Poor body conditions during the breeding period in a seabird population with low breeding success

Breeding has a high energetic cost and, in central place foragers, also restricts the ability of parents for self-foraging, with affecting on their body and immunological conditions. In migratory species, breeding overlaps the period of recovery of body conditions before the following outbound migration. Those factors may reduce the breeding success and even lead parents to abandon their offspring if parental body condition cannot guarantee the success of the next migration. We studied the body and immunological conditions of a low breeding success population of Brown Skuas (Stercorarius antarcticus lonnbergi) throughout their breeding season. We evaluated changes in their body mass, blood parameters and heterophil:lymphocyte ratios (H:L), at three stages of their breeding period: Incubation, Early rearing and Late rearing. The body mass of female Brown skuas decreased during the initial stages of the breeding period and recovered towards the end. The changes in metabolites suggest a use of protein body reserves but not of lipid reserves. The H:L ratio did not indicate changes in the immune condition of the birds. Our results suggest that this low breeding success population begins its breeding period with a poor body condition that it continues to decline throughout the season. Those poor body conditions added to the need for recovery for the next migration may be the cause of the abandonment of almost totally feathered chicks. Our evidence highlights the importance of food availability, in either the breeding or wintering sites, or both, on the health of the populations of migratory top predators.


Introduction
The body condition of birds is the result of a dynamic interaction between the physiological state, health, energetic requirements and the amount and quality of the food (Murphy 1996). Breeding is a stage in the life cycle with a high degree of stress and energetic cost (Williams 1966) caused, among others, by the physiological cost of egg production and the activities related to caring and feeding the offspring (Kalmbach et al. 2004;Harding et al. 2011;Mitchell et al. 2012). Because of the physiological cost of egg production in females and, due to differences in the rearing activities and investment between sexes (Weimerskirch and Lys 2000;Weimerskirch et al. 2009;Hennicke et al. 2015), the cost of reproduction may differ in males and females. In addition, the movements of birds feedback with the body condition, through both the influence over each other and on the food that parents are able to obtain (Tveraa et al. 1998;Nathan et al. 2008). Therefore, in central place foragers, breeding also influences the body condition through the restriction that imposes to the movements of parents (Bertoldi Carneiro et al. 2015), as well as through the reduction in their time for self-feeding (Weimerskirch 1990).
Energy invested by parents in the breeding activities increases as the offspring grow and increase their energetic requirements (Drent and Daan 1980;. Such investment in offspring has negative effects on the body condition of parents, in a magnitude that increases throughout the breeding period and with the Responsible Editor: Y. Cherel. Reviewed by J-P. Robin and an undisclosed expert.  (Drent and Daan 1980;Weimerskirch 1990;Kalmbach et al. 2004;Mitchell et al. 2012), and may also compromise the future survivorship of parents (Harding et al. 2011). In some species, particularly in long-lived ones, the parents may abandon their offspring when their body conditions drop below a limit that compromises their survivorship and future breeding (Chaurand and Weimerskirch 1994;Gauthier-Clerc et al. 2001;Groscolas et al. 2008).
In migratory birds, the breeding period is preceded and followed by a migration, which often means that in addition to the energetic expenditure related to breeding, the birds must recover their body condition after the previous migration and prepare for the next one. Such migratory movements are energetically demanding, involve the use of body reserves, and require a recovery period both after the migration and during stopovers (Guglielmo et al. 2005;Lyons et al. 2008;Seewagen et al. 2011). The breeding success is influenced by the body conditions of parents both before and during the breeding period. Therefore, migratory species may be under more demanding conditions for reproduction than resident ones (Drent and Daan 1980;Boersma and Ryder 1983;Gustafsson et al. 1994;Selman and Houston 1996;Machado-Filho et al. 2010). Energy destined to supply those demands of the reproductive cycle can be subsidised by food resources obtained during the current breeding period-income breeders-, or by body reserves stored before the breeding period-capital breeders (Stephens et al. 2009) at different distances from the breeding grounds-distant capital or local capital breeders (Klaassen et al. 2006). The stress generated by those different moments in the life cycle as well as by environmental factors stimulates the release of glucocorticoid hormones to the blood (Harvey et al. 1984;Ouyang et al. 2011), which produces a negative effect on the body and immunological conditions (Marra and Holberton 1998;Cirule et al. 2012;Kulaszewicz et al. 2017). Body condition can be evaluated through different aspects such as morphological features as fat and muscle reserves (Carrascal et al. 1998;Rae et al. 2009), hematological and biochemical parameters (Masello and Quillfeldt 2004;Smith et al. 2007;Minias 2014). Fat and muscle stores are energy reserves that can be used in case of an insufficient energetic income from the diet. Fat reserves have fast formation and utilisation according to immediate needs, while protein reserves are more difficult to mobilise and are, therefore, considered as long-term indicators of body conditions (Carrascal et al. 1998). Energy readily available from either the diet, mobilisation of body reserves, or both is found in the blood as glucose. The immunological condition is also an indicator of general body condition (Norris and Evans 2000;Machado-Filho et al. 2010;Kulaszewicz et al. 2017) as the immune activation requires energetic reserves (Moreno-Rueda 2011) and is negatively influenced by food shortage as well as by breeding activities (Norris and Evans 2000;Kulaszewicz et al. 2017).
Potter Peninsula, at King George Island, Antarctica, is a breeding ground for Brown Skuas (Stercorarius antarcticus lonnbergi). The population breeding there probably winters in a wide range in the west-Atlantic, reaching as north as the South Brazil Shelf (Krietsch et al. 2017). Throughout the breeding period, parents must recover their body conditions after the spring migration in order to successfully and to be prepared for the following outbound migration. At the moment of this study, the breeding success of Brown Skuas at this site was very low (Graña Grilli 2014) and some chicks were seen abandoned to starvation even after complete feathering (pers. obs.). Low breeding success in a population can be related with suboptimal conditions of breeders, and the aim of this work was to study the body and immunological conditions of Brown Skuas throughout their breeding season at this locality.

