Vocal Repertoire of Sooty Mangabeys (Cercocebus Torquatus Atys) in The

Ethology 110, 301—321 (2004) 2004 Blackwell Verlag, Berlin ISSN 0179–1613

Vocal Repertoire of Sooty Mangabeys (Cercocebus torquatus atys)in the Taı¨ National Park

Friederike Range* & Julia Fischer *Department of Psychology, University of Pennsylvania, Philadelphia, PA, USA; Max-Planck Institute for Evolutionary Anthropology, Leipzig, Germany

Abstract Three factors, environment, quality and type of interaction, and phylogeny have been hypothesized to influence the structure of signal repertoires in primates. Much is known about vocal repertoires of terrestrial, savannah-dwelling species and arboreal, rainforest-dwelling species, but very little is known about terrestrial rainforest species.To fill this knowledge gap and to further elucidate how the three factors influence vocal repertoires of primates, we designed a study on sooty mangabeys (Cercocebus torquatus atys), a terrestrial Old World primate that lives in dense rainforests.Recordings of sooty mangabeys in their natural environment were used to compile the vocal repertoire of this species.All calls are described according to the basic acoustical structure and the behavioral context in which they occurred.Descriptions are supplemented by quantitative measurements of call occurrence in all age–sex classes.For the most frequently produced vocalizations, a preliminary acoustical analysis was conducted to test for individual and contextual differences.Finally, we compare vocalizations of sooty mangabeys with vocalizations of several other primate species and discuss how the factors mentioned above might influence the vocal repertoire of sooty mangabeys. Corresponding author: Friederike Range, Department of Psychology, University of Pennsylvania, 3720 Walnut Street, Philadelphia, PA, 19104-6196, USA. E-mail: [email protected]

Introduction One key means of understanding the evolution of non-human primate vocal behavior is to investigate the selective pressures that shape the acoustic structure of vocalizations.Several factors appear to aﬀect the structure of a signal repertoire.First, sound propagation, as measured by attenuation, reverberation and amplitude, depends on characteristics of the local habitat (Marten & Marler 1977; Waser & Waser 1977; Waser & Brown 1984, 1986).Thus, it is likely that the

U. S. Copyright Clearance Center Code Statement: 0179-1613/2004/11004–301/$15.00/0 www.blackwell-synergy.com 302 F.Range & J.Fischer acoustic properties of vocalizations will be shaped to reduce distortion in the environment through which the signal is normally transmitted (the Ôlocal adaptation hypothesisÕ (Gish & Morton 1981)).Secondly, the quality and type of interactions among group members may affect signal diversity (Maestripieri 1996).However, the way individuals interact is in turn influenced by the species Õ habitat.Marler (1975, 1976 hypothesized that continuous acoustic variation between and/or within signal types should evolve when individuals inhabit relatively open habitat and interact at high rates and in face-to-face interactions with conspecifics.In contrast, signals with no intermediates between call types should be favored when auditory signals must operate without accompanying visual or other contextual cues; for example, in forest habitats or when being broadcast over long distances (Marler 1975, 1976).Thirdly, phylogenetic descent is expected to play a substantial role as the structure of primate calls is largely under genetic control (Ju¨ rgens 1992). To date, much is known about the vocal repertoires of terrestrial monkey species living in open habitat, such as baboons and vervet monkeys (e.g. Struhsaker 1967; Cheney & Seyfarth 1980; Seyfarth et al.1980a; Cheney & Seyfarth 1982a; Cheney et al.1996; Rendall et al.1999; Fischer et al.2001), and of arboreal monkey species living in rainforests, such as Diana monkeys, grey- cheeked mangabeys and spider monkeys (Chalmers 1968; Waser & Waser 1977; Chapman & Weary 1990; Zuberbu¨ hler 1995; Teixidor & Byrne, 1999).Some clear differences in vocal repertoires between these two groups are apparent, but we do not yet know whether the differences are due primarily to habitat (open terrain vs. forest), to locomotion (terrestrial vs.arboreal), or both.To date, very little information is available on the vocal repertoire of terrestrial rainforest species (but see Waser 1982).To fill this gap, we collected data on the vocalizations of the sooty mangabey (Cercocebus torquatus atys), a terrestrial, forest-dwelling monkey species that lives in West Africa. Traditionally, all mangabey species have been combined into a single genus, Cercocebus, (Schwartz 1928; Booth 1956; Napier & Napier 1967) that was commonly divided into two species: the torquatus-group and the albigena-group. However, recent molecular studies suggest that these species-groups are paraphyletic (Barnicott & Wade 1970; Barnicott & Hewett-Emmett 1972; Cronin & Sarich 1976; Groves 1978; Disotell 1994, 1996; van der Kuyl et al.1995; Fleagle & McGraw 1999) and should be placed in two separate genera.Accordingly, the torquatus-group, including the sooty mangabey, is assigned to the genus Cercocebus, which is most closely related to forest-dwelling mandrills and drills (Mandrillus spp.), whereas the albigena-group belongs to the genus Lophocebus, which is most closely related to open-terrain baboons (Papio spp.) and geladas (Theropithecus).Species of both genera are mainly forest dwelling, but while the Cercocebus group is mainly terrestrial, the Lophocebus group is arboreal. As virtually nothing is known about the vocal repertoire of sooty mangabeys, and even less about the communication of their closest relatives, the mandrills and drills, this study fills a gap in our knowledge about the vocal behavior of an important group of Old World monkeys.We attempt to understand how the Vocal Repertoire of Sooty Mangabeys 303 structure of the signal repertoire of free-ranging sooty mangabeys is influenced by environment, quality and type of interaction, and phylogeny, by (1) describing the vocal repertoire and presenting sonograms of the most frequently produced vocalizations, (2) examining the behavioral context in which specific call types are produced, (3) conducting a preliminary acoustic analysis of the most frequent vocalizations (grunts, twitters and copulation calls) to test if vocalizations vary between the sexes, between individuals and between contexts, thus giving us an idea about the quality of interactions, and finally (4) by comparing these aspects of sooty mangabey vocalizations with the same aspects of vocalizations of other, arboreal and terrestrial, savannah- and forest-dwelling non-human primate species.

