Demography of the Neotropical austral migrant Elaenia albiceps chilensis (Tyrannidae) on its breeding grounds: climate and food off er eff ects

Th e White-crested Elaenia (Elaenia albiceps chilensis) is a Neotropical austral migrant that breeds in Patagonian forests and overwinter in South America tropics. We investigated White-crested Elaenia demography and the environmental conditions the specie is subjected at the northwest of Patagonia as a fi rst attempt to elucidate its population dynamic. We studied the species abundance, individuals' body condition, survival, productivity of young and site fi delity according to changes in rainfall patterns and food supply, during three breeding seasons (2009 to 2012). We color banded 361 adult and juvenile White-crested Elaenias. During the fi rst studied season, precipitations were remarkably above the mean and ripe fruit supply was lower than in the other seasons, whereas the abundance of arthropods remained similar along the study. Th e species abundance decreased 50% across the study, while adult apparent survival rate, the number of breeding females, young productivity and site fi delity were 50% lower in the fi rst season in comparison to the following seasons. Females presented lower body condition at arrival in the fi rst two studied periods. Site fi delity diminished after unfavorable breeding season (i.e. low young productivity and fruit supply), especially for males, although they were more faithful than females and had a slightly higher apparent survival rate. Our results suggest that White-crested Elaenia demography is aff ected by changes in its main food resource supply (fl eshy fruits) during the breeding season, indicating that this part of the annual cycle could be fundamental for its population dynamic. KEY-WORDS: body condition, breeding season, frugivory, Patagonian forest, population dynamic, site fi delity, survival, Whitecrested Elaenia. greater parental care) that is compensated with higher adult survival (see review in Martin 1996 & Russell et al. 2004). Th roughout South America information is contrasting, with some studies indicating a higher adult survival from tropical to temperate latitudes (Martin 1996) and others pointing to a lack of such trend (Karr et al. 1990, França & Marini 2010), or even to variations among phylogeny and regions for tropical birds (Blake & Loiselle 2013). Although this trend has been mainly suggested for resident species, it also applies to species that perform regional or long distance migrations (Martin 1996, Russell et al. 2004, Blake & Loiselle 2013). Among behavioral features site fi delity is one that can be aff ected by species life history as well as by environmental quality, conditioning migratory birds' abundance (Holmes & Sherry 1992). Site fi delity will be favorable if the area allows individuals to pair a previous successful partner or be more competitive because of prior territory ownership, as well as if they are able to Demography of the migratory White-crested Elaenia Presti et al. 241 Revista Brasileira de Ornitologia 26(4): 2018 identify the distribution and availability of resources as food, refugees or nestling sites (Newton 2008). Contrary, site fi delity will be disadvantageous for migrants if the site diminishes in quality or represents an unpredictable habitat (Switzer 1993). In addition, site fi delity can diff er according to sex, age and competitive condition, generating diff erences in dispersal and habitat use patterns (Latta & Faaborg 2002). Environmental factors are relevant causes of demographic variation in resident and migratory bird populations. Food limitation can be a fundamental factor aff ecting body condition and, consequently, breeding success and survival (Martin 1987). Factors such as precipitation, habitat quality and resource spatial distribution, will aff ect food abundance and availability, aff ecting diff erently the population (Wilson et al. 2011). Th e eff ects of those environmental factors on the population can be quantifi ed by measuring diff erent external variables that indicates habitat quality (Benson & Bednarz 2010) as well as estimating the individuals' body condition by their body mass or fat deposits (Labocha & Hayes 2012). On the other hand, unusual weather conditions, sometimes related with periodic climatic events, aff ect the sizes of bird populations (Sillett et al. 2000). Some studies suggest that southern temperate birds' survival either is reduced after periods with unusual precipitation, rainy (Th omson & Estades 2012) or dry (Sagario 2010). Th erefore, it is expected that climate variations also aff ect migratory birds' demography at