Deep-sea ascending predation by king penguin and its prey reaction observed by animal-borne video camera

Certain air-breathing animals, which spend much of their lives near the ocean surface or on land, are able to capture prey in the deep ocean where the energetics and the time spent foraging are limited by their lung capacity and the amount of dissolved oxygen in their blood. Animal-borne video loggers provide a valuable tool for investigating underwater foraging behavior (Davis et al., 1999). However, capturing video footage at dark ocean depths remains challenging, particularly for penguins. While animal-borne cameras with light sources have been successfully used on large-bodied animals such as seals (Adachi et al., 2021; Foster-Dyer et al., 2023; Heaslip & Hooker, 2008), the smaller body size of penguins restricts the types of devices that can be attached. Although some studies have investigated penguin foraging behavior using video recordings (e.g., Ponganis et al., 2000; Thiebault et al., 2019; Tokunaga et al., 2023), these recordings were not obtained in the low-light environment of the deep sea.

Here, we report a successful recording of deep-sea foraging behavior of a king penguin (Aptenodytes patagonicus) and the prey's response using a video logger with an LED light source (Figure 1). King penguins, the second largest species of penguin (Spheniscidae), rely on visual cues for foraging (Ainley & Wilson, 2023; Bost et al., 2002; Martin, 1999). They can dive to depths exceeding 100 m, well below the photic zone, in search of food (max: 343 m, Pütz & Cherel, 2005). Their diet primarily consists of myctophid fish (Myctophidae) (Bost et al., 2002; Cherel & Ridoux, 1992), the most abundant mesopelagic fish in the ocean (Catul et al., 2011; de Busserolles et al., 2015). Myctophids also serve as the main food source for other marine species in sub-Antarctic marine regions including southern elephant seals (Mirounga leonina) (Cherel et al., 2008) and Antarctic fur seals (Arctocephalus gazella) (Cherel et al., 1997). As king penguins forage within the constraints of diving duration, limited light availability, and interspecies competition, their foraging strategy provides a unique model for studying the complex challenges of deep-sea foraging. Additionally, video footage of predation events allows for the analysis of prey response behavior in myctophids to king penguins, which remains largely unexplored (Ainley & Wilson, 2023).

In February 2024, we fitted five king penguins breeding at the Ratmanoff colony, Kerguelen Archipelago (southern Indian Ocean, 49°14'S; 70°33'E), with digital video loggers that included red LED light sources (LoggLaw-CAM, Biologging Solutions Inc., Japan, 38 mm diameter, 67 mm length, 128 g, 30 fps, full HD). These loggers also recorded depth every second and were programmed to start video recording upon reaching 80 m depth using the depth sensor trigger function. They were attached to the penguin's back using waterproof tape (Tesa, Germany), Loctite glue (Henkel, Germany), and plastic cable ties (Hellermann Tyton, UK). See Appendix S1: Section S1 for details on the potential effects of the light source on penguin and prey behavior. Four penguins were recaptured, and their loggers were retrieved. One of these penguins repeatedly dived for food during the video recording period (1 h 32 min), which occurred at dawn when king penguins typically begin foraging actively (Watanabe et al., 2023). The penguin was on a seven-day foraging trip, with the video recording occurring on the third day, midway through its trip. Although penguins are highly hydrodynamic and the attached logger may have affected the individual's performance, its trip duration, dive depth (mean: 145 m, n = 704), and dive duration (mean: 4.5 min, n = 704) derived from the full trip data were comparable to those reported in previous studies (Charrassin et al., 2002; Pütz et al., 1998; Scheffer et al., 2016). The other loggers failed to function the depth trigger correctly, and no foraging dives were recorded.

During the video recording period, the penguin made 12 foraging dives (mean dive depth ± SD: 128 ± 31 m) including 10 complete dives and 2 that were only partially recorded (Figure 2a). The penguin exhibited 136 foraging attempts, successfully capturing prey in 118 cases (86.8%). In unsuccessful attempts, the penguin either missed its prey or else detected it and stretched its neck but did not attempt to peck. According to the depth profile, 111 foraging attempts occurred during the penguin's ascent (Figure 2b,c), accounting for 81.6% of total attempts.

