Sex, moulting dynamics, and reproductive status in Atlantic Forest tanagers and their collective impact on body condition

ABSTRACT Body condition in individuals is intricately linked to behaviour, physiology, and immunity. This study investigated how seasonal factors, such as reproduction and moulting, influence the body condition of 21 tanager species in the Atlantic Forest. Utilising publicly available data, I employed phylogenetic generalised linear mixed models to dissect these influences. Notably, females with brood patches exhibited concurrent body and feather moult, suggesting simultaneous energy-costly constraints. Furthermore, females with brood patches displayed higher body condition, indicating proactive preparation or increased resource availability during breeding. Lastly, individuals moulting body feathers exhibited enhanced body condition, suggesting heightened foraging activity. These data suggest complex connections between sex, moulting, and reproductive status in tanagers in the Atlantic Forest.

The relationship between body condition and survival can vary with age and sex.Juveniles typically manifest lower body conditions relative to adults (Jones et al. 2002).Furthermore, either males or females or neither sex can show higher body condition: (a) males exhibit superior body condition relative to females in Red-Legged Partridges (Alectoris rufa) (Nadal et al. 2018), whereas (b) females display elevated body condition in comparison to males in Willow Flycatchers (Empidonax traillii) (Owen et al. 2005).In contrast, there can also be a lack of conspicuous differences between sexes (Scherer et al. 2014;Angel et al. 2015;Wojczulanis-Jakubas et al. 2015), or a complicated interaction between sex and age, exemplified in Swallow-Tailed Manakins (Chiroxiphia caudata), in which juvenile males surpass females in body condition, while adult females outperform adult males (Souza Penha and da Silva Rodrigues 2022).
These observed variations underscore the potential influence of ancillary factors, notably regional differences between temperate and tropical regions.Tropical bird species potentially face heightened threats, including a higher risk of extinction (Reif and Štěpánková 2016) and lower heat tolerance (Pollock et al. 2021).Consequently, they are important species for investigating body condition dynamics, particularly in the context of climate change scenarios (Şekercioğlu et al. 2012).
Feather moult constitutes a systematic and seasonally recurring process in avian life, representing a crucial aspect for birds of diverse age groups (Gill 2007).This process is energetically demanding and is closely linked to physiological trade-offs, where factors such as migration and reproduction can impose constraints on feather moult (Murphy 1996).Given its substantial energy expenditure, moult holds the potential to impact the overall body condition of birds.In temperate bird species, empirical evidence supports this notion, for example in Barnacle Geese CONTACT Victor Aguiar de Souza Penha victoraspenha@gmail.com Supplemental data for this article can be accessed at https://doi.org/10.1080/01584197.2024.2328675.
(Branta leucopsis) (Portugal et al. 2007) and Captive Starlings (Sturnus vulgaris) (Swaddle and Witter 1997).However, it is important to determine whether similar relationships are manifest in tropical avian species.For example, previous studies have found a negative link between stress, measured through plasma corticosterone, and body condition in tropical populations of Long-Tailed Finches during the moult period (Poephila acuticauda, and P. personata) (Maute et al. 2013).
In tandem with the moulting season, reproduction represents another energetically demanding phase in the avian life cycle.Some species strategically moult their feathers before the breeding season (Wolfe et al. 2009(Wolfe et al. , 2010(Wolfe et al. , 2012)).Subsequently, the breeding season itself involves a suite of resource-intensive behaviours, including decisions on breeding location, nest construction, surveillance of nest and nestlings, and foraging activities (Smith et al. 2007;Slagsvold and Wiebe 2011;Mainwaring and Hartley 2013;Morinay et al. 2018;Williams 2018;Martínez et al. 2022).A study encompassing 20 bird species in Finland found that females tended to exhibit a lower body condition towards the conclusion of the breeding season, suggesting a potential constraint on female body condition imposed by reproductive efforts (Andersson et al. 2018).However, moult and breeding periods may also overlap, as demonstrated in tropical Amazonian bird furnariids and thamnophilids (Johnson et al. 2012), suggesting that there might even be a synergistic detrimental effect of breeding and moulting on the body condition of birds.
In this study, I examine the impact of sex, feather moult (body and flight feathers), and brood patch presence on body condition in 21 different tanagers (Passeriformes: Thraupidae) in the Atlantic Forest Biome.While previous research in Amazonian tanagers suggests a lack of overlap between moult and breeding (Johnson et al. 2012), I explored the potential synergistic effects of simultaneous moulting and breeding on tanager body condition.This study utilised a comprehensive database of tanager species in the Atlantic Forest Biome, captured across various Brazilian states and years (Rodrigues et al. 2019).Predictions included a negative effect of moult on body condition due to the energy trade-off between feather replacement and fat reserve deposition.Additionally, individuals with a brood patch were expected to exhibit a higher body condition to cope with the energy trade-offs associated with reproduction.