southern South America. All these theories about the factors controlling population dynamics remain slightly evaluated in passerine birds of South America, been mainly studied in tropical species (Karr et al. 1990, Martin et al. 2000, Parker et al. 2006, Blake & Loiselle 2013). Th us, there is still little information for temperate latitudes of South America and migratory birds of the Neotropical austral system, despite some advances for diff erent passerine families in survival (Willson & Pendleton 2008, Sagario 2010, Th omson & Estades 2012), site fi delity (Jahn et al. 2009) and life history traits (e.g. parental care, Llambías et al. 2015). Th e White-crested Elaenia (Elaenia albiceps chilensis) is the most abundant Neotropical austral migrant passerine at the Patagonian forest (Grigera et al. 1994, Ippi et al. 2009). Th is tyrant fl ycatcher arrives up to mid-October to breed and start its migration to the tropics between mid-February and March (Fitzpatrick 2004, Bravo et al. 2017). It is a frugivorous-insectivorous species with an important role as the main disperser of seed of endemic plants of the Patagonian forest (Armesto et al. 1987, Amico & Aizen 2005) and as a controller of herbivorous insects of Nothofagus trees (Mazia et al. 2009). Th ere are studies of White-crested Elaenia abundance (Brown et al. 2007), diet (Grigera 1982, Brown et al. 2007), foraging behavior (Chust et al. 2012, Cueto et al. 2016a), functional role (Mazia et al. 2009, Cavallero et al. 2013, Bravo et al. 2015), and migration routes (Marini & Cavalcanti 1990, Capllonch et al. 2011, Cueto et al. 2016b, Bravo et al. 2017), but its survival has never been studied, as well as the relationship between its demographic parameters with environmental factors. Th e aim of our study was to analyze the demography of the White-crested Elaenia through the analysis of changes in abundance, productivity of young, survival, site fi delity and body condition, related with food off er and precipitation variability during the breeding season. We aim to improve our knowledge about the factors aff ecting its population at temperate latitudes of South America.


INTRODUCTION
Th e study of demography and the factors that regulate populations are among the most important approaches to understand population trends. Birds demographic (abundance, survival and reproductive success) and behavioral features (site fi delity, space use and dispersal) are regulated by both intrinsic and extrinsic factors (e.g. physical and nutritional traits of individuals and food or weather variations, respectively; Martin 1996, Newton 1998. Th ese factors interact with each other in positive and negative ways, and their implications in a population are expected to vary according to the species life history (Newton 1998).
Among demographic features, diff erences in life history between North and South Hemisphere birds suggest that Southern hemisphere species have lower reproductive success (as indicated by smaller clutch sizes, higher nest predation rates, more nesting attempts and

Demography of the Neotropical austral migrant Elaenia albiceps chilensis (Tyrannidae) on its breeding grounds: climate and food off er eff ects
greater parental care) that is compensated with higher adult survival (see review in Martin 1996& Russell et al. 2004. Th roughout South America information is contrasting, with some studies indicating a higher adult survival from tropical to temperate latitudes (Martin 1996) and others pointing to a lack of such trend (Karr et al. 1990, França & Marini 2010, or even to variations among phylogeny and regions for tropical birds (Blake & Loiselle 2013). Although this trend has been mainly suggested for resident species, it also applies to species that perform regional or long distance migrations (Martin 1996, Russell et al. 2004, Blake & Loiselle 2013. Among behavioral features site fi delity is one that can be aff ected by species life history as well as by environmental quality, conditioning migratory birds' abundance (Holmes & Sherry 1992). Site fi delity will be favorable if the area allows individuals to pair a previous successful partner or be more competitive because of prior territory ownership, as well as if they are able to identify the distribution and availability of resources as food, refugees or nestling sites (Newton 2008). Contrary, site fi delity will be disadvantageous for migrants if the site diminishes in quality or represents an unpredictable habitat (Switzer 1993). In addition, site fi delity can diff er according to sex, age and competitive condition, generating diff erences in dispersal and habitat use patterns (Latta & Faaborg 2002).