Methods Study Site and Subjects The study was conducted over a 16-month period (April 2000–December 2001) on free-ranging sooty mangabeys in the Taı¨ National Park in Ivory Coast (620¢Nto510¢N and 420¢Wto650¢W).The park is the last remaining major block of primary forest in West Africa and covers approximately 4 54 000 ha. Visibility ranges from 5 to 20 m throughout the home range of our study group. During the study period, group size ranged between 98 and 122 animals.The group has been under study since 1997.All animals are well habituated to human observers and can be recognized individually.Because of political problems in Ivory Coast, this study was terminated earlier than planned and thus sample sizes are small.However, we feel that the present data give a good preliminary overview of the vocal repertoire of the sooty mangabey.

Data Collection Behavioral data to determine use and hourly rates of vocalizations were collected by focal animal sampling (Altmann 1974).Focal samples were 15 min long with at least 60 min between consecutive samples of the same individual. During observations, we used instantaneous sampling (Altmann 1974) to record the individual’s general activity (foraging, traveling, social interaction or resting). Social interactions and vocalizations were recorded continuously according to a detailed ethogram (Range & Noe¨ 2002). Behavioral data were recorded by F.Range on 24 adult females from April– August 00 and on 12 juveniles (seven females, ﬁve males) from May– December 01.Y.Meystre also collected behavioral data on 12 adult males (October 00–April 01).As there was no temporal overlap between the two observers, no inter-observer reliability tests were possible. Vocalizations were tape-recorded only from May–December 2001 using a Sony-DAT PCM-M1 recorder and a Sennheiser directional microphone (ME 68). Vocalizations were collected opportunistically.Whenever a call was recorded, we 304 F.Range & J.Fischer also noted the caller’s identity, its activity and social behavior (for details of the ethogram see Range & Noe¨ 2002).

Data Analysis For each call type, we calculated vocalization time as the percentage of minutes during focal animal sampling when the focal individual used a specific call type at least once.Hourly rates of vocalizations for individuals were calculated by dividing the total number of calls given by the individual during focal samples by the sum of the total observation time for that individual.Spectrograms were generated with Avisoft-SASLab (R.Specht, Berlin, Germany). To document the acoustic features of the sooty mangabey vocal repertoire we used only vocalizations tape-recorded from known adult individuals.For a summary of all call types used for acoustic analysis and sample sizes see Table 1. All statistical analyses were performed on individual means. Tape-recorded vocalizations of sooty mangabeys were categorized by ear. For male grunts and copulation calls we measured the number of units per bout and bout duration by hand using Cool Edit (Syntrillium, Phoenix, AZ, USA). Call types used for the acoustic analyses (grunts, twitters and copulation calls) were submitted to the SIGNAL sound analysis system (Beeman, 1996) and a fast Fourier transform was conducted (grunts: 1024-pt FFT, time step: 3 ms, frequency resolution: 11 Hz; twitter: 1024-pt FFT, time step: 3 ms, frequency resolution: 46 Hz).The resulting frequency time spectra were analyzed with a custom software program, LMA 8.4 (Hammerschmidt, 1990). The program extracts different call parameters that describe the acoustic structure of each call. The parameters used in the present analysis are defined in the footnote of Table 3. Parameters that were used for analyzing the different call types were determined after a preliminary data analysis.If a vocalization type was given in bouts (male grunts) or consisted of several syllables (twitters and copulation calls), we conducted acoustic analyses on the third syllable or call in the sequence.It is a somewhat arbitrary choice, but it confers the advantage of avoiding variation caused by warming up or running out of breath.The decision was made after inspecting the spectrograms, and finding that the third syllable seems to be least variable within call types.

Table 1: Distribution of tape-recorded vocalizations used for the acoustic analyses per call and per individual

Total calls used for No.of calls per Type of vocalization the acoustic analysis Individuals individual (range)

Grunts Adult males 38 5 7–8 Adult females 46 5 9–10 Twitter 145 31 1–18 Copulation calls 34 5 6–8 Vocal Repertoire of Sooty Mangabeys 305

Statistical Analysis We hypothesized that grunts would vary between the sexes and between individuals, that copulation calls would differ between individuals, and that twitters would vary with social context.To test these hypotheses, we used discriminant function analyses.Discriminant function analysis identifies a linear combination of quantitative predictor variables that best characterize the differences among groups.Variables are combined into one or more discriminant functions.Based on these discriminant functions, the classification procedure assigns each call to its appropriate group (correct assignment) or to another group (incorrect assignment).For external validation we used the cross-validation classification procedure, in which each case is classified by the functions derived from all cases other than that one. For the analyses of grunts and copulation calls, we tested six to ten calls per individual; for twitters the number varied between one and six calls per female per context.All analyzed grunts were recorded during foraging of the subject to control for contextual differences.Calls were selected based on quality. We also conducted multivariate analyses of variance (manova) to test for statistical significance of the observed differences between groups.Alpha was set at 0.05, trends were reported for 0.1 > alpha > 0.05. SPSS (version 7.5.1.) was used to perform the discriminant function analysis, and JMP (version 4.0.3) was used to perform the manova.

