Antarctica is a cold continent that typically possesses marine ecosystems with still low human presence, although it can be exposed to global anthropogenic activities (^{Bargagli, 2008}). Although metals in Antarctica are linked to a natural phenomenon caused by the different Antarctic areas geochemical characteristics (^{Sánchez-Hernández, 2000}), several anthropo-genic sources (oil spills, paints, open field garbage burning, or fuel combustion) contribute to increasing these natural levels (^{Tin et al., 2009}).

Before the Protocol on Environmental Protection to the Antarctic Treaty (ratified in 1998), many human activities were carried out without considering this region's environmental health (^{Curtosi et al., 2010}). Also, the increasing human activity in the Antarctic Peninsula area (which concentrates most of the research and tourism) could increase the accumulation of metals in the Antarctic biota (^{Lynch et al., 2010}; ^{Jerez et al., 2011}, ^2013a, ^b; ^{Celis et al., 2015}).

Aquatic birds can accumulate trace elements in different tissues; thus, they can be used indirectly to evaluate the marine ecosystem's toxicological status under study (^{Savinov et al., 2003}). Particularly, penguins can be useful indicators of regional environmental health because they are superior predators, long-lived species, and present wide distribution ranges with abundant populations (^{Espejo et al., 2017}). Additionally, penguins are extremely interesting as bioindicators because of their intense molting process (^{Carravieri et al., 2014}), they can be finicky eaters with a restricted diet (^{Lescroël et al., 2004}; ^{García Borboroglu & Dee Boersma, 2013}), and they represent an important part of the avian biomass in this region (^{Dee Boersma, 2008}). Penguins inhabit exclusively in Southern Hemisphere, and about two-thirds of penguin species are seriously threatened by contamination, climate change, fishing, alterations of ecosystems, diseases, and even tourism (^{García-Borboroglu & Dee-Boersma, 2013}; ^{IUCN, 2016}). The Adélie penguin (Pygoscelis adeliae) is a seabird species that only inhabit the entire coast of the Antarctic continent, feeding mainly of krill and fish (^{García-Borboroglu & Dee-Boersma, 2013}).

It is well known that human activities can affect the health condition of birds (^{Burton et al., 2006}; ^{Johnson et al., 2006}; ^{Costantini et al., 2007}; ^{Madsen & Riget, 2007}). Pollutants, such as heavy metals, oil, synthetic organic contaminants, and plastics, can cause oxidative stress and, consequently, affect seabirds' immune system (^{Briggs et al., 1996}; ^{Thompson & Hamer, 2000}; ^{Rainio & Eeva, 2010}). Monitoring pollutants using feathers of seabirds has attracted increased attention among scientists (^{Honda et al., 1986}; ^{Ancora et al., 2002}; ^{Jerez et al., 2011}, ^2013a,^b) because feathers are non-invasive for the bird and because they may be useful for evaluating metal levels in internal tissues of birds (^{Burger, 1993}; ^{Agusa et al., 2005}). Metals can adversely affect captive birds' condition by reducing their body weight, damaging organs, altering their metabolism, or causing behavioral changes (^{Furness, 1996}; ^{Eisler, 2000}). Some studies have been done on relating body condition to metal concentrations in different tissues (e.g., muscles) of aquatic birds (^{Debacker et al., 2000}; ^{Takekawa et al., 2002}; ^{Wayland et al., 2002}). There is evidence relating the body condition of birds to their survival (^{Bergan & Smith, 1993}; ^{Blums et al., 2005}), reproduction (^{Blums et al., 2002}; ^{Bustnes et al., 2002}), and behavior (^{Dufour & Weatherhead, 1991}; ^{Bachman & Widemo, 1999}).

Some metals have a high affinity for the sulfhydryl groups of the feather's structural proteins (^{Metcheva et al., 2006}). Although the accumulation of As, Pb, Hg, Cu, Zn, and Cd has received more attention in penguins, only a few studies have dealt with some other trace elements (^{Espejo et al., 2017}). In birds, it has previously observed that there is a relationship between muscle metal levels and morphological alterations (^{Eisler, 2000}; ^{Ohlendorf & Heinz, 2011}), and also a positive correlation between muscle metal levels and feathers (^{Agusa et al., 2005}; ^{Espejo et al., 2017}). To our knowledge, there are no studies in birds that show a relationship between metal concentrations in feathers and body condition. Although there is likely influence of other contaminants, the biological effects of metal contaminant mixtures are poorly understood and difficult to predict (^{Ohlendorf & Heinz, 2011}). This study is the first approximation to understanding metals' effects on penguins' morphological conditions using feathers as a non-invasive matrix.