Environmental factors are relevant causes of demographic variation in resident and migratory bird populations. Food limitation can be a fundamental factor aff ecting body condition and, consequently, breeding success and survival (Martin 1987). Factors such as precipitation, habitat quality and resource spatial distribution, will aff ect food abundance and availability, aff ecting diff erently the population (Wilson et al. 2011). Th e eff ects of those environmental factors on the population can be quantifi ed by measuring diff erent external variables that indicates habitat quality (Benson & Bednarz 2010) as well as estimating the individuals' body condition by their body mass or fat deposits (Labocha & Hayes 2012). On the other hand, unusual weather conditions, sometimes related with periodic climatic events, aff ect the sizes of bird populations (Sillett et al. 2000). Some studies suggest that southern temperate birds' survival either is reduced after periods with unusual precipitation, rainy (Th omson & Estades 2012) or dry (Sagario 2010). Th erefore, it is expected that climate variations also aff ect migratory birds' demography at southern South America.
All these theories about the factors controlling population dynamics remain slightly evaluated in passerine birds of South America, been mainly studied in tropical species (Karr et al. 1990, Martin et al. 2000, Parker et al. 2006, Blake & Loiselle 2013. Th us, there is still little information for temperate latitudes of South America and migratory birds of the Neotropical austral system, despite some advances for diff erent passerine families in survival (Willson & Pendleton 2008, Sagario 2010, Th omson & Estades 2012, site fi delity (Jahn et al. 2009) and life history traits (e.g. parental care, Llambías et al. 2015).
Th e White-crested Elaenia (Elaenia albiceps chilensis) is the most abundant Neotropical austral migrant passerine at the Patagonian forest (Grigera et al. 1994, Ippi et al. 2009). Th is tyrant fl ycatcher arrives up to mid-October to breed and start its migration to the tropics between mid-February and March (Fitzpatrick 2004, Bravo et al. 2017. It is a frugivorous-insectivorous species with an important role as the main disperser of seed of endemic plants of the Patagonian forest (Armesto et al. 1987, Amico & Aizen 2005 and as a controller of herbivorous insects of Nothofagus trees (Mazia et al. 2009). Th ere are studies of White-crested Elaenia abundance (Brown et al. 2007), diet (Grigera 1982, Brown et al. 2007, foraging behavior (Chust et al. 2012, Cueto et al. 2016a, functional role (Mazia et al. 2009, Cavallero et al. 2013, and migration routes (Marini & Cavalcanti 1990, Capllonch et al. 2011, Cueto et al. 2016b, Bravo et al. 2017), but its survival has never been studied, as well as the relationship between its demographic parameters with environmental factors.
Th e aim of our study was to analyze the demography of the White-crested Elaenia through the analysis of changes in abundance, productivity of young, survival, site fi delity and body condition, related with food off er and precipitation variability during the breeding season. We aim to improve our knowledge about the factors aff ecting its population at temperate latitudes of South America.

Study site
We conducted our study from October to March of three consecutive breeding seasons (2009/10, 2010/11 and 2011/12) in a Patagonian mixed forest of Nothofagus dombeyi (Coihue) and Austrocedrus chilensis (Ciprés de la Cordillera) at Steff en Lake (41 o 30'S; 71 o 35'W, 550 m a.s.l). Th is is a representative and well protected area of the temperate forest of southern South America inside Nahuel Huapi National Park, at the northwest of Argentinean Patagonia (Mermoz et al. 2009). At the study site, the forest is characterized by a canopy above 20 m of height and an understory (< 10 m) with fl eshy fruit shrubs, mainly Aristotelia chilensis (Maqui) and Schinus patagonicus (Laura).
Th e climate is cold-temperate with precipitations concentrated in autumn and winter (April-September). Annual mean precipitation in the study area is 1264 mm (for 1993-2012 period) and annual mean temperature is 9.1 o C (for 2000-2012 period), with a maximum mean temperature of 15 o C (January) and a minimum of 3.5 o C (July) (data from Station N o 2300 "Lago Steff en", Red Hidrológica Nacional, Subsecretaría de Recursos Hídricos de la Nación).
We established two permanent plots on the northern margin of the lake, separated 1000 m by a small bay. Th e plots area were 4 ha (Plot 1) and 7 ha (Plot 2) and we gridded them in 25 x 25 m squares, for bird and food resource surveys (68 squares for P1 and 109 squares for P2). Both plots had the same vegetation composition and structure, representing a well-preserved mixed forest with three types of patches: open areas, young forest and old forest (see Bravo et al. 2015), therefore we pooled the data for the analysis.

