cues. However, environmental features suitable as visual cues can be occluded by conspecifics in colonial species. Global features of the landscape, on the other hand, might not provide enough resolution for fine scale positioning (Cheng and Spetch 1998). As a result, the use of visual cues for navigation inside colonies may also be problematic. To better understand the role of visual cues for navigation in a crowded environment, we studied king penguins (Aptenodytes patagonicus). Several characteristics of this species render them an ideal model for the study of visual orientation. First, they form very large and densely populated colonies, some of which can number over 300,000 breeding pairs (Weimerskirch et al. 1992; Aubin and Jouventin 1998). Second, their unusual breeding style makes the use of established routes in the colony very difficult. King penguins do not build nests but incubate a single egg on their feet. Partners alternate their parental duties. Breeding pairs defend a small part of the colony, known as an attachment place, which is approximately 1 m 2 (Bried and Jouventin 2001). Locations of attachment places can shift slightly through the breeding season, with an average displacement of 4.4 m (Lengagne et al. 1999). Because of this, returning individuals may find a new arrangement of neighboring couples as they return to the colony following foraging trips. Third, these foraging trips can last for a couple of days to several months (Stonehouse 1960), and birds have to consider any changes that might take place during their absence. Also, heavy rains, storms, and waves can dramatically change the shape of the colony due to flooding (personal observations; Viera et al. 2006). Finally, since penguins do not fly, they cannot survey the colony from above to aid their orientation. King penguins are long-lived species with an average life span of ca 20 years (see Gauthier-Clerc et al. 2004), and they attempt to breed each year of their adult life (mean annual proportion of non-breeding adults at Possession Island = 13%; see Le Bohec et al. 2007). Breeding individuals repeatedly face the challenge of finding a particular place in the colony. In the beginning of the breeding season (incubation and brooding stages), birds search for their partners at the attachment place. Later, they start looking for their chicks at the rendezvous zone : chicks that are several weeks old and capable of thermoregulation are left unattended by the parents (Stonehouse 1960; Barrat 1976). Chicks form groups known as crèches, and the rendezvous zone is the location within the crèche where a chick was last fed (Stonehouse 1960; Dobson and Jouventin 2003b). Once within the colony, breeders are likely to proceed directly to their previous attachment place or rendezvous zone, which makes them particularly interesting for navigational studies.

Behav Ecol Sociobiol (2010) 64:1145 1156 1147 The destinations of non-breeders, on the other hand, are less obvious. They might enter the colony to train in courtship while they are still juveniles, or to acquire social information, such as the quality of a breeding site (Nocera et al. 2006). The objectives of this study were to describe the terrestrial homing paths of adult king penguins and to test whether visual cues are important for penguin orientation within a colony. To address this, we examined paths of king penguins moving through the colony. To investigate the importance of visual cues, we took advantage of the natural daily light and dark cycle. At night, the visibility of visual cues decreases. If penguins rely on visual cues, they should be less likely to enter the colony in the dark. However, older, more experienced, birds should be more familiar with the landscape within and around a colony, so we proposed that they would be better able to cope with limited visual cues. Accordingly, we compared the arrivals and departures of young and old birds through the colony during the hours of darkness. Methods Study sites and animals Kerguelen Island The paths of breeding king penguins were recorded at a colony at Cape Ratmanoff, Courbet Peninsula, Kerguelen Island (49 42 S, 70 33 E) in November 2007. This large colony stretches for over a kilometer on a flat sandy beach along a north to south axis. The landscape around the beach allows penguins to come ashore >3 km south or north from each end of the colony. We focused our efforts on the birds located at the southern edge to minimize human disturbance to one part of the colony. Individuals returning from foraging trips were spotted as they exited the water. A human observer equipped with a GPS (TechnoSmArt) followed focal individuals at a distance of approximately 15 m, which was sufficient to prevent any sign of disturbance to the bird being followed. Positional data were collected every second. Such a methodology allowed data collection without animal capture and has been successfully used before on Magellanic penguins (Boersma, D., personal communications). Animals were followed during daylight hours. Observations were not conducted during sunrise, sunset, or nighttime hours because light levels were not sufficient for a human observer to follow an unmarked penguin among thousands of conspecifics. At the time of data collection, we saw no animals in the colony incubating eggs due to the late start of 2007 breeding season, and it was possible to follow individuals inside the colony with minimal disturbance. Once a parent bird reached a crèche and gave a call, it was observed for up to 1 h. Usually parents found their chicks within the first 5 min. After parent and chick reunited, the observer walked the last 10 m of the penguin trajectory. The observer waited for 15 min before discontinuing the observation. Except for two individuals, all penguins found their chicks when they entered crèches. Two individuals that did not find their chicks were observed until the end of the hour. These adults called repeatedly