Are the diets of sympatric Pygoscelid penguins more similar than previously thought?

In recent years, functional changes in Southern Ocean are becoming more noticeable, due to climate change and increasing human impacts, including a growing fishery that is concentrating in the Antarctic Peninsula (AP) region. Antarctic krill Euphausia superba is often the primary prey species for animals such as Pygoscelis penguins, a sentinel species for ecosystem monitoring and management. During the last two decades in the AP gentoo penguin numbers (Pygoscelis papua) have increased and their range has shifted southward, in contrast to the decline in numbers of Adélie (P. adeliae) and chinstrap (P. antarcticus) penguins. Given divergent population trends, the goal of this study was to examine differences in their diet, and size structure of Antarctic krill recovered from penguin diet samples. The study is based on diet samples collected during the austral summers on King George Island (South Shetland Islands) where P. adeliae, P. antarcticus, and P. papua breed in mixed colonies. Results indicate that the penguins consumed krill of similar sizes during the breeding period. In contrast to prior diet studies, we found higher proportions of krill in the gentoo diet and changes in the percentage of krill in the diet relative obtained during 1970s. The similarity in diets among all three species suggests that the availability prey items (e. g., fishes) may be changing and driving higher dietary overlap. Moreover, we also check differences in krill length among penguin individuals and we did not find any statistically significant differences. We also found plastic debris in penguin stomachs during both summers.


Introduction
The marine ecosystem in the Antarctic Peninsula region is strongly dependent on the availability of a small, swarming crustacean, Euphausia superba, (hereafter krill), which is the food base for numerous whale, seal and seabird species (Hill et al. 2006). For some of the predators, such as Pygoscelid penguins, krill can constitute up to 99% of their diet (Tierney et al. 2008;Juáres et al. 2018;Panasiuk et al. 2020). Krill biomass is estimated at around 380 million tons (Atkinson et al. 2009) and its harvesting is regulated by the Commission for the Conservation of Antarctic Marine Living Resources (CCAMLR). Krill are predicted to be affected by harvesting and the effects of climate change, including warming sea-surface temperatures, melting of sea ice, and changes in phytoplankton composition and abundance (Hill et al. 2006). In particular, the effects of climate warming on krill growth may have significant implications for krill biomass which can cause negative consequence for dependent predators (Klein et al. 2018).
Among the important krill-dependent predators are penguins, which constitute nearly 90% of Antarctic avian biomass (Croxall et al. 2002). As the main component of Antarctic avifauna, Pygoscelid penguins exhibit extensive overlap in their distributions and nesting sites (Volkman 1 3 et al. 1980;Lishman 1985). They exhibit different foraging locations and foraging-depth ranges during the breeding season, but also during winter period (Hinke et al. 2019;Korczak-Abshire et al. 2021), which helps maintain their co-occurrence throughout the AP region. Despite such differences in niche, the similarity of diet among the Pygoscelid penguins suggests a shared constraint on their performance (Hinke et al. 2007). Their diet is heavily dependent on krill (70 to 100% of the diet) (Volkman et al. 1980;Hinke et al. 2007;Panasiuk et al. 2020), thus their foraging distribution and abundance trends are usually linked with krill biomass variability (Trivelpiece et al. 2011a, b;Stryker et al. 2020).
Relatively few research sites afford the opportunity to examine the diet of the three Pygoscelid penguin species simultaneously. Long-term research on breeding populations of Adélie (Pygoscelis adeliae), gentoo (P. papua), and chinstrap (P. antarcticus) penguins in Admiralty Bay, King George Island included study of their trophic niches. Volkman et al. (1980) and Trivelpiece et al. (1987) demonstrated that gentoo diets contained more fish than Adélie or chinstrap diets. More recently, Miller et al. (2010) highlighted local differences in diets, reporting that from 1997-1998 to 2003-2004 fish constituted 28.8% of a diet of gentoo and 0.6% of chinstrap at Cape Shirreff (Livingston Island), while in Admiralty Bay (King George Island) fish represented 1.6% of a diet of gentoo and 0.1% of chinstrap. Within the last three decades in Admiralty Bay, the breeding populations of Adélie and chinstrap penguins have decreased, while the population of gentoo penguins has increased (Watters et al. 2020), consistent with regional trends (Lynch et al. 2012). Such large-scale changes in the abundance and distribution of each species may affect their competitive relationships for food resources.
The main goal of our study was to analyze the diet composition and krill sizes consumed by sympatric Adélie, gentoo and chinstrap, and to compare our results with historical diet data. As a null hypothesis, we expected similar diet compositions of Adélie and chinstraps, but a higher proportion of fish in gentoo diets. We also expected similar distributions of the sizes, sex, and maturity stages of krill in the diets of three sympatric penguin species that initiate foraging trips from Admiralty Bay, leading to more similar diet characteristics that previously observed. We used two consecutive years of diet data to address our hypotheses related to dietary composition of each species.

