Behavioral syndromes are associated with social plasticity and competence in a wild primate

doi:10.21203/rs.3.rs-6571987/v1

Download PDF

Research Article

Behavioral syndromes are associated with social plasticity and competence in a wild primate

https://doi.org/10.21203/rs.3.rs-6571987/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 19 Nov, 2025

Read the published version in Behavioral Ecology and Sociobiology →

You are reading this latest preprint version

The ability to optimize social behavior to varying socioecological circumstances has been termed “social competence” and relies on behavioral plasticity. Behavioral syndromes, i.e. consistent individual differences in intraindividual correlations among behavioral traits, appear to be a constraint on social competence, yet studies exploring this have largely been limited to experimental laboratory settings. Here, we tested the importance of behavioral syndromes to social competence in wild Barbary macaques (Macaca sylvanus), an endangered primate with established links between positive social relationships and survival. We studied two groups (n = 27 individuals) in which behavioral syndrome phenotypes were established in a previous study. “Shyer” (i.e., risk avoidant) individuals showed greater plasticity compared to “bolder” individuals in affiliative responses to the immediate social environment, being more likely to initiate grooming with larger numbers of conspecific bystanders present. Shyer individuals also increased their grooming social network connectivity to a greater degree than bold individuals in periods of higher anthropogenic pressure. In periods of colder weather, “shyer” individuals tended to concentrate their grooming network into fewer ties, whereas “bolder individuals” slightly increased their number of connections. In general, “shyer” individuals were more central in the grooming networks and any changes in network connectivity in relation to socioecology were small, reflecting the fact that stability in social network position is a common phenomenon in various taxa. Overall, we show that behavioral syndrome phenotypes influence plasticity in affiliative behavior and thus may be a key factor in individual responses to the rapidly changing socioecologies of the Anthropocene.

personality

plasticity

sociality

fitness

networks

Climate change and other anthropogenic factors are anticipated to increase the instability of the social and ecological environments of all group living animals, including humans. Positive social relationships are key for survival in many species, therefore, there is a growing interest in the flexibility of individuals in managing their social relationships against a rapidly fluctuating socioecological setting. Here we show that in wild Barbary macaques (Macaca sylvanus), a species that relies on social ties to survive in an extreme mountainous environment, behavioral syndromes (i.e., personality) predicted individual ability to change their levels of social engagement and network connectivity across various environmental gradients (i.e., changing numbers of social partners, human pressure, and shifts in temperature). This link between behavioral syndrome phenotypes and social plasticity has previously largely only been seen in laboratory settings. Therefore, our study is significant in that we show that this translates and is relevant in a wild setting and in a species with known links between its social behavior and its survival.

Among group-living animals, including humans, more social individuals tend to have improved fitness outcomes (Cameron et al., 2009; Holt-Lunstad et al., 2010; Lehmann et al., 2016; Nuñez et al., 2015; Silk, 2007). Despite the benefits of being social, time and energy are limited resources for all animals (Dunbar et al., 2009; Grove, 2012; Korstjens et al., 2010). Therefore, individuals often face choices about when to allocate resources to social interactions at the expense of other vital behaviors such as resting, foraging, parental care, travelling or vigilance (Dunbar, 1992; Dunbar et al., 2009; Lehmann et al., 2007). Furthermore, to maximize the benefits of these social behaviors, individuals must also choose appropriate social partners to avoid social rejection or antagonism (Mielke et al., 2018; Taborsky & Oliveira, 2012).

The social environment is vulnerable to fluctuation arising from changes in group-level factors, such as group size (Lehmann et al., 2007), sex ratios (L’Italien et al., 2012), and the stability of dominance hierarchies (Flack et al., 2006). Social time investment strategies may also be influenced by ecological variation. For example, in periods of low food availability or high predation risk, individuals may prioritize behaviors more closely related to survival (e.g. foraging or vigilance) rather than social interactions (Foster et al., 2012; Lehmann et al., 2007; Marshall et al., 2016). Lastly, individuals need to adjust social choices based on variation in the immediate social environment, i.e., the availability of social partners and potential bystanders to that social interaction (Kaburu & Newton-Fisher, 2016; Mielke et al., 2018). Therefore, it maybe adaptive for individuals to be flexible in their own social choices and strategies to ensure dedicating appropriate time to socialization, socializing with appropriate partners or with the appropriate numbers of partners to maximize fitness in different socioecological settings. An individual’s ability to adjust social behavior to the socioecological environment has been termed “social competence” (Taborsky and Oliveira 2012). Given the potential adaptive benefits of optimizing social living, there is growing interest in understanding which traits facilitate or constrain social competence in fluctuating environments (Jolles et al., 2019).

Behavioral syndromes, i.e. consistent individual differences in intraindividual correlations among behavioral traits (Dall et al., 2004; Sih et al., 2004), are anticipated to influence social decision making and thus social relationships and network formation (Krause et al., 2010). Behavioral syndromes may influence social relationships through homophily, whereby individuals of a particular behavioral phenotype occupy the same social niches as other individuals that behave in the same consistent fashion (Bergmüller & Taborsky, 2010). Supporting this, social homophily based on behavioral phenotypes has been observed in humans and other animal taxa (Aplin et al., 2015; Ebenau et al., 2019; Massen & Koski, 2014; Noë et al., 2016, 2018). Furthermore, studies across taxa reveal that individuals have relatively consistent social phenotypes that can persist over the course of the lifespan (Aplin et al., 2015; Fisher et al., 2016; Jacoby et al., 2014; P. J. Tkaczynski et al., 2020).

Behavioral syndromes may predict flexibility to fluctuations in the environment. In a number of fish studies, “bold” (consistently risk-prone and behaviorally proactive) individuals have been shown to have lower behavioral plasticity to changes in socioecological settings compared to “shy” (risk avoidant, reactive) individuals (Jolles et al., 2014, 2015, 2019; Kareklas et al., 2016; Kurvers et al., 2010; Magnhagen & Bunnefeld, 2009; Ólafsdóttir & Magellan, 2016). These results confirm assumptions of the “coping style” framework of behavioral syndromes, which distinguishes individuals as either behaviorally and physiologically “proactive” or “reactive” to stressful conditions (Coppens et al., 2010; Koolhaas et al., 1999, 2010). Here, proactive/bold individuals are expected to have low levels of physiological stress activity and reactivity compared to reactive/shy individuals, meaning they are also expected to be less physiologically sensitive and behaviorally responsive to socioecological variation when compared reactive/shy individuals (Coppens et al., 2010; Koolhaas et al., 1999, 2010). The relationship between behavioral syndromes and behavioral plasticity to the socioecological environment has typically been examined through experimental manipulations in laboratory conditions (Jolles et al., 2014, 2015, 2019; Kareklas et al., 2016; Kurvers et al., 2010; Ólafsdóttir & Magellan, 2016). Therefore, whether behavioral syndromes affect social plasticity and competence in natural settings, and in circumstances where social choices have established fitness consequences, merits further investigation.

In our study, we specifically test the relationship between behavioral syndromes and social plasticity and competence in a wild primate, the Barbary macaque (Macaca sylvanus). Specifically, we examine (a) whether behavioral syndromes are associated with the degree to which individuals adjust social decision making in relation to changes in the immediate socioecological conditions (social partner choice, bystander composition, human presence), and (b) whether behavioral syndromes are associated with the degree to which individuals adjust their social connectivity in relation to changes in broader socioecology (reproductive seasonality, anthropogenic presence, food availability, and temperature).

Barbary macaques are seasonal breeders (Fooden, 2007; Young et al., 2013) and face extreme variation in climate with cold, snowy winters associated with peaks in mortality (Majolo, McFarland, et al., 2013; McFarland & Majolo, 2013). Furthermore, certain Barbary macaque groups occupy areas of varying anthropogenic activity and disturbance (Campbell, 2019; Campbell et al., 2018a, 2016; Maréchal et al., 2011; Waterman et al., 2019). Importantly, in this species, individuals with strong social relationships, measured by social grooming interactions, have attenuated physiological stress (Young et al., 2014), while centrality in grooming social networks is associated with greater access to thermoregulatory huddles (Campbell, Tkaczynski, Lehmann, et al., 2018) and probability of survival during cold winters (Lehmann et al., 2016). Therefore, this species occupies a highly heterogeneous environment in which the cost-benefit ratio of social interactions, particularly grooming interactions, is likely to fluctuate, and in which social relationships are known to have fitness benefits.

Three behavioral syndromes have been quantified in Barbary macaques: Excitability, Sociability and Tactility (Tkaczynski et al. 2019a). On a behavioral level, Excitability is structurally similar to the “bold-shy” axis commonly studied in the behavioral syndrome literature (Sih et al. 2004; Tkaczynski et al. 2019a), with more “excitable” individuals behaviorally similar to bold individuals, and less “excitable” individuals behaviorally similar to shy individuals. In male Barbary macaques, more “excitable” or “bold” individuals also have lower hypothalamic pituitary adrenal axis activity compared to less “excitable” or “shy” individuals (Tkaczynski et al. 2019b), matching physiological predictions regarding the aforementioned “bold-shy/proactive-reactive” framework (Koolhaas et al., 1999). For the present study, only the Excitability syndrome was considered. Although the Sociability (characterized by time in proximity with other individuals and in spatially central positions within the group) and Tactility (characterized by rates of self- and allogrooming) behavioral syndromes identified in Barbary macaques may influence social plasticity in this species, they are not readily comparable to those more common in behavioral syndrome literature (P. J. Tkaczynski et al., 2019). Furthermore, for Tactility, the analysis may result in issues of circularity when examining how grooming decisions vary in different circumstances. Therefore, we focus on Excitability and refer to this behavioral syndrome as Excitability/Boldness in the current study to be consistent with former research on this species yet make the results relevant and comparable to the broader behavioral syndrome literature.