The study was carried out during the austral summer between January and February 2016. Samples consisted of feathers from adult Adélie penguins of four locations of the Antarctic Peninsula area, western Antarctica (Fig. 1): Ardley Island (62°13'S, 58°56'W) from the King George Island, Base O'Higgins (63°19'S, 57°53'W) and Kopaitic Island (63°19'S, 57°55'W) from the west coast of Graham Land, and Arctowski (62°09'S, 58°28'W) from the King George Island. At each penguin rookery, the birds were captured using a long-handled net during molting and before foraging. All feather samples were collected using disposable plastic gloves, kept in sealed plastic bags, and stored at −4°C for transport until their laboratory analyses. Besides, birds captured were measured morphologically based on body length (cm) and body mass (kg).

Figure 1 Sampling locations of Adélie penguin colonies from the western Antarctic Peninsula area: a) South Shetland Islands; b) Graham Land. 

In the laboratory, samples were washed with ultrapure water (18.2 MΩ cm⁻¹), dried at room temperature, and ground with IKA^® A11 Basic Auto-miller for sieving (24 mesh dm⁻²). Sub-samples (0.02 and 0.45 g) were digested in the microwave using high purity grade (GR) nitric acid, hydrochloric acid, and perchloric acid. All the reagents used were Suprapur (Merck^®). The concentrations of metals were deter-mined by using electrothermal atomic absorption spectrometry (ETAAS) ZEEnit 60 (Analytik Jena, equipped with Zeeman-effect BG correction system) at the Radioisotopes Lab. Biophysics Institute, Federal University of Rio de Janeiro (Brazil). Measurements were performed in triplicate, and then the values averaged. Quality control was carried out through blanks, which were proceeded through in the same way as the samples, using certified reference material Dolt-4 (dogfish liver), Dorm-3, and Dorm-4 (fish protein) NRCC. The relative condition factor (ReCF) was calculated using a regression of residuals of logarithmically transformed body mass (W) against body length (L), according to ^{Labocha & Hayes (2012)}. The value of ReCF indicates whether a penguin has a good body condition (positive value) or a less than good body condition (negative value) (^{Brown, 1996}).

An analysis of variance (ANOVA) was used to analyze the database, normality, and homogeneity of variance tests. Evaluated parameters included contaminant levels and analyzed morphological values (weight and length) in all four-study locations. As the data failed to pass the normality tests, a nonparametric statistical test (Kruskal-Wallis) was performed, and correlations for metals and condition factors calculated using Spearman's correlations. A value of P ≤ 0.05 was considered significant. All analyses used InfoStat software (^{Di Rienzo et al., 2009}).

The concentrations of metals in Adélie penguin feathers of the four locations studied here are shown in Table 1. Considering all the sampling sites together, metal burden in feathers varied as follows: Zn > Al > Fe > Cu > Se > Mn > Cr > Cd > Ni > V > Pb. The highest Al, V, Fe, and Se mean concentrations were obtained from Arctowski, while the highest Cr, Ni, Cu, Cd, and Pb were found in Ardley Island. In contrast, the highest mean levels of Zn were determined at Kopaitic Island (ANOVA, P ≤ 0.05). When comparing the metal concentrations in Adélie penguin feathers between the locations of the South Shetland Islands and Graham Land (Table 2), we noted that the levels of Al, V, Fe, and Cd levels were significantly higher in South Shetland Islands (P ≤ 0.05). Higher concentrations of Al, V, Fe, and Cd in penguin feathers found in King George Island (South Shetland Islands) may be due to increased human presence. Here, the Antarctic scientific bases, heavy traffic vessels, airplanes, tourists, scientists, and support personnel are most concentrated (^{Tin et al., 2009}). Similarly, ^{Jerez et al. (2011)} reported higher levels of Al, Cr, Mn, Fe, Se, and Pb at King George Island than some Graham Land locations. When comparing with other studies on metals in Adélie penguin feathers, our Al, Cu, Cr, Fe, Ni, Pb, and Se levels were lower than those previously reported by previous studies in Antarctica (Table 5). Our Zn levels observed in the four localities were similar to those previously reported in Admiralty Bay, Avian Island, and King George Island, but below those reported in Zhongshan Station (140 μg g⁻¹, Table 5), a locality in the sub-Antarctic area of the Atlantic Ocean (^{Yin et al., 2008}). For Cd, this metal showed higher concentrations than previously reported (Table 5), and only the levels reported by ^{Ancora et al. (2002)} at Edmonson Point (0.30 μg g⁻¹) are similar to those we found at Artowski (0.32 ± 0.12 μg g⁻¹). Our Mn concentrations (0.20-6.38 μg g⁻¹) are higher than the levels previously reported in Adélie penguins (<0.01-2.01 μg g⁻¹) (Table 5), even though lower than the Mn levels detected in feathers of adult seabirds from industrialized and populated areas, such as the Brazilian coasts (11.4 μg g⁻¹, ^{Barbieri et al., 2010}). Our increasing Mn levels in Adélie penguins seem to be related to its current use as an additive in combustibles in replacement of Pb (^{Burger & Gochfeld, 2000}). At the same time, recent evidence shows that Mn levels in hepatic tissues of Antarctic penguins (^{Jerez et al., 2013b}) are slightly higher than those detected two decades ago (^{Honda et al., 1986}; ^{Szefer et al., 1993}). ^{Agusa et al. (2005)} found V levels of 0.076 μg g⁻¹ dry weight in feathers of black-tailed gulls (Larus crassirostris) from Japan, but in general, the levels of V in feathers of seabirds have been poorly investigated. At present, it is not possible to detect a geographical pattern in the concentrations of metals because data are fragmented, in which case any spatial analysis must be carried out with caution, following the recommendation indicated by ^{Espejo et al. 2017}.