Results Grunts Grunts (Fig.1) are short (102–188 ms, see Table 3), low frequency vocalizations given in a variety of contexts.They are the most frequent vocalization given by sooty mangabeys (Table 2).A clear difference in the temporal structure of call bouts was observed between adult males and adult females.Adult males usually grunted several times in a row with regular intervals of about 207 ms between grunts (six males, 45 call bouts), whereas inter-call intervals in female grunts ranged from a few seconds to a few minutes, and varied between contexts. Most grunts were produced while animals were engaged in foraging activities such as searching or feeding.Sooty mangabeys also grunted during social interactions such as approaching, embracing or, in the case of adult males and some juveniles, dominance interactions.A few times, animals were also observed to grunt while sitting or traveling.Table 2 summarizes frequencies (rates per hour) and proportions of grunts recorded in various contexts during focal animal sampling for different sex–age classes. To test for individual distinctiveness between male and female grunts of sooty mangabeys, we first conducted a discriminant function analysis with ÔsexÕ as the grouping variable.The average correct assignment was 97.6%; the cross- validation yielded an average correct assignment of 97.6%. We found significant multivariate differences between the sexes (F ¼ 5.57, d.f. ¼ 1, p < 0.021). 306

Table 2: Rates and proportions of vocalizations given across diﬀerent contexts presented separately for adult males, adult females and juveniles. N refers to the number of individuals analyzed in each category

Sex/age class

Call type Parameters Adult females (N ¼ 24) Adult males (N ¼ 12) Juveniles (N ¼ 12)

Grunts Average rate per hour (range) 2.12 ± 1.41 (0.43–6.33) 2.98 ± 2.86 (0.20–10.62) 1.33 ± 1.08 (0.39–4.45) % (±SD) of all vocalizations 69.2 ± 14.5 78.1 ± 25.2 47.7 ± 14.1 % (±SD) of grunts given when foraging 69.0 ± 14.0 46.0 ± 14.9 63.3 ± 15.1

% (±SD) of grunts given in the context 7.93 ± 8.52 20.3 ± 10.6 13.8 ± 13.8 J.Fischer & F.Range of greeting % (±SD) of grunts given during 0 5.6 ± 10.1 2.8 ± 5.8 dominance interactions Twitters Average rate per hour (range) 0.46 ± 0.39 (0–1.29) 0 0.86 ± 0.82 (0–2.10) % (±SD) of all vocalizations 15.8 ± 12.6 0 24.5 ± 19.1 % (±SD) of twitters given when foraging 59.3 ± 38.2 0 65.0 ± 27.9 Screams Rates per hour (range) 0.06 ± 0.07 (0–0.18) Not recorded once 0.35 ± 0.28 (0–0.86) during focal sampling % (±SD) of all vocalizations 3.0 ± 4.3 0 13.5 ± 11.4 % (±SD) of screams given in the context 92.3 ± 18.8 0 16.7 ± 27.7 of contact aggression % (±SD) of screams given in the context 7.7 ± 18.8 0 83.3 ± 27.7 of redirected aggression Growls Rates per hour 0.16 ± 0.17 (0–0.67) Not recorded once 0.30 ± 0.16 (0–0.55) during focal sampling %(±SD) of all vocalizations 6.2 ± 7.9 0 23.3 ± 9.6 Table 3: Acoustic parameters of grunts, twitters and copulation calls (for the latter two, only the ﬁrst elements were analyzed)

Grunts (foraging) Twitter

Adult females Adult males Foraging Social Social/infant Watching Copulation (N ¼ 5) (N ¼ 5) (N ¼ 20) (N ¼ 16) (N ¼ 11) (N ¼ 13) calls (N ¼ 5)

Parameter Mean SD Mean SD Mean SD Mean SD Mean SD Mean SD Mean SD

No.of elements M 11.8 4.6 11 3 Grunt rate (s)MC 0.26 0.05 0.49 0.15 Df1chfreSFTC 36.58 18.8 52 22 44.2 6.4 43.1 11.8 41.8 12.9 45.5 11.0 51.1 14.6 Mangabeys Sooty of Repertoire Vocal Df1fretrFMTC 11.1 4.7 10 4 18 5 21 7 20 8 19 8 10 5 Df1malocFCT 0.64 0.37 0.52 0.35 0.27 0.26 0.27 0.24 0.25 0.36 0.24 0.27 0.44 0.33 Df1mean (Hz)FMT 248 69 311 41 2106 334 1934 318 1849 394 2170 368 389 79 Df2mean (Hz)T 444 224 – – 3609 629 3508 567 3159 873 3731 653 1282 1009 Q1mean (Hz)SFMCT 350 84 285 40 2364 287 2185 354 2101 391 2398 458 384 71 Q2mean (Hz)SFCT 488 114 385 55 3237 578 3036 647 2832 538 3124 707 548 186 Ranmax (Hz)FT 978 666 997 1126 6329 1548 6603 1631 5539 1467 6064 1815 2125 1848 Ranmean (Hz)FMT 704 283 530 66 4371 1053 4322 1021 3624 1076 3983 1045 948 242 Duration (ms)SFMT 113 18 187 24 51 31 35 9 38 7 52 28 115 45

N ¼ Number of females or males that provided calls for the analysis. SParameters used for analysis of sex differences in grunts. FParameters used for female grunt analysis. M Parameters used for male grunt analysis. TParameters used for twitter analysis. CParameters used for copulation call analysis. No.of elements ¼ Number of elements per bout; Rate ¼ Number of elements divided by the total duration of the bout; Df1chfre ¼ Number of times the course of the first dominant frequency band crosses the smoothed average, measure for the local modulation; Df1fretr ¼ Regression function describing the trend of the 1st dominant frequency band; Df1maloc ¼ Location of maximum of the 1st dominant frequency band, ranges between 0–1; Df1mean (Hz) ¼ Average of the first dominant frequency band across all time segments; Df2mean (Hz) ¼ Average of the second dominant frequency band across all time segments; Q1mean (Hz) ¼ Frequency at which the amplitude distribution reaches the first quartile of the total distribution, mean across all time segments; Q2mean (Hz) ¼ Frequency at which the amplitude distribution reaches the second quartile of the total distribution, mean across all time segments; Ranmax (Hz) ¼ Maximum frequency range (difference between lowest and highest frequency); Ranmean (Hz) ¼ Average frequency range; Duration (ms) ¼ Time between beginning and end of call. 307 308 F.Range & J.Fischer