Using detailed behavioral data from one autumn and two winter periods, we examined whether Excitability/Boldness is associated with plasticity in grooming social decision making and social connectivity (grooming social network centrality). We predicted that more challenging conditions (e.g., breeding vs non-breeding season, greater anthropogenic presence in the form of higher numbers of tourists, lower availability of high-quality food, lower temperatures, greater numbers of bystanders or higher-ranking bystanders) would inhibit social engagement, reduce the probability to engage in grooming, and lead to reductions in grooming social network centrality. We also predicted that “shy” macaques will demonstrate greater plasticity in social behavior compared to “bold” individuals based on similar studies conducted in laboratory fish studies (Jolles et al., 2014, 2019; Magnhagen & Bunnefeld, 2009; Ólafsdóttir & Magellan, 2016) and predictions regarding plasticity based on the aforementioned “bold-shy/proactive-reactive” framework for behavioral syndromes (Coppens et al., 2010; Koolhaas et al., 1999, 2010).

All data collection was conducted following ethical approval by University of Roehampton and the receipt of research permits from Haut-Commissariat aux Eaux et Forêts et à la Lutte Contre la Désertification, Royaume du Maroc.

Study site and subjects

We collected data from two groups of wild Barbary macaques located in Ifrane National Park, Morocco (33˚ 24’N, 05˚ 12’W; elevation 1,500-2,000m above sea level). Both groups were habituated to human researchers and were observed from approximately 7 meters, with all subjects individually recognized. The subjects of this study were the adult males and females of each group (n = 12 [5 males] in the Blue group; n = 15 [6 males] in the Green group). Adults were defined as sexually mature individuals, which was based on body size in both sexes, the presence of anogenital swellings during mating season for females, and descended testicles and large canines for males (Fooden, 2007).

Behavioral data collection

Behavioral data were collected to quantify behavioral syndromes, determine dominance hierarchies, analyze variation in behavioral time allocation and construct social networks. We used a combination of focal, scan, and ad libitum sampling (Altmann, 1974), collected between October 2013 and March 2014, and between February 2015 and April 2015. Behavioral data were collected from near-dawn to near-dusk. Focal observational samples lasted 30 minutes and the order of subjects was pseudorandomized, with no individual sampled more than once on the same day.

During focal observations, the duration of state behaviors (foraging, feeding, resting, traveling, vigilance, play and grooming) were recorded continuously (P. J. Tkaczynski, 2016), while the frequency of contact, proximity, agonistic, dominance and solitary behaviors, as well as facial displays (open mouth, bearing teeth) and vocalizations, (grunts, barks, screams, teeth chatters) were recorded as point events (Fischer & Hammerschmidt, 2002; P. J. Tkaczynski, 2016). For feeding behavior, we also noted what the individuals were eating, first at a broad categorical level (leaves/grass; insects; mushrooms; fruit/nuts; human-produced), then specific species noted if they were possible to identify (e.g., acorns from Quercus ilex and Quercus faginea oaks). At the start and end of the focal samples, proximity scans were conducted to record the number of conspecific group members, as well as humans (specifically tourists and vendors at eco-tourism sites) within 0-1m, 1-5m and 5-10m of the subject of the focal sample. Observers also visually assessed whether subjects were central or peripheral within the group in terms of spatial position; individuals were considered peripheral if they were the outermost individual at the front, rear, or side of the group, and otherwise were considered to be central. The proximity and spatial data were incorporated into the behavioral syndrome quantification; the numbers of humans in proximity were used as a predictor variable in our models (see below). Additional aggression and dominance interactions were collected ad libitum; these aggression and dominance data were used in the dominance interactions and to assess whether focal individuals had received aggression on a given day prior to focal sampling.

All observation data was collected by a team including PJT, LADC and six research assistants. The data were collected using a Psion handheld computer and The Observer XT software version 8.0 (Noldus Information Technology, 2008). Inter-observer reliability tests using intra-class coefficients (ICC; (Shrout & Fleiss, 1979)) were conducted for all researchers: researchers collected behavioral data only once they had recorded two consecutive tests where the frequency and duration of variables recorded were significantly reliable compared to those recorded by PJT (ICC > 0.95; p < 0.05).

Socioecological variables

To account for seasonality, we differentiated between mating and non-mating seasons. The dates of the mating seasons were determined by the first and last observed copulation with ejaculation within each group (Young et al., 2013): mating season 1 (Blue group: 11-Oct-2013:20-Jan-2014; Green group: 9-Oct-2013:9-Jan-2014), non-mating season 1 (Blue group: 21-Jan-2014:5-Mar-2014; Green group: 10-Jan-2014:6-Mar-2014) and non-mating season 2 (4-Feb-2015:18-Apr-2015 for both groups).

To quantify anthropogenic presence, at the start of each focal observation we estimated the number of humans (excluding researchers doing the observations) present within the radius of the group spread of the study group. We recorded the number of humans using a logarithmic index: 0 (no humans), 1 (between 1 and 10 humans), 2 (between 11 and 100 humans), and 3 (above 101 humans) humans present (Price et al., 2014). Using this index allowed for rapid and thus practical assessments of human presence concurrent to behavioral data collection (Price et al., 2014). In addition, during the proximity scans at the start and end of focal samples, we recorded the number of humans within 10 meters of the focal subject.

To quantify high quality food in diet, we calculated the proportion of high quality food items in the diet of subjects (Thompson & Wrangham, 2008). Barbary macaques are dietary generalists that occupy a typically nutrient poor habitat and rely predominantly on herbaceous plants, shrubs, and grasses (Ménard, 2002). However, during autumn and winter seasons, they preferentially feed on more nutrient-rich items, namely the acorns of oak trees (Quercus ilex and Quercus faginea), and the berries of hawthorn bushes (Crataegus spp) (Ménard, 2002; Waterman et al., 2019). Therefore, for each day, we recorded the proportion of the total time spent feeding that was spent feeding on either acorns or hawthorn berries.

As daily minimum temperature increases physiological stress (Young et al., 2014) and the likelihood of the macaques to form larger thermoregulatory huddles (Campbell et al., 2018b), we used this measure as our temperature variable in our analyses. During 2013 and 2014, air temperature was recorded hourly when in the presence of subjects using a 3500 Kestrel Pocket Weather Station. In 2015, due to a change in protocol relating to associated projects, air temperature was recorded only when the subjects left or entered sleeping trees in the morning and evening, respectively. In all time periods, temperatures were recorded in the shade and with the device 1.5 m above the ground. In total, temperature recordings were made on 111 days during the entire study period, an average of 13.75 (± 6.43) daily temperature recordings per study month.

Statistical analysis

All data preparation, analysis, and visualization were performed in R 3.6.3 (R Core Team, 2020) using the RStudio interface (RStudio Team, 2021). For data cleaning, preparation, and the generation of figures, we used a combination of base R functions and the tidyverse suite of packages (Wickham et al., 2019).

Dominance rank calculations

Dyadic interactions of aggression (contact and non-contact) and submission were used for rank calculations. We used a likelihood-based adaptation of the Elo rating approach (Foerster et al., 2016; Mielke et al., 2018; Neumann et al., 2011). Elo ratings reflect the probability of an individual to ‘win’ in a dominance interaction with another individual, with traditional approaches applying a fixed starting value, k, to all individuals (Neumann et al., 2011). The likelihood modification uses a maximum-likelihood estimation to optimize the starting k for all individuals, such that all individuals enter the hierarchy with individual k values (Foerster et al., 2016; Mielke et al., 2018). We assigned continuous Elo ratings to subjects for each day of observation; each score was standardized between 0 (lowest rank) and 1 (highest rank) within each group.

Behavioral syndrome quantification

In a previous study (P. J. Tkaczynski, 2016; P. J. Tkaczynski et al., 2019), three independent, non-correlated and consistently expressed behavioral syndromes have been identified in our study subjects: Excitability, Sociability and Tactility. In brief, behavioral syndrome quantification involved two stages: (i) testing the repeatability of behavioral variables across the three seasons; (ii) identifying correlations among the repeatable behavioral variables. The repeatability of behavioral variables was estimated using the rptr package (Stoffel et al., 2018); factor analyses using the psych package (Revelle, 2018) identified correlations among repeatable behaviors. The sums of the mean values of the behavioral variables that loaded onto a particular factor were used to create individual (subject-specific) scores for a particular behavioral syndrome. One score per behavioral syndrome per season was allocated to each subject, with higher scores indicating greater expression of the syndrome (P. J. Tkaczynski et al., 2019). For the reason outlined above, for the present study, only scores for the Excitability syndrome were considered

Social plasticity in decision making in relation to current social environment

In total, we collected 1,236 hours of behavioral data, a mean ± SD of 46 ± 2 hours per subject.

To determine individual choices about social grooming in relation to the current social environment, we first merged the activities in our focal data into two categories: grooming (included the former variables of giving grooming and of simultaneous grooming) and other activities (co-feed, feed, foraging, foraging-feeding, rest, self-groom, receiving grooming, travel, vigilance). Note, we included “receiving grooming” as separate from our “grooming” category, as we wanted to characterize variation in proactive social decisions about social time allocation from the perspective of the focal animal, i.e., an individual can be resting, approached by another individual and groomed, without making effort to participate in that social interaction.

We then converted the first 15 minutes of each focal sample into a single datapoint, recording whether or not the subject groomed an adult partner during this interval. We focused on adult grooming partners as relationships among adults have established fitness outcomes in this population (Campbell et al., 2018b; Lehmann et al., 2016; Young et al., 2014). The first 15 minutes was chosen as the duration of the interval, as this allowed us to use the proximity scan at the start of the focal to calculate, and thus include in our models the potential social partners and bystanders within 10m of the subject, while allowing sufficient time for the subject to make a social decision to groom another individual or not. We excluded all intervals where the subject was out-of-sight for more than 1 minute and any focal sample in which the focal subject was already engaged in grooming at the start of the focal. We further excluded all datapoints for which we were missing data on our main predictors of variation in social time allocation.

To test variation in social plasticity to the current social environment, we used a generalized linear mixed effect model (model 1) with a Bernoulli error structure and a logit link function; the dependent variable was each interval/choice and whether grooming with an adult partner occurred during that interval.

Our main predictors were the Excitability/Boldness score for each individual in interaction with the following socioecological variables:

anthropogenic presence (number of humans within 10m),
number of conspecific bystanders within 10m of subject, and
maximum rank difference (Elo score difference on day of observation) between the focal and the highest-ranking conspecific within 10m of the subject.

We included control variables of the dominance rank (Elo score of subject on day of focal sample) and sex of the focal subject, the time of day of the interval (calculated as proportion of day, e.g. 12:00 = 0.50; including as a linear and a quadratic term to account for potential peaks in activity around dawn and dusk), the minimum temperature (°C) recorded that day, and group membership (Blue or Green) of the individual.

The random effect of individual identity was included in the model, as well as random slopes for anthropogenic presence, number of conspecifics within 10m, maximum rank difference, dominance rank, time of day (liner and quadratic terms) and the minimum temperature within individual identity. These time allocation models also included a random effect (random intercept) of day of sampling to account for potential autocorrelation or non-independence of response, predictor, or control variables.

Plasticity in social connectivity in relation to fluctuating socioecology

Network construction and metrics

For the network analyses, for each group, directed and weighted social networks based on dyadic grooming data were constructed, here focusing on all grooming bouts between adult partners. We aimed to generate grooming networks spanning short enough periods to determine potential short-term adjustments in social connectivity to fluctuations in socioecology. However, we also wanted to generate networks with sufficient observation time that individuals had the opportunity to interact with social partners. To determine the most appropriate temporal aggregation of our grooming data, we used the netTS R package (Bonnell & Vilette, 2021). This package uses a “moving window” and bootstrapping approach to assess the robustness of networks aggregated at different length timespans.

On average, each individual was observed at least once every 2.29 (± 0.07) days; the maximum interval between two focal samples for an individual within each field season was 17 days. Therefore, we specified our grooming data to be aggregated at 15-, 25-, and 40-day timespans, with a 5-day moving window. Using the ‘check.windowsize’ function of netTS, we performed 1,000 bootstrap comparisons between node degree (the number edges connected to the node) calculated in observed and bootstrapped networks for each of the three timespans. This function calculates the cosine similarity between the observed and bootstrapped metrics. The sensitivity to missing edges was assessed by bootstrapping from either 100, 80 or 70% of the original data.

The results of this procedure were visually assessed and are presented in figures S1 and S2 of the supplementary materials. The 25-day aggregation was clearly preferable to the 15-day aggregations, with no marked improvement with a 40-day aggregation. Therefore, for the network analyses, we split the observational data into nine, 25-day time windows. This equated to roughly one month of data collection and guaranteed each individual was observed at least once, even in time windows with fewer observation days due to field conditions (extreme weather, difficulty finding groups). Subjects were observed for mean ± SD of 5.69 ± 1.05 hours per time window (Table S3).

For each group and for each time window, we used Bayesian inference to quantify edge weights (grooming given and received) between dyads, and then construct social networks (Hart et al., 2021; Redhead et al., 2021). This approach allowed us to incorporate and measure uncertainty in the edge weights, which may be susceptible to variable sampling effort or measurement errors in observing each individual within the dyad, or interactions between particular dyads (Hart et al., 2021).

For each group and each time window, we fitted a mixed-effect binomial model (i.e., 9 models per group). In each model, the response variable was how many focal observations individuals within a dyad groomed one another given the opportunity (i.e., the total number of focal observations each individual was observed in) within a given time window. We included actor and recipient identities as a random effect, with this identity incorporating the direction of the grooming interaction. For example, if individual A was grooming individual B, the dyad identity for that interaction was “A-B”; if B was grooming A, the identity would be “B-A".

From these models (one for each group), we used the posterior draws (n = 3,000; see Model Fitting and Validation below) for the dyad identities to calculate the median edge weights for each dyad in each time window, i.e., the proportion of time individuals within a particular dyad either gave or received grooming given the opportunity to groom one another within a particular time window. Using the median edge weights of the posterior distribution, which are equivalent to the “simple ratio index” (Hart et al., 2021; Whitehead, 2008), we constructed directed and weighted social networks. Figures S4-S21 in the supplementary materials illustrate each network within each time window (n = 9) for the Blue and Green groups, as well as the uncertainty of edge estimates.

For each individual in each of the networks generated from the posterior draws, we calculated two metrics: (a) out-degree, the number of edges connected to the node (scaled between 0 and 1 to account for differences in group sizes), i.e. the number of partners to which the individual directed grooming; and (b) out-strength, the sum of all edge weights connected to the node, i.e. the total grooming rate per individual to all partners (Farine & Whitehead, 2015). We chose these particular metrics as they have been related to fitness outcomes in this species and population previously (Campbell, Tkaczynski, Lehmann, et al., 2018; Lehmann et al., 2016). As these metrics were derived from the median posterior edge weights between each node, they represent the most probable out-degree or out-strength value for each individual based on sampling effort and potential measurement error for each individual in the study (Hart et al., 2021).

Social networks were constructed and illustrated, as well metrics calculated, using the igraph R package (Csardi & Nepusz, 2006).

Social connectivity models

For each time window of each group, we calculated the mean anthropogenic presence index, mean daily proportion of high-quality food in the diet, and mean daily minimum temperature (°C) during this observation period. Additionally, each network of each group was demarked as occurring during either in a mating or non-mating season.

During time window 4, the Blue group transitioned from a mating season to a non-mating season (55.85% of observations occurred during the non-mating period), therefore, we did not include the metrics derived from this network in the analysis.

For both social connectivity models ((a) out-degree, (b) out-strength), individuals social network metrics (out-degree or out-strength) were the dependent variables and our main predictors were the interactions between Excitability/Boldness score and mean anthropogenic presence index, mean daily proportion of high-quality food in the diet, and mean daily minimum temperature (°C), and seasonality (mating or non-mating) during each time window. The interactions were fit to measure plasticity (rate of change) in social connectivity given the socioecological gradients for different behavioral syndrome phenotypes. We included the mean rank (Elo score) of each individual for each network as a control variable. The random effect of individual identity was included in the model, as well as random slopes for each of our socioecological predictors within individual identity.

Both the degree (model 2) and strength (model 3) models were fitted with Beta error structure and a logit link function.

Social network metrics are inherently non-independent variables and, therefore, do not meet the assumptions of responses in most frequentist statistical approaches, especially using null hypothesis testing (Croft et al., 2011; Franks et al., 2021). Using Bayesian inference rather than null hypothesis testing, the probability of a value occurring is determined not by its observed frequency alone, and we can calculate the uncertainty about any such probability (Brent et al., 2017).

Model fitting and validation

For our analyses we employed a Bayesian approach with all models fitted and estimated using Hamiltonian Monte Carlo methods and Stan software (Stan Development Team, 2021) with the brms package (Bürkner, 2018) within R.

Prior to model fitting, any potential issues of covariation or collinearity between our fixed effects were inspected via pairwise plots, pairwise correlations, and variance inflation factors (VIFs). Pairwise plots and correlation coefficients were generated using the ‘covees’ function of the GGally package (Schloerke et al., 2020); VIFs were generated using the ‘VIF’ function of the car package in R (Fox & Weisberg, 2011). In the final presented models, there were no issues with collinearity or covariance among the variables (maximum VIF < 4.5; maximum rho < 0.7; Table S22 and figures S23-24 in the supplementary materials). In the social connectivity models, mean anthropogenic presence was collinear with group identity (VIF > 10.0); therefore, we included only the mean anthropogenic presence variable in those models.

For all models, numeric variables were standardized as z-scores. We fit models with weakly regularizing priors for the fixed effects (β ~ Normal(0,1)). For the priors for the components of the random effects, we used the default priors provided by the get_prior function of brms, namely a weakly regularizing half student-t prior (df = 3, scale parameter = 10) for the random intercepts, and a uniform LKJ Cholesky prior (η = 1) for covariance matrices of the random slopes.

For all models, we specified three chains of 5000 iterations, 2000 of which were devoted to the warm-up. Sampling diagnostics (Rhat < 1.01) and trace plots confirmed chain convergence for all models. Effective sample sizes confirmed no issues with autocorrelation of sampling for all models (Tables S25-27 list all effective samples sizes and Rhat values for the main models presented).

To interpret the strength and uncertainty of the associations between our predictor variables and outcomes, we report the model estimate, 90% credible intervals and the proportion of posterior (p + or p−) supporting the direction (positive or negative) of the model estimate of the associations (Martin et al., 2020; McElreath, 2020; McShane et al., 2019). We considered effects whose direction was supported by more than 95% of the posterior distribution to be well supported, and those whose direction was supported by more than 90% of the posterior distribution to be weakly supported.

All models were validated using posterior predictive checks (see figures S28-32 in the supplementary materials for plots of these checks).

Model 1: Plasticity in social decision making in relation to current social environment

A total of 1,467 datapoints from 27 subjects were included into the social decision making analysis. Overall, subjects engaged in grooming in 164 (11.179%) of these samples.

Table 1 details the results of Model 1. The probability for subjects to initiate grooming with another adult partner increased with the number of conspecifics within proximity of the subject, and “shy” individuals (those with lower Excitability/Boldness score) clearly responded more strongly to the number of conspecifics within proximity compared to “bold” individuals (those with higher Excitability/Boldness scores) (p- = 0.982; Fig. 1). Using the median values across all draws from the posterior distribution, “bold” individuals saw a 10.865% increase (from 0.054 to 0.060) in their probability to initiate grooming when going from having 0 to 5 conspecifics within proximity, whereas for “shy” individuals, the probability to groom increased by 55.168% (from 0.078 to 0.138).

The model also found that the probability for subjects to interact socially was lower for higher ranking individuals (p- = 0.908), males were less likely to initiate grooming (p- = 0.963), while across all individuals, the probability to initiate peaked in the middle of day (p- = 0.998; Figure S30 in supplementary materials).

To check our results were not driven by the length of social choice window interval, we also re-ran this analysis using a shorter interval length of 5 minutes, or using the whole focal observation of 30 minutes (i.e. the whole focal sample length). Tables S33 and S34 in the supplementary materials provide the model estimates from these models; Figure S35 in the supplementary materials shows the posterior predictive checks for both models. For both models, the direction of the estimate for the interaction between Excitability/Boldness score and the number of conspecifics within 10m was the same as in our main model; for the 5-minute interval model the direction of the estimate had weaker support than our main model (p- = 0.934), whereas the direction of the estimate had stronger support in the 10-minute interval model (p- = 0.992).

Table 1

Fixed and random effect/slope estimates from “Plasticity in social decision making in relation to current social environment” model. Estimates reflect conditional probabilities of focal individual initiating grooming with another adult. Fixed effect estimates supported by more than 95% of the posterior are highlighted in bold; fixed effect estimates supported by more than 90% of the posterior are highlighted in italics. For categorical variables, the reference level is in parentheses.
Coefficient	Estimate	Est.Error	Q5	Q95
Intercept	-1.910	0.343	-2.504	-1.375
Excitability/Boldness	0.230	0.198	-0.095	0.560
Number of humans within 10m	-0.204	0.242	-0.665	0.100
Number of conspecifics within 10m	0.163	0.128	-0.042	0.371
Maximum rank difference with bystander	0.145	0.156	-0.108	0.410
Daily minimum temperature	0.098	0.118	-0.094	0.288
Individual rank	-0.286	0.219	-0.649	0.071
Time of day	-0.347	0.122	-0.553	-0.152
Time of day2	-0.364	0.140	-0.611	-0.155
Sex (female)	-0.912	0.513	-1.737	-0.070
Group (Blue)	0.105	0.360	-0.478	0.705
Excitability/Boldness : Number of humans within 10m	-0.001	0.163	-0.271	0.255
Excitability/Boldness : Number of conspecifics within 10m	-0.228	0.118	-0.428	-0.046
Excitability/Boldness : Maximum rank difference with bystander	0.023	0.126	-0.192	0.224
Random effects and slopes
Day of observation: Intercept	0.305	0.178	0.034	0.613
Individual identity: Intercept	0.612	0.231	0.227	1.003
Individual identity: Slope for Number of humans within 10m	0.323	0.279	0.023	0.864
Individual identity: Slope for Number of conspecifics within 10m	0.268	0.165	0.030	0.562
Individual identity: Slope for Maximum rank difference with bystander	0.399	0.200	0.072	0.733
Individual identity: Slope for Daily minimum temperature	0.289	0.162	0.040	0.571
Individual identity: Slope for Individual rank	0.252	0.195	0.020	0.620
Individual identity: Slope for Time of day	0.181	0.137	0.014	0.444
Individual identity: Slope for Time of day2	0.274	0.159	0.037	0.548

Models 2 & 3: Plasticity in social connectivity

Out-degree

A total of 231 network metrics from 27 subjects were included into model 2, examining changes in network out-degree in response to socioecological predictors.

Table 2a details the results of “Model 2: Plasticity in social connectivity – out-degree”. “Bold” and “shy” individuals (those with higher and lower Excitability/Boldness scores respectively) showed slightly different responses to changes in anthropogenic presence and mean minimum temperatures (Fig. 2a and b).

“Shy” individual in general had slighter higher network out-degrees, while both “bold” and “shy” individuals showed slight increases in network connectivity with increasing levels of anthropogenic presence (p + = 0.966). Using median values from posterior draws, “bold” individuals had network out-degrees 9.115% higher in time windows of high anthropogenic presence (i.e., above the mean anthropogenic index across the study period) compared to time windows with low anthropogenic presence (i.e., below the mean). “Shy” individuals had network out-degrees 21.396% higher across the same gradient.

Although associated with some uncertainty (p- = 0.932), the interaction between Excitability/Boldness and mean minimum temperature within a time window showed that “bold” and “shy” individuals differed in the direction of their response to this gradient. When comparing network out-degrees in time windows with lower temperatures (below the mean minimum temperature across the study period) to higher temperatures (above the mean), “bold” individuals decreased their network out-degrees by 14.337%, whereas “shy” individuals increased their network out-degrees by 5.401%. It is worth noting that in both “shy” and “bold” individuals these changes in network out-degree were small in magnitude (changes of 0.013 and 0.022 respectively).

In model 2, we also observed that across all individuals, network out-degree was lower in non-mating seasons (p- = 0.998) and in males (p- = 0.963).

Out-strength

Again, a total of 231 network metrics from 27 subjects were included into model 3, examining changes in network strength in response to socioecological predictors.

Table 2b details the results of “Model 3: Plasticity in social connectivity – strength”. In this model, we found no meaningful differences in the responses between “bold” and “shy” individuals. However, across all individuals, network strength was lower in non-mating seasons (p- = 1.000) and in males (p-= 0.943) but increased in time windows with warmer minimum temperatures (p + = 0.991).

Table 2

Fixed and random effect/slope estimates from models exploring “Plasticity in social connectivity”. Part (a) shows the estimates from the model with out-degree as the response, (b) the estimates with strength as the response. Fixed effect estimates supported by more than 95% of the posterior are highlighted in bold; fixed effect estimates supported by more than 90% of the posterior are highlighted in italics. For categorical variables, the reference level is in parentheses.
Coefficient	Estimate	Est.Error	Q5	Q95
(a) Out-degree
Intercept	-0.875	0.205	-1.213	-0.532
Excitability/Boldness	0.060	0.148	-0.187	0.307
Season (mating)	-0.616	0.188	-0.920	-0.310
Mean anthropogenic index	-0.021	0.146	-0.262	0.218
Food availability	0.011	0.106	-0.161	0.187
Mean minimum temperature	-0.020	0.068	-0.131	0.090
Rank	-0.127	0.122	-0.320	0.074
Sex (female)	-0.626	0.356	-1.225	-0.056
Excitability/Boldness : Season (mating)	-0.048	0.173	-0.330	0.245
Excitability/Boldness : Mean anthropogenic index	0.241	0.138	0.020	0.472
Excitability/Boldness : Food availability	0.078	0.115	-0.112	0.265
Excitability/Boldness : Mean minimum temperature	-0.098	0.068	-0.210	0.007
Random effects and slopes	Estimate	Est.Error	Q5	Q95
Individual identity: Intercept	0.530	0.148	0.301	0.784
Individual identity: Slope for Season (mating)	0.528	0.213	0.151	0.877
Individual identity: Slope for Mean anthropogenic index	0.257	0.165	0.028	0.547
Individual identity: Slope for Food availability	0.166	0.118	0.015	0.387
Individual identity: Slope for Mean minimum temperature	0.179	0.091	0.032	0.336
(b) Strength
Intercept	-1.360	0.135	-1.584	-1.141
Excitability/Boldness	-0.006	0.101	-0.169	0.163
Season (mating)	-0.616	0.126	-0.825	-0.407
Mean anthropogenic index	-0.062	0.064	-0.169	0.045
Food availability	-0.004	0.063	-0.111	0.097
Mean minimum temperature	0.080	0.034	0.025	0.135
Rank	-0.033	0.067	-0.143	0.076
Sex (female)	-0.274	0.175	-0.558	0.014
Excitability/Boldness : Season (mating)	0.063	0.104	-0.106	0.237
Excitability/Boldness : Mean anthropogenic index	0.058	0.057	-0.036	0.154
Excitability/Boldness : Food availability	0.004	0.061	-0.097	0.104
Excitability/Boldness : Mean minimum temperature	0.011	0.032	-0.041	0.062
Random effects and slopes
Individual identity: Intercept	0.509	0.089	0.375	0.662
Individual identity: Slope for Season (mating)	0.449	0.085	0.319	0.595
Individual identity: Slope for Mean anthropogenic index	0.064	0.050	0.005	0.163
Individual identity: Slope for Food availability	0.104	0.066	0.012	0.222
Individual identity: Slope for Mean minimum temperature	0.110	0.035	0.058	0.170

In our study, Barbary macaques adjusted their social decision making based on the composition of bystanders and showed marginal changes in their social connectivity across different socioecological gradients. The extent to which individuals adjusted either their social decision making or connectivity was in part related to their behavioral syndrome phenotypes. The “coping style” framework of behavioral syndromes (Coppens et al., 2010; Koolhaas et al., 1999, 2010) predicts that “shyer” individuals should be more responsive than “bold” individuals to changes in socioecology, a pattern that has been observed in a range of laboratory-based studies in non-mammals (Jolles et al., 2019; Kareklas et al., 2016; Ólafsdóttir & Magellan, 2016). In Barbary macaques, individuals were more likely to initiate grooming when in the presence of a larger number of bystanders, with “shyer” individuals far more responsive to changes in bystander composition compared to “bold” individuals, lending further support to the “coping style” framework and replicating findings in other taxa.

When considering how behavioral syndrome phenotype was associated with changes in social connectivity, the results were more nuanced and less clear. Bold individuals decreased slightly their grooming network out-degrees (number of ties within the network) in warmer weather, while shy individuals showed marginal increases. Behavioral syndrome phenotype was not associated with changes in grooming network strength (mean strength of grooming associations); instead, all individuals increased the mean strength of their affiliative ties in warmer weather, and in the mating rather than non-mating seasons.

Studies of the links between behavioral syndromes and social plasticity in the wild are rare (Couchoux et al., 2021), and more, our study has added value in that the social traits studied have established links with fitness outcomes in our study species (Campbell, Tkaczynski, Lehmann, et al., 2018; Lehmann et al., 2016). As all wild animals (and indeed humans) are living in rapidly changing environments (Otto, 2018), our study lends weight to the significance of behavioral syndromes in mediating individual ability or motivation to adjust behaviors to shifting social and ecological settings.

In the short-term, animals should adjust their decisions to engage in social interactions based on immediate social and ecological circumstances in order to optimize the cost-benefit ratio of any social engagement (Bshary et al., 2001; Mielke et al., 2018). In our study, Barbary macaques were generally more likely to initiate an affiliative grooming bout when there were more partners available, with shyer individuals more sensitive to partner availability. Both the rank differentials and degree of social bonds with bystanders are predictors of grooming initiation and interventions (interrupting grooming bouts of others) in several primate species (Kaburu & Newton-Fisher, 2016; Mielke et al., 2018, 2021), including Barbary macaques (McFarland & Majolo, 2012; Roubová et al., 2015). In our study, we did not have long-term affiliative data with which to calculate bonds; however, it is notable that we did not observe any sensitivity to the rank of bystanders (although higher ranking individuals were generally less likely to initiate grooming). Barbary macaques are considered to have a relatively egalitarian social system, which may mean rank is less influential in partner choice compared to more despotic species (Balasubramaniam et al., 2018; De Moor et al., 2025). Meanwhile, our observed effect of grooming being more likely the higher the number of available social partners might reflect larger bystander sizes increasing the probability of bond partners being part of the available social cohort. Future work in this species can hopefully incorporate longer-term affiliative data to address this gap in our study.

Adjusting social behavior to changes in the socioecological environment has been termed “social competence”(Taborsky & Oliveira, 2012). “Shyer” Barbary macaques are more responsive to changes in bystander composition, yet whether this is a fitness-enhancing or strategic adjustment remains unclear. Shyer individuals may be taking more opportunities to reinforce bonds with key social partners when these opportunities present themselves. Similarly, in larger aggregations of individuals, shyer individuals may be more likely to initiate grooming to offset increased risks of agonism. In our study, aggressions were rare prior to observations meaning probabilities to initiate grooming are unlikely to have been driven by post-conflict dynamics (McFarland & Majolo, 2011). As such, it remains unclear what advantage (if any), shyer individuals gain from their sensitivity and plasticity compared to bold individuals.

In contrast, for our measures of social connectivity we can make more potential predictions about the fitness benefits of behavioral syndrome-mediated adjustments in the face of changing socioecological settings. For Barbary macaques, human presence increases anxiety and stress (Maréchal et al., 2011), while grooming (both giving and receiving) is associated with reduced physiological stress levels (Shutt et al., 2007; Sonnweber et al., 2015), although not necessarily reduced anxiety (Semple et al., 2013), in this species. Unsurprisingly then, in our study, individuals increased the number of grooming ties in periods with higher amounts of anthropogenic presence, with shyer individuals showing marginally more sensitivity to this gradient. As in our social decision making analysis, we again see some suggestion of shy individuals showing greater social plasticity (or competence), although the differences between the behavioral syndrome phenotypes were much less marked here.

For Barbary macaques, cold and wet winters in their high-altitude environment are challenging and are associated with high mortality risks (Majolo, Lavieren, et al., 2013; McFarland & Majolo, 2013). Individuals with more grooming social ties have access to larger thermoregulatory huddles (Campbell, Tkaczynski, Lehmann, et al., 2018) and are more likely to survive the winter (Lehmann et al., 2016; McFarland & Majolo, 2013). All individuals tended to decrease their grooming network strength in cold weather, however, shy individuals showed a tendency to concentrate this into fewer ties, whereas bold individuals had marginally more ties in colder months. Both strategies may be adaptive: for shy individuals, building ties in warmer and more favorable weather may be a good long-term strategy to ensure key partners in winter, while for bold individuals (which overall had lower network out-degree compared to shy individuals), investing in more ties only when essential (i.e., winter) may be an efficient social tactic. Indeed, these divergent but potentially equally adaptive strategies may explain how diversity in behavioral syndromes can be maintained within a population (Wolf & Weissing, 2010). That said, in our study, the changes in network connectivity in relation to socioecological gradients observed in our study were very small, reflecting the fact that individual stability in social network position is a relatively common phenomenon in various taxa (Aplin et al., 2015; Blumstein et al., 2013; Finger et al., 2017; Fisher et al., 2016; Krause et al., 2010).

For all animals, including humans, social and ecological environments are becoming increasingly less stable (Blumstein et al., 2022; Otto, 2018; Testard et al., 2021). Therefore, there is an urgent need to understand whether and to what extent individuals adjust to these fluctuations (Jolles et al., 2019). Our study supports a growing body of laboratory-based work that shows that behavioral syndrome phenotype is a trait influencing both short- and long-term social plasticity, and thus competence. We also found that different behavioral syndromes show divergent social responses to certain environmental gradients (here changes in temperature): this suggests that maintaining behavioral syndrome diversity may be a key factor for species and population resilience and sustainability in the Anthropocene.

Data availability

All data and code used to produce this manuscript are available from the following repository: https://github.com/ptkaczynski/Social_comp_behav_syn

Funding

This study was funded by the University of Roehampton Vice Chancellor Scholarship awarded to P.J.T. (January 2013).

Competing interests

The authors have no competing interests.

Acknowledgments

We thank Ifrane National Park, the Haut Commissariat aux Eaux et Forêts et à la Lutte Contre la Désertification (now Agence Nationale des Eaux et Forets), Ecole Nationale Forestiere d'Ingeniuers and Institut Scientifique de Rabat for research permission and facilitation. We especially thank Prof Mohamed Mouna (formerly of Institut Scientifique de Rabat) for his support with this project – we would have included him as a co-author but have been unable to contact him in recent years. We also thank Els Van Lavieren for initially putting us in contact with Prof Mouna. This project would not have been possible without both of their support. We thank Melanie LaCava, Kevin Remeuf, Natalie Miller, Marin Deith, and Jamie Canepa for their contribution to collecting data. We also thank James Waterman for support in the field. We thank Balbir Singh Josen, Mary Mackenzie and Voulla Bergmann for their assistance with the laboratory work. Thank you also to Chris Herridge for his technical assistance managing the data.

Altmann, J. (1974). Observational Study of Behavior: Sampling Methods. Behaviour, 49(3/4), 227–267.
Aplin, L. M., Firth, J. A., Farine, D. R., Voelkl, B., Crates, R. A., Culina, A., Garroway, C. J., Hinde, C. A., Kidd, L. R., Psorakis, I., Milligan, N. D., Radersma, R., Verhelst, B. L., & Sheldon, B. C. (2015). Consistent individual differences in the social phenotypes of wild great tits, Parus major. Animal Behaviour, 108, 117–127. https://doi.org/10.1016/j.anbehav.2015.07.016
Balasubramaniam, K. N., Beisner, B. A., Berman, C. M., Marco, A. D., Duboscq, J., Koirala, S., Majolo, B., MacIntosh, A. J., McFarland, R., Molesti, S., Ogawa, H., Petit, O., Schino, G., Sosa, S., Sueur, C., Thierry, B., Waal, F. B. M. de, & McCowan, B. (2018). The influence of phylogeny, social style, and sociodemographic factors on macaque social network structure. American Journal of Primatology, 80(1), e22727. https://doi.org/10.1002/ajp.22727
Bergmüller, R., & Taborsky, M. (2010). Animal personality due to social niche specialisation. Trends in Ecology & Evolution, 25(9), 504–511. https://doi.org/10.1016/j.tree.2010.06.012
Blumstein, D. T., Hayes, L. D., & Pinter-Wollman, N. (2022). Social consequences of rapid environmental change. Trends in Ecology & Evolution, 0(0). https://doi.org/10.1016/j.tree.2022.11.005
Blumstein, D. T., Petelle, M. B., & Wey, T. W. (2013). Defensive and social aggression: Repeatable but independent. Behavioral Ecology, 24(2), 457–461. https://doi.org/10.1093/beheco/ars183
Bonnell, T. R., & Vilette, C. (2021). Constructing and analysing time-aggregated networks: The role of bootstrapping, permutation and simulation. Methods in Ecology and Evolution, 12(1), 114–126. https://doi.org/10.1111/2041-210X.13351
Brent, L. J. N., Ruiz-Lambides, A., & Platt, M. L. (2017). Persistent social isolation reflects identity and social context but not maternal effects or early environment. Scientific Reports, 7(1), 1–11. https://doi.org/10.1038/s41598-017-18104-4
Bshary, R., Noë, R., Hammerstein, P., & others. (2001). Economics in nature. Cambridge University Press,.
Bürkner, P.-C. (2018). Advanced Bayesian Multilevel Modeling with the R Package brms. The R Journal, 10(1), 395. https://doi.org/10.32614/RJ-2018-017
Cameron, E. Z., Setsaas, T. H., & Linklater, W. L. (2009). Social bonds between unrelated females increase reproductive success in feral horses. Proceedings of the National Academy of Sciences of the United States of America, 106(33), 13850–13853. https://doi.org/10.1073/pnas.0900639106
Campbell, L. A. D. (2019). Fostering of a wild, injured, juvenile by a neighbouring group: Implications for rehabilitation and release of Barbary macaques confiscated from illegal trade. Primates, 60(4), 339–345. https://doi.org/10.1007/s10329-019-00729-w
Campbell, L. A. D., Tkaczynski, P. J., Lehmann, J., Mouna, M., & Majolo, B. (2018). Social thermoregulation as a potential mechanism linking sociality and fitness: Barbary macaques with more social partners form larger huddles. Scientific Reports, 8(1), 6074. https://doi.org/10.1038/s41598-018-24373-4
Campbell, L. A. D., Tkaczynski, P. J., Mouna, M., Derrou, A., Oukannou, L., Majolo, B., & van Lavieren, E. (2018). Behavioral thermoregulation via microhabitat selection of winter sleeping areas in an endangered primate: Implications for habitat conservation. Royal Society Open Science, 5(12), 181113. https://doi.org/10.1098/rsos.181113
Campbell, L. A. D., Tkaczynski, P. J., Mouna, M., Qarro, M., Waterman, J., & Majolo, B. (2016). Behavioral responses to injury and death in wild Barbary macaques (Macaca sylvanus). Primates; Journal of Primatology, 57(3), 309–315. https://doi.org/10.1007/s10329-016-0540-4
Coppens, C. M., Boer, S. F. de, & Koolhaas, J. M. (2010). Coping styles and behavioral flexibility: Towards underlying mechanisms. Philosophical Transactions of the Royal Society of London B: Biological Sciences, 365(1560), 4021–4028. https://doi.org/10.1098/rstb.2010.0217
Couchoux, C., Garant, D., Aubert, M., Clermont, J., & Réale, D. (2021). Behavioral variation in natural contests: Integrating plasticity and personality. Behavioral Ecology, 32(2), 277–285. https://doi.org/10.1093/beheco/araa127
Croft, D. P., Madden, J. R., Franks, D. W., & James, R. (2011). Hypothesis testing in animal social networks. Trends in Ecology & Evolution, 26(10), 502–507. https://doi.org/10.1016/j.tree.2011.05.012
Csardi, G., & Nepusz, T. (2006). The igraph software package for complex network research. InterJournal Complex Systems. Http://Www.Interjournal.Org/Manuscript_abstract.Php?361100992. https://ci.nii.ac.jp/naid/10030147770/
Dall, S. R. X., Houston, A. I., & McNamara, J. M. (2004). The behavioral ecology of personality: Consistent individual differences from an adaptive perspective. Ecology Letters, 7(8), 734–739. https://doi.org/10.1111/j.1461-0248.2004.00618.x
De Moor, D., Skelton, M., MacaqueNet, Amici, F., Arlet, M. E., Balasubramaniam, K. N., Ballesta, S., Berghänel, A., Berman, C. M., Bernstein, S. K., Bhattacharjee, D., Bliss-Moreau, E., Brotcorne, F., Butovskaya, M., Campbell, L. A. D., Carosi, M., Chatterjee, M., Cooper, M. A., Cowl, V. B., … Brent, L. J. N. (2025). MacaqueNet: Advancing comparative behavioral research through large-scale collaboration. Journal of Animal Ecology, n/a(n/a). https://doi.org/10.1111/1365-2656.14223
Dunbar, R. I. M. (1992). Time: A hidden constraint on the behavioral ecology of baboons. Behavioral Ecology and Sociobiology, 31(1), 35–49. https://doi.org/10.1007/BF00167814
Dunbar, R. I. M., Korstjens, A. H., & Lehmann, J. (2009). Time as an ecological constraint. Biological Reviews, 84(3), 413–429. https://doi.org/10.1111/j.1469-185X.2009.00080.x
Ebenau, A., von Borell, C., Penke, L., Ostner, J., & Schülke, O. (2019). Personality homophily affects male social bonding in wild Assamese macaques, Macaca assamensis. Animal Behaviour, 155, 21–35. https://doi.org/10.1016/j.anbehav.2019.05.020
Farine, D. R., & Whitehead, H. (2015). Constructing, conducting and interpreting animal social network analysis. Journal of Animal Ecology, 84(5), 1144–1163. https://doi.org/10.1111/1365-2656.12418
Finger, J. S., Guttridge, T. L., Wilson, A. D. M., Gruber, S. H., & Krause, J. (2017). Are some sharks more social than others? Short- and long-term consistencies in the social behavior of juvenile lemon sharks. Behavioral Ecology and Sociobiology, 72(1), 17. https://doi.org/10.1007/s00265-017-2431-0
Fischer, J., & Hammerschmidt, K. (2002). An Overview of the Barbary Macaque, Macaca sylvanus, Vocal Repertoire. Folia Primatologica, 73(1), 32–45. https://doi.org/10.1159/000060417
Fisher, D. N., Rodríguez-Muñoz, R., & Tregenza, T. (2016). Wild cricket social networks show stability across generations. BMC Evolutionary Biology, 16(1), 151. https://doi.org/10.1186/s12862-016-0726-9
Flack, J. C., Girvan, M., Waal, F. B. M. de, & Krakauer, D. C. (2006). Policing stabilizes construction of social niches in primates. Nature, 439(7075), 426. https://doi.org/10.1038/nature04326
Foerster, S., Franz, M., Murray, C. M., Gilby, I. C., Feldblum, J. T., Walker, K. K., & Pusey, A. E. (2016). Chimpanzee females queue but males compete for social status. Scientific Reports, 6, 35404. https://doi.org/10.1038/srep35404
Fooden, J. (2007). Systematic Review of the Barbary Macaque, Macaca Sylvanus (Linnaeus, 1758). Fieldiana Zoology, 1–60. https://doi.org/10.3158/0015-0754(2007)113[1:SROTBM]2.0.CO;2
Foster, E. A., Franks, D. W., Morrell, L. J., Balcomb, K. C., Parsons, K. M., van Ginneken, A., & Croft, D. P. (2012). Social network correlates of food availability in an endangered population of killer whales, Orcinus orca. Animal Behaviour, 83(3), 731–736. https://doi.org/10.1016/j.anbehav.2011.12.021
Fox, J., & Weisberg, S. (2011). An R Companion to Applied Regression. SAGE.
Franks, D. W., Weiss, M. N., Silk, M. J., Perryman, R. J. Y., & Croft, D. P. (2021). Calculating effect sizes in animal social network analysis. Methods in Ecology and Evolution, 12(1), 33–41. https://doi.org/10.1111/2041-210X.13429
Grove, M. (2012). Space, time, and group size: A model of constraints on primate social foraging. Animal Behaviour, 83(2), 411–419. https://doi.org/10.1016/j.anbehav.2011.11.011
Hart, J. D. A., Weiss, M. N., Franks, D. W., & Brent, L. J. N. (2021). BISoN: A Bayesian Framework for Inference of Social Networks (p. 2021.12.20.473541). bioRxiv. https://doi.org/10.1101/2021.12.20.473541
Holt-Lunstad, J., Smith, T. B., & Layton, J. B. (2010). Social Relationships and Mortality Risk: A Meta-analytic Review. PLOS Medicine, 7(7), e1000316. https://doi.org/10.1371/journal.pmed.1000316
Jacoby, D. M. P., Fear, L. N., Sims, D. W., & Croft, D. P. (2014). Shark personalities? Repeatability of social network traits in a widely distributed predatory fish. Behavioral Ecology and Sociobiology, 68(12), 1995–2003. https://doi.org/10.1007/s00265-014-1805-9
Jolles, J. W., Briggs, H. D., Araya-Ajoy, Y. G., & Boogert, N. J. (2019). Personality, plasticity and predictability in sticklebacks: Bold fish are less plastic and more predictable than shy fish. Animal Behaviour, 154, 193–202. https://doi.org/10.1016/j.anbehav.2019.06.022
Jolles, J. W., Fleetwood-Wilson, A., Nakayama, S., Stumpe, M. C., Johnstone, R. A., & Manica, A. (2014). The role of previous social experience on risk-taking and leadership in three-spined sticklebacks. Behavioral Ecology, 25(6), 1395–1401. https://doi.org/10.1093/beheco/aru146
Jolles, J. W., Fleetwood-Wilson, A., Nakayama, S., Stumpe, M. C., Johnstone, R. A., & Manica, A. (2015). The role of social attraction and its link with boldness in the collective movements of three-spined sticklebacks. Animal Behaviour, 99, 147–153. https://doi.org/10.1016/j.anbehav.2014.11.004
Kaburu, S. S. K., & Newton-Fisher, N. E. (2016). Bystanders, parcelling, and an absence of trust in the grooming interactions of wild male chimpanzees. Scientific Reports, 6(1), 20634. https://doi.org/10.1038/srep20634
Kareklas, K., Arnott, G., Elwood, R. W., & Holland, R. A. (2016). Plasticity varies with boldness in a weakly-electric fish. Frontiers in Zoology, 13(1), 22. https://doi.org/10.1186/s12983-016-0154-0
Koolhaas, J. M., de Boer, S. F., Coppens, C. M., & Buwalda, B. (2010). Neuroendocrinology of coping styles: Towards understanding the biology of individual variation. Frontiers in Neuroendocrinology, 31(3), 307–321. https://doi.org/10.1016/j.yfrne.2010.04.001
Koolhaas, J. M., Korte, S. M., De Boer, S. F., Van Der Vegt, B. J., Van Reenen, C. G., Hopster, H., De Jong, I. C., Ruis, M. A., & Blokhuis, H. J. (1999). Coping styles in animals: Current status in behavior and stress-physiology. Neuroscience and Biobehavioral Reviews, 23(7), 925–935.
Korstjens, A. H., Lehmann, J., & Dunbar, R. I. M. (2010). Resting time as an ecological constraint on primate biogeography. Animal Behaviour, 79(2), 361–374. https://doi.org/10.1016/j.anbehav.2009.11.012
Krause, J., James, R., & Croft, D. P. (2010). Personality in the context of social networks. Philosophical Transactions of the Royal Society B: Biological Sciences, 365(1560), 4099–4106. https://doi.org/10.1098/rstb.2010.0216
Kurvers, R. H. J. M., Oers, K. V., Nolet, B. A., Jonker, R. M., Wieren, S. E. V., Prins, H. H. T., & Ydenberg, R. C. (2010). Personality predicts the use of social information. Ecology Letters, 13(7), 829–837. https://doi.org/10.1111/j.1461-0248.2010.01473.x
Lehmann, J., Korstjens, A. H., & Dunbar, R. I. M. (2007). Group size, grooming and social cohesion in primates. Animal Behaviour, 74(6), 1617–1629. https://doi.org/10.1016/j.anbehav.2006.10.025
Lehmann, J., Majolo, B., & McFarland, R. (2016). The effects of social network position on the survival of wild Barbary macaques, Macaca sylvanus. Behavioral Ecology, 27(1), 20–28. https://doi.org/10.1093/beheco/arv169
L’Italien, L., Weladji, R. B., Holand, Ø., Røed, K. H., Nieminen, M., & Côté, S. D. (2012). Mating Group Size and Stability in Reindeer Rangifer tarandus: The Effects of Male Characteristics, Sex Ratio and Male Age Structure. Ethology, 118(8), 783–792. https://doi.org/10.1111/j.1439-0310.2012.02073.x
Magnhagen, C., & Bunnefeld, N. (2009). Express your personality or go along with the group: What determines the behavior of shoaling perch? Proceedings of the Royal Society B: Biological Sciences, 276(1671), 3369–3375. https://doi.org/10.1098/rspb.2009.0851
Majolo, B., Lavieren, E. van, Maréchal, L., MacLarnon, A., Marvin, G., Qarro, M., & Semple, S. (2013). Out of Asia: The Singular Case of the Barbary Macaque. In The Macaque Connection (pp. 167–183). Springer, New York, NY. https://doi.org/10.1007/978-1-4614-3967-7_11
Majolo, B., McFarland, R., Young, C., & Qarro, M. (2013). The Effect of Climatic Factors on the Activity Budgets of Barbary Macaques (<Emphasis Type="Italic">Macaca sylvanus</Emphasis>). International Journal of Primatology, 34(3), 500–514. https://doi.org/10.1007/s10764-013-9678-8
Maréchal, L., Semple, S., Majolo, B., Qarro, M., Heistermann, M., & MacLarnon, A. (2011). Impacts of tourism on anxiety and physiological stress levels in wild male Barbary macaques. Biological Conservation, 144(9), 2188–2193. https://doi.org/10.1016/j.biocon.2011.05.010
Marshall, H. H., Sanderson, J. L., Mwanghuya, F., Businge, R., Kyabulima, S., Hares, M. C., Inzani, E., Kalema-Zikusoka, G., Mwesige, K., Thompson, F. J., Vitikainen, E. I. K., & Cant, M. A. (2016). Variable ecological conditions promote male helping by changing banded mongoose group composition. Behavioral Ecology, 27(4), 978–987. https://doi.org/10.1093/beheco/arw006
Martin, J. S., Ringen, E. J., Duda, P., & Jaeggi, A. V. (2020). Harsh environments promote alloparental care across human societies. Proceedings of the Royal Society B, 287(1933), 20200758. https://doi.org/10.1098/rspb.2020.0758
Massen, J. J. M., & Koski, S. E. (2014). Chimps of a feather sit together: Chimpanzee friendships are based on homophily in personality. Evolution and Human Behavior, 35(1), 1–8. https://doi.org/10.1016/j.evolhumbehav.2013.08.008
McElreath, R. (2020). Statistical Rethinking: A Bayesian Course with Examples in R and STAN. CRC Press.
McFarland, R., & Majolo, B. (2011). Grooming Coercion and the Post-Conflict Trading of Social Services in Wild Barbary Macaques. PLOS ONE, 6(10), e26893. https://doi.org/10.1371/journal.pone.0026893
McFarland, R., & Majolo, B. (2012). The occurrence and benefits of postconflict bystander affiliation in wild Barbary macaques, Macaca sylvanus. Animal Behaviour, 84(3), 583–591. https://doi.org/10.1016/j.anbehav.2012.06.010
McFarland, R., & Majolo, B. (2013). Coping with the cold: Predictors of survival in wild Barbary macaques, Macaca sylvanus. Biology Letters, 9(4), 20130428. https://doi.org/10.1098/rsbl.2013.0428
McShane, B. B., Gal, D., Gelman, A., Robert, C., & Tackett, J. L. (2019). Abandon Statistical Significance. The American Statistician, 73(sup1), 235–245. https://doi.org/10.1080/00031305.2018.1527253
Ménard, N. (2002). Ecological plasticity of Barbary macaques (Macaca sylvanus). Evolutionary Anthropology: Issues, News, and Reviews, 11(S1), 95–100. https://doi.org/10.1002/evan.10067
Mielke, A., Bruchmann, C., Schülke, O., & Ostner, J. (2021). Grooming interventions in female rhesus macaques as social niche construction. Animal Behaviour, 173, 105–114. https://doi.org/10.1016/j.anbehav.2021.01.001
Mielke, A., Preis, A., Samuni, L., Gogarten, J. F., Wittig, R. M., & Crockford, C. (2018). Flexible decision-making in grooming partner choice in sooty mangabeys and chimpanzees. Royal Society Open Science, 5(7), 172143. https://doi.org/10.1098/rsos.172143
Neumann, C., Duboscq, J., Dubuc, C., Ginting, A., & Engelhardt, A. (2011). Assessing dominance hierarchies: Validation and advantages of progressive evaluation with Elo-rating. Animal Behaviour, 82, 911–921. https://doi.org/10.1016/j.anbehav.2011.07.016
Noë, N., Whitaker, R. M., & Allen, S. M. (2018). Personality Homophily and Geographic Distance in Facebook. Cyberpsychology, Behavior, and Social Networking, 21(6), 361–366. https://doi.org/10.1089/cyber.2017.0615
Noë, N., Whitaker, R. M., Chorley, M. J., & Pollet, T. V. (2016). Birds of a feather locate together? Foursquare checkins and personality homophily. Computers in Human Behavior, 58, 343–353. https://doi.org/10.1016/j.chb.2016.01.009
Nuñez, C. M. V., Adelman, J. S., & Rubenstein, D. I. (2015). Sociality increases juvenile survival after a catastrophic event in the feral horse ( Equus caballus ). Behavioral Ecology, 26(1), 138–147. https://doi.org/10.1093/beheco/aru163
Ólafsdóttir, G. Á., & Magellan, K. (2016). Interactions between boldness, foraging performance and behavioral plasticity across social contexts. Behavioral Ecology and Sociobiology, 70(11), 1879–1889. https://doi.org/10.1007/s00265-016-2193-0
Otto, S. P. (2018). Adaptation, speciation and extinction in the Anthropocene. Proceedings of the Royal Society B: Biological Sciences, 285(1891), 20182047. https://doi.org/10.1098/rspb.2018.2047
Price, A. R. G., Ghazi, S. J., Tkaczynski, P. J., Venkatachalam, A. J., Santillan, A., Pancho, T., Metcalfe, R., & Saunders, J. (2014). Shifting environmental baselines in the Red Sea. Marine Pollution Bulletin, 78(1), 96–101. https://doi.org/10.1016/j.marpolbul.2013.10.055
R Core Team. (2020). R: A language and environment for statistical computing. R Foundation for Statistical Computing, (Version 3.6.3) [Computer software]. https://www.R-project.org
Redhead, D., McElreath, R., & Ross, C. T. (2021). Reliable network inference from unreliable data: A tutorial on latent network modeling using STRAND. PsyArXiv. https://doi.org/10.31234/osf.io/mkp2y
Revelle, W. (2018). psych: Procedures for Psychological, Psychometric, and Personality Research. https://CRAN.R-project.org/package=psych
Roubová, V., Konečná, M., Šmilauer, P., & Wallner, B. (2015). Whom to groom and for what? Patterns of grooming in female Barbary macaques (Macaca sylvanus). PloS One, 10(2), e0117298. https://doi.org/10.1371/journal.pone.0117298
RStudio Team. (2021). RStudio: Integrated Development Environment for R. (Version 1.4.1106) [R]. RStudio.
Schloerke, B., Cook, D., Larmarange, J., Briatte, F., Marbach, M., Thoen, E., Elberg, A., Toomet, O., Crowley, J., Hofmann, H., & Wickham, H. (2020). GGally: Extension to ‘ggplot2’ (Version 2.0.0) [Computer software]. https://CRAN.R-project.org/package=GGally
Semple, S., Harrison, C., & Lehmann, J. (2013). Grooming and Anxiety in Barbary Macaques. Ethology, 119(9), 779–785. https://doi.org/10.1111/eth.12119
Shrout, P. E., & Fleiss, J. L. (1979). Intraclass correlations: Uses in assessing rater reliability. Psychological Bulletin, 86(2), 420–428. https://doi.org/10.1037/0033-2909.86.2.420
Shutt, K., MacLarnon, A., Heistermann, M., & Semple, S. (2007). Grooming in Barbary macaques: Better to give than to receive? Biology Letters, 3(3), 231–233. https://doi.org/10.1098/rsbl.2007.0052
Sih, A., Bell, A., & Johnson, J. C. (2004). Behavioral syndromes: An ecological and evolutionary overview. Trends in Ecology & Evolution, 19(7), 372–378. https://doi.org/10.1016/j.tree.2004.04.009
Silk, J. B. (2007). Social Components of Fitness in Primate Groups. Science, 317(5843), 1347–1351. https://doi.org/10.1126/science.1140734
Sonnweber, R. S., Ravignani, A., Stobbe, N., Schiestl, G., Wallner, B., & Fitch, W. T. (2015). Rank-dependent grooming patterns and cortisol alleviation in Barbary macaques. American Journal of Primatology, 77(6), 688–700. https://doi.org/10.1002/ajp.22391
Stan Development Team. (2021). Stan Modeling Language Users Guide and Reference Manual [Computer software].
Stoffel, M., Nakagawa, S., & Schielzeth, H. (2018). rptR: Repeatability Estimation for Gaussian and Non-Gaussian Data. https://CRAN.R-project.org/package=rptR
Taborsky, B., & Oliveira, R. F. (2012). Social competence: An evolutionary approach. Trends in Ecology & Evolution, 27(12), 679–688. https://doi.org/10.1016/j.tree.2012.09.003
Testard, C., Larson, S. M., Watowich, M. M., Kaplinsky, C. H., Bernau, A., Faulder, M., Marshall, H. H., Lehmann, J., Ruiz-Lambides, A., Higham, J. P., Montague, M. J., Snyder-Mackler, N., Platt, M. L., & Brent, L. J. N. (2021). Rhesus macaques build new social connections after a natural disaster. Current Biology, 0(0). https://doi.org/10.1016/j.cub.2021.03.029
Thompson, M. E., & Wrangham, R. W. (2008). Diet and reproductive function in wild female chimpanzees (Pan troglodytes schweinfurthii) at Kibale National Park, Uganda. American Journal of Physical Anthropology, 135(2), 171–181. https://doi.org/10.1002/ajpa.20718
Tkaczynski, P. J. (2016). The behavioral ecology of personality in wild Barbary macaques [Ph.D., University of Roehampton]. https://pure.roehampton.ac.uk/portal/en/studentthesis/the-behavioural-ecology-of-personality-in-wild-barbary-macaques(023582d2-2214-448c-bf12-c1bef7d5549e).html
Tkaczynski, P. J., Mielke, A., Samuni, L., Preis, A., Wittig, R. M., & Crockford, C. (2020). Long-term repeatability in social behavior suggests stable social phenotypes in wild chimpanzees. Royal Society Open Science, 7(8), 200454. https://doi.org/10.1098/rsos.200454
Tkaczynski, P. J., Ross, C., MacLarnon, A. M., Majolo, B., & Lehmann, J. (2019). Measuring personality in the field: An in situ comparison of personality quantification methods in wild Barbary Macaques (Macaca sylvanus). Journal of Comparative Psychology, in press.
Tkaczynski, P., Ross, C., Lehmann, J., Mouna, M., Majolo, B., & MacLarnon, A. (2019). Repeatable GC expression is associated with behavioral syndromes in males but not females in Barbary macaques. https://doi.org/10.6084/m9.figshare.7701701.v3
Waterman, J. O., Campbell, L. a. D., Maréchal, L., Pilot, M., & Majolo, B. (2019). Effect of human activity on habitat selection in the endangered Barbary macaque. Animal Conservation, n/a(n/a). https://doi.org/10.1111/acv.12543
Whitehead, H. (2008). Analyzing Animal Societies: Quantitative Methods for Vertebrate Social Analysis. University of Chicago Press.
Wickham, H., Averick, M., Bryan, J., Chang, W., D’Agostino McGowan, L., François, R., Grolemund, G., Hayes, A., Henry, L., Hester, J., Kuhn, M., Lin Pederson, T., Bache, S. M., Müller, K., Ooms, J., Robinson, D., Seidel, D. P., Spinu, V., Takahashi, K., … Yutani, H. (2019). Welcome to the tidyverse. Journal of Open Source Software, 4(43), 1686. https://doi.org/10.21105/joss.01686
Wolf, M., & Weissing, F. J. (2010). An explanatory framework for adaptive personality differences. Philosophical Transactions of the Royal Society B: Biological Sciences, 365(1560), 3959–3968. https://doi.org/10.1098/rstb.2010.0215
Young, C., Majolo, B., Heistermann, M., Schülke, O., & Ostner, J. (2013). Male mating behavior in relation to female sexual swellings, socio-sexual behavior and hormonal changes in wild Barbary macaques. Hormones and Behavior, 63(1), 32–39. https://doi.org/10.1016/j.yhbeh.2012.11.004
Young, C., Majolo, B., Heistermann, M., Schülke, O., & Ostner, J. (2014). Responses to social and environmental stress are attenuated by strong male bonds in wild macaques. Proceedings of the National Academy of Sciences, 111(51), 18195–18200. https://doi.org/10.1073/pnas.1411450111