within the same crèche but in different locations. Repeated calling indicated their strong motivation to find a chick. The reasons why these adults did not find their chicks were unknown, but presumably their chicks were dead. The paths of 23 penguins were analyzed. When analyzing GPS data, we considered each penguin path in its entirety as well as breaking the path into two parts outside and inside of the colony core. The colony core was defined as an area where breeding birds were defending their attachment places. The other areas of the beach were considered to be outside the colony core. Possession Island Information on the birds' arrivals to and departures from this colony was collected at the La Grande Manchotière colony located on Possession Island (46 25 S, 51 45 E), Crozet Archipelago. About 22,000 pairs of king penguins bred in 2007 (Beaune and Le Bohec, unpublished data). King penguin colony arrivals and departures (N total arrivals/ departures=74,274 for 1,919 individuals) were analyzed over an 8-month period (August 1st March 31st) using data from the Antavia automatic identification system (Fig. 1). This system was installed at the penguin colony in January 1998 and has been used since then for the long-term monitoring of the population. Birds in the colony are implanted subcutaneously with individual passive integrated transponder tags (PIT tags), and their movements are registered automatically by pairs of antennas buried on the three natural pathways available to the birds, allowing for the continuous monitoring of birds entering or exiting the colony (for more details, see Gendner et al. 2005; Le Bohec et al. 2007). Since 1991, 6,005 penguins have been monitored with PIT tags. Among these birds, 5,000 were implanted with PIT tags as chicks starting in 1998 when the Antavia system was installed. The antennas are located approximately 40 and 70 m inside of the colony core and therefore allow for monitoring of moving birds that are already inside the colony. Extreme aggression from birds with eggs or young chicks discourages moving individuals from wandering in the colony (Cote and Dewasmes 1999; Cote 2000), and they are likely to proceed directly to their attachment places. Previous video recording studies by Challet et al. (1994) demonstrated that breeding birds

1148 Behav Ecol Sociobiol (2010) 64:1145 1156 Fig. 1 La Grande Manchotière king penguin colony and the Antavia detection system (Possession Island). The diagram shows the study area at Possession Island. Arrivals to and departures from the colony are detected by pairs of antennas buried underground at three natural passages (stars) to the study zone of the breeding colony (black dashed line). The antennas are 40 and 70 m away from the edge of the colony that enter the colony relieve their partners both during the day and night. Therefore, passing through the antenna areas at any time of the day would provide a good indication of a bird moving through the colony. To investigate the effect of light levels on penguin orientation, we focused only on the colony arrivals and departures of breeders (N breeders =1,004). Breeding activities of tagged birds were inferred from their arrivals to and departures from the colony during the breeding season. Breeders have a characteristic schedule associated with their presence in the colony (caring for the egg/young) and their absence (foraging trips) that is especially apparent during incubating and brooding stages. Non-breeders, on the other hand, do not show such sequences of arrival and departures. The breeding status of each individual was also confirmed by analyzing video recordings for body and plumage condition (see details in Descamps et al. 2002; Gendner et al. 2005). Breeders have a strong motivation to find a specific place within a colony. They must walk to their attachment places to relieve a partner from caring for an egg/newborn chick or to a crèche to feed an older chick. To investigate the effect of experience on navigational abilities, we looked at two different age groups. King penguins start breeding when they are 3 5 years old (Le Bohec 2007). Therefore, we split the population into two groups: young breeders ( 8 years old) that had relatively few years of breeding experience and old breeders (>8 years old) that had several years of breeding experience. Luminosity data were recorded by a Météo France station located 1 km away from the Antavia system. Global radiation per hour (RgH) was measured in Joule per square meter. Statistical analysis From GPS trajectory data (Kerguelen Island data), we extracted the duration of the trip (t), the total length of the path (Dt), and the B-line (Ds) between the starting point (the place where a penguin exited water) and the end point of the path (crèche). As a measure of the straightness (optimality) of a penguin's path, we used the linearity index (LI) defined as LI ¼ Ds=Dt 1 (Batschelet 1981). Thus, animals following a straight line to their goal would have an LI value of 1. The LI of the path outside the colony core was compared to the LI inside the core using a Wilcoxon signed-rank test. To analyze frequencies of colony arrivals and departures (Crozet Island data), we used Pearson's chi-squared test. Arrivals and departures of breeding penguins were also analyzed at several luminosity levels: low luminosity level (0 5.6 RgH), night (0 0.8 RgH) and day (0.9 432.6 RgH) luminosity levels, and all luminosity level (0 432.6 RgH). The low luminosity levels span the range of values that were obtained at night and during twilight. The night luminosity levels correspond to the ambient light present during night hours and without a full moon, starting 1 h after sunset and continuing until 1 h before sunrise. Arrivals and departures at the low luminosity levels are of particular interest. These include luminosities at which no landmarks are visible at all (darkness) and at which most or all visual landmarks are visible (twilight).