Bird surveys
Scientifi c nomenclature and common name follow the recommendation by the South American Checklist Committee (Remsen-Jr. et al. 2018). To get demographic data of the White-crested Elaenia we used capture-markrecapture method. We captured birds using 24 mist nets of 12 m × 3 m, 38 mm mesh size (12 per plot). Nets were distributed along the area separated at least 50 m from each other, arranged in fi xed places at every capture session to register the mist net of every captured and recaptured bird in order to study site fi delity. During every breeding season, nets were opened for two consecutive days per month in both plots (in similar month dates along every session), since sunrise and for at least fi ve hours.
Every captured bird was banded with numbered aluminum and a unique combination of color bands. We weighted birds to the nearest 0.1 g with a 30 g spring scale in every capture occasion and measured their tarsus length to estimate their arrival and departure body condition (details below). We aged individuals by the presence (adult) or lack (juvenile) of the characteristic white crown (Schulenberg 2009). To determine the sex and reproductive status of birds, we registered the development of the incubation patch (females) and cloacal protuberance (males) following Ralph et al. (1993), and used the diff erences in wing and tail length between females and males for sexing non-breeding birds following Cueto et al. (2015). When an individual did not match with those criteria of sexing it was classifi ed as "unsexed adult".
From November to February, after net sampling, we surveyed each plot to relocate those banded birds that remained at the study site to complete their capture history for survival and site fi delity estimates. Searches were carried out systematically by walking through the study area during fi ve hours after sunrise and three hours before sunset, at least once per day for a mean of 10 days or until no new banded birds were seen after searching for 10 h. We registered the identity of each banded bird found and its location on the grid. Th e position of every relocated bird was recorded estimating the distance and angle respecting a reference point of the grid, using a compass. We also recorded eventual encounters while collecting environmental data.

Arthropods and fruit abundance
Arthropod abundance was estimated using blue sticky traps (Young 2005). We used blue because it is neutral in terms of attracting or repelling insects (Ausden & Drake 2006). Each trap was of 10 × 20 cm, attached inside a plastic transparent cylinder of 15 cm diameter to prevent that leaves, branches and rain aff ect its capture effi ciency. We used 28 blue sticky traps (14 per plot) that were randomly placed, tided horizontally on an exposed branch of a shrub or tree, two meters above the ground, operated for an average of 15 days per month depending on weather conditions. After the survey, we counted the number of captured individuals at each trap.

Precipitation patterns
To assess the potential eff ect of climate on the Whitecrested Elaenia's demography and their food abundance, we analyzed the variations on precipitation at the study site. Precipitation data were available from the weather station N° 2300 -"Lago Steff en" (Red Hidrológica Nacional, Subsecretaría de Recursos Hídricos de la Nación), for the period 1993-2012. We considered wet or dry months to be those with monthly rainfall above or below one standard deviation of the average monthly rainfall for a 20-years period of climatic data (Ropelewski & Folland 2000).

Data analysis
Bird abundance and productivity of young: In order to compare bird abundance among breeding seasons, we calculated the season capture rate for each bird category (female and male in breeding and non-breeding conditions; and juveniles) as the total number of captures per 100 h of net eff ort. We used the data from November to February, and excluded from the analysis birds captured in October and March that did not remain in the study site in the summer (i.e. those individuals that were just passing through). Juveniles capture rate in every season was used as a measure of productivity of young (du Feu & McMeeking 1991).
Relative body condition: We generated an index of body condition using the residuals of the linear regression between every bird tarsus length and their body mass (Labocha & Hayes 2012). Positive residuals indicated that birds were in relatively good condition, i.e. they were heavy for their size, and negative residuals indicated relatively poorer condition. We used a two-way ANOVA to examine the diff erences of body condition between breeding seasons, arrival from spring migration (October-November) and departure from the breeding study site (February-March) of females and males. Given the unequal sample sizes, we used Type III method to for all returning White-crested Elaenias individuals and by sexes, through a diff erence of proportions test (Agresti 2002). We used a Bonferroni correction to adjust  when more than one test was performed on related data. We considered the data from November to March for the analysis and used the free software InfoStat (2014).