Dataset
This study utilised data from the Atlantic bird traits dataset, a compilation of data from 711 bird species across South America (Rodrigues et al. 2019) To ensure data quality, I adhered to the scientific names outlined by Jetz et al. (2012), and excluded data records lacking data on essential variables: body mass, total tail length, sex, age, presence of flight feather moult, presence of body feathers moult, capture date, presence of brood patch, and Brazilian state.Data were also excluded from individuals from states and years with a sample size below 10, as well as juveniles due to limited sample size in certain Brazilian states (Supplementary Tables S1, 2, and 3 -more information in the Body condition section).
Age determination was based on plumage patterns, commissure presence, iris colouration, or ossification level, while sex determination considered age and plumage (Rodrigues et al. 2019).Flight feather moult was identified by actively growing symmetrical or asymmetrical flight feathers, and body feather moult was recognised by growing feathers in any body area except the wings.I included both flight and body feather moults as separate variables because of their differential functional significance.Finally, a breeding season variable was established based on capture date.Captures from October to March were considered to be in the breeding season, and those from April to September as nonbreeding, aligning with neotropical passerine patterns (IBAMA 1994;Gill 2007).
Finally, I conducted a thorough investigation of the overlap between the presence of brood patch, body moult, and flight feather moult, separated by sex.
Therefore, I found 238 males that were labelled as having a brood patch (reproductive_state variable in the dataset).The dataset also provides specifically the stage in which the brood patch was found (brood_patch variable in the dataset).I found that 225 individuals did not have information on the brood patch stage, 12 individuals had a stage 0, meaning that the brood patch was absent, and one individual had a brood patch stage as 5. Brood patch stage followed the regulations for bird banding procedures from the Centro Nacional de Pesquisa e Conservaҫão de Aves Silvestres (CEMAVE), which stablished 6 stages: 0 absence of brood patch; 1: few lost feathers and low level of vascularisation; 2: clear vascularisation; 3: extreme vascularisation with a thick skin; 4: loss of vascularisation but still with a thick skin; 5: absence of vascularisation and the presence of growing feather sin the region (Souza and Serafini 2020).The male individual labelled with a brood patch at stage 5 was Tangara cayana, which is mostly known for the female only having a brood patch (Duca and Marini 2011).Therefore, due to the uncertainty of these 238 male individuals, I decided to remove them from the statistical analysis.After, I checked the proportion of females that had a brood patch, regardless of the stage, and that were moulting the flight and body feathers.Male individuals removed from statistical analysis can be found in Supplementary Table 5.

Study groups
Tanagers (Passeriformes: Thraupidae) stand as denizens of the Neotropics, spanning from Central to South America, and occur in diverse habitats, including lowland rainforests, grasslands, and high-altitude fields (Rosenberg et al. 1999;Rodríguez-Ruíz et al. 2011;Eisermann et al. 2011a;Rodrigues et al. 2019;Winkler et al. 2020;Aguiar de Souza Penha et al. 2022, 2023).Tanagers are believed to be socially monogamous, with both sexes participating in parental care, with clutch size typically ranging from 1-4 eggs and an incubation period lasting 12 to 14 days, but with a high species variation (Aguiar de Souza Penha et al. 2022, 2023).The incubation is usually performed by the female (Klatt et al. 2008;Dos Santos and Marini 2010;Duca and Marini 2011;Eisermann et al. 2011b;Snchez Martínez and Londoño 2012;Sánchez-Martínez and Londoño 2017;Cerón-Cardona et al. 2018;Mortensen and Reed 2018;Veloso et al. 2018;Rodrigues et al. 2019;Loaiza-Muñoz and Londoño 2020).We still have little knowledge about the moulting dynamics of tanagers, but studies in the region analysing passerines found that flight feather moult usually takes place at the end or after the breeding season (Marini and Durães 2001;Magalhães et al. 2007;de Andrade et al. 2018;Faccio et al. 2018;Rodrigues et al. 2019), but with body moult happening throughout the year (Rodrigues et al. 2019).While the majority of tanager species are currently categorised as least concerned by the IUCN, the escalating threats of habitat loss and introduction of exotic species have contributed to an increased number of threatened species in recent years (Winkler et al. 2020).The phylogeny used in this study follows Jetz et al. (2012).