Sampling
The study was carried out in Potter Peninsula, King George Island, Antarctica (62°15′S, 58°40′W) during the austral summers 2012/13 (11 pairs) and 2013-14 (10 pairs). The breeding period was divided into three stages: (1) incubation (In), (2) early rearing (Er) and (3) late rearing (Lr) (Ibañez et al. 2018). For In, the sampling was done 10-13 days after the egg was found. For Er, the trapping was done about 10-13 days after the chick was found (chicks were found no later than 3 days after hatching). The Lr stage started when the chicks were completely feathered but still not flying (about 40 days old). During the season 2012/13, 6 pairs lost their eggs at the In stage while 5 got to hatch chick and only 3 of them got to fledge chicks. During the season 2013-14, 6 of the studied pairs lost their eggs at In, 4 got to hatch chicks and 2 of them fledged chicks (see Graña Grilli 2014).
At each one of the defined stages, both members of each breeding pair were captured for extraction of a 2 ml blood sample from the brachial vein and record of their body mass (± 20 g). One drop of blood was used to make a smear and the remaining was kept in tubes with sodic heparin to avoid coagulation. At laboratory, plasma was obtained by centrifugation at 3500 rpm and kept frozen (− 20 °C) until analysis in Laboratorio de Análisis Clínicos y Bioquímicos (HIGA Gandulfo). Body mass was also recorded at the beginning (S) of incubation, as soon as the eggs were found. The birds were sexed by molecular techniques (Griffiths et al. 1998).
To assess an effect of the location of the nest relative to the main source of food, the proximity of the nest of each studied skua pair to the penguin colonies also was measured and compared among birds that lost their nest during the In stage and those that succeed to hatch chicks. Because the sampled birds also were tagged with GPS loggers and the home range was calculated for each bird, the proximity was measured using ArcGIS 10.1, as the lineal distance from the nest to the nearest border of the closest penguin sub-colony included in the kernel density estimation 95% of each bird during the In stage (unpub. data).