We estimated the displacement distance between the fi rst capture and the fi rst recapture or relocation of individuals in every breeding season. For this, the position of the mist nets and relocation position of every bird were georeferenced in a metric coordinate system, using the free software Quantum Gis 1.8.0-Lisboa (QGIS Development Team, 2012). We performed a T test (Agresti 2002) to compare the displacement distance between sexes and between returning periods: 2009/10 to 2010/11 versus 2010/11 to 2011/12, and 2009/10 to 2010/11 versus 2009/10 to 2011/12 displacement.
Arthropods and fruit variations: Abundance of arthropods was assessed by calculating a relative abundance index per trap (RAI = number of caught individuals / number of trapping days) and then estimated the average of all traps for the season. We evaluated diff erences in arthropod abundance among seasons by using one factor Analysis of Variance with Repeated Measures (Winer 1971), because we counted arthropods in the same points at every season. We used the Tukey multiple comparison test for comparing means when season factor was signifi cant (Winer 1971). We tested equality and symmetry of covariance matrices (Winer 1971). Data violated the symmetry assumption, so we used the Greenhouse-Geiser procedure (Lower Bound Epsilon correction = 0.5) to adjust the degrees of freedom of within-subject eff ects (Winer 1971). For ripe fruit off er, we calculated the proportion of individuals that off ered no ripe fruits (index 0), less than 50 ripe fruits (indexes 1 and 2), and those that off ered more than 50 ripe fruits (indexes 3 to 5). For each category, we evaluated the diff erence in abundance of ripe fruit between seasons using a test for diff erence of proportions (Zar 2010). ANOVAs, the Tukey comparisons and all assumption tests were performed on the free software InfoStat (2014).

RESULTS
Bird abundance and productivity of young: We banded 361 adult and juvenile White-crested Elaenias and recaptured 44 individuals during 5394 mist netting hours (Table  1). Complementary we relocated 25 adults that were not recaptured in nets but detected by active search. We did not recapture or relocate any marked juvenile after its hatching year. Male capture during October-November represented 69% of all captures. Th e latest captured individuals towards the end of the breeding compute the sum of squares (Shaw & Mitchell-Olds 1993). Comparisons of treatments after ANOVA were performed using the Kramer modifi cation of Tukey's test (Day & Quinn 1989) using the free software InfoStat (2014).
Survival: We estimated apparent annual survival rate between two breeding season intervals (2009/10 to 2010/11 and 2010/11 to 2011/12) as the probability of an individual survive and return to the same breeding area. Th is is an approximation of the real survival rate, because it does not account for permanent or temporal emigration (Lebreton et al. 1992). We used captures, recaptures and relocation data of 269 birds, considering only adult birds as we never recaptured or relocated any marked juvenile. We excluded from the analysis the individuals that were captured just once and were not relocated during the season (i.e., likely transient individuals). We fi tted Cormack-Jolly-Seber (CJS) markrecapture models to assess apparent annual survival rate and capture probability (Lebreton et al. 1992). We confi rmed models goodness-of-fi t and lack of overdispersion using the median c procedure (Burnham & Anderson 2002). We modeled apparent survival probability () as constant, varying in time (years) and between sexes, and examined the infl uence of these two covariates together, as additive and interactive eff ects. We considered capture-recapture-relocation probability (p) constant in time and equal among sexes. We based our decision on sampling eff ort as we systematically sampled the plots with mist nets and active search since the arrival of the birds till their departure, compensating possible diff erences in detection caused by changes in females and males behavior during their breeding period (as incubation and territorial defense). Th e relative importance of each model was compared using: Akaike›s Information Criterion adjusted for small sample size (AICc), the diff erence of AICc between models (ΔAICc), their weights (wi) and deviance (Burnham & Anderson 2002). Models with a ΔAICc < 2 of the top model were considered to have equal support, so we model-averaged survival estimates (Burnham & Anderson 2002). Model averaging generated an annual mean apparent survival rate for each sex and breeding period, with their standard errors and confi dence intervals. We performed this analysis in R (version 3.1.1 2014, Th e R Foundation for Statistical Computing) using the RMark package (version 2.1.8; Laake & Rexstad 2008) that provides an interface to the software MARK (version 7.2; White & Burnham 1999).