Sampling collection and laboratory studies
Diet samples were collected in the Austral summers (December-February) of 2011/2012 and 2012/2013. Fieldwork was conducted at King George Island, South Shetland Islands (62° 10' S, 58° 30'W), within the Antarctic Specially Protected Area (ASPA) No. 128 where Adélie, chinstrap, and gentoo penguins breed in five colonies (Fig. 1). Following procedures of the Conservation of Antarctic Marine Living Resources (CCAMLR) Ecosystem Monitoring Program (CEMP), the population size, expressed in the total number of occupied nests, and breeding success, expressed as the mean number of chicks alive during the crèche stage per nest were calculated. Diet samples were collected from adult penguins after their foraging trips using the water-offloading technique (Wilson 1984) with CEMP standard methods modifications (see more in Panasiuk et al. 2020). Diet samples were sorted to identify the frequency of occurrence by number and weight of krill, fish, and other, unidentifiable diet items. All krill individuals collected in the diet samples were counted (based on the pairs of eyes for digested animals). Whole, undamaged individuals were sexed, and measured for total length, according to CCAMLR standard protocol (https:// www. ccamlr. org/): from the anterior side of the eyeball to the tip of the telson (Makarov and Denys 1980). Diet samples from 2011/12 were used to estimate the difference between krill sizes consumed by penguins and to compare with krill sizes presented by Panasiuk et al. (2020). Diet samples from both seasons were used for comparing krill sizes and maturity stages. We only used diet samples that contained at least 30 whole individuals for comparing krill sizes and maturity stages. Sexed animals were grouped into the following clusters: F-females, Fs-sexually active females defined by presence of spermatophores, and Mmales. Additional groups-Es and J-were used for individuals without clear sex indication and individuals with damaged bodies (Es), and juvenile individuals.

Statistical analyses
Statistical analyses were carried out using Rstudio software package, Rversion 4.0.3. We used a one-way ANOVA with a Tukey's HSD post hoc test to compare differences in krill length and maturity stages in the penguin diet samples. Additionally, we used a nested ANOVA to compare differences inn krill lengths in which krill length was a response variable and years and penguin species were factors with year nested within species. Differences in krill sizes between years were statistically significant (p < 0.0001), however, because our main goal was to check differences between penguin species and inter-individual differences within species we pooled samples to examine sex and maturity stages of the krill population eaten by penguins. We also checked interindividual differences in krill lengths using the Kruskal-Wallis test.