Hematological parameters
The concentration of metabolites associated with metabolic pathways of protein and lipids as well as that of glucose as a source of readily available energy was determined in plasma. To this end, enzymatic assays were carried out on simple samples using a high precision automatic analyser Architect i2000 (Abbott Diagnostics, Illinois, USA). Total proteins, urea and uric acid were the parameters related to the metabolic path of proteins analysed; and triglycerides and non-esterified fatty acids (NEFA, Randox, USA) were those related to the metabolic path of lipids.
Urea and uric acid are products of protein catabolism and can result from either the use of internal reserves or the metabolism of food proteins (Cherel and Le Maho 1985;Alonso-Alvarez et al. 2002). On the other hand, total proteins are positively related to nutritional conditions (Gustafsson et al. 1994). Therefore, the coupled analysis of these three parameters allows determining whether the bird is using its own protein body reserves or getting proteins from the diet. On the other hand, fat ingested in food as triglycerides (Guglielmo et al. 2005) is split up in the gut producing NEFA and monoacilglycerol that cross the gut wall and in the blood stream re-synthesise triglycerides. Contrariwise, triglycerides contained in the cells of fat stores are hydrolysed producing NEFA that are released to the blood stream and transported to the muscle cells (Price 2010). Therefore, the concentrations of triglycerides and NEFA in plasma are indicators of fat stores formation and utilisation (Guglielmo et al. 2005;Price 2010).

Leucocyte counts
Blood smears were fixed by immersion in methanol for 3 min and air dried and then stained with Wright colourant. Stained smears were examined in microscope (1000×) under immersion oil and a sample of 100 leucocytes was counted and classified as lymphocytes, heterophils, eosinophils, basophils and monocytes (Campbell 1994). From those numbers, the ratio heterophil:lymphocyte (H:L) was calculated (Davies et al. 2008;Lebigre et al. 2012).
The H:L ratio obtained from the leucocitary profile is an indicator of the health status of birds (Cirule et al. 2012). Glucocorticoids induce the release of heterophils into the bloodstream and preclude them of leaving it and, at the same time, induce the adhesion of lymphocytes to the blood vessels wall and then their pass to other tissues (Davies et al. 2008;Cirule et al. 2012). For this reason, the H:L ratio increases under stressing conditions, such as during breeding (Norris and Evans 2000;Antonio-Garcia et al. 2010) and low food availability (Machado-Filho et al. 2010).

Data analysis
We used GLMM to analyse the variation of each variable according to the stage of the breeding period and the sex of the birds. The identity of the bird and the season were considered random effects. A different effect of the breeding stage according to the sex of the bird was also studied. We used the lme function in the nlme package of the software R (R Core Team. 2014). The influence of the distance between the nest and the penguin colony was analysed through a GLMM for binomial distribution using the identity of the pair as the random effect in the glmer function in the lme4 package (R Core Team 2014). Results are presented as mean ± SD and were considered significant at alpha values ≤ 0.05.

Results
Laying dates were similar in both seasons, taking place between 23 November and 22 December in the season 2012/13 and between 20 November and 28 December in the season 2013/14. The date of laying was not related to the body mass of males (t = 1.70, P = 0.13) or females (t = 0.40, P = 0.70) (Fig. 1). The mean distance between the nest and the nearest border of a penguin colony included in the home range of the bird was 7.42 m (± 4.82 m) and did not change between birds that lost their eggs during the In stage and those that succeeded to hatch chicks (Z = 1.53, P = 0.12). Because of the high occurrence of nesting failure the number of birds sampled at each stage decreased severely, and 5 birds for which the technique did not get to determine the sex were excluded from the body mass analysis (S: n = 33, In: n = 30, Er: n = 15, Lr: n = 10). Females were heavier than males during the entire breeding period (est = − 123.5, t = − 3.35, P < 0.01), and, in both years, heavier females tended to pair with heavier males despite the Pearson's correlation was significant only for the first season (2012/13: r 2 = 0.48, P = 0.04; 2013/14: r 2 = 0.18, P = 0.11) (Fig. 2). The body mass at the In stage was not different between those birds that lost their nest during the In stage and those that succeeded to hatch chicks neither for males (t = 0.70, P = 0.51) nor females (t = − 0.22, P = 0.83) (Fig. 3). None of the blood parameters was related to the probability of incubation success.