Site fi delity: We analyzed it as the percentage of adult birds banded in a previous fi eld season that were recaptured or relocated in a subsequent fi eld season. We compared the diff erences in annual return rates between breeding seasons (2009/10 to 2010/11 vs. 2010/11 to 2011/12) ^ season (March) were mainly juveniles. Captures of new individuals decreased almost 50% between the fi rst and last studied seasons (Table 1). During 2009/10 we registered the lowest percentage of breeding females (28%) and the lowest productivity of young, while for the following seasons the percentage of breeding females exceeded that of non-breeders (68% for 2010/11 and 73% for 2011/12) and the productivity of young increased (Fig. 1).
Survival: Th e best fi tted model indicated a variation in apparent annual survival rate between breeding seasons, being 50% lower for 2009/10-2010/11 interval in relation to 2010/11-2011/12 (0.41 ± 0.07 vs. 0.80 ± 0.15). Models that considered survival vary among seasons and sexes had a great support following AICc and  ΔAICc criterion (Table 2). Results from model averaging revealed that the mean apparent annual survival rate was equally low for females and males for the fi rst interval but males tended to increase its survival more than females towards the following breeding season (Table 3).
Site fi delity: From 231 banded and sexed adult White-crested Elaenias, 69 individuals returned the following breeding seasons. Within those faithful birds, 14 individuals returned all the studied seasons and 12 skipped from the fi rst season (2009/10) to the third (2011/12), not being recapture or relocated during the second season (2010/11). Annual return rate varied signifi cantly between breeding seasons, being higher for 2010/11 to 2011/12 period (Zc = -0.13, P = 0.028) in Table 2. Models results for the apparent annual survival rate () and recapture (p) probability of the migratory Whitecrested Elaenia in three breeding seasons at Nahuel Huapi National Park, Patagonia, Argentina (n = 269). Models are arranged by Akaike's Information Criterion adjusted for sample size (AICc). (.) Parameter is constant; k: number of parameters; wi: model weight.  (Table 4). Considering the birds that returned across the entire study and those that skipped the second studied season, site fi delity of 2011/12 season increased from 33% to 49% (54 individuals), increasing the above-mentioned diff erence with the fi rst interval. Fidelity did not vary signifi cantly between sexes (Zc = -0.11, P = 0.084), but males tended to return in greater proportion than females towards 2011/12 (Table  4). With respect to territory fi delity, only three males moved between plots along the study, less than 1500 m. Th e mean displacement between seasons was: 94 ± 100 m for 2009/10 to 2010/11 seasons, 85 ± 71 m between 2010/11 to 2011/12 , and 102 ± 84 m for birds that were banded in 2009/10 and only relocated in 2011/12. Th ere were no signifi cant diff erences between the mean displacement of the fi rst and second period (t = 0.42, P = 0.676), neither between the fi rst and the third (t = -0.29, P = 0.772). Females tended to displace more than males when they returned (n = 30, 118 m ± 86 m vs. n = 54, 76 ± 78 m on average, respectively. t = 2.25, P = 0.027).

Model
Precipitation variability: Th e fi rst study year was an unusual rainy year, with an important input during the October-March period (638 mm) that was almost two standard deviations above the historical record (371 mm). At the following breeding seasons, precipitation was close to average (347 and 390 mm, respectively; Fig. 3).
Arthropods and fruit off er. Arthropod abundance was similar among seasons (F 2, 27 = 3.027; P = 0.093; Fig. 4) but ripe fruit production varied notably between seasons. Th e proportion of plants without ripe fruits was higher in the fi rst study season than the two following breeding seasons (Fig. 5, test for more than two proportions, Zc = 22.86, P = 0.003), while plants with more than 50 ripe fruits were higher in the second and third seasons (Zc = 6.94, P = 0.031; Fig. 5).