Body condition
I derived estimates of body condition for all species by analysing residuals from a mass-tail length regression model (Peig and Green 2010) implemented through generalised linear mixed models (GLMM).This analysis was conducted using the lmer function from the lme4 package (Bates et al. 2015) in the R software (R Core Team 2019).In the modelling process, I incorporated breeding season, year of capture, age, and sex as random factors whenever applicable.Also, to eliminate undesired spatial-temporal variability, I computed body condition values for each species and within each Brazilian state, following established methodologies (da Silva Rodrigues et al. 2021; Souza Penha and da Silva Rodrigues 2022).

Statistical analysis
First, I conducted a thorough investigation of the overlap between the presence of brood patch, body moult, and flight feather moult in females only.Then, to understand the predictors of body condition in tanagers, I conducted a visual examination of the body condition distribution through a histogram and employed the normalize function from the BBmisc package (Bischl et al. 2022) to achieve a normalised distribution of the body condition.The multicollinearity among predictors was assessed by calculating the variance inflation factor (VIF) using the VIF function from the regclass package (Petrie 2020).A GVIF (1/2df) value of two was considered indicative of multicollinearity.I used a generalised linear mixed model with the lmer function from the lme4 package (Bates et al. 2015), with the normalised body condition variable as the response, and the brood patch presence, sex, flight feather moult, and the body feather moult as predictors.Models included the interaction between sex and flight feather moult, body and flight feather moults, and presence of brood patch and body feather moult.Random terms were also included, namely: the longitude, latitude, the breeding status, year, and the species identity.The interaction between flight feather moult and body moult (VIF = 4.05), and between sex and flight feather moult (VIF = 2.61) were highly collinear, so I removed those interactions from the model.After that removal, no collinearity was found among the tested predictors (VIF values: Presence of brood patch = 1.32;Sex = 1.18;Flight moult = 1.18;Body moult = 1.28;Interaction between presence of brood patch and body moult = 1.19).For final modelling, a phylogenetic generalised linear mixed model (PGLMM) was employed utilising the pglmm function from the phyr package (Ives et al. 2020), using the final model after verifying for multicollinearity.The response variable was body condition, with the predictors including sex, occurrence of flight moult, occurrence of body moult, presence of brood patch, and the interaction between presence of brood patch and body feather moult.The interaction terms in the model were reported only if statistically significant.To address potential spatial autocorrelation and temporal variability, the same random factors, namely latitude, longitude, breeding season (outside the breeding season and inside the breeding season), and year of capture were incorporated into the model.Additionally, the species phylogeny was utilised to account for species relatedness (Jetz et al. 2012).Model comparison involved evaluating the full model against a null model, with the full model deemed statistically relevant only if significantly different from the null model (lower AIC value).To assess model performance, a plot of average residuals versus average fitted values was generated using the binnedplot function from the arm package (Gelman and Su 2022) (Supplemental Figure S1).Finally, to consider a variable as statistically significant, I consider the 95% confidence interval as not including zero, and the p-value as being lower than 0.05.To produce the 95% confidence interval, a bootstrap procedure with 1000 iterations was used.For visualisation of statistically significant variables, I used the ggplot2 package with predicted values from the model, using the predict function form R (R Core Team 2019), as well as the forestplot function from the forestplot package (Gordon and Lumley 2022).Mapping of sampling locations was accomplished using the shape files from IBGE (IBGE -Instituto Brasileiro de Geografia e estatística 2013), and the ggplot2 package (Wickham 2016).All the analysis were conducted in the R software (R Core Team 2019).