Results
Krill, by mass and number, constituted more than 99% of the diet in all samples (Table 1), with only small contributions of other prey, including amphipods, and other euphausiids (Thysanoessa macrura, and Euphausia frigida) and fish. In case of frequency of occurrence-tiny fish remains or otoliths were reported in 70% of gentoo penguin stomachs in season 2012/2013 and 20% in 2011/2012. For Adélie and chinstrap, however, fish prey was present in first researched season in only 10% of individuals.
Most krill individuals in the diet samples were classified as adults with a median length of 40 mm (mean 39.76 ± 6.93). Juvenile krill were rare (Fig. 2, Table 2) and were most abundant in the Adélie diet. On average, the largest krill were observed in the diet samples of gentoo penguins, while diet samples from Adélie and chinstrap penguins contained smaller krill (Fig. 2, Table 1). The mean sizes of krill consumed by Adélie penguins was 38.40 ± 7.28 mm and chinstrap penguins was 40.68 ± 5.55 mm, while in gentoo diets was 41.7 ± 6.22 mm (Fig. 3, Table 1). The differences between the length of a krill in the diets were biggest between Adélie and chinstrap penguins and Adélie and gentoo penguins (Tukey HSD, p < 0.0001). Smaller difference was noted between gentoo and chinstrap penguins (Tukey HSD, p = 0.015). We also checked interindividual differences, and results of the Kruskal-Wallis   test showed no statistically significant differences among Adélies (p = 0.4616), gentoos (p = 0.4478), and chinstraps (p = 0.4289).
Female krill formed the majority of the diet for all penguin species (Table 2). The largest average size of krill in each sex and maturity stage was found in the gentoo diet ( Table 2). The biggest krill was a 73 mm female with spermatophore (Fs), and this group represented, on average, the largest animals in all samples with mean lengths that varied from 44.43 to 46.45 mm (Table 2). Male krill individuals were the second largest group in terms of size (both by median and mean value) ( Table 2).
We found algae, stones, feathers and other debris in the Pygoscelis penguin stomachs (Fig. 4). Some net fragments were identified. They were accompanied by plastic film elements of uncertain origin, although some of them looked like fragments of plastic bags. Most of the plastic debris could be classified as microplastics (1-5 mm; 10 plastic parts), mesoplastics (< 5-20 mm; 23 plastic parts) (Barnes et al. 2009) and classified as user plastic (only non-industrial remains of

Krill dominance and size in Pygoscelis diet
Krill constituted almost 100% of the penguin diets, in weight and by number, during the 2011/2012 summer season, similar to other reports (Miller et al. 2010;Panasiuk et al. 2020). Fishes, in all cases, constituted < 0.1% of mass in the diet ( Table 1). The similarities in diet composition among sympatric Pygoscelid penguins observed during our study differ from those observed roughly for decades ago. For example, Trivelpiece et al. (1987) noted that krill constituted 86.5% of in the gentoo penguin diet during the period 1976-1985. Similar proportions were also noted earlier by Volkman et al. (1980), with the krill comprising 84.5% of the gentoo    (Table 3). Jabłoński (1985) recorded even 48.6% of fish in the gentoo diet (Table 3). We explore possible reasons for the increased similarity in diet below. Adélie and chinstrap penguins in the broader Antarctic Peninsula region typically exhibit a summer diet-based almost exclusively of krill (Lishman 1985). Such high amounts of krill in the diet of all three species suggest sufficient local availability of krill in the environment during the summer breeding period (Hinke et al. 2007;Nardelli et al. 2021). The apparent increase in the occurrence of krill in the diet of gentoo penguins since the early 1980s may imply a decline in availability of secondary prey items, particularly fish. While we noted that 70% of gentoos consumed fish in 2012/2013 season which would corresponds with Miller et al. (2010) findings in Cape Shirreff (Livingston Island), most of noted presence in our study was based on otoliths presence in stomach content. Gentoo penguins were known to consume Pleurogramma antarctica in King George Island vicinity (Volkman et al. 1980). A study of the distribution of Pleurogramma on the Western Antarctic Peninsula shelf exposed a gap in its occurrence which could indicate a local collapse of this fish (no individuals smaller than 150 mm were recorded, and after 1990s smaller contribution in predators diet has been noted) (Parker et al. 2015). Evidence of a decade's long decline in the occurrence of fish in Antarctic fur seal diets (Klemmendson et al. 2020) provides supporting evidence for potential changes in the preyscape available to penguins in this region. The relative dominance of krill in the diet is also supported by recent results from a molecular study of penguin feces (Zeng et al. 2022) that suggests a diet dominated by Euphausiid species. Therefore, further study is necessary to explain the switch to a diet with higher proportions of krill, especially with increasing population of gentoo penguins in the region, and the concurrent decrease in Adélie and chinstrap penguin populations numbers (Lynch et al. 2012). We note that the stomach contents can be related to sample timing relative to the period of a day or night foraging. Jansen et al. (1998) andTrivelpiece (2008) observed during their studies that the diet of chinstrap penguins consists of more fish during nights, and samples for this study were collected only in the afternoons.
In our study, the mean length of krill in all penguin stomach contents was 38-42 mm and adult individuals were the majority (Figs. 2, 3). These sizes are comparable to the findings of Volkman et al. (1980) and Coria et al. (1995), but smaller than the animals noted by Trivelpiece et al. (2011a) from the same area of research. Differences in the length of krill consumed by penguins may be caused by differences in the availability of certain age classes, different foraging areas, or year-to-year differences in prey availability, recruitment, and abundance (Coria et al. 1995;Hewitt et al. 2003;Kokobun et al. 2015;Reiss 2016). It is worth noting that our study area is recognized as a region of occurrence of juvenile krill (Fevolden and George 1984) and cycles of krill recruitment are evident throughout the WAP (Miller and Trivelpiece 2007;Saba et al. 2014;Reiss 2016;Conroy et al. 2020). However, few krill less than 30 mm in length were observed in the diet samples, which suggests that presumably krill recruitment in the area was not high in the preceding year.
Our two-year study confirmed that the largest krill were consumed by gentoo penguins, while the smallest animals were typically found in Adélie diets (Figs. 2, 3; Table 1). In general however, the krill sizes in the diets of the three penguin species were similar. It is evident that strong dietary overlap exists between Pygoscelis penguins in the study area, and that dietary overlap may be much more pronounced than reported in the past (Volkman et al. 1980;Jabłoński 1985;Lishman 1985;Trivelpiece et al. 1987).