Hematological parameters
The sex of the bird and the stage of the breeding period did not interact for any of the analysed hematological parameters and none was affected by the sex. Therefore, both sexes were pooled and non-sexed birds were included (In: n = 34, Er: n = 16, Lr: n = 8).

Leukocyte count
In all the blood stains, there was a high predominance of heterophils and lymphocytes, with a high variability among them and an almost complete absence of basophils ( Table 2). The H:L ratio did not show changes throughout the stages of the breeding period (Er: est = − 2.44, t = − 1.17, P = 0.25; Lr: est = − 3.42, t = − 1.45, P = 0.16); however, the H:L ratio was higher in females than in males (est = − 3.74, t = − 2.21, P = 0.04) (Fig. 5).

Discussion
Our work suggests that the low breeding success in the population studied coincides with poor and declining body condition of the breeding birds throughout their breeding period. This is evidenced by the trends found in the concentration of blood metabolites suggesting a decline in the ingestion of lipids which is not compensated with the catabolism of body fat reserves, but possibly, with the use of protein body reserves. The removal of protein stores instead of the more easily available and calorically denser fat stores suggests that this population starts the breeding period with a general low body condition without the extra availability of fat reserves to cover the energetic requirements. The removal of protein stores from muscle is generally reflected by a low concentration in plasma of total proteins and a high concentration of the products of protein catabolism, uric acid and urea (Alonso-Alvarez et al. 2002;Ferrer et al. 2013). Contrarily, high plasma concentrations of all of these metabolites would indicate protein ingestion. Despite we did not find changes in the concentration of plasmatic uric acid, the decrease in total proteins and the increase in urea concentration throughout the breeding period suggest the use of protein stores by breeding Brown skuas (Alonso-Alvarez et al. 2002;Ferrer et al. 2013). On the other hand, the use of fat stores is indicated by a low plasma concentration of triglycerides and high concentration of NEFA (Cherel et al. 1988;Price 2010), whereas fat-store formation is reflected by high concentration of triglycerides (Smith et al. 2007;Antonio-Garcia et al. 2010;Price 2010). In the studied Brown skuas, there was a decrease in the concentration of triglycerides towards the end of the breeding season, suggesting a decrease in lipid ingestion. However, we did not find an increase in NEFA concentration-instead, we found a decreasing trend-; therefore, there is no evidence of a consumption of body fat stores during the breeding period.
Lipid reserves are used more than protein reserves as a source of energy due to their higher economy for storing and energetic content (McWilliams et al. 2004). The decrease in triglycerides concentration suggests that in the course of the breeding period the lipid intake from the diet decreases. This, added to the lack of evidence for the use of fat stores suggests that Brown skuas may already be in energetic deficit at the beginning of the breeding period and that subsequent reductions in the energy obtained from the diet would be compensated by the consumption of protein stores. The stability found in the plasma level of glucose is normal, even in fasting birds (Cherel et al. 1988), and suggests the idea that possibly protein degradation (for energy obtaining and gluconeogenesis) coupled to the glucose directly obtained from the diet allow to keep the level of energy readily available for tissues. Declines in body conditions, particularly after chicks hatching, have been found in other bird species (Croll et al. 1991;Weimerskirch and Lys 2000). However, results indicating that protein reserves are probably being used as a source of energy, despite do not indicate that fat reserves have been completely depleted, can be compared to the condition of birds in an advanced fasting state (Cherel et al. 1988).
Body mass depends not only on lean and fat content but also on the hydric balance. Body water content has an inverse relationship with fat content (Ellis and Jehl 1991) and the use of protein reserves can be mediated by an increase in the secretion of corticosterone under stressing conditions (Hayashi et al. 1994), which also has antidiuretic effects (Holmes et al. 1983). Therefore, changes in corticosterone secretion may induce mechanisms which can explain the increase in body mass recorded towards the end of the breeding period, despite the decrease in plasma protein and triglycerides. The decrease found in total protein in plasma would be due to the decrease found in the concentration of globulins (Ibañez et al. 2018), while albumin-protein that in contrast to globulins is related to body reserves (Lumeij 1987)-increased its concentration in plasma in the last stage of the breeding period (Ibañez et al. 2018). This suggests that Brown skuas may be finishing their catabolic state towards the end of the breeding period, leading to the observed recovery of body mass.