Behav Ecol Sociobiol (2010) 64:1145 1156 1149 Day luminosity levels correspond to the hours of peak visibility. To understand how time of day and luminosity level influence breeders' arrivals and departures, we employed the generalized additive model (GAM) approach (Schimek 2000). In the analysis, we used the number of detections (penguins passing the antennas) during the months when breeders were present (5-month period) as a response variable, with time and luminosity as predictors. As appropriate for the count data, Poisson distributions and log link functions were used. Smoothing functions were applied to time and luminosity variables. To estimate significance of each term, the full model was compared using a χ 2 test to a model where one predictor was omitted. Including penguin identity as a random effect to account for repeated detections of some individuals had no effect; we therefore used the total number of detections (arrivals and departures) for each hour at specific luminosity levels to run the time and luminosity dependent models. Arrival and departure times for breeders during sunrise hours (0400 0600 hours) were compared with a Welch two sample t test (Welch 1947). We specifically looked at sunrise hours because they were characterized by the peak in colony arrivals and departures. To compare bird arrivals/ departures between the two age groups ( 8 and >8 years old), we used a Mann Whitney U test (Siegel and Castellan 1988). Non-parametric tests were chosen when data did not satisfy the normality assumption. All tests were two-tailed with the significance level set at α=0.05. Statistics were computed using R 2.6.1 (http://www.r-project.org) with the lme4 package (Bates and Sarkar 2006) and SPSS statistical package, version 7 (the Predictive Analytics Company). Results Homing paths at the Ratmanoff colony, Kerguelen Island At Ratmanoff colony, adult king penguins made extensive trips across the beach before they found their chicks in the crèches (Fig. 2). The length of trips varied greatly between individuals (Table 1). The median duration of the trips was 32 min, and median speed was 0.24 m/s. Penguins did not proceed to their crèche directly (median LI total trip =0.68). The straightness of the paths depended on whether penguins were outside or inside of the colony core. The paths became less direct once animals entered the core (Wilcoxon signed-rank test, n=23, Z=-3.771, p< 0.001). Once inside, birds proceeded in bouts. They moved quickly from one place to the next, paused, and then moved quickly again. While proceeding towards the crèche, the birds did not vocalize. They called only when they reached their crèche. There is no evidence that Scale 100 m Fig. 2 Examples of king penguin paths from the ocean to crèche (Kerguelen Island). Black thick line represents the outline of the colony core where breeding adults defend their attachment places, light gray thick lines represent the outline of the lake and ocean shoreline at high tide, thin gray and thin black lines represent tracks of six adult penguins proceeding from the water to their chick in a crèche penguins arriving at this colony used pathways commonly used by many individuals. Bird arrivals to and departures from La Grande Manchotière colony, Possession Island Penguins (breeders and non-breeders) arrived and left the colony at all times during the 24-h day (Fig. 3). Over 8 months, there were consistently large numbers of arrivals and departures at sunrise; though, departures significantly outnumbered arrivals (Table 2). The relative proportion of arrivals and departures varied across months. In August, more birds arrived than left. The pattern was reversed in September, October, February, and March. No differences were observed between number of arrivals and departures

1150 Behav Ecol Sociobiol (2010) 64:1145 1156 Table 1 King penguins' paths parameters (Kerguelen Island) Total duration (min) Total length (m) Overall speed (m/s) Total LI Outside colony core length (m) Inside colony core length (m) Outside colony core LI Inside colony core LI Median 32 399.0 0.24 0.68 169.0 137.1 0.80 0.61 IR 17.0 464.8 0.07 0.20 182.8 165.9 0.16 0.29 Min 4 92 0.14 0.26 24.7 31.4 0.58 0.19 Max 86 1,153 0.38 0.84 1,120.3 512.1 0.90 0.84 Medians, interquartile ranges (IR), minimum (min), and maximum (max) values of 23 penguin paths are listed for total trip duration, total trip length, overall speed, total linearity index (LI), length of the trip inside and outside of the colony core, and linearity index for the portions of the paths inside and outside of the colony during November, December, and January (Table 2). During the course of a day, fewer animals moved in and out of the colony at night (12%), more animals entered or left the colony during the day hours (51%), and a distinctive peak in arrivals/departures was observed between 0400 and 0800 hours (37%). We found a significant dependence of breeder arrivals and departures on both time and luminosity levels (Figs. 3 and 4; Table 3). When all luminosity levels were considered, GAM total was significant (χ 2 test, n=7,714, p< 0.0001) and explained 21.2% of the variance in number of penguin arrivals/departures. Many birds moved in and out of the colony even at extremely low luminosity levels, including 0 RgH. Early sunrise hours with low luminosity levels were characterized by a dramatic increase in penguin arrivals to and departures from the colony. Similar results were obtained when only low luminosity levels were considered (Fig. 4). GAM low was significant (χ 2 test, n= 2,158, p<0.0001) and explained 36.9% of the variance in number of arrivals/departures based on time and luminosity. The general pattern of arrivals/departures was