Krill sex and stages in Pygoscelis diet
During this study, a preference for female krill was evident across penguin species (Table 2). In over 3800 measured krill individuals, only 6.5% were recognized as males. Similar results were observed by Volkman et al. (1980), Reid et al. (1996), and Lynnes et al. (2004). In our study female krill with attached spermatophores constituted more than 20% of all measured females. The disproportionate occurrence of female krill in the diet of predators also was surprising for Reid et al. (1996) around the South Georgia, who surmised that predators opted to catch the largest available krill-gravid females. Sexually active females were the most numerous group in predator diet and net samples in researched done by Reid et al. (1996). Hill et al. (1996) suggested that for macaroni penguins, the predominance of female krill in the diet could be a result of capture avoidance by faster swimming male krill. The higher proportion of female krill in penguin diet may also point to selectivity for the largest individuals available. In the Bransfield Strait,  Table 3 Comparisons of composition of pygoscelid stomach contents by weight (%) at King George Island populations in different years (Amp. amphipods) Type of prey Volkman et al. (1980) King George Island: 1977-1978Jabłoński (1985 King George island: 1978-1979, 1979-1981Miller et al. (2010) King George Island: 1997-1998to 2003-2004gentoo, 1997-1998  juvenile krill generally form the majority of the population, while adult females represent a smaller fraction of population (Siegel 1983;Fevolden and George 1984). Therefore, the dominance of female krill in penguin stomachs affirms selectivity for the largest krill individuals available. The female krill carrying spermatophores were also the largest krill in the available diet samples with a mean length of 45 mm for all species (Table 2). The preponderance of females in the diet may also affected by digestion. Irvine (2002) suggested that smaller males are usually much more affected by the sampling methods and preservation than females, and tend to break. Those factors may lead to the underestimations of smaller males in samples and could explain the relatively small number of males < 40 mm present in the diet samples.