DISCUSSION
Our results suggest that the demography of migratory White-crested Elaenia is aff ected by environmental changes on its breeding grounds. Th e three seasons of our study had contrasting climatic conditions, with the fi rst season having a remarkable increase in rainfall respecting the historical average for the study area. Th is climatic variation was refl ected in changes in the supply of ripe fruits of Aristotelia chilensis, but not in the abundance of understory arthropods. Th e increase in rainfall during the fi rst season resulted in few days of full sun, likely leading to the inability of Aristotelia chilensis shrubs to produce ripe fruits, a relationship widely documented for others species (Aalders et al. 1969, Patten & Proebsting 1986, Spayd et al. 2002. During that fi rst season a lower proportion of females became reproductive (28%), which was refl ected in a low productivity of young. Besides, birds' relative body condition tended to be lower than the following seasons and both survival and site fi delity diminished by 50% after that period. Table 4. Site fi delity (annual return rate) of the migratory White-crested Elaenia in three breeding seasons at Nahuel Huapi National Park, Patagonia, Argentina. Estimation is presented as the percentage and number of adult birds that returned from a previous banding season, by sex-class and for the complete studied population. Total estimation includes birds captured at 2009/10 that returned at 2011/12, not being recaptured or relocated during 2010/11. Returns include both recaptures in nets and relocations of color-banded individuals.    Previous studies have shown the impact of food limitation in reproductive status and success of resident and migratory birds, as birds adapt their reproductive cycle according to food availability (Martin 1987, Sherry & Holmes 1995, Marshall et al. 2002, Jahn et al. 2014. For example, females of Parus major are able to modify their laying date so it coincides with periods of higher caterpillar abundance (their main food item) and higher temperatures, promoting nesting success (van Noordwijk et al. 1995). With respect to the infl uence of arthropods supply on White-crested Elaenias demography, we found that slightly variations in foliage arthropods abundance did not aff ect bird's reproductive status and productivity of young, neither body condition or survival. Similar results were reported for Nearctic-Neotropical migrants (Vireo olivaceus, Marshall et al. 2002, andSeiurus aurocapilla, Vernouillet et al. 2014) and for White-crested Elaenias breeding at central Chile (González et al. 2004). Consistent with Cueto et al. (2016a) that found changes in the species foraging behaviour related with fruit off er variations at the same study site, we conclude that fruit is the main food resource aff ecting White-crested Elaenia reproductive success and demography during the breeding period. Our work represents a primary approximation of the eff ect of food limitation in White-crested Elaenia demography and reinforces the importance of more detailed study for a complete understanding of the role of food resources on the species population dynamic.

Banding and returning seasons
Decrease in new individuals capture rate along our study years could be considered as an indirect measure of low immigration rate, caused by a poor infl ux of new birds to the population (Newton 1998). Th is could be due to events of lower reproductive success (as in the fi rst studied season), or higher mortality during the nonbreeding period. As White-crested Elaenia present high migratory connectivity between their non-breeding and breeding areas (Bravo et al. 2017), the decrease in their captured rate could also be related with environmental variations at their wintering sites on the Brazilian Cerrado and Atlantic Forest (for example, see Wilson et al. 2011).
Our study represents the fi rst survival estimation for a Neotropical austral migrant species and revealed that survival varied across seasons. For Nearctic-Neotropical migrants breeding at temperate latitudes of North America and wintering in the tropics, the temporal variation in survival has been attributed to weather anomalies across the areas used by birds along their life cycle (Mazerolle et al. 2005), as they promote decreases in food availability diminishing survival and productivity of young (Sillett et al. 2000, Mazerolle et al. 2005. Our results are consistent with those studies, as survival of White-crested Elaenia decreased after years with rainfall anomalies that aff ected ripe fruit supply. Unfortunately, there are not similar studies at the species wintering sites that allows linking the entire annual cycle survival process. Nevertheless, recent studies revealed White-crested Elaenias returns to their wintering areas in the Brazilian Cerrado and Atlantic Forest by routes with diff erent habitat structure and refueling opportunities (Bravo et al. 2017), indicating that a good body condition at departure could be fundamental to survive the journey between breeding and wintering sites. According to that, the poor relative body condition of birds during arrival and departure of the fi rst season could enhance birds' mortality during autumn migration, diminishing survival values between breeding seasons. Th is relationship between rainy breeding seasons with the reduction in the species departure body condition has also been reported by previous studies (González et al. 2004). Our results reinforce the need of demographic studies across the entire annual cycle to improve our knowledge of White-crested Elaenia population dynamic.