Overlap between brood patch and moult
The mean body condition across all species was 0.002 ± 2.32 (mean ± standard error) (Figure 1).The range in values of condition was dramatic, with one individual of Trichothraupis melanops exhibiting the highest condition in São Paulo state (24.76), while one individual of Tachyphonus rufus had the lowest condition value in Pernambuco (−14.4).Very few females moulted feathers during the breeding season (Table 1).Only 3% of females that had a brood patch were also moulting the flight feathers, with Haplospiza unicolor having the highest number of individuals (3 individuals), whereas 12% of females that had a brood patch were also moulting the body feathers, with Sicalis flaveola having the highest number of individuals (6 individuals).In addition, I found 1% of females moulting both flight and body feathers while having a brood patch, being found only in two species (number of individuals in parenthesis), namely Haplospiza unicolor (2), and Sporophila caerulescens (1).A complete summary can be found in Supplementary Table S1.In addition, most females with a brood patch were distributed in the breeding season, from October to March, whereas females without a brood patch were captured through the whole year, with a slightly higher number outside the breeding season, from April to September (Supplementary Table S7).

Predictors of body condition
The full model (AIC = 5924.66)exhibited a lower AIC value compared to the null model (AIC = 5971.80).The presence of a brood patch and body moult were associated with body condition, such that females that had a brood patch and those individuals moulting the body feathers had a higher body condition compared to females without a brood patch and non-moulting individuals, respectively (Table 2, Figure 2).There was no significant association of the interaction between brood patch presence and body moult, as well as for sex and flight feather moult (Table 2).

Discussion
In this study, there was an observed association of female tanagers undergoing moult (both body and flight feathers) and those exhibiting a brood patch.This observation diverges from a prior study on Amazonian tanagers (Johnson et al. 2012), where the analysis focused solely on primary feather moult.Despite methodological differences, that study covered eight tanager species with a substantial sample size (708 individuals), including Cyanocompsa cyanoides, Coereba flaveola, Lanio fulvus, Oryzoborus angolensis, Ramphocelus carbo, Tachyphonus cristatus, Tachyphonus surinamus, and Volatinia jacarina.By narrowing our comparison to species shared between both studies (Tachyphonus cristatus and Volatinia jacarina), our findings align, as in the present study, these species did not show an overlap between flight feather moult and the presence of a brood patch.However, the varying number of females exhibiting concurrent brood patch and moulting -whether flight feather, body feather, or both -suggests potential differences among tanager populations.
I observed that individuals displaying a brood patch exhibited a higher body condition compared to those without.This outcome suggests that females with a brood patch during the breeding season may enjoy enhanced resource availability.It is worth noting that the breeding season may coincide with the wet season and peaks of flowering, leaf flushes, and resource abundance in the Atlantic Forest (Develey and Peres 2000;Morellato et al. 2000).As an example of other tropical groups, in years marked by increased rainfall and potentially higher resource availability, female White-Rumped Munias (Lonchura striata), experienced heightened overall body conditions, leading to consequential increases in clutch size and female fecundity (Oppel et al. 2013;Hidalgo Aranzamendi et al. 2019).Females may strategically enhance their body condition before and during the breeding season, possibly to better cope with the various stressors associated with reproduction.For instance, in temperate species, females with stage 2 brood patches exhibited superior body conditions during the breeding season (Redfern 2010).Similarly, in Common Bulbuls (Pycnonotus barbatus), a tropical species, individuals experienced a peak in body mass during the incubation period, even though it was not specifically correlated with brood patch development (Nwaogu et al. 2017).Female tanagers in Atlantic Forest may derive benefits from synchronising the breeding season with the peak of resource availability, as reflected in their elevated body condition.I observed that individuals undergoing body feather moult exhibited a higher body condition compared to those not moulting, a finding contrary to my initial expectations.However, a parallel trend was found in a study of House Finches (Haemorhous mexicanus) inhabiting the Arizona desert, USA.In that study, male House Finches with elevated body conditions displayed more intense body feather moulting compared to their urban counterparts (Hutton et al. 2021).A similar association was noted among migratory Barn Swallows (Hirundo rustica) wintering in Nigeria (Saino et al. 2013).These results suggest two non-mutually exclusive hypotheses.First, individuals undergoing body feather moult may actively seek energy-rich resources.Second, it's plausible that individuals attain a higher body condition before the onset of moulting to better cope with the energy expenditures associated with replacing body feathers.Finally, no discernible associations between sex and body condition emerged in my study.In earlier studies, findings varied between males exhibiting higher body condition than females (Nadal et al. 2018), females demonstrating higher body condition than males (Owen et al. 2005), and even the potential influence of age influencing the dynamics between body condition and sex (Souza Penha and da Silva Rodrigues 2022).This suggests that in tanagers, both males and females might share similar behavioural constraints linked to securing energy-rich resources or encountering physiological trade-offs associated with reproduction, migration, and moulting.Interestingly, individuals undergoing flight feather moult displayed a comparable body condition to nonmoulting counterparts, hinting that the moulting process might not exert substantial costs.However, recent studies have proposed vital connections between moult, plasma proteins, and circulating triglyceride levels -significant indicators of nutrient status in individuals (Podlaszczuk et al. 2017;Drake and McGraw 2023).To enhance our understanding of the energy dynamics during moulting periods in tanagers, I recommend that future studies measure critical circulating determinants of body condition.
It's essential to acknowledge some limitation in my findings -the inability to distinguish between asymmetrical and symmetrical flight feather moults.Asymmetrical moults are typically associated with feather loss, while symmetrical moults are linked to the seasonal process of feather replacement (Gill 2007;Rodrigues et al. 2019).Furthermore, given the limited sample size of juveniles in certain Brazilian states, I was unable to explore the relationship between age and sex in relation to body condition.Therefore, future research endeavours should prioritise investigating how feather moult impacts body condition while considering various age groups, particularly juvenile birds.
In summary, this study revealed a noteworthy convergence between female tanagers undergoing moult and those exhibiting a brood patch, suggesting potential strategic advantages tied to the breeding season.Additionally, individuals with a brood patch displayed elevated body condition, indicating an adaptive response linked to resource availability during the Atlantic Forest's wet season.This pattern aligns with strategic behaviours observed in diverse temperate and tropical bird species during their respective breeding periods, underscoring the significance of timing and resource optimisation.Unexpectedly, individuals undergoing body feather moult exhibited higher body condition, challenging our initial expectations and implying resource-seeking behaviours or proactive readiness for the energy-intensive feather replacement process.To extend these findings, future research could delve into tanager dietary patterns, energy allocation during different moult and breeding stages, and the timing and duration of moult events to gain deeper insights into their ecological implications.