High variability in the H:L ratio is common among birds (D' Amico et al. 2010;Lebigre et al. 2012). This parameter did not indicate changes in the stress level in the different stages of the breeding period of Brown skuas. However, a decrease in γ-globulin and IgY indicated a decline in the immunological conditions of these same birds throughout their breeding period (Ibañez et al. 2018), probably related to the effort experienced by Brown skuas during their breeding period. This difference in results may indicate that the leucocyte proportion is more resistant to show the effects of an immunological decline than other portions of the immune system. However, the immunological condition is poorer in females than in males, which can be related to a higher investment in breeding, which is not reflected in other body parameters.
In contrast to previous findings (Phillips et al. 2002), in the studied population of Brown skuas, heavier females tended to breed with heavier males. But in the same way than a close population (Carneiro et al. 2015), the studied one did not show a relationship between the body mass of males or females and the hatching success probability. The fixed inversion hypothesis predicts that long-lived birds would maintain their physiological conditions fixed by passing on the excessive breeding costs to their partners or their offspring, minimising on themselves the impact of, for instance, reduced food availability or accessibility (Croll et al. 1991;Gauthier-Clerc et al. 2001). It is possible that Page 7 of 9 142 Brown skuas started their breeding activities, after migration, in a body condition close to a minimum tolerable limit. This poor initial condition added to the need for the stocking of fat stores ready to be used in the following outbound migration (Cherel et al. 1988;Price 2010;Mitchell et al. 2012) may determine that parents prioritise their own energetic recovery and the abandonment of the offspring if the inversion in rearing chicks leads them to an even greater body condition decline. Such prioritisation of the parents' body conditions recovery preparing for the next migration is suggested in our results by the increase in body mass in the last stage of the breeding period.
That general poor body condition can explain the death of almost totally feathered chicks, without evidence of either injuries nor strong climatic events, as well as the low number of breeding attempts and hatching success in this population (Graña Grilli 2014), which is repeated in recent years in another close location (Krietsch et al. 2016). This situation indicates changing conditions from those reported in previous studies for the same area, as not only a much smaller number of Brown skuas bred but also South polar skuaswhich before outnumbered Brown skuas-did not breed during the study period (Ritz et al. 2006;Hahn et al. 2007). In addition to higher numbers of breeding attempts, also higher breeding success was reported in the area 10 years before this study-0.17 and 0.57 in 2002/03 and 2003/04, respectively-even for birds with a lower mean body mass than that of the birds studied in this work (Ritz et al. 2006;Hahn et al. 2007).
We cannot attribute this high breeding failure to the location of nests, as most of them were inland relative to the penguin rockeries. Therefore, egg trampling by penguins walking between the nest and the sea cannot be the cause of the high nest failure at the incubation stage as found in other skuas populations (Hagelin and Miller 1997). Our results failed also in finding a relationship between laying date and breeding success as previous work did (Pezzo et al. 2001). In the same way, we did not find a pattern of higher breeding failure in nests farther away from penguin colonies (Pezzo et al. 2001). However, a general shortage on food resources, poor body condition, or both, may induce longer foraging time reducing nest attendance and, therefore, increasing nest predation, probably even intraspecific predation. This situation may be caused by a low amount or quality of food in either the breeding or the wintering sites that this top predator is not able to overcome. It is important to study whether this situation is also taking place at other breeding sites, in other species breeding in the same region, and its progression over the time, in order to analyse its causes and to take actions at the species or the ecosystem level.