Adult penguins at Arctowski exhibited a lower average weight (W) than the penguins in the colonies at Ardley Island and Base O'Higgins (ANOVA, P ≤ 0.05, Table 3). Also, body condition values (ReCF) did not differ between Ardley Island and Kopaitic Island and between Base O'Higgins and Arctowski. However, ReCF values on both sites (Ardley and Kopaitic islands) did differ significantly from the ReCF values found at Base O'Higgins and Arctowski. There was no significant difference in mean length (L) between the sites studied (P > 0.05). Of all analyzed trace elements across the four study sites, only ReCF showed significant correlations to both Cu and Zn levels in feather samples (Table 4). Among locations, feather Cu and Zn in adult Adélie penguins were related positively (P ≤ 0.01) to their body condition (ReCF) only at Arctowski (Fig. 2). On this site, penguins also exhibited a significantly lower weight (Table 1); possibly, both metals stores decreased by some physiological stress (^{Wayland et al., 2002}), caused by low food availability (^{Wiebkin, 2012}). Considering that penguin rockeries at Arctowski are near the highest concentration of Antarctica's human activities (^{Tin et al., 2009}; ^{Espejo et al., 2014}), food availability for these penguins colonies could negatively be affected (^{García-Borboroglu & Dee-Boersma, 2013}). Studies in Antarctica have shown that impacted near-shores sites generally have lower species richness, biodiversity, and variability than control sites (^{Stark et al., 2006}). In any case, a low body condition of seabirds often implies little subcutaneous fat deposits and reduced pectoral muscle mass (^{Debacker et al., 2000}).

Figure 2 Correlation of relative condition factor (ReCF) to metal concentrations in Adélie penguins (Pygoscelis adeliae) at Arctowski, South Shetland Islands. 

Some studies have found a relationship between feathers and muscle tissues for some trace elements in birds (^{Del Hoyo et al., 1992}), reaching up to 85% of similarity between the concentrations of Al, As, Cd, Pb, Hg, Cu, Zn and Mn in feathers and muscles of penguins (^{Espejo et al., 2017}). Additionally, we found significant positive correlations between Al-Fe, Cr-Cu, Cr-Pb, V-Se, Fe-Pb, Ni-Cu, Ni-Cd, Cu-Zn, Cu-Cd, Cu-Pb, Zn-Se, Se-Pb, and Cd-Pb, while a significant negative correlation between V-Pb. There is evidence showing that Se and Zn have a detoxifying effect on Cd and Hg in seabirds from the Arctic and Antarctica (^{Norheim, 1987}; ^{Smichowski et al., 2006}). Moreover, Se has shown to have some degree of detoxifying effect on Hg, Cd, and Pb in birds (^{Ohlendorf & Heinz, 2011}) and penguins (^{Kehrig et al., 2015}). Our results indicate that probably Cu acts as a detoxifying agent for Cr, Cd, and Pb, whereas Se could be for V and Pb, which needs to be more investigated.

In conclusion, as indicated in Table 5, this study is the first report of Cr, V, Fe, Ni, and Se concentrations in Adelie penguin feathers. Further studies are needed to assess other metals in Adelie penguins. The levels of Al, Cd, Cr, Cu, Fe, Mn, Ni, Pb, Se, V, and Zn found in this study indicate these metals are relevant in the Antarctic environment to bioaccumulate in Adélie penguin feathers. The results showed that, except for Cu and Zn, most of the metals in Adélie penguins' feathers showed no significant relationship with the body condition of the penguin rockeries studied here. However, this relationship may be site-specific, and metals might be affecting the organisms at low biological levels (e.g., molecular, biochemical). Although feathers are a pathway for detoxification of the organism of penguins (^{Espejo et al., 2017}), little is known about the interaction of metals with penguins' health; hence, genetic, molecular, and biochemical studies are required to assess those possible biological effects. Since feathers are non-invasive to birds, this study provides valuable information to validate them as a biological matrix for measuring birds' body health.