All prey silhouettes in the recordings resembled fish, consistent with findings that the summer diet of king penguins in this region primarily consists of myctophids, specifically Krefftichthys anderssoni (Bost et al., 2002; Cherel & Ridoux, 1992). The silhouettes were approximately 40–50 mm in length based on a comparison with the king penguin's beak (~85 mm). Small particles, often referred to as marine snow, were consistently present in the field of view (Video S1).

From the video footage, the exact moment when the penguin began targeting prey was distinguishable in 16 cases (Video S2). The penguin started targeting 0.5–1.5 s before capture by directing its head toward the prey, corresponding to a distance of 1–3 m given that the mean cruising speed of king penguins is approximately 2.0 m/s (Culik et al., 1996; Sato et al., 2010). At the point of capture, the penguin extended its long, flexible neck (Guinard et al., 2010) by approximately 10–20 cm to make adjustments by straightening its head toward the prey regardless of direction.

The recorded fish were either stationary or swimming at a slower speed than the penguin, resulting in minimal positional changes within the frame. Regardless of movement, the fish exhibited similar frequencies of evasive (55 cases) and non-evasive (57 cases) behaviors (Appendix S1: Figure S1; Video S3). The remaining 24 cases were unclear, preventing observation of the prey's reaction. Evasive responses, if exhibited, initiated either within the same video frame as capture or 1–3 frames (i.e., 0.03–0.10 s) before capture, typically too late for the fish to escape. In these instances, just before capture, the fish's body assumed a C-shape, a commonly observed rapid evasive maneuver (Domenici & Blake, 1997; Noda et al., 2014). In 9 of the 55 evasive cases, the fish successfully avoided capture. However, even without evasive behavior by the fish, the penguin occasionally missed its prey (8 of 57 cases). In most of the instances, the penguin extended its neck but did not attempt to capture the prey, rather than failing due to an unsuccessful strike. After missing prey, the penguin did not pursue further or persisted in re-attempting to capture the fish (cf. Tokunaga et al., 2023). Based on the stability of the camera angle and the sound recorded by the microphone in the logger, the penguin frequently adjusted its swimming direction. Although swimming speed can vary from 1.0 to 3.0 m/s (Ropert-Coudert et al., 2000), such variation was not clearly discernible from the video footage.

Prey–predator interactions in low-light conditions provide crucial insights into the predation strategy of air-breathing animals and the survival tactics of the prey in the deep sea. Although some prey capture attempts undoubtedly occurred outside the camera's view, the high foraging success rate observed in the recorded footage (86.8%) is remarkable, highlighting the need for further video-based approaches.

From the analyzed video footage, particularly the timing at which the penguin oriented its neck toward the prey, we inferred that king penguins recognize the presence of fish from a distance of at least 1–3 m. Moreover, since targeting occurs after the initial recognition of prey, penguins may detect fish even earlier because they can dilate their pupils as widely as nocturnal owls to maintain vision in dark environments (Martin, 1999).

The idea that king penguins tend to forage during their ascents has been suggested in previous research (Ainley & Wilson, 2023; Ropert-Coudert et al., 2000); however, visual evidence indicating the precise moments of foraging has been lacking. Although data were obtained from only one individual, our video footage strongly supports this idea. The dive profile obtained in our study shows that the penguin repeatedly ascended and descended during the bottom phase, a behavior described as “wiggle” (Bost et al., 2007), allowing the penguin to remain within a specific depth range where prey is abundant even while foraging during ascent.

This strategy provides several benefits. One is the counter-shading effect, which helps penguins detect fish silhouettes against the bright surface (Ropert-Coudert et al., 2000, 2001). Additionally, approaching from the darker, deeper side of the ocean may offer the advantage of reduced visual detection by fish, potentially allowing predators to remain unnoticed for longer periods. Furthermore, the air trapped in the penguin's air sacs and feathers creates positive buoyancy, making it easier to capture prey from below while ascending, as less effort is required compared to descending (Ainley & Wilson, 2023), even though they must descend before ascending again.