In addition, our mean apparent annual survival estimate (60%) is higher than the mean reported for other migratory Flycatchers of comparable size and behavior, breeding at temperate North America (50%; Karr et al. 1990, Koronkiewicz et al. 2006) but similar to tropical resident fl ycatchers (Johnston et al. 1997, França & Marini 2010, Blake & Loiselle 2013. Our results reinforce the hypothesis of higher survivorship of southern hemisphere tropical and temperate birds respecting to northern hemisphere ones (Martin 1996). Nevertheless, this trend requires more research including the hypothesis of reproduction cost, nest predation rate and adult mortality, that have been mainly verifi ed for North American and tropical birds (Martin et al. 2000, Martin 2015. Regarding migratory behavior, the higher capture rate of White-crested Elaenia males comparing to females at the beginning of the breeding season suggest that males could be arriving earlier, a behavior that is expected in territorial birds as a manner of territory acquisition (Kokko et al. 2006). Th e later departure of juveniles than of adults was also reported in studies at the austral extreme distribution of the species (Brown et al. 2007), and is usual in immature birds without prior migration experience (Whitaker & Warkentin 2010). Concerning site fi delity, comparisons between seasons showed a tendency for greater dispersal after breeding periods with low reproductive success, suggesting that Whitecrested Elaenias could be more predisposed to relocate after seasons with low productivity, as were reported for Nearctic-Neotropical migrants (Haas 1998).
On the other hand, our results of site fi delity are higher than previous studies that just considered recaptured birds on its estimations reporting returning rates only up to 11% (e.g. Espinosa & Egli 1997, Amico & Aizen 2005, Brown et al. 2007), even in long term studies (Rozzi & Jiménez 2013). Include data of relocations prevent underestimations of returning rates and improves apparent annual survival estimation (Martin et al. 2017).
However, if we considered only recaptured birds our result of site fi delity remains higher for Steff en's Lake (19% at 41 o S) comparing to the estimation of a long term study at the species austral distribution in Navarino Island (10% at 56°S). Th e lower faithfulness at Navarino could respond to its habitability conditions, as summer precipitations and temperatures are lower than our study area, the vegetation structure varies widely in short distances and the main fruit item consumed by elaenias (Drymis winteri fruits) are available in autumn during their departure (Rozzi & Jiménez 2013). Th ese factors could be promoting higher replacement rates of birds breeding at that higher latitude, indicating that Navarino Island could be a marginal habitat for the species.

Conclusions and future perspectives
Previous studies have emphasized the importance for species with low reproductive success to compensate this demographic feature with higher adult and juvenile survival (Martin 1996, Stahl & Oli 2006. However, migratory individuals face variable conditions along their life cycle that will impact on their survival indicating that there must be other mechanisms acting in their demographic dynamics. White-crested Elaenia has a diet and foraging behavior plasticity that allows them to adapt their food consumption to what is available according to the environment and rainfall variations, all along their geographical distribution (Johnson 1967, Grigera 1982, Celis-Diez 2002, Cueto et al. 2016a. Th erefore, the compensation to a low productivity of young with a higher survival rate could be mediated by the Whitecrested Elaenia feeding plasticity, allowing the species to live in diverse and changing habitats.
Nevertheless and beyond the species plasticity, there is a connectivity between breeding and wintering areas as well as migratory routes (Bravo et al. 2017) that can force changes in the species demography, as those areas are being negatively aff ected by climatic change and human activities aff ecting the annual cycle of the birds. For example, the species' main wintering and stopover biomes (Atlantic Forest, Yungas and Cerrado, respectively) are highly threatened by deforestation for agricultural purposes (Malizia et al. 2012, Ferreira et al. 2014. Th erefore to improve our knowledge of Whitecrested Elaenia population dynamic it is essential to promote long term studies that connect studies in demography, behavior, habitat use and food supply, with processes developed in wintering areas and throughout the routes the species uses during migration.