Fig. 1: (a) Grunt emitted by an adult female during foraging.(b) Grunt of the same females as in a, emitted during grooming of another female.(c) Another grunt of the same female who is approaching and embracing a female with an infant.(d) One grunt from an adult male during foraging.(e) A male grunting bout.In comparison with adult females, males usually grunted several times in a row with regular intervals between grunts, whereas inter-call intervals in female grunts ranged from a few seconds to a few minutes and varied between contexts

We then separated males and females and tested whether grunts were individually distinctive.About 78.3%of the selected female grunts were classified correctly (Fig.2).However, cross-validation revealed only 45.7%correct assignment.The manova did not yield significant differences between individual females (F ¼ 1.53, d.f. ¼ 4, NS).For adult males, the classification procedure yielded an average correct assignment of 84.2% (Fig. 2). Cross-validation assigned 68.4% correctly, and the manova revealed significant differences among males (F ¼ 3.02, d.f. ¼ 4, p < 0.032). Table 3 shows the acoustic parameters used for these analyses.

Twitters Mangabey twitters ranged from soft and melodic sounds to harder, almost harsh sounds (Fig.3).Twitters given by adult females during foraging and social interactions were always given as trains with up to 23 syllables.Each syllable consisted of several, irregular modulated frequency bands ranging between 1 and 20 kHz.Twitters were only heard from adult females and juveniles of both sexes and occurred mostly in similar behavioral contexts as grunts (for behavioral Vocal Repertoire of Sooty Mangabeys 309

Fig. 2: Distribution of the discriminant scores for (a) grunts of five individual adult females (two letter codes) and (b) grunts of five adult males (three letter codes) along two canonical discriminant functions established to discriminate among the five individuals in each sex class.Group centroids are represented by gray squares.The plots illustrate that the grunts given by individual females (a) and individual males (b) differ from each other 310 F.Range & J.Fischer measurements see Table 2).Most twitters occurred when animals were searching for food.Seven females twittered exclusively in this context whereas five others never twittered at all.Twitters produced by adult females during searching were

(a)

(b)

(c)

(d)

Fig. 3: (a) Twitter of a foraging adult female.(b) Twitter of an adult female handling an infant. (c) and (d) Twitters from two adult females sitting between two subgroups Vocal Repertoire of Sooty Mangabeys 311 answered by nearby animals either by emitting the same call type or by grunting in 21.8% of 177 calls. Furthermore, juveniles and adult females twittered when approaching or grooming other individuals, especially in the presence of infants (see above). Finally, twitters were recorded when adult females sat between subgroups. During low-food availability, groups of sooty mangabeys are widely dispersed when foraging and distances between small groups of individuals can reach up to 100 m (own data).In these circumstances, adult females were observed producing a certain type of twitter (Fig.3c, d) not heard in any other context.Typically, this vocalization was accompanied by the signaler looking in the direction of another subgroup.As soon as one subgroup approached the other, the vocalizing female would stop calling and join the other animals. For six females we had at least one high quality recording in a social context with or without an infant (12 calls with an infant and 16 calls without). Discriminant analysis revealed a correct assignment to the two contexts in 89.3%. Cross-validation yielded an average correct assignment of 64.3%. Moreover, we analyzed differences between the foraging and social contexts without an infant (see above) for 10 females with at least one call in each context (32 and 30 calls respectively).Only 66.1%of the original calls were correctly assigned and cross-validation assigned only 54.8% correctly. Finally, we tested for contextual differences between the foraging context and the social context with an infant.Analysis revealed 96.3%correct assignment of twitters from five females with a total of 16 calls in the foraging context and 11 calls in the social context.Cross-validation assigned 74.1calls correctly.The data set for twitters given between subgroups was too small for any analysis.A summary of the acoustic parameters for all twitters recorded in the various contexts is given in Table 3.

Screams Screams are noisy vocalizations produced only in agonistic interactions (Fig.4).They comprised both tonal and harsh elements.The tonal structure – fundamental with several accompanying overtones – was often overlaid with wide-band noise up to 15 kHz.The intensity of this noise differed between calls so that in some calls the harmonic structure was not visible at all (Fig.4a).In other calls, harmonic structure became visible in the beginning of a scream (Fig.4b) or appeared occasionally throughout the call (Fig.4c). Screams were recorded from juveniles and adult females but only once by an adult male (see Table 2).Even though conflicts between adult males were frequent, adult males were heard screaming only five times during the entire study.Screaming by adult and juvenile animals occurred in two sub-categories of aggressive interactions.In context 1, screams occurred during contact aggression when, after a conflict, the aggressor took the tail of the loser into its mouth.During the tail bite, the loser crouched to the ground and screamed (Fig.4a).Screaming in this context was observed in all sex–age classes except 312 F.Range & J.Fischer

Fig. 4: (a) Adult female screaming after tail bite from another, higher ranking female.(b) Adult female screaming at a sub-adult male (c) Adult female screaming at an adult male (d) Adult male screaming at another adult male.(e) Infant chattering and then screaming at its mother who resists the attempts of her infant to nurse adult males.Overall rates were rather low for adult females (0.06h )1) and slightly higher for juveniles (0.07 h)1).In context 2, individuals, as a response to an attack, screamed at the aggressor.Most frequently, this behavior was observed of juveniles challenging either other juveniles or adult females (0.28 h)1).Adult females always screamed when they responded aggressively towards an aggressor, but the overall rate was very low (0.01 h)1).Most of the attacks were directed against adult or sub-adult males (Fig.4b, c).In contrast Vocal Repertoire of Sooty Mangabeys 313 to older animals, infants screamed in a wider variety of behavioral contexts (Fig.4e).