Krill availability in Pygoscelis diet and their breeding success
Many reports suggest a direct linkage between reproductive success of predators and fluctuations in krill abundance and structure (Lynnes et al. 2004;Trivelpiece et al. 2011b). In the 2011/2012 breeding season, the abundance of Adélie and chinstrap penguins was larger than during the 2012/13 breeding season (Table 1). The situation was in opposite for gentoo penguins. The breeding success of all three penguin species, however, was higher in 2011/2012 than in 2012/2013 (Table 1), similar to other breeding locations like the Lions Rump (Korczak-Abshire personal observations) (ASPA 151), and the Stranger Point (Juáres et al. 2020). This suggests that the 2011/12 season may have been better for the penguins. During our study we did not find any considerable differences in diet composition, krill size, or numerical of krill between both seasons. However, in a krill dominated system, we may not expect the diet composition of penguins to change considerably, but the performance of birds can be negatively affected if foraging trip durations are long (e.g., Hinke et al. 2007). Episodic phenomena, including heavy snowfall events (Spring 2012 and January 2013-personal observation), can also negatively affect reproductive performance (see Hinke et al. 2012). For example, the strong wind, rainfalls, and thick snow events affect the increase of chicks' energy input and lower thermal insulation factors. Thus, the observed difference in breeding success between years may reflect meteorological anomalies during incubation and guard stages.

What is the reason for dietary overlap?
Historical data shows a distinction between chinstrap and gentoo trophic niche based on diet differences (Volkman et al. 1980) and foraging ranges (Trivelpiece et al. 1987). However, climate change in Antarctic Peninsula region (Meredith and King 2005) Ciaputa and Sierakowski (1999) and they suggested that penguin population trends and krill abundance were linked. Population changes among gentoo and chinstrap penguins are related to temperature and its anomalies (Petry et al. 2016), while the decline in Adélie penguin numbers is correlated with reduction of sea ice during winter season (Juáres et al. 2015). Taking into consideration increases in air temperature from 1981 to 2011 during summer, autumn and winter seasons, along with increases in sea-surface temperatures and increased variability in sea ice extent (Kejna et al. 2013), we should expect a response in birds population numbers. Yet, between the 2011/2012 and 2012/2013 summer seasons, we didn't detect substantial differences in krill size composition in Pygoscelis stomachs. Similarity in diet composition across the penguin species can be connected with a lower availability of krill or other alternative prey items with an increasing competition between them. If the numbers of Adélie and chinstrap penguins are in decline, while gentoo populations rise, it could indicate that gentoo penguins are expanding their trophic niche to occupy the vacancy created by decreasing populations of its congeners, and this possibility was also discussed by Miller et al. (2010).

Plastic debris in penguin stomachs
One of the first observations of ingested plastic by marine vertebrates south of the Polar Front was in a snow petrel (Pagodroma nivea) in the 1980s (Van Franeker and Bell 1988). Similar findings for other Antarctic predators remain rare (Waller et al. 2017). We found plastic debris in the Pygoscelis penguins stomachs in two consecutive seasons. Among Antarctic penguins, evidence for plastic ingestion was noted from penguin scats (Bessa et al. 2019) in the regions of Bird Island (South Georgia) and Signy Island (South Orkney Islands), and in regurgitate (Panasiuk et al. 2020) collected from birds at King George Island (South Shetland Islands). Although the Southern Ocean has smaller concentration of micro-and macroplastic than other parts of the World Oceans (Suaria et al. 2020), observations of plastic debris in the penguin stomachs during two breeding seasons (now one decade ago) suggest that its occurrence in diets is not a spurious finding. We found that Pygoscelid penguins are also able to ingest larger pieces of plastic (Fig. 4) and this could indicate additional potential mechanisms to impact penguins. While there are grounds for supposition, answer the question: are the diets of sympatric Pygoscelid penguins more similar than previously thought? definitely needs more data. It is possible that widely studied belief about the decline of E. superba (e.g., Atkinson et al. 2009) could be now facing its turning point with confirmed big aggregations of returning whales, as fin whales noted in 2018 by Herr et al. (2022). Such situation could suggest that if there is enough of krill to support dietary needs of some predators groups this could also be a reason for higher percentage of dietary overlap of pygoscelid penguins. However, to regard this as a trend-such observation as in presented study need to be continued, as well as studies on what drives these changes.