Figure 1 .
Figure 1.Map of Brazil and the capture sites of all 2,127 individuals belonging to 21 tanager species (Passeriformes: Thraupidae) incorporated in this study.The black dots on the map denote the sampling locations.Detailed information regarding each location site and the number of captured individuals per species is provided in SupplementaryTable 6 for comprehensive reference.

Figure 2 .
Figure 2. Results of the phylogenetic generalized linear mixed model (PGLMM) examining body condition in 2,127 individuals belonging to 21 tanagers species.The PGLMM model had body condition as the response variable, and the presence of a brood patch (reference level: absent), sex (reference level: female), flight feather moult (reference level: absent), and body feather moult (reference level: absent) as predictors.The tanager phylogeny was included in the model to account for species-relatedness.I also included the latitude, longitude, Brazilian state, and year as random terms, to account for potential spatial-temporal variability in the model.Here, I show the 95% confidence intervals (C.I.), with the black box indicating the average C.I., and the straight line indicating the lower (beginning) and upper (end) C.I. Results indicate that females with a brood patch have a higher body condition compared to females without a brood patch and that individuals moulting the body feathers also had a higher body condition compared to non-moulting individuals.
Table 6 for comprehensive reference.

Table 1 .
Summary data of 953 female individuals belonging to 21 tanager species.Here I show the number of females per species, detailing the total count of captures (N), the number of females undergoing flight feather moult (FM), body moult (BM), or both, categorised by the presence or absence of a brood patch.

Table 2 .
Results of the Phylogenetic Generalized Linear Mixed Model (PGLMM) of 2,127 individuals belonging to 21 tanager species with body condition as the response variable, while predictors included sex, presence of body feather moult, flight feather moult, and breeding status.I present fixed-term effects, along with estimates, standard errors (SE), 95% confidence intervals (95% C.I.), and corresponding p-values.Random-effect terms (longitude, latitude, Brazilian, state, breeding season, and year) are delineated by variance and standard deviations.A phylogenetic effect was also taken into consideration to account for species relatedness (phylogenetic effect), showing the variance and standard deviation of the species effect within the model.