analogous even when arrivals and departures were considered separately (Table 3). For breeders, departures and arrivals took place at different times (Welch two sample t test, N=2,368, t= -5.186, df=2,229.028, p<0.0001). The arrivals (mean= 5.4 h) preceded the departures (mean=5.7 h, Fig. 5). The age of the birds had an effect on the arrivals/departures at night luminosity levels (Table 4). More of the older breeders arrived or left the colony at low light levels. These differences in arrivals/departures disappeared at the higher luminosity levels (0.9 432.6 RgH) when the proportions of young and old breeders entering and exiting the colony were similar. Discussion Our observations suggest that king penguins can cover distances of over 1 km on land to reach their chicks in Fig. 3 King penguin colony arrivals and departures (Possession Island). Penguin (breeders and non-breeders) arrivals to and departures from the colony during 24-h period for the 8 months of observations. The time periods of low light intensity are shaded. Different months are represented by the various intensity of gray. N arrivals/departures =74,274, for 1,919 individuals

Behav Ecol Sociobiol (2010) 64:1145 1156 1151 Table 2 Monthly comparison of king penguin colony arrivals and departures (Possession Island) Months N Arrivals (%) Pearson's χ 2 p value August 889 76.72 253.7964 <0.00001* September 1,006 45.43 8.4135 0.00372* October 2,872 47.42 7.6267 0.00575* November 4,039 50.90 1.3194 0.25070 December 2,633 48.54 2.2518 0.13350 January 2,235 50.02 0.0004 0.98310 February 2,574 44.72 28.7428 <0.00001* March 3,188 42.66 68.7026 <0.00001* Total 19,436 48.69 13.2776 0.00027* Mean proportions of arrivals to the colony between 0400 and 0600 hours (sunrise hours) for each month of the study period (August, 2007 March, 2008) is compared to the mean proportion of departures using a Pearson's χ 2 test. The statistically significant differences between numbers of arrivals and departures are indicated with an asterisk. N is the number of detections of tagged birds (breeders and non-breeders) crossing the antennas crèches. The birds did not always proceed to their attachment places in the colony along a route that approximated the shortest path available. The numbers of colony arrivals and departures were affected by the time of day and luminosity levels. Individuals were able to enter the colony even in total darkness. Most arrivals and departures took place at sunrise. At nighttime luminosities, older breeding birds were more likely to enter or leave the colony. The homing paths of king penguins were diverse in their shapes and durations of travel. We saw no indication that the birds tried to exit the sea at the place nearest to their crèches. King penguins are very fast swimmers, averaging 2.1 m/s in water (Kooyman et al. 1992). However, on land, we found their speed to be much slower (0.24 m/s). The flat relief of Ratmanoff site allows animals to come ashore at any place on the beach. It is then unclear why some individuals choose to walk along the beach, given their superior swimming abilities. One explanation is that the birds cannot establish the location of their crèche from the water. Also, it could be because adult penguins are more vulnerable to predation at sea than on land (Stonehouse 1960; Guinet 1992; Yoda et al. 1999). Individuals might exit the sea at what they consider the closest and safest position to their attachment places, given the aquatic travel constraints such as currents, tide, other penguins, or predators encountered. Similar measurements of land traveling paths were conducted on Magellanic penguins (Wilson et al. 1999). They walked shorter distances to their nests (145 m) at faster speeds (0.73 m/s) when compared to king penguins. Magellanic penguins, like king penguins, also did not exit water at the point closest to their nest. Although on a larger scale penguins seem to go directly to their crèches, at a smaller scale it is clear that they do not always proceed along straight paths. Once inside the colony, their paths become even less direct. When an individual crosses the colony, it experiences intense aggression from conspecifics, leading these birds to purposefully pass by sleeping individuals to avoid aggressive encounters (Cote and Dewasmes 1999). Therefore, one would expect penguins to minimize the amount of walking inside the colony. However, the relatively big detours that some individuals made in the colony (for example, light grey track in Fig. 2) are unlikely to be explained by a route that passes close to sleeping birds. Together, the long walking trips and wandering paths inside the colony suggest that finding a particular place in the colony is not a trivial task. The relative proportion of penguin (breeders and nonbreeders) arrivals and departures at the colony at Possession Island during sunrise hours (0400 0600 hours) depended on the month. Previously, no difference was found between colony arrivals and departures (Challet et al. 1994). This is most likely due to the different duration of the studies and the time periods analyzed: Challet et al. (1994) considered only 3 months of observations (January March) while our analysis spans 8 months. Also, we analyzed only sunrise hours when the majority of arrivals and departures takes place (Fig. 3). We do not necessarily expect the observed pattern to hold over different years. Meteorological conditions can vary, which can lead to an early or late onset of Fig. 4 King penguin breeders' arrivals to and departures from the colony (Possession Island). Breeders' arrivals to and departures from the colony at the low luminosity levels (0 5.6 RgH) between 1700 hours (when light levels start to decrease) and 1600 hours. The number of penguins arriving to and departing from the colony at the given time and luminosity is indicated by the grayscale. There is a distinctive peak in activity at the low luminosity levels. N arrivals/ departures=2,158 for 1,004 breeders