In addition to penguin behavior, we also examined the behavior of the preyed-upon myctophids, an abundant species in low-light ocean zones. From the video footage, we observed that either the fish did not evade the penguin at all or else their evasion behavior started only shortly before being captured. One possible reason for this lack of evasion, or the delayed evasion, is that they were unable to detect the approaching penguin. Myctophids are also known to exhibit torpor at deep depths during the day (Gon & Heemstra, 1990). Although some myctophid species can detect bioluminescence from tens of meters away (Turner et al., 2009), this does not apply to penguins, which do not emit light. Detecting penguins becomes even more challenging when they approach from the darker and deeper parts of the ocean. The reactions of the fish just before capture may instead be triggered by sensing turbulence or vibrations in the surrounding water caused by the approaching penguin's head.

It is important to consider that the prey's response behavior could be a strategic attempt to avoid capture (some evasions in our observations were successful). For fish with considerably slower swimming speeds (0.1–0.3 m/s, Ignatyev 1996) than those of penguins, the optimal evasion strategy may be to dodge the penguin rather than swim away (Wilson et al., 2015). If evasion occurs too early, it allows the penguin to adjust its direction; thus, remaining stationary until the last moment may be advantageous. However, based on the analyzed video footage, we consider it unlikely that the fish timed their evasions, as nearly all fish displaying evasive behavior were captured immediately after adopting a C-shaped posture, too late to dodge the penguin's strike.

The foraging behavior of king penguins, as observed in this study, can be described as follows: They dive to a certain depth, ascend while capturing prey, and then descend to repeat the process multiple times. During ascent, they actively adjust their swimming direction and speed to seek prey, targeting stationary or slow-moving fish (or their aggregations) from several meters away before capturing them from below. At the moment of capture, they extend their necks by approximately 10–20 cm to precisely adjust their strike. The fish remain unaware until the moment of capture or just before. If the penguins miss a prey item, they do not pursue it but instead shift their focus to another nearby fish, suggesting an abundance of fish schools in the water column. Although fish occasionally succeed in evasive maneuvers, their attempts are generally ineffective. Their primary survival strategy appears to be blending into a large group rather than escaping individually.

This study's contribution to understanding the foraging strategies of king penguins raises new questions about how animals sharing the same habitat and competing for the same food sources develop their foraging strategies based on their respective physical and locomotive characteristics. Recent studies on southern elephant seals and Antarctic fur seals, which also inhabit the sub-Antarctic marine region and feed on myctophids, have reported distinct foraging strategies (Chevallay et al., 2023, 2024). Antarctic fur seals actively chase evading fish, whereas southern elephant seals stealthily approach their prey from distances greater than 10 meters. Both strategies differ from that of king penguins, particularly in the timing of prey targeting and the evasive behavior of their prey. Notably, the behavior of approaching prey from below during ascent, leading to the formation of the “wiggle” pattern in the diving profile, is unique to king penguins among these three species. These differences in foraging strategies result from variations in body size, sensory capabilities, movement capabilities, and maneuverabilities. The repeated ascent-and-descent pattern employed by king penguins may be best aligned with their high maneuverability, fast swimming speed, and relatively small body size compared to other mammals. At broader spatial scales, an important question is how species differ in movement distance and prey encounter rates in the highly reduced light environment, factors that ultimately determine their overall foraging efficiency. Accumulating data from more king penguins could provide deeper insights into predator–prey behavior, allowing for a better understanding of the strategic differences among species competing for the same prey.

Leo Uesaka: Conceptualization, resources, formal analysis and investigation, writing original draft. Charles André Bost: Conceptualization, resources, review and editing, project administration, funding acquisition. Katsufumi Sato: Conceptualization, resources, review and editing, project administration. Kentaro Q. Sakamoto: Conceptualization, resources, review and editing, project administration, funding acquisition.

The authors declare no conflicts of interest.

Animal care was performed humanely following the rules issued by the Réserve Nationale des Terres Australes. The experiment was conducted with permission from the Animal Experimental Committee of the Atmosphere and Ocean Research Institute, University of Tokyo (permit number: P23-28), and the Préfet des Terres Australes et Antarctiques Françaises, France, after approval from the Comité National de la Protection de la Nature (CNPN 2024 annual decision).