No competing interests reported.

suppsocialcompetencebes.docx

Download PDF

Journal Publication

published 19 Nov, 2025

Read the published version in Behavioral Ecology and Sociobiology →

Editorial decision: Revision requested
03 Jul, 2025
Reviews received at journal
26 Jun, 2025
Reviews received at journal
25 Jun, 2025
Reviewers agreed at journal
12 Jun, 2025
Reviewers agreed at journal
22 May, 2025
Reviewers invited by journal
16 May, 2025
Editor assigned by journal
07 May, 2025
Submission checks completed at journal
07 May, 2025
First submitted to journal
01 May, 2025

You are reading this latest preprint version

Behavioral syndromes are associated with social plasticity and competence in a wild primate

Status:

Journal Publication

Version 1

Abstract

Figures

Significance statement

Introduction

Methods

Study site and subjects

Behavioral data collection

Socioecological variables

Statistical analysis

Dominance rank calculations

Behavioral syndrome quantification

Social plasticity in decision making in relation to current social environment

Plasticity in social connectivity in relation to fluctuating socioecology

Network construction and metrics

Social connectivity models

Model fitting and validation

Results

Model 1: Plasticity in social decision making in relation to current social environment

Models 2 & 3: Plasticity in social connectivity

Out-degree

Out-strength

Discussion

Declarations

Data availability

Competing interests

References

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1