1152 Behav Ecol Sociobiol (2010) 64:1145 1156 Table 3 Dependence of king penguin colony arrivals and departures on time and luminosity levels (Possession Island) Significance of smoothed predictor variables (time and luminosity) at all and low luminosity levels. The statistically significant results are indicated with an asterisk. Numbers of detections (breeders) at each luminosity levels are indicated in the table Variable Estimated df χ 2 p value Total arrivals/departures: GAM at all luminosity levels (0 432.6 RgH), N=7,714 Luminosity 7.994 256.20 <0.00001* Time 8.857 1,529.40 <0.00001* Total arrivals/departures: GAM at low luminosity levels (0 5.6 RgH), N=2,158 Luminosity 6.995 40.11 <0.00001* Time 8.437 1,151.12 <0.00001* Arrivals: GAM at low luminosity levels (0 5.6 RgH), N=997 Luminosity 5.150 25.06 0.00290* Time 7.741 487.44 <0.00001* Departures: GAM at low luminosity levels (0 5.6 RgH), N=1,140 Luminosity 5.254 26.89 0.00146* Time 8.250 680.74 <0.00001* breeding activities at any particular year. This, in turn, is associated with a different proportion of arrivals and departures each month. In our study, detection of more arrivals than departures in August was probably due to the massive return of birds after their long winter foraging trips to the Marginal Ice Zone (Bost et al. 2004). A higher number of departures in September and October can be accounted for by the recently molted birds that depart for 3-week long foraging trip to recover body reserves before they start breeding (Barrat 1976; Cherel et al. 1994; Gauthier-Clerc et al. 2002). During November, December, and January, the number of birds entering and exiting the colony was similar. During this time, the feeding trips are usually relatively short, and partners exchange their parental duties frequently (Descamps et al. 2002). Finally, more departures than arrivals in February and March are likely the result of failed breeders leaving the colony for the sea. To test the hypothesis that visual cues are important for penguin navigation on land, we looked at the number of colony arrivals and departures throughout the 24-h day. We focused on breeding individuals because they must find their attachment places or rendezvous zones within the colony to ensure their reproductive success. Moreover, breeders may need to locate their destinations quickly so their partners can leave on foraging trips to replenish their body reserves. If the body mass of an incubating/brooding bird drops below a certain threshold, it will abandon the egg/chick (Gauthier-Clerc et al. 2001). In our study, relatively few breeders entered or exited the colony during nighttime hours. More animals entered and exited the colony during daylight hours, but the majority of the arrivals and departures took place around sunrise (Fig. 5). Challet et al. (1994) also observed a similar pattern of penguin arrivals and departures. The decrease in the arrivals and departures at night is unlikely to be explained by a decrease in activity. The amount of time penguins devote to sleeping and resting is not different between day and night (Challet et al. 1994). At night it might be harder to find an attachment place/rendezvous zone/colony exit due to the limited visibility, and individuals prefer to wait for sunrise before moving. This would explain fewer arrivals and departures at night, the peak in the arrivals and departures during sunrise, and the relatively high but stable level of arrivals and departures during the day (Fig. 5). The activity peak during sunrise could also be partially due to the switching of partner duties. A massive arrival of birds to the colony triggers a massive departure of their partners since egg and chick exchanges are relatively fast due to the extreme aggression from neighbors (Cote 2000). As expected, in our study the mean arrival time was earlier than mean time for departures. Synchronization of arrivals and departures is most likely due to the grouping behavior Fig. 5 Breeder arrivals and departures (Possession Island). Total number of colony arrivals and departures over a 24-h period for the 5 months of observations. Arrows indicate the peaks of arrivals and departures during the 24-h period. N arrivals =3,602 and N departures = 3,993 for 1,004 breeders

Behav Ecol Sociobiol (2010) 64:1145 1156 1153 Table 4 Differences in king penguin colony arrivals and departures based on their age (Possession Island) Old ( 8years old), N=561 Young (<8years old), N=443 Mann Whitney test, p value (U) Day luminosity levels (0.9 432.6 RgH) Arrivals 0.3402 0.3560 0.5051 (122,357) Departures 0.3609 0.3578 0.5364 (126,806) Arrivals + departures 0.3733 0.3907 0.4667 (122,030) Night luminosity levels (0 0.8 RgH) Arrivals 0.2520 0.1955 <0.0001* (35,652) Departures 0.2356 0.2169 0.0041* (34,131) Arrivals + departures 0.2483 0.2143 0.0002* (35,214) Mean proportion of penguin arrivals, departures, or arrivals + departures for old breeder ( 8 years old) and young breeders (<8 years old) at day (0.9 432.6 RgH) and night (0 0.8 RgH) luminosity levels. Mean proportion of arrivals was calculated as the ratio of arrivals at daytime (or nighttime) luminosity levels over arrivals + departures. Analogous calculations were done for the mean proportion of departures. N total =1,004 breeders. Statistically significant differences in arrivals/departures between different age groups are indicated with an asterisk of king penguins. This is especially apparent for departing