American Journal of Primatology 71:333–342 (2009) RESEARCH ARTICLE Dialects in Pygmy Marmosets? Population Variation in Call Structure STELLA DE LA TORRE1 AND CHARLES T. SNOWDON2 1 College of Biological and Environmental Sciences, Department of Ecology, Universidad San Francisco de Quito, CumbayáQuito, Ecuador 2 Department of Psychology, University of Wisconsin, Madison, Wisconsin Population variation in primate vocal structure has been rarely observed. Here, we report significant population differences in the structure of two vocalizations in wild pygmy marmosets (Trills and J calls). We studied 14 groups of pygmy marmosets Callithrix (Cebuella) pygmaea pygmaea from five populations in northeastern Ecuador. We analyzed the acoustic structure of Trills and J calls recorded from two adult animals in each group through focal samples. Although individuals and groups within a population differed in call structure, we found consistent structural differences at a population level for Trills and J calls. Pair-wise comparisons for the two call types point to San Pablo and Amazoonico as the populations that differed the most, whereas Hormiga and Zancudococha showed the least differences. Discriminant function analysis indicates that calls from each population could be classified accurately at rates significantly above chance. Habitat acoustics, social factors and genetic drift may explain interpopulation vocal differences. This is the first evidence of within-subspecies vocal differences, or dialects, in wild populations of a neotropical primate species. Am. J. Primatol. 71:333–342, 2009. r 2009 Wiley-Liss, Inc. Key words: vocal variability; pygmy marmoset; Callithrix (Cebuella) pygmaea; Ecuadorian Amazon; dialects INTRODUCTION Bird song has long fascinated those interested in the development of communication because the variability seen in different populations of the same species (or subspecies) suggests that song might be learned from others [reviewed by Catchpole & Slater, 1995]. However, less evidence exists of similar processes in nonhuman primates [Egnor & Hauser, 2004; Janik & Slater, 1997]. Variation in vocal structure between populations has been described only in few primate species, wild chimpanzees Pan troglodytes [Crockford et al., 2004; Marshall et al., 1999; Mitani & Brandt, 1994; Mitani et al., 1999], wild silvery gibbons Hylobates moloch [Dallmann & Geissmann, 2001], captive Barbary macaques Macaca sylvanus [Fischer et al., 1998], provisioned Japanese macaques M. fuscata [Green, 1975; Masataka, 1992] and between subspecies of squirrel monkeys Saimiri oerstedi and of saddle-back tamarins Saguinus fuscicollis [Boinski & Newman, 1988; Hodun et al., 1981]. This raises the question of what has led to such flexible communication in birds and in humans in contrast to the relatively inflexible vocal development of nonhuman primates. Snowdon et al.  argued that this perceived lack of vocal variability could be an artifact of the lack of appropriate quantitative data for this order. In this study, we present evidence to support their argu- r 2009 Wiley-Liss, Inc. ment, documenting interpopulation differences in the physical structure of two separate vocalizations of the pygmy marmoset Callithrix (Cebuella) pygmaea pygmaea. We discuss how habitat acoustics, vocal learning and genetic drift could be influencing the reported differences. The pygmy marmoset is an arboreal primate restricted to river-edge forests in the Upper Amazon basin [de la Torre, 2000; Hershkovitz, 1977; Soini, 1988]. Visibility in these forests is generally poor, therefore, pygmy marmosets, as other arboreal primates, are likely to be highly dependent on vocal communication [Marler, 1965; Seyfarth, 1987]. Like other callitrichines, pygmy marmosets are cooperative breeders that live in stable, heterosexual groups varying in size from two to nine individuals, with small home range areas from 0.1 to 1.2 ha Contract grant sponsor: USPHS Grant; Contract grant number: MH029775; Contract grant sponsor: National Geographic Society; Contract grant number: 5806-96. Correspondence to: Stella de la Torre, Universidad San Francisco de Quito, Av. Interoceanica y Jardines del Este, Cumbayá-Quito, Ecuador. E-mail: email@example.com Received 18 July 2008; revised 25 November 2008; revision accepted 25 November 2008 DOI 10.1002/ajp.20657 Published online 8 January 2009 in Wiley InterScience (www. interscience.wiley.com). 334 / de la Torre and Snowdon [Ferrari & Lopes Ferrari, 1989; Soini, 1988; Yépez et al., 2005]. Group cooperation in the care of young requires close coordination between group members, and this is mainly achieved through vocal communication [de la Torre & Snowdon, 2002; de la Torre et al., 2000; Snowdon & Cleveland, 1984; Soini, