Introduction

In mammals, vocal signals can encode information about the caller’s identity, sex, genetic quality, dominance status, and emotional state [1–8], and vocalizations often play an important role in synchronizing the activities and movements of group members, maintaining contact between widely dispersed individuals, and reducing uncertainty about the outcome of social interactions [9–13]. For example, African elephants (Loxodonta africana) use low-frequency harmonically-rich rumbles to maintain contact with distant members of their social networks [11]. These calls contain acoustic information about individual sex, size, identity, and emotional state [14,15]. Northern elephant seals (Mirounga angustirostris) recognize the rhythm and timbre of other males’ threat vocalizations, and use this information to adjust their behavioral responses [16]. Meerkats (Suricata suricatta) give moving calls before they start to move, and group moves are more likely to occur when at least three individuals join in a calling chorus [17]. Female rhesus macaques (Macaca mulatta) give contact calls (grunts, girneys, coos) as they approach other members of their groups, and these calls are reliable signals that the callers will not behave aggressively [18]. Thus, vocal signals are a rich source of information for animals in social groups.

Playback studies have revealed “the exquisite skills of receivers, who take in and weigh information from different sources, to make adaptive decisions” ([19] page 76). Listeners’ responses to calls are influenced by a variety of factors, including their knowledge of the caller’s identity, the type of call that is given, and the history of their recent interactions with the caller [13]. For example, when female vervet monkeys (Chlorocebus pygerythrus) hear the scream of an infant, they look toward the infant’s mother [20].

Less is known about the factors that influence the production and usage of calls, partly because it is more difficult to experimentally manipulate call production than call perception. However, in an experiment in which chimpanzees were exposed to a snake model, chimpanzees were more likely to give alarm calls in the presence of individuals who were aware that the snake was present than in the presence of individuals who were unaware of the snake’s presence [21]. Female Japanese macaques (Macaca fuscata) are more likely to exchange contact calls (coos) with the females that they groom most often than with other females [22]. Nonetheless, our knowledge of the factors that influence call usage in most taxa remains limited.

There have been a number of studies of the acoustic structure, function, and usage of contact calls (grunts) by chacma baboons (Papio ursinus). Grunts are low amplitude, harmonically rich calls [23] which are most often given as the group begins to move through an open area [24, 25] and in social contexts [25–31]. Grunts are individually identifiable [23,32] and listeners seem to be able to determine when they are the target of others’ grunts [33]. As in rhesus macaques [18], grunts given in social contexts seem to function as signals of benign intent and facilitate nonaggressive social interactions, such as infant handling [28,31]. Grunts are also effective in reconciling conflicts with former opponents [26–28] and their close kin [29]. Analyses of naturally occurring sequences of interactions show that female chacma baboons take a variety of contextual factors into account when they decide whether to grunt or remain silent after approaching other females [34]. Females are more likely to grunt as they approach lower ranking females whom they might intimidate, threaten, or attack, than as they approach higher ranking females, whom they are unlikely to harass. In addition, females are less likely to grunt as they approach their own mothers and daughters, with whom they have highly affiliative and nonaggressive relationships, than they are to grunt as they approach other females. These data suggest that grunts reduce uncertainty about the likely outcome of an interaction between pairs of females whose relationship is not consistently affiliative.

Call structure is strongly influenced by phylogeny, but the usage of calls may differ among closely related species [35]. For example, the call repertoire of guinea baboons (Papio papio) is identical to the call repertoire of chacma baboons, but there are a number of differences in the ways that calls are used [36,37]. Male chacma baboons often emit loud calls in agonistic conflicts with other males and these calls provide reliable information about male fighting ability [38,39], while male guinea baboons (Papio papio) rarely use long calls in this context [37]. Female chacma and yellow baboons (Papio cynocephalus) routinely vocalize after copulating, and these calls are thought to incite male-male competition [40], but female guinea baboons rarely call after copulating [37]. These differences in the usage of vocalizations may reflect species-specific differences in social organization and reproductive strategies, as male guinea baboons form stronger social bonds and compete less vigorously for access to females than male chacma baboons do [37]. Studies closely related species can provide valuable insights about the selective factors that underly variation in call production and usage.

Here, we examine the factors that influence the function and usage of grunts by female olive baboons. There are a number of similarities in the social organization and behavior of olive and chacma baboons [41,42]. For example, both species form multi-male, multi-female groups from which males disperse at puberty. Like female chacma baboons, olive baboons form their strongest bonds with their own mothers and mature daughters and establish linear matrilineal dominance hierarchies. Females in both species are attracted to newborn infants and seem strongly motivated to handle them [30,43,44]. These similarities provide an important opportunity to determine whether the patterns observed in chacma baboons are robust. Based on the similarities in social organization between olive and chacma baboons, we predicted that olive baboons would use grunts in much the same way that chacma baboons do. Thus, we predicted that grunts would function as signals of benign intent in olive baboons, and would be associated with a higher probability that peaceful interactions would occur and a lower probability that aggression would occur. We also expected females to grunt to mothers of young infants more than they grunted to other females. We predicted that grunts would be more likely to be given when the outcome of approaches was uncertain. Thus, females would grunt more to lower ranking females than to higher ranking females, and less to their own mothers and daughters than to other females. We also tested a novel prediction that proximity bouts that began with grunts would last longer than proximity bouts that did not begin with grunts. Finally, we extended the analysis to consider how grunts affected the likelihood that recipients of approaches would initiate aggression, infant handling, and affiliation toward females that approached them.

Results

Results

Approximately 12% of approach sequences began with grunts by the actor (0.12 ± 0.01, n = 822 dyads), and these grunts typically occurred within the first five seconds after the approach (average time elapsed: 5.35 ± 0.63).

Discussion

When female olive baboons grunt after they approach other females, they are less likely to behave aggressively toward their partners or intimidate them and less likely to be harassed or intimidated themselves. When females grunt as they approach, they are also more likely to handle their partners’ infants, and more likely to initiate affiliation or receive affiliation from their partners. Females are also likely to remain in proximity longer if they approach and grunt than if they approach and remain silent. Thus, grunts seem to act as reliable signals of benign intent, which are effective in inhibiting aggression from partners, and facilitating nonaggressive social interactions. These results support and extend previous findings derived from observations of chacma baboons [34], and suggest that this is a robust phenomena.

In both chacma and olive baboons, decisions about whether to grunt or remain silent after approaches to other females seem to depend in part on their partners’ likely response to their approach. This is influenced by the presence of infants, partners’ relative rank, and their genetic relationship. In both species, females are substantially more likely to grunt as they approach mothers of young infants than as they approach other females, which suggests that there may be a need for signalers to communicate to mothers of young infants that their intentions are benign. Similarly, females are also more likely to grunt to females who are lower ranking than themselves, and potential victims of aggression, than to those that are higher ranking than themselves and relatively safe from harassment. They are also less likely to grunt to their own mothers and daughters, with whom they maintain close and tolerant social bonds, than they are to grunt to other females.

Although female olive baboons and female chacma baboons seem to be influenced by similar factors when they decide whether to grunt or remain silent, there are some differences in their behavior as well. Overall, the females in our study groups were less likely to grunt after they approached other females than the chacma baboons were (this study: 14% vs 28%, ref. 34). In addition, the magnitude of the odds ratios associated with the predictor variables were all lower for the olive baboons than for the chacma baboons. It is possible these differences reflect subtle variation in the dynamics of social relationships among females in these two species or populations. For example, nepotistic biases in social behavior are less pronounced in our study groups than in chacma baboons or yellow baboons [42], and this might influence females’ decisions about whether to grunt as they approached their relatives. It is also possible that methodological differences may have affected the results. In this study, approaches to within 1 m were recorded while approaches to within 2 m were recorded for the chacma baboons [34].

In summary, female olive and chacma baboons seem to use grunts to signal their intention to behave peacefully. Decisions about whether to call or to remain silent after an approach, are influenced by contextual factors, among which we have identified the relative rank of the two females, whether an infant is present, and genetic relationship between the actor and recipient. The data support the hypothesis that signals of benign intent are particularly important for pairs of females whose relationships are not predictably affiliative and there is uncertainty about the likely outcome of an interaction. Taken together, these data also support the view that baboons and other primates may “exhibit a flexibility in call usage that is similar to the flexibility that they display when responding to the calls of others.” ([13]:1975). At the same, however, subtle differences in the usage of grunts may reflect differences in the dynamics of social relationships across populations or species. Further research is needed on other species to assess the extent of flexibility in call usage and identify the factors that influence animals’ decisions about whether or not to call.

Acknowledgments

We thank the Office of the President of the Republic of Kenya and the Kenya Wildlife Service for permission to conduct this field research. We thank Kate Abderholden, Megan Best, Megan Cole, Moira Donovan, Alexandra Duchesneau, Jessica Gunson, Molly McEntee, Sam Patterson, Laura Peña, Vance Reeds, and Leah Worthington for their contributions to data collection. We thank the staff of the Uaso Ngiro Baboon Project, particularly Jeremiah Lendira, James King’au, Joshua Lendira, and Frances Molo for their help in the field; David Muiruri for invaluable assistance with logistics and data management; and the African Conservation Centre for facilitating the UNBP project and assisting us with our work. We thank Dorothy Cheney for useful discussions of the data, and two reviewers for useful criticisms of a previous version of the paper. The research on which this paper was based was supported with funds to JBS from Arizona State University. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.