Self-Directed Behaviors Prior to Approaching Conspecifics in Wild Tibetan Macaques (Macaca thibetana)
ABSTRACT
Self-directed behaviors (SDBs) are commonly used as reliable indicators of anxiety, frustration, and stress in nonhuman primates. These behaviors are typically observed in stressful situations, such as being in close proximity to dominants or during post-conflict. However, there is limited evidence on whether individuals increased their SDBs before engaging in potentially risky behaviors. In this study, we used a focal sampling method to record and analyze SDB rates in wild Tibetan macaques (Macaca thibetana) before approaching another group member (as a proxy for potentially risky behavior). Specifically, we examined the 5-min period (selected from focal samplings) before approaching a conspecific to address two main questions: (1) how rank and kinship affect SDB rates before approaching behavior, and (2) when individuals exhibit the highest anxiety, as indicated by peak SDB rates. Our results show that: (1) both males and females exhibited higher rates of SDBs before approaching a higher- than a lower-ranking group member of the same sex; (2) females showed no difference in SDB rates before approaching a female kin (sister or mother/daughter) compared to approaching a non-kin; (3) lower-ranking males exhibited higher SDB rates when approaching a receptive female, but not when approaching a non-receptive female; (4) females showed the highest SDB rates in the last minute (minute 5) before approaching a higher-ranking female. These findings suggest that Tibetan macaques increase their SDBs before engaging in potentially risky behaviors, providing an insight into the psychological state of an individual who is about to engage in a social interaction.
Abbreviations
-
- ANOVA
-
- one-way analysis of variance
-
- APP
-
- approaching behavior
-
- CBI
-
- Clutton-Brock et al.'s index
-
- GLMM
-
- generalized linear mixed model
-
- LSD
-
- least square differences
-
- SDBs
-
- self-directed behaviors
-
- VIFs
-
- variance inflation factors
1 Introduction
Self-directed behaviors (hereafter SDBs), such as self-scratching in response to changes or uncertainty in the social environment, are widely used as indicators of anxiety, internal conflict, and stress in nonhuman primates (Maestripieri et al. 1992; Manson and Perry 2000; Daniel et al. 2008; Reamer et al. 2010; Yamanashi and Matsuzawa 2010; Ellis et al. 2011; Zhang et al. 2014). Physiological and pharmacological studies have demonstrated a correlation between SDB rates and anxiety (Ninan et al. 1982; Insel 1984; Crawley et al. 1985; Schino et al. 1996). Ethological studies on primate sociality have generally indicated an increase in SDBs when subordinate individuals are in close proximity to dominants or in post-conflict situations (Castles et al. 1999; Manson and Perry 2000; Daniel et al. 2008; Zhang et al. 2014). Therefore, in this context, dominance relationships are thought to influence their frequency of SDBs. In species such as rhesus macaques (Macaca mulatta) and Japanese macaques (M. fuscata), dominance rank appears to follow a linear hierarchy, with males being dominant to females, and each sex exhibiting a distinct, sex-based hierarchy (Thierry 2007). For species exhibiting a linear dominance hierarchy, close proximity to dominant group members—especially during periods of resource scarcity (e.g., food or mates)—can create uncertainty and risk for lower-ranking individuals (Daniel et al. 2008). In some studies, researchers have suggested that SDB rates decline more rapidly to baseline levels following reconciliation between former opponents in post-conflict situations than when no reconciliation occurs (Aureli et al. 1989; Aureli and Schaik 1991; Castles and Whiten 1998). In female-philopatric primate species, related females tend to form strong alliances characterized by frequent social interactions. These individuals are more likely to reconcile after conflicts, particularly in despotic species (Aureli 1997). This suggests that interactions between related individuals may reduce SDB rates. Aside from the influence of existing stressors (e.g., close proximity to dominants or post-conflict scenarios), Maestripieri et al. (1992) proposed that self-directed behaviors are also generally associated with potentially risky or competitive situations. For example, Kummer (1968) observed that male hamadryas baboons (Papio hamadryas) increased their scratching rates during the initial stages of group movement coordination, driven by their conflicting impulses from the centripetal movement of females and the actions of neighboring males. Similarly, Smuts (1985) reported a female olive baboon (Papio anubis) scratched several times, seemingly undecided about which individual to approach for feeding. However, such observations were mostly anecdotal, with limited quantitative data directly examining whether individuals increase their SDB rates before engaging in potentially risky behaviors.
In the present study, we investigated the relationship between SDBs and potentially risky behaviors in one group of wild adult Tibetan macaques (M. thibetana), which exhibit a linear dominance hierarchy and despotic dominance style (Berman et al. 2004; Li and Kappeler 2020). Tibetan macaques live in multi-male, multi-female social groups, characterized by female philopatry and male dispersal (Li and Kappeler 2020). Females, but not natal males, display strong affiliative kin bias and social tolerance (Balasubramaniam et al. 2020). In this study, we examined the SDB rates in Tibetan macaques before approaching a conspecific (as a proxy of potentially risky behavior). Accordingly, we propose four predictions as follows.
P1: Adult females will exhibit higher rates of SDBs before approaching (1) a higher-ranking same-sex member compared to a lower-ranking one, and (2) a non-kin same-sex member compared to related one (sister or mother/daughter). Considering that female Tibetan macaques exhibit linear, despotic hierarchy, females are expected to show higher SDB rates before approaching an higher-ranking, related female than lower-ranking, related same-sex member.
P2: Males will exhibit higher SDB rates before approaching a higher-ranking than a lower-ranking same-sex member.
P3: Given that dominant males are generally intolerant of subordinate males' mating behavior (Li and Kappeler 2020), lower-ranking males will exhibit higher SDB rates before approaching a receptive female. In contrast, male dominance rank will not affect SDB rates before approaching a non-receptive female.
P4: Assuming that anxiety peaks in the final minute before approaching a higher-ranking conspecific, both males and females are expected to exhibit higher SDB rates in the last minute (minute 5) compared to earlier minutes (minute 1, minute 2, minute 3, minute 4).
2 Methods
2.1 Ethical Statement
All research protocols reported in this manuscript were approved by the Chinese Wildlife Management Authority, and adhere to the American Society of Primatologists Ethical Principles for the Treatment of Nonhuman Primates. Since this study solely involves observational research on wild primates without any invasive procedures, it does not require review by an institutional ethics committee in China. All research activities complied with China's Wildlife Protection Law and local regulations (Huangshan Forestry and Garden Bureau).
2.2 Study Site and Subjects
From August 2017 to January 2019, we collected behavioral data from a free-ranging group of Tibetan macaques at Mt. Huangshan National Reserve in Anhui province, China. The study group is known as Yulingkeng1 (YA1) and has been continually observed since 1986. Matrilineal kin relationships (mother, daughter, half siblings) are known for all females in the study group. However, such relationships are not known for adult males, as they migrate from their natal group. In this study, we included only mother-daughter dyads (YM-YH, YM-YCY, YM-YCLA, YH-YXX, TH-TXX, TH-TXH, TT-TR, TT-TRY, TR-THX, TR-THY) and sister-sister dyads (YH-YCY, YH-YCLA, YCY-YCLA, TXX-TXH, TR-TRY, THX-THY) as kin relationships, while all other dyads were defined as non-kin (refer to Xia et al. 2012). During the study, there were 46–48 group members including 15 adult females and 8-10 adult males. The female dominance hierarchy remained stable throughout our study period (Table 1A). In contrast, due to rank replacements and male immigration and emigration, the male dominance hierarchy was divided into four distinct periods (Table 1B).
| Female ID | Dominance rank | Number of valid samples (APP) |
|---|---|---|
| YH | 1 | 23 |
| YXX | 2 | 24 |
| YCY | 3 | 19 |
| YM | 4 | 6 |
| YCL | 5 | 7 |
| TXH | 6 | 12 |
| HH | 7 | 5 |
| TH | 8 | 15 |
| TXX | 9 | 19 |
| TR | 10 | 9 |
| TRY | 11 | 4 |
| THX | 12 | 7 |
| THY | 13 | 7 |
| TT | 14 | 6 |
| YZ | 15 | 4 |
| TOTAL | 167 |
| Male ID | Period 1 (2Aug-15Nov) | Period 2 (16Nov-26Mar) | Period 3 (Mar27-26Aug) | Period 4 (27Aug-15Jan) | Number of valid samples (male-male APP) | Number of valid samples (male-female APP) |
|---|---|---|---|---|---|---|
| HXM* | 1 | 2 | 2 | 3 | 3 | 7 |
| TG* | 2 | EMI | EMI | EMI | 0 | 0 |
| ZB | 3 | 3 | 3 | 4 | 2 | 10 |
| YRB* | 4 | 4 | EMI | EMI | 4 | 6 |
| BT | 5 | 5 | 4 | 5 | 2 | 7 |
| DS | 6 | 6 | 5 | 6 | 0 | 11 |
| HM | 7 | 7 | 6 | 7 | 1 | 7 |
| TRG* | 8 | 1 | 1 | 2 | 6 | 7 |
| YCLO* | 9 | 8 | 7 | 1 | 11 | 7 |
| YRQ* | 10 | EMI | EMI | EMI | 0 | 0 |
| ZF | — | — | — | 8 | 0 | 2 |
| HL | — | — | 8 | 9 | 1 | 3 |
| TOTAL | 30 | 67 |
- Note: “EMI” indicates that the male emigrated from the YA1 group during that study period.
- * Males were born into the group. Individual “HL” immigrated into the YA1 group during the period 3, and “ZF” immigrated into the YA1 group during the period 4.
2.3 Data Collection
We conducted focal animal sampling over a continuous 30-min period on a single adult male or adult female and recorded their behavior and the identity of other individuals they interacted with (Altmann 1974). To balance our data, we sampled each focal individual randomly, with no subject observed twice before all others were recorded. Approaching behavior (hereafter APP) was defined as one individual moving to a position within 1 m of another individual. We selected a distance of one meter because it represents a range within which one individual could quickly touch or come into contact with another. During our focal sampling, self-directed behaviors (SDBs) including self-scratching (defined as the movement of the hand or foot during which the tips of the manual or pedal digits are drawn across one's fur or skin) and self-grooming (defined as picking through and/or slowly brushing aside fur with one or both hands) were recorded and grouped together for analysis (Manson and Perry 2000; Zhang et al. 2014). We scored SDBs in terms of frequency rather than duration. If, after > 2 s of engaging in a SDB, the focal animal engaged in another type of SDB (e.g., switching from self-scratching to self-grooming), this was scored as a new bout (Manson and Perry 2000; Daniel et al. 2008; Zhang et al. 2014). When the focal individual exhibited an SDB, we recorded its occurrence to calculate the rate of SDBs during each minute across a maximum period of 5 min. However, we did not record other self-directed behaviors such as yawning, self-touching, and shaking, which can also be associated with anxiety or tension. This decision was based on a previous study of this Tibetan macaque group, which found that self-scratching (68.6%) and self-grooming (22.2%) account for 90.8% of all self-directed behaviors (Zhang et al. 2014).
- 1.
Any 5-min period of continuous observation before the focal animal “A” approaching another adult individual “B”.
- 2.
During this 5-min period, “A” did not approach any other individual, and “A” and “B” were not located within 1 m of each other before the initial approach. If “A” approached a second individual “C” during a 5-min period before approaching “B”, we excluded this from our analysis. This was done because the dominance and/or kin relationship between “A” and “C” might have affected the SDB rates of “A” during the 5-min period before approaching “B”.
- 3.
We excluded instances in which the focal individual “A” was involved in any conflict within 5 min of approaching “B.” This was done because the rate of SDBs during the first 5 min after a conflict is significantly higher than in the absence of a conflict (Zhang et al. 2014).
And, (4) if the focal individual approached several adult individuals at a distance of < 1 m, we assumed that the nearest individual was the target of the approach and included this interaction in our analysis.
We determined the dominance ranks of all adult group members using David's score, which classifies aggressive-submissive interactions. David's scores reflect an individual's social rank within the group, with higher scores indicating a higher position in the dominance hierarchy. For individuals with equal scores, we applied Clutton-Brock et al.'s index (CBI) using Pij values (the proportion of individual i winning interactions with individual j) to refine their ranks (for additional details, see Clutton-Brock et al. 1979; Gammell et al. 2003 and Zhang et al. 2014). Aggressive behavior was scored when an individual stared, raised its eyebrows, chased, slapped, grabbed, or bit another individual (for detailed definitions, see Berman et al. 2007; Li 1999). Submission was scored when one individual in a conflict bared its teeth, cowered, attempted to leave, avoided the other individual, or screamed (for detailed definitions, see Berman et al. 2004; Li 1999). During the study, we recorded 330 aggressive-submissive interactions among females and 130 aggressive-submissive interactions among males.
All behavioral data were collected by a single observer (QXZ) using a digital voice recorder (ICD-PX470, SONY China, Beijing, China).
2.4 Data Analysis
For the total focal samples, we used an independent t-test to compare SDB rates between males and females. To test P1 and P2, we used the 5-min samples to examine the effect of the relationship between the focal individual and the approached same-sex individual on SDB rates. We ran two generalized linear mixed models (GLMMs), with focal individual ID, approached individual ID, and dyads as random factors. One GLMM was run for adult females. In this model, the rate of SDBs during each 5-min period before approaching behavior was set as the dependent variable, with predictor variables including the relative dominance rank of A to B (binomial: higher/lower) and the kin relationship between A and B (binomial: kin/non-kin). We also included the interaction between relative dominance rank and kin relationship to explore whether the influence of kinship on SDB rates varied with relative dominance rank. Kin relationships were categorized as close kin (mother-daughter dyads, sister-sister dyads) or non-kin (more distant kin relationship or non-kin relationship). A second GLMM was run for adult males. In this model, we set the rate of SDBs during each 5-min period before approaching behavior as the dependent variable, with the relative dominance rank of the approached individual (binomial: higher/lower) as predictor variable.
To test P3, we used 5-min samples to investigate whether a higher-ranking male imposes a stressful situation for a male when approaching a receptive female. We ran a GLMM with the focal male's SDB rates during each 5-min period before approaching a female as the dependent variable. Predictor variables included the dominance rank of the focal male and the receptivity of the approached female (scored binomially as receptive or non-receptive based on copulation or visible signs such as moist, pink genitals, sometimes with residual semen). We also included the interaction between male dominance rank and female receptivity to examine whether the influence of female receptivity on SDB rates changed across the range of male dominance ranks. Focal individual ID, the ID of the individual approached, and dyads were included as random factors. Male dominance rank was centered for interpretation of main effects.
All GLMMs were built based on a Poisson error structure and the log-link function in R v.4.1.3 (R Development Core Team, 2022) using the glmer function of the R package lme4 (Bates et al. 2012). Likelihood ratio tests were performed using the R function ANOVA to assess the significance of the full model versus the null model (only random factors included). To address potential collinearity among variables, we checked variance inflation factors (VIFs, Field 2005) using the function vif in the R package car (Fox and Weisberg 2010). All VIFs were below 3 (range: 1.019–2.029), indicating acceptable collinearity (Bowerman and O'Connell 1990; Myers 1990). Additionally, when interactions in the GLMMs were or nearly significant, we performed post hoc analysis by splitting the data frames and re-running the models, as suggested by Sonnweber et al. (2015).
To test P4, which examined changes in SDB rates during the 5-min period before approaching another individual, we divided the 5-min period into five equal 1-min intervals. We compared SDB rates between these intervals using a one-way analysis of variance (one-way ANOVA). If significant differences were found, post-hoc tests (Least Significant Differences, LSD) were used to identify which time interval exhibited the highest SDB rates. Additionally, to better reflect SDB rates across the 5-min sampling period, we used paired t-tests to compare SDB rates for each minute during the 5-min sampling period with baseline SDB rates for both females and males. All analyses were conducted in R v.4.1.3 (R Development Core Team, 2020), and significance was set as p < 0.05.
3 Results
Based on data from all focal samplings, the mean ± SE rate of self-directed behaviors (SDBs) in adult female Tibetan macaques was 0.207 ± 0.010 per observation minute, while in adult males, it was 0.186 ± 0.013 per observation minute. There was no significant difference in SDB rates between males and females (t = 1.245, n = 25, p = 0.226).
We tested the prediction (P1) that adult females exhibited higher rates of SDBs before approaching a higher-ranking than lower-ranking, non-kin compared to related same-sex group members. As the interaction between kinship and dominance rank in the GLMM did not significantly affect the focal female's rate of SDBs and was consequently removed, we reran the reduced GLMM model without the interaction. The results showed that females significantly increased their SDBs during the 5-min period before approaching a higher-ranking female compared to a lower-ranking female (standardized estimate = 0.6219 ± 0.2037, p = 0.002; Table 2). However, kinship did not significantly influence the focal females' SDB rates (standardized estimate = 0.0392 ± 0.2675, p = 0.883; Table 2). Thus, P1 was partially supported.
| Independent variable | Estimate | SE | Z value | p |
|---|---|---|---|---|
| Intercept | 0.2572 | 0.2111 | 1.218 | 0.223 |
| Rank (Higher) | 0.6219 | 0.2037 | 3.053 | 0.002 |
| Kinship (Kin) | 0.0392 | 0.2675 | 0.147 | 0.883 |
- Note: Results for a GLMM with 167 samplings, focal female ID (N = 15) and approached female ID (N = 15) as random factors. Significant results are in bold.
For prediction P2, males showed a significant increase in SDB rates before approaching a higher-ranking male compared to a lower-ranking male (standardized estimate = 1.4115 ± 0.5019, p = 0.005; Table 3). Thus, P2 was supported.
| Independent variable | Estimate | SE | Z value | p |
|---|---|---|---|---|
| Intercept | 0.2289 | 0.4116 | 0.556 | 0.578 |
| Rank: Higher | 1.4239 | 0.5086 | 2.800 | 0.005 |
- Note: Results for a GLMM with 30 samples of focal male ID (N = 8 adult males) and approached male ID (N = 11 adult males) as random factors. Samplings selected did not include YRQ, TG, and ZF as the focal member. Significant results are in bold.
We next tested whether lower-ranking males exhibited higher rates of SDBs than higher-ranking males before approaching a receptive female (P3). As shown in Table 4, the interaction between male rank and female receptivity approached significance in influencing male SDB rates. A post hoc analysis revealed that the lower a male's rank, the higher the rates of SDBs he exhibited when approaching a receptive female (χ2 = 7.96, p = 0.005). In contrast, male dominance rank did not significantly affect SDB rates when approaching a non-receptive female (χ2 = 0.06, p = 0.780; Figure 1). These results provide support for prediction P3.
| Independent variable | Estimate | SE | Z value | p |
|---|---|---|---|---|
| Intercept | −0.1252 | 0.2263 | −0.553 | 0.580 |
| Rank | 0.0084 | 0.0838 | 0.100 | 0.920 |
| Receptive state | −0.0450 | 0.3122 | −0.144 | 0.886 |
| Rank* Receptive state | 0.2257 | 0.1235 | 1.828 | 0.068 |
We hypothesized that anxiety would increase as lower-ranking individuals approach higher-ranking conspecifics, particularly in the final minute of the approach (minute 5), compared to earlier intervals within the 5-min period (minute 1, 2, 3 or 4; P4). A one-way ANOVA revealed a significant difference in SDB rates among the five 1-min intervals when females approached higher-ranking females (F = 2.975, p = 0.041, see Table 5 and Table 6). Post-hoc tests (LSD) showed that SDB rates during the final minute (minute 5) were significantly higher than in earlier intervals, except for minute 3 (see Table 6 and Figure 2; minute 5 vs. minute 1: p = 0.006; minute 5 vs. minute 2: p = 0.010; minute 5 vs. minute 3: p = 0.068; minute 5 vs. minute 4: p = 0.022).
| Terms | SDBs rates | ||||
|---|---|---|---|---|---|
| Minute 1 | Minute 2 | Minute 3 | Minute 4 | Minute 5 | |
| Females approach higher-ranking females | 0.165 ± 0.076 | 0.193 ± 0.071 | 0.314 ± 0.106 | 0.242 ± 0.078 | 0.582 ± 0.151 |
| Females approach lower-ranking females | 0.075 ± 0.031 | 0.137 ± 0.051 | 0.224 ± 0.089 | 0.178 ± 0.040 | 0.163 ± 0.067 |
| Males approach higher-ranking males | 0.167 ± 0.096 | 0.389 ± 0.056 | 0.278 ± 0.278 | 0.167 ± 0.096 | 0.611 ± 0.278 |
| Males approach lower-ranking males | 0.178 ± 0.097 | 0.000 | 0.222 ± 0.222 | 0.133 ± 0.133 | 0.111 ± 0.111 |
| Terms | One sample ANOVA | Post-hoc tests (LSD) | ||||||
|---|---|---|---|---|---|---|---|---|
| F value | df | p | 1 | 2 | 3 | 4 | 5 | |
| Females approach higher-ranking females | 2.759 | 44 | 0.041 | a | a | ab | a | b |
| Females approach lower-ranking females | 0.859 | 54 | 0.495 | — | ||||
| Males approach higher-ranking males | 0.982 | 14 | 0.460 | — | ||||
| Males approach lower-ranking males | 0.394 | 14 | 0.808 | — | ||||
- Note: Post-hoc tests (LSD) were performed if the one-way ANOVA results were significant. Significant results are in bold.
Next, we compared SDB rates in the 5-min period before a female approached a higher-ranking female with baseline SDB rates using paired t-tests. The results showed that only in the fifth minute did the SDB rates (mean ± SE = 0.58 ± 0.15) significantly exceed baseline SDB rates (minute 1: t = −0.630, df = 8, p = 0.546; minute 2: t = −0.352, df = 8, p = 0.734; minute 3: t = 0.946, df = 8, p = 0.372; minute 4: t = 0.322, df = 8, p = 0.756; minute 5: t = 2.379, df = 8, p = 0.045). In contrast, no significant differences in SDB rates were observed across the five 1-min intervals before a higher-ranking female approaching a lower-ranking female (one-way ANOVA, F = 0.859, p = 0.495).
Adult males did not show differences in SDBs rates across the five 1-min intervals when a lower-ranking male approached a higher-ranking male (one-way ANOVA, F = 0.982, p = 0.460) or when a higher-ranking male approached a lower-ranking male (one-way ANOVA, F = 0.394, p = 0.808). Furthermore, paired t-tests comparing SDB rates per minute during the 5-min focal period when a male approached a higher-ranking male with baseline SDB rates showed no significant differences (minute 1: t = −0.301, df = 2, p = 0.792; minute 2: t = 2.807, df = 2, p = 0.107; minute 3: t = 0.276, df = 2, p = 0.808; minute 4: t = −0.340, df = 2, p = 0.766; minute 5: t = 1.425, df = 2, p = 0.290). Thus, P4 was supported for adult females but not for adult males.
4 Discussion
In the present study, we examined self-directed behaviors (SDBs), an indicator of anxiety, in the 5-min period before adult male and adult female Tibetan macaques approaching a conspecific. This allowed us to explore the association between self-directed behavior and the potential risks involved in approaching a higher-ranking individual. Our results showed that females exhibited increased SDB rates before approaching a higher-ranking than lower-ranking same-sex group member. However, they did not experience an increase in SDB rates when approaching a nonkin compared to approaching a related female (mother, daughter, or sister) (P1 was partially supported). For adult males, significantly higher SDB rates were observed before approaching a higher-ranking compared to a lower-ranking same-sex group member (P2 was supported). Furthermore, lower-ranking males exhibited higher rates of SDBs before approaching a receptive female compared to higher-ranking males (P3 was supported). This finding reflects the strict linear male dominance hierarchy and reproductive competition for access to fertile females. Finally, while adults females showed the highest SDB rates in the final minute (minute 5) before approaching a higher-ranking same-sex conspecific, adult males exhibited relatively consistent SDB rates throughout the 5-min period (P4 was supported for adult females but not for adult males).
These findings provide direct evidence that Tibetan macaques experience higher SDB rates before approaching a higher-ranking same-sex conspecific, with females, in particular, showing increased anxiety in the final minute before the approach. This pattern of behavior is similar to that observed in hamadryas baboons (Papio hamadryas), as mentioned in the introduction (Kummer 1968), where individuals face a internal conflict between the potential risks of aggression and the benefits of affiliating with higher-ranking individuals. Tibetan macaques, with their despotic and linear dominance hierarchy, may be particularly sensitive to the risks posed by higher-ranking individuals, yet still recognize the potential benefits of interacting with them (Berman et al. 2007; Xia et al. 2012).
Contrary to expectations, we found no reduction in SDBs before females approaching a close kin compared to non-kin females, despite the strong kin bias in affiliation behaviors such as grooming and co-feeding (Berman et al. 2004). This suggests that the predictability of the relationship, as indicated by SDB rates (Cords 1988; Cords and Aureli 1993; Castles et al. 1999), is primarily influenced by relative dominance rank rather than kinship. This finding adds a new dimension to our understanding of female-female relationships in Tibetan macaques, urging us to move beyond the simple concept of “kin bias” and consider the complex interplay of dominance and kinship.
Our results also revealed that lower-ranking males exhibited higher SDB rates before approaching a receptive female. This is supported by a study indicating that lower-ranking males are often attacked by higher-ranking males if they attempt to mate with a receptive female (Li and Kappeler 2020). Tibetan macaques have a promiscuous mating system, with both males and females interrupting copulations of others, particularly higher-ranking males interrupting lower-ranking males' mating attempts (Li and Kappeler 2020). Berman et al. (2007) reported that, in the presence of an alpha male, beta and gamma males were tolerated to mate with receptive females during the mating season. However, we did not observer any such tolerance throughout our study period, indicating a strict linear dominance hierarchy among males. In this study, due to the limited visibility in the forest, we did not distinguish whether the approach of lower-ranking males toward receptive females occurred within or outside the sight of high-ranking males. However, this does not affect the accuracy of these results, as in the highly gregarious groups of Tibetan macaques (Li 1999; Xia et al. 2012), particularly during the mating season, high-ranking males often move around the group to monitor and prevent lower-ranking males from mating with receptive females (Li 1999). Therefore, even when not directly within the sight of high-ranking males, lower-ranking males remain nervous when attempting to approach and mate with receptive females, as high-ranking males may appear nearby at any moment and attack them.
Female Tibetan macaques showed a peak in SDB rates during the final minute (minute 5) before approaching a higher-ranking female conspecific, suggesting that females might plan their behavior in the short-term but are less likely to engage in long-term planning. This contrasts with the relatively consistent SDB rates seen in males throughout the 5-min period, which might suggest that males do not exhibit the same level of anticipatory anxiety or planning. One possible explanation for this difference is that, while both males and females exhibit strong linear dominance hierarchies, females experience more despotic social conditions with fewer conciliatory tendencies and lower rates of counter-aggression (Berman et al. 2004), leading to higher anxiety when approaching a higher-ranking same-sex individual.
In conclusion, our study provides evidence that both male and female Tibetan macaques increase their SDB rates before approaching a higher- rather than lower-ranking same-sex conspecific, with females exhibiting the highest SDB rates in the final minute before approach. This study is one of the few to investigate the SDB rates in the period preceding a risky behavior, offering new insights into the psychological state of an individual who is about to engage in a social interaction.
Author Contributions
Qi-Xin Zhang: conceptualization (lead), formal analysis (lead), funding acquisition (supporting), writing – original draft (lead), writing – review and editing (equal). Xi Wang: writing – original draft (supporting), writing – review and editing (equal). Yong Zhu: conceptualization (supporting), formal analysis (supporting), writing – review and editing (equal). Paul A. Garber: conceptualization (supporting), writing – review and editing (equal). Dong-Po Xia: conceptualization (supporting), funding acquisition (supporting), writing – review and editing (equal). Jin-Hua Li: conceptualization (supporting), funding acquisition (lead), writing – review and editing (equal).
Acknowledgments
This study was conducted at Mt. Huangshan China; we thank the Huangshan Monkey Management Center and the Huangshan Garden Forest Bureau for their permission for us to conduct research at the field site. We give special thanks to Mr. HB Cheng's family for their outstanding logistic support of our study. Natural Science Foundation of China (32370535; 32070455), Natural Science Research Project of Anhui Educational Committee (2023AH051279), and Research Foundation for Talented Scholars of Hefei Normal University (2022rcjj47).
Conflicts of Interest
The authors declare no conflicts of interest.
Open Research
Data Availability Statement
The data set generated and analysed during the study is available in the open figshare repository, 10.6084/m9.figshare.25377766.