Growl All tape-recorded growls comprised several acoustically similar syllables that were sometimes combined with other call types (see below).Each growl syllable consisted of a low fundamental frequency band with several accompanying overtones (Fig.5a, b).Growls occurred in agonistic interactions with other group members (Table 1).Typically, the higher-ranking individual in a conﬂict raised its eyebrows, stared and growled at the opponent.Aggressors would sometimes alternate this behavioral sequence with rapid turns of its head towards other high-ranking individuals, which could result in agonistic support.

Other Threat Vocalizations Four other vocalizations were recorded during agonistic interactions: the grumble (Fig.5c), the hoo (Fig.5d), the intense threat (Fig.5e), and the wau (Fig.5f).All were tape-recorded only 3–5 times during this study.The grumble lasted between 266 and 482 ms and had a fundamental frequency between 100 and 200 Hz (three calls of three different individuals).It had a rich harmonic structure with up to 12 overtones, was produced by individuals in all sex and age classes, and was heard in combination with growls during the same type of agonistic interactions described above.The hoo call had a rich harmonic structure and was also heard in combination with growling, but only from juveniles or sub- adult animals.The intense-threat call sounded like a combination of both the twitter and the alarm call and, in comparison with the grumble and the hoo call was heard only a few times during intense fights.The wau call type was heard only five times during the entire study.Adult males emitted the call as they watched an intense fight between other adult males.

Copulation Call Copulation calls could last up to 10 s and were given as phrases with up to 51 units (Fig.6). In Taı ¨ , sooty mangabeys exhibited a distinct mating season. Females emitted copulation calls mainly during copulations, but occasionally also during defecation.Males did not vocalize during copulation, but occasionally grunted after ejaculation. We performed a discriminant function analysis to test for individual diﬀerences in the copulation calls of ﬁve females for whom we had at least six high quality recordings of copulation calls when the females copulated with adult or sub-adult males.About 97.1%of the copulation calls were assigned to the correct group, cross-validation assigned 82.4% correctly. The statistical analysis showed a trend towards individual distinctiveness of copulation calls (manova: 314 F.Range & J.Fischer

(a) (b)

(e) (f)

Fig. 5: (a) Adult female growling.(b) Third syllable of the growl.(c) Adult female grumbling.(d) Call element ÔhooÕ recorded during growling of a 5-yr old male.(e) The intense threat call given in intense ﬁghts.(f) ÔWauÕ call of adult male

F ¼ 2.42, d.f. ¼ 4, p < 0.072). Table 3 shows the acoustic parameters used for these analyses.

Whoop Gobble Whoop gobbles are loud, low frequency, long calls given exclusively by adult males (Fig.7).All long calls started with a low-pitched introductory note, the whoop, which is separated from the rest of the call by several seconds (up to 6.7 s). After the whoop, the call is preceded by a repetitive series of one or two types of tonal syllables, the gobble.The ﬁrst type was lower-pitched, longer and Vocal Repertoire of Sooty Mangabeys 315

Fig. 6: Copulation call of an adult female with maximum sexual swelling

Fig. 7: Whoop gobble of the alpha male more frequency modulated than the second type.Whoop gobbles were heard in the morning or when another mangabey group was nearby.The call was also heard in combination with sightings or actual attacks of predators.In all contexts, the males would usually call several times and during predator encounters they would often alternate whoop gobbles with alarm calls (see below).

Alarm Calls Both males and females produced loud alarm calls.Sooty mangabeys gave alarm calls mainly towards three different predators: snakes, eagles and leopards (Fig.8). Sooty mangabeys encountered Gabon vipers (Bitis gabonica) two to three times per week and these snakes always elicited alarm calls.Upon hearing an alarm call, other mangabeys approached the vocalizing animal, climbed 1 or 2 m up into a tree and scanned the forest floor. One of the main predators of sooty mangabeys in Taı¨ is the crowned eagle (Shultz & Noe¨ , 2002; Shultz, 2001).Although predation attempts were observed infrequently (three times in 6 months), the eagle was detected at least four or five times per week perching in a tree or flying over the canopy close to the group, eliciting alarm calls by mangabeys.If the eagle flew over the group, mangabeys ran down the trees and looked towards the sky.If, however, a perched eagle was detected, mangabeys gave alarm calls while approaching the eagle until the eagle flew off. Leopards have been observed to attack members of our observation group four times.Typically, sooty mangabeys jumped up into a tree upon hearing or seeing a leopard and gave alarm calls. Sooty mangabeys give different alarm calls to the three predator types. Although we do not have enough tape recordings for a statistical analysis, 316 F.Range & J.Fischer

Fig. 8: (a) Snake (adult female).(b) Snake (adult male).(c) Eagle (adult female).(d) Eagle (adult male).(e) Leopard (adult female).(f) Leopard (adult male).(g) Group encounter (adult male) humans can learn to distinguish the alarm calls by ear.The sonograms presented in Fig.8 also indicate an acoustical diﬀerence between those calls.

Discussion The sooty mangabey vocal repertoire consists of 19 vocalizations that differ from each other either in their acoustical structure or in the behavioral context with which they are associated.The number of audibly distinct calls given by sooty mangabeys is similar to the number of audibly distinct call types given by other primate species (e.g. baboons (Papio spec.): 15 (Rowell & Hinde 1962); Macaca fuscata: 37 (Itani 1963); Cercopithecus aethiops: 36 (Struhsaker 1967); Gorilla gorilla: 23 (Schaller 1963); Pan troglodytes: 25 (Goodall 1965)).However, repertoire size is difficult to measure in any species when calls are classified solely Vocal Repertoire of Sooty Mangabeys 317 by ear or acoustic features.Playback experiments on grunts of vervets (Seyfarth et al.1980b) and baboon (Rendall et al.1999) and on the alarm calls of vervets (Cheney & Seyfarth 1981), baboons (Fischer et al.2001), Barbary macaques (Fischer 1998; Fischer & Hammerschmidt 2001) and Diana monkeys (Zuberbu¨ hler 1995) – to name just a few – have shown that the animals themselves distinguish between different call subtypes in ways that are not initially apparent to humans. Moreover, the way the calls are analyzed, and the authorsÕ tendency to split or lump, also affects the interpretation of repertoire size.Therefore, the number of call types of a species may not be very meaningful. Some vocalizations of sooty mangabeys resemble vocalizations described for a wide range of species, both in their acoustical structure and behavioral context, suggesting that neither phylogeny nor environmental factors (e.g. habitat or locomotion) predominately influence the signal structure and function of these calls.For example, sooty mangabeys produce screams and growls in agonistic conflicts as do many other primate species (Gouzoules et al.1984; Gouzoules & Gouzoules 1990; Fischer & Hammerschmidt 2002).Grunts also seem to be similar in their acoustical structure and contexts of use (foraging, grooming, infant handling) to those in both terrestrial, savannah-dwelling (baboons, vervets) and arboreal, forest-dwelling species (grey-cheeked mangabeys (Cercocebus albigena), blue monkeys (Cercophithecus mitis), mustached guenon (Cercopithecus cephus), greater spot-nosed guenon (Cercopithecus nictitans) (Cheney & Seyfarth 1982a; Gautier & Gautier-Hion 1988; Brown 1989; Rendall et al.1999)).Some other arboreal rainforest monkey species, however, produce very different close-range vocalizations often referred to as trills (e.g. Cercophithecus or Colobus monkey species) (Uster & Zuberbu¨ hler 2001; personal observation). However, the basic acoustical structure and function of some other sooty mangabey vocalizations do suggest some influence of phylogenetic descent.Alarm calls of adult males closely resemble the ÔkarouÕ call given by Cercocebus galeritus and the two-phase bark or wahoo given by closely related baboon males (Waser 1982).They differ completely from alarm calls given by other arboreal rainforest primate species including Diana monkeys (Zuberbu¨ hler et al.1996) and Camp- bell’s monkeys, Cercopithecus campbelli (Zuberbu¨ hler 2001). Even though alarm calls and whoop gobbles may be influenced by phylogeny, phylogenetic descent cannot explain all acoustic features of these calls as they differ in several aspects from those of closely related species.In comparison with the Lophocebus alarm calls, for example, sooty mangabey alarm calls are differentiated into the initial set of staccato pulses and the low-pitched end unit, whereas in Lophocebus no clear end unit is apparent.Moreover, whoop gobbles have been described for both the Lophocebus and the Cercocebus mangabey species and a homologous call – the grating roar or roar grunt – has been described for baboons (Waser 1982).Although, baboons have a homologous call, the whoop gobble is rather different, which could suggest an adaptation to the local environment (savannah vs.forest). Finally, some sooty mangabey vocalizations seem rather species-specific and thus could be adaptations to the habitat and mode of locomotion.For example, 318 F.Range & J.Fischer although Barbary macaques (Macaca sylvanus) and some arboreal new-world monkeys (e.g. Samiri sciureus; Winter et al., 1966; Schott, 1975) produce calls comprising multiple elements that resemble the sooty mangabeysÕ twitter, the characteristics of the calls can differ appreciably.Whereas the units in twitters of sooty mangabeys most often comprise clear frequency sweeps, the pant-bark of the Barbary macaque consists of several noisy and unmodulated units. The last factor thought to influence the structure of vocalizations is the quality and type of interactions.Sooty mangabeys live in large multi-male/multi-female groups with well-differentiated social relationships.Our analysis of grunts suggests that calls differ between males and females, and that male grunts, at least, are individually distinguishable.At present, we have no good explanation for the sexual differentiation of call individuality.Possibly, some of the individual variation in male calls can be attributed to differences in body size, which is more pronounced than that among females.Copulation calls also showed a trend towards differences between individuals, allowing for the possibility of individual recognition. We also offer evidence that twitters differ somewhat across contexts, although cross-validations for the discrimination of twitters given in Ôsocial contexts with infantsÕ vs. Ôsocial contexts without infantsÕ and ÔsocialÕ vs. ÔforagingÕ were low. Moreover, alarm calls also seem to differ according to the predator causing the calls. Unfortunately, no detailed acoustical analysis is possible yet, but the notion that even humans can learn the difference between alarm calls given in response to snakes, eagles and leopards suggests that mangabeys convey semantic information about the predator present as has been reported for Diana monkeys, Campbell’s monkeys, vervet monkeys and Barbary macaques (Struhsaker 1967; Seyfarth et al. 1980a,b; Zuberbu¨ hler et al.1996; Fischer 1998; Zuberbu¨ hler 2001).These preliminary indications of individual and contextual differences within call types are in line with results from other primate and mammal species (Cheney & Seyfarth 1982b; Chapman & Weary 1990; Holekamp et al.1999; Fischer et al.2001; Rendall et al., 1996; Semple et al.2002; Teixidor & Byrne, 1999; Sayigh et al.1999). The goal of this paper was to increase our understanding of how the structure of the signal repertoire is influenced by the three factors environment, quality and type of interaction, and phylogeny.Our results suggest that phylogeny and/or the amount of terrestrial locomotion largely determine the crude acoustical structure of at least some vocalizations.A detailed comparison of the fine structure of acoustic parameters within call types across species might reveal adaptations to the local habitat within the basic acoustic structure of vocalizations, as has been shown for loud calls in mangabeys and baboons (Waser 1982).The information content of sooty mangabey vocalizations seems similar to information conveyed by vocalizations of other monkey species studied in this respect.

Acknowledgements We thank the Ministe` re de la Recherche Scientiﬁque and the Ministe` re de l’Agriculture et des Ressources Animales of Coˆ te d’Ivoire for permission to conduct our research in the Taı¨ National Park. We are grateful to the CSRS and the staﬀ of the CRE research station in Taı¨ for logistical support. Vocal Repertoire of Sooty Mangabeys 319

F.R. was supported by the ÔDAAD Doktorandenstipendium im Rahmen des gemeinsamen Hochschulsonderprogramms III von Bund und LaendernÕ.Richard Peho and Gerard Gha provided invaluable assistance in the ﬁeld by locating the study group and by collecting data.We thank the members of the Taı¨ Monkey Project for support in the ﬁeld and R.Noe¨ , R.Seyfarth, D.Kitchen and J.Crawford for comments on earlier drafts of the manuscript.

Literature Cited Altmann, J.1974: Observational study of behaviour: Sampling methods.Behaviour 49, 227—265. Barnicott, N.A.& Hewett-Emmett, D.1972: Red cell and serum proteins of Cercocebus, Presbytis, Colobus and certain other species.Folia.Primatol.(Basel) 17, 442—457. Barnicott, N.A.& Wade, P.T.1970: Protein structure and the systematics of Old World monkeys.In: The Old World Monkeys (Napier, J.P.& Napier, P.H.,eds).Academic Press, London, pp. 227—260. Beeman, K.1996: SIGNAL Operation Manual: Engineering Design, Belmont. Booth, A.H.1956: The Cercopithecidae of the Gold and Ivory Coast: geographic and systematic observations.Ann.Mag.Nat.Hist. 9, 476—480. Brown, C.H.1989: The active space of blue monkey and grey-cheeked mangabey vocalizations.Anim. Behav. 37, 1023—1034. Chalmers, N.R.1968: The social behaviour of free-living mangabeys in Uganda.Folia primatol. 8, 263—281. Chapman, C.A.& Weary, D.M.1990: Variability in spider monkeys vocalizations may provide basis for individual recognition.Am.J.Primatol. 22, 279—284. Cheney, D.L.& Seyfarth, R.M.1980: Vocal recognition in free-ranging vervet monkeys.Anim. Behav. 28, 362—367. Cheney, D.L.& Seyfarth, R.M.1981: Selective forces aﬀecting the predator alarm calls of vervet monkeys.Behaviour 76, 25—61. Cheney, D.L.& Seyfarth, R.M.1982a: How vervet monkeys perceive their grunts.Anim.Behav. 30, 739—751. Cheney, D.L.& Seyfarth, R.M.1982b: Recognition of individuals within and between groups of free- ranging vervet monkeys.Am.Zool. 22, 519—529. Cheney, D.L.,Seyfarth, R.M.& Palombit, R.1996: The function and mechanisms underlying baboon ÔcontactÕ barks.Anim.Behav. 52, 507—518. Cronin, J.E.& Sarich, V.M.1976: Molecular evidence for dual origin of mangabeys among Old World monkeys.Nature 260, 700—702. Disotell, T.R.1994: Generic level relationships of the papionini (Cercopithcoidea).Am.J.Phys. Anthropol. 94, 47—57. Disotell, T.R.1996: The phylogeny of old world monkeys.Evol.Anthropol. 5, 18—24. Fischer, J.1998: Barbary macaques categorize shrill barks into two call types.Anim.Behav. 55, 799—807. Fischer, J.& Hammerschmidt, K.2001: Functional referents and acoustic similarity revisited: The case of Barbary macaque alarm calls.Anim.Cogn. 4, 29—35. Fischer, J.& Hammerschmidt, K.2002: An overview of the Barbary Macaque, Macaca sylvanus, vocal repertoire.Folia Primatol.(Basel) 710, 1—14. Fischer, J., Hammerschmidt, K., Cheney, D. L. & Seyfarth, R. M. 2001: Acoustic features of female chacma baboon barks.Ethology 107, 33—54. Fleagle, J.G.& McGraw, W.S.1999: Skeletal and dental morphology supports diphyletic origin of baboons and mandrills.Proc.Natl Acad.Sci.USA 96, 1157—1161. Gautier, J.-P. & Gautier-Hion, A. 1988: Vocal quavering: a basis for recognition in forest guenons. In: Primate Vocal Communication (Todt, D., Goedeking, P. & Symmes, D., eds). Springer-Verlag, Heidelberg, pp.15—30. Gish, S.L.& Morton, E.S.1981: Structural adaptations to local habitat acoustics in Carolina wren songs.Zeitschrift fuer Tierpsychologie 56, 74—84. Goodall, J.1965: Chimpanzees in the Gombe stream reserve.In: Primate Behavior: Field Studies of Monkeys and Apes (DeVore, I., ed). Holt, Rinehart and Winston, New York, pp. 425—474. 320 F.Range & J.Fischer

Gouzoules, H.& Gouzoules, S.1990: Matrilineal signatures in the recruitment screams of pigtail macaques, Macaca nemestrina.Behaviour 115, 327—347. Gouzoules, H., Gouzoules, S. & Marler, P. 1984: Rhesus monkey (Macaca mulatta) screams: representational signalling in the recruitment of aid.Anim.Behav. 32, 182—193. Groves, C.P.1978: Phylogenetic and population systematics of the mangabeys (Primates: Cercopithecoidea).Primates 19, 1—34. Hammerschmidt, K.1990: Individuelle Lautmuster bei Berberaffen ( Macaca sylvanus): Ein Ansatz Zum Verstaendnis ihrer vocalen kommunikation.Berlin: Fu Berlin. Holekamp, K. E., Boydston, E. E., Szykman, M., Graham, I., Nutt, K. J., Birch, S., Piskiel, A. & Singh, M.1999: Vocal recognition in the spotted hyaena and its possible implications regarding the evolution of intelligence.Anim.Behav. 58, 383—395. Itani, J.1963: Vocal communication of the wild Japanese monkey.Primates 4, 11—66. Ju¨ rgens, U.1992: On the neurobiology of vocal communication.In: Nonverbal Vocal Communication (Papousˇ ek, H., Ju¨ rgens, U.& Papous ˇ ek, M., eds). Cambridge University Press, Cambridge, pp. 31—43. van der Kuyl, A. C., Kuiken, C. L., Dekker, J. T. & Goudsmit, J. 1995: Phylogeny of African monkeys based upon mitochondrial 12S rRNA sequences.J.Mol.Evol. 40, 173—180. Maestripieri, D.1996: Primate cognition and the bared-teeth display: A reevaluation of the concept of formal dominance.J.Comp.Psychol. 110, 402—405. Marler, P.1975: On the origin of speech from animal sounds.In: The Role of Speech in Language (Kavanaugh, J.F.& Cutting, J.E.,eds).MIT Press, Cambridge, Massachusetts, pp.11—37. Marler, P.1976: Social organization, communication and graded signals: the chimpanzee and the gorilla.In: Growing Points in Ethology (Bateson, P.P.G.& Hinde, R.A.,eds).Cambridge University Press, Cambridge, UK, pp.239—280. Marten, K.& Marler, P.1977: Sound transmission and its significance for animal vocalizations, 1: temperate habitats.Behav.Ecol.Sociobiol. 3, 271—290. Napier, J.R.& Napier, P.H.1967: Handbook of Living Primates.Academic Press, London. Range, F.& Noe¨ , R.2002.Familiarity and dominance relations in female sooty mangabeys in the Taı ¨ National Park.Am.J.Primatol. 56, 137—153. Rendall, D., Rodman, P. S. & Emond, R. E. 1996: Vocal recognition of individuals and kin in free- ranging rhesus monkeys.Animal Behaviour 51, 1007—1015. Rendall, D., Seyfarth, R. M., Cheney, D. L. & Owren, M. J. 1999: The meaning and function of grunt variants in baboons.Anim.Behav. 57, 583—592. Rowell, T.E.& Hinde, R.A.1962: Vocal communication by the rhesus monkey ( Macaca mulatta). Proc.Zool.Soc.Lond. 138, 279—294. Sayigh, L. S., Tyack, P. L., Wells, R. S., Solow, A. R., Scott, M. D. & Irvine, A. B. 1999: Individual recognition in wild bottlenose dolphins: a field test using playback experiments.Anim.Behav. 57, 41—50. Schaller, G.B.1963: The Mountain Gorilla: Ecology and Behavior.The University of Chicago Press, Chicago, Illinois. Schott, D.1975: Quantitative analysis of the vocal repertoire of squirrel monkeys ( Saimiri sciureus). Zeitschrift fuer Tierphysiologie 38, 225—250. Schultz, S.2001: Notes on interactions between crowned eagles and monkeys in Taı ¨ National Park, Ivory Coast.Folia Primatologica 72, 248—250. Schultz, S.& Noe, R.2002: The consequences of crowned eagle central-plane foraging on predation risk in monkeys.Proceedings of the Royal Society of London Series B-Biological Sciences 269, 1797—1802. Schwartz, E.1928: The species of the genus Cercocebus.Ann.Mag.Nat.Hist. 10, 664—670. Semple, S., McComb, K., Alberts, S. & Altmann, J. 2002. Information content of female copulation calls in yellow baboons.Am.J.Primatol. 56, 43—56. Seyfarth, R., Cheney, D. & Marler, P. 1980a: Vervet monkey alarm calls: Semantic communication in a free-ranging primate.Anim.Behav. 28, 1070—1094. Seyfarth, R.M.,Cheney, D.L.& Marler, P.1980b: Monkey responses to 3 different alarm calls - evidence of predator classification and semantic communication.Science 210, 801—803. Vocal Repertoire of Sooty Mangabeys 321

Struhsaker, T.T.1967: Auditory communication among vervet monkeys ( Cercopithecus aethiops).In: Social Communication among Primates (Altmann, S.A.,ed).University of Chicago Press, Chicago, pp.128—161. Teixidor, P.& Byrne, R.W.1999: The ÔWhinneyÕ of spider monkeys: Individual recognition before situational meaning.Behaviour 136, 279—308. Uster, D.& Zuberbu ¨ hler, K.2001: The functional significance of Diana monkey ÔclearÕ calls. Behaviour 138, 741—756. Waser, P.M.1982: The evolution of male loud calls among mangabeys and baboons.In: Primate Communication (Snowdon, T., Brown, C. H. & Peterson, M. R., eds). Cambridge University Press, Cambridge, pp.117—144. Waser, P.M.& Brown, C.H.1984: Is there a sound window for primate communication? Behav.Ecol. Sociobiol. 15, 73—76. Waser, P.M.& Brown, C.H.1986: Habitat acoustics and primate communication.Am.J.Primatol. 10, 135—156. Waser, P.& Waser, M.S.1977: Experimental studies of primate vocalizations: Specifications for long- distance propagation.Zeitschrift fuer Tierpsychologie 43, 355—363. Winter, P., Plood, D. & Latta, J. 1996: Vocal repertoire of squirrel monkey (Saimiri Scureus).Its analysis and significance.Experimental Brain Research 1, 359–370. Zuberbu¨ hler, K.1995: Diana Monkey loud calls: messages for conspecifics and predators.Thesis, Department of Psychology, University of Pennsylvania, USA. Zuberbu¨ hler, K.2001: Predator-specific alarm calls in Campbell’s monkeys, Cercopithecus campbelli. Behav.Ecol.Sociobiol. 50, 414—522. Zuberbu¨ hler, K., Noe¨ , R.& Seyfarth, R.M.1996: Diana monkey long-distance calls: messages for conspecifics and predators.Anim.Behav. 53, 589—604.

Received: October 2, 2002

Resubmitted: September 18, 2003

Initial acceptance: November 25, 2003

Final acceptance: February 7, 2004 (S. A. Foster)