individuals. Birds form small groups in the water before they leave shore waters (personal observations), possibly to minimize predation risks. Similar grouping behavior has been documented in several other species of penguins. For example, rockhopper penguins (Eudyptes chrysocome) and Adélie penguins (Pygoscelis adeliae) forage at sea in groups (Tremblay and Cherel 1999; Takahashi et al. 2004). Also, little penguins (Eudyptula minor) form groups as they arrive at and depart the colony (Reilly and Cullen 1981; Daniel et al. 2007). One of the most remarkable findings of this study is that many individuals enter the colony at very low night luminosity levels, including 0 RgH. While king penguin eyes are adapted to low light intensities (Martin 1999), the extent of their vision at night is not known. Experiments on Humboldt penguins (Spheniscus humboldti) established that their eyes are 1.9 times more sensitive than humans (Martin and Young 1984). Similar calculations for king penguins based on estimates of their eye parameters suggest that they are 1.5 times more sensitive than humans (Martin, G. R., personal communication). Consequently, king penguins can see only slightly better than humans in the dark. Thus, the fact that birds enter and exit the colony even in complete darkness strongly suggests that visual cues are not essential for orientation, at least for some individuals. King penguins are not the only species that can depart from their colony at night. Gentoo penguins (Pygoscelis papua) were found to sometimes depart on foraging trips between 0200 and 0600 hours. However, there was no indication that gentoo penguins return to the colony before 0600 hours (however, the foraging trips of only 19 birds were followed) (Williams and Rothery 1990). Unlike king and gentoo penguins, chinstrap penguins (Pygoscelis antarctica) were found to avoid crossing coastlines during nighttime hours (Jansen et al. 1998). The authors hypothesized that it could be due to the risk of predation or lack of prominent visual cues necessary for navigation. Another study that suggested the importance of visuals cues for penguin navigation showed that little penguins (Eudyptula minor) did not return to the colony when low level fog was present (Chiaradia et al. 2007). Young king penguin breeders, similarly to gentoo, chinstrap, and little penguins, depend on the availability of visual cues. At nighttime luminosity levels, they are less likely than older breeders to enter or leave the colony. The difference between young and old breeders also explains the activity peak at sunrise. As visual cues increase with daylight, animals that could not enter or leave at night were able to proceed to their destinations. Orientation based on visual cues, especially distant ones, can be a challenging process and may require several years of experience. For example, in Columbian ground squirrels (Urocitellus columbianus), older individuals use distant visual cues for orientation while young squirrels tend not to (Nesterova 2007). Over the years, older penguins have more opportunities to learn the layout of the colony, and therefore, should be better at visual orientation inside the colony. King penguins are known to return to the same section of the colony year after year (Descamps et al. 2009). Moreover, they exhibit a high site fidelity. A 2-year study found that 69% of the birds returned to the exact same location to breed in the second year, 19% settled a few meters away, and only 12% established an attachment place in a different part of the colony (La Grande Manchotière colony, n=172, Gauthier- Clerc, unpublished data). Experience breeding at the same location for multiple years may therefore improve knowledge of the surrounding landscape and be helpful for orientation. The present findings on adult navigation are in accordance with previous work done on king penguin chicks (Nesterova et al. 2009). Homing experiments with 10-

1154 Behav Ecol Sociobiol (2010) 64:1145 1156 month old chicks demonstrated that visual cues are important for orientation, but not essential. After passive displacement (100 m) from the colony, fewer chicks found their crèches during the night than during the day. It appears that the ability to orient within a colony develops from a very early age, but several years of experience are required to allow precise homing when visual cues are limited or not available. The mechanisms that king penguins use to orient in the darkness have not yet been investigated. Ten-month old chicks attend to information carried by the wind (presumably acoustic or olfactory cues), especially at night, when picking the general direction of travel (Nesterova et al. 2009). It is possible that adults also rely on olfactory, acoustic, or magnetic cues. However, the usefulness of these types of cues for precise positioning inside the crowded colony still needs to be determined. Over the years, king penguins might memorize a general route to their attachment place. In addition, given their site fidelity and their superior auditory abilities, it is possible that with time they learn the acoustic landscape created by their neighbors around their attachment place and use this to aid orientation. One of the important implications of this study is the use of distant (global) visual cues for orientation in the colony. If the presence of many conspecifics obstructs local cues, then penguins must rely on the landscape outside of the colony. Some colonies can stretch over several kilometers. As a result, the nearest visible features of the landscape can be several kilometers away. There are other animals that rely on distant cues for precise positioning. Columbian ground squirrels need to see the outline of the forest edge and mountains to locate their burrows (Vlasak 2006). Meerkats (Suricata suricatta) might rely on distant landscape features as well. Their desert habitat is limited in unique local cues; nevertheless, they are successful at finding more than a thousand escape burrows within extensive territories (Manser and Bell 2004). Studies on wood ants (Formica japonica) and Central Australian desert ants (Melophorus bagoti) also provide evidence for the use of distant cues to locate a nest (Fukushi 2001; Cheng et al. 2009). While multiple laboratory studies also demonstrate the use of distant cues, due to limitations of the experimental room, distant cues are usually only several meters away from a goal (rats (Rattus norvegicus): Suzuki et al. 1980; Leonard and McNaughton 1990; black-capped chickadees (Parus atricapillus): Brodbeck 1994; Clark's nutcrackers (Nucifraga Columbiana): Gould-Beierle and Kamil 1996; Gould-Beierle and Kamil 1999). This study is the first step in unveiling the mysteries of king penguin orientation on land. Adult king penguins are capable of locating a particular square meter space within an area of hundreds of meters of densely populated colony in complete darkness. This ability seems to develop over years, with older birds more likely to navigate at night than young adults. Future experiments should investigate the potential mechanisms that allow such precise positioning with limited local visual cues. Acknowledgements We are extremely grateful to G. R. Martin for the stimulating discussion regarding penguin vision and calculations of the estimate of king penguin visual threshold. We want to thank our field assistants for their help in the field. We are indebted to F.S. Dobson and J.D. Whittington for their valuable comments on the earlier versions of the manuscript. Many thanks are due to the Institute Polaire Français Paul-Emile Victor (IPEV) and Terres Australes et Antarctiques Françaises for the logistical support provided on the field. Also, we are indebted to Météo France for sharing their meteorological data with us. The research was funded by a grant from the French Ministry for Foreign and European Affairs Lavoisier to C. Le Bohec, I.P.E.V. grants to F. Bonadonna (ETHOTAAF 354) and to Y. Le Maho (ECOPHY 137) and National Science Foundation International Research Fellowship (#0700939) to A. P. Nesterova. This study was performed according to IPEV and Centre National de la Recherche Scientifique guidelines for the Ethical Treatment of Animals and complied with current French regulations. References Alves C, Boal JG, Dickel L (2008) Short-distance navigation in cephalopods: a review and synthesis. Cogn Process 9:239 247 Aubin T, Jouventin P (1998) Cocktail-party effect in king penguin colonies. Proc R Soc Lond Ser B-Biol Sci 265:1665 1673 Aubin T, Jouventin P, Hildebrand C (2000) Penguins use the two-voice system to recognize each other. Proc R Soc Lond Ser B-Biol Sci 267:1081 1087 Aubin T, Mathevon N, Staszewski V, Boulinier T (2007) Acoustic communication in the kittiwake Rissa tridactyla: potential cues for sexual and individual signatures in long calls. Polar Biol 30:1027 1033 Balcombe JP (1990) Vocal recognition of pups by mother Mexican free-tailed bats, Tadarida brasiliensis mexicana. Anim Behav 39:960 966 Balcombe JP, McCracken GF (1992) Vocal recognition in Mexican free-tailed bats do pups recognize mothers? Anim Behav 43:79 87 Barrat A (1976) Quelques aspects de la biologie et de l'ecologie du manchot royal (Aptenodytes patagonicus) des Iles Crozet. Com Nat Fr Rech Ant 40:9 52 Bates D, Sarkar D (2006) lme4: Linear mixed-effects models using S4 classes (R package version 0.9975-10). Available at http://cran.rproject.org/src/contrib/descriptions/lme4.html. Accessed January 25, 2007 Batschelet E (1981) Circular statistics in biology. Academic, London Benhamou S, Poucet B (1998) Landmark use by navigating rats (Rattus norvegicus): contrasting geometric and featural information. J Comp Psychol 112:317 322 Bonadonna F, Spaggiari J, Weimerskirch H (2001) Could osmotaxis explain the ability of blue petrels to return to their burrows at night? J Exp Biol 204:1485 1489 Bonadonna F, Cunningham GB, Jouventin P, Hesters F, Nevitt GA (2003) Evidence for nest odour recognition in two species of diving petrel. J Exp Biol 206:3719 3722 Bost CA, Charrassin JB, Clerquin Y, Ropert-Coudert Y, Le Maho Y (2004) Exploitation of distant marginal ice zones by king penguins during winter. Mar Ecol Prog Ser 283:293 297

Behav Ecol Sociobiol (2010) 64:1145 1156 1155 Bried J, Jouventin P (2001) The king penguin (Aptenodytes patagonicus), a non-nesting bird which selects its breeding habitat. Ibis 143:670 673 Brodbeck DR (1994) Memory for spatial and local cues a comparison of a storing and a nonstoring species. Anim Learn Behav 22:119 133 Caduff D, Timpf S (2008) On the assessment of landmark salience for human navigation. Cogn Process 9:249 267 Challet E, Bost CA, Handrich Y, Gendner JP, Lemaho Y (1994) Behavioral time budget of breeding king penguins (Aptenodytes patagonica). J Zool 233:669 681 Cheng K (2000) How honeybees find a place: lessons from a simple mind. Anim Learn Behav 28:1 15 Cheng K, Spetch ML (1998) Mechanisms of landmark use in mammals and birds. In: Healy S (ed) Spatial representation in animals. Oxford University Press, Oxford, pp 1 17 Cheng K, Spetch ML, Kelly DM, Bingman VP (2006) Small-scale spatial cognition in pigeons. Behav Processes 72:115 127 Cheng K, Narendra A, Sommera S, Wehner R (2009) Traveling in clutter: navigation in the Central Australian desert ant Melophorus bagoti. Behav Processes 80:261 268 Cherel Y, Charrassin JB, Challet E (1994) Energy and protein requirements for molt in the king penguin Aptenodytes patagonicus. Am J Physiol 266:R1182 R1188 Chiaradia A, McBride J, Murray T, Dann P (2007) Effect of fog on the arrival time of little penguins Eudyptula minor: a clue for visual orientation? J Ornithol 148:229 233 Clark JA, Boersma PD, Olmsted DM (2006) Name that tune: call discrimination and individual recognition in Magellanic penguins. Anim Behav 72:1141 1148 Cote SD (2000) Aggressiveness in king penguins in relation to reproductive status and territory location. Anim Behav 59:813 821 Cote SD, Dewasmes G (1999) Do sleeping king penguins influence the movement of conspecifics through a colony? Polar Biol 22:13 16 Daniel TA, Chiaradia A, Logan M, Quinn GP, Reina RD (2007) Synchronized group association in little penguins, Eudyptula minor. Anim Behav 74:1241 1248 Descamps S, Gauthier-Clerc M, Gendner J-P, Le Maho Y (2002) The annual breeding cycle of unbanded king penguins Aptenodytes patagonicus on Possession Island (Crozet). Av Sci 2:87 98 Descamps S, Le Bohec C, Le Maho Y, Gendner JP, Gauthier-Clerc M (2009) Relating demographic performance to breeding-site location in the king penguin. Condor 111:81 87 Dobson FS, Jouventin P (2003a) How mothers find their pups in a colony of Antarctic fur seals. Behav Processes 61:77 85 Dobson FS, Jouventin P (2003b) Use of the nest site as a rendezvous in penguins. Waterbirds 26:409 415 Fukushi T (2001) Homing in wood ants, Formica japonica: use of the skyline panorama. J Exp Biol 204:2063 2072 Gauthier-Clerc M, Le Maho Y, Gendner JP, Durant J, Handrich Y (2001) State-dependent decisions in long-term fasting king penguins, Aptenodytes patagonicus, during courtship and incubation. Anim Behav 62:661 669 Gauthier-Clerc M, Le Maho Y, Gendner JP, Handrich Y (2002) Moulting fast and time constraint for reproduction in the king penguin. Polar Biol 25:288 295 Gauthier-Clerc M, Gendner JP, Ribic CA, Fraser WR, Woehler EJ, Descamps S, Gilly C, Le Bohec C, Le Maho Y (2004) Long-term effects of flipper bands on penguins. Proc R Soc Lond Ser B-Biol Sci 271:S423 S426 Gendner JP, Gauthier-Clerc M, Le Bohec C, Descamps S, Le Maho Y (2005) A new application for transponders in studying penguins. J Field Ornithol 76:138 142 Gibson BM, Kamil AC (2001) Search for a hidden goal by Clark's nutcrackers (Nucifraga columbiana) is more accurate inside than outside a landmark array. Anim Learn Behav 29:234 249 Gould-Beierle KL, Kamil AC (1996) The use of local and global cues by Clark's nutcrackers, Nucifraga columbiana. Anim Behav 52:519 528 Gould-Beierle KL, Kamil AC (1999) The effect of proximity on landmark use in Clark's nutcrackers. Anim Behav 58: 477 488 Guinet C (1992) Predation behavior of killer whales (Orcinus orca) around Crozet Islands. Can J Zool-Rev Can Zool 70:1656 1667 Healy S (1998) Spatial representation in animals. Oxford University Press, Oxford Insley SJ (2000) Long-term vocal recognition in the northern fur seal. Nature 406:404 405 Jansen JK, Boveng PL, Bengtson JL (1998) Foraging modes of chinstrap penguins: contrasts between day and night. Mar Ecol Prog Ser 165:161 172 Kooyman GL, Cherel Y, Lemaho Y, Croxall JP, Thorson PH, Ridoux V (1992) Diving behavior and energetics during foraging cycles in king penguins. Ecol Monogr 62:143 163 Le Bohec C (2007) Life history strategies in long-lived birds: the king penguin. Ph.D. thesis, Louis Pasteur University, p 290 Le Bohec C, Gauthier-Clerc M, Gremillet D, Pradel R, Bechet A, Gendner JP, Le Maho Y (2007) Population dynamics in a longlived seabird: I. Impact of breeding activity on survival and breeding probability in unbanded king penguins. J Anim Ecol 76:1149 1160 Lengagne T, Jouventin P, Aubin T (1999) Finding one's mate in a king penguin colony: efficiency of acoustic communication. Behaviour 136:833 846 Leonard B, McNaughton BL (1990) Spatial representation in the rat: conceptual, behavioral, and neurophysiological perspectives. In: Kesner RP, Olton DS (eds) Neurobiology of comparative cognition. Lawrence Erlbaum, Hillsdale, pp 363 422 Loughry WJ, McCracken GF (1991) Factors influencing femalepup scent recognition in Mexican free-tailed bats. J Mammal 72:624 626 Manser MB, Bell MB (2004) Spatial representation of shelter locations in meerkats, Suricata suricatta. Anim Behav 68:151 157 Mardon J, Bonadonna F (2009) Atypical homing or self-odour avoidance? Blue petrels (Halobaena caerulea) are attracted to their mate's odour but avoid their own. Behav Ecol Sociobiol 63:537 542 Martin GR (1999) Eye structure and foraging in king penguins Aptenodytes patagonicus. Ibis 141:444 450 Martin GR, Young SR (1984) The eye of the humboldt penguin, Spheniscus humboldti visual fields and schematic optics. Proc R Soc Lond Ser B-Biol Sci 223:197 222 Nesterova AP (2007) Age-dependent use of local and global landmarks during escape: experiments using Columbian ground squirrels. Behav Processes 75:276 282 Nesterova AP, Mardon J, Bonadonna F (2009) Orientation in a crowded environment: can King Penguin (Aptenodytes patagonicus) chicks find their creches after a displacement? J Exp Biol 212:210 216 Nocera JJ, Forbes GJ, Giraldeau LA (2006) Inadvertent social information in breeding site selection of natal dispersing birds. Proc R Soc B-Biol Sci 273:349 355 Phillips AV, Stirling I (2000) Vocal individuality in mother and pup South American fur seals, Arctocephalus australis. Mar Mamm Sci 16:592 616 Reilly PN, Cullen JM (1981) The little penguin Eudyptula minor in Victoria. II. Breeding. Emu 81:1 19 Renouf D (1985) A demonstration of the ability of the harbor seal Phoca vitulina (L) to discriminate among pup vocalizations. J Exp Mar Biol Ecol 87:41 46 Schimek MG (ed) (2000) Smoothing and regression. Approaches, computation, and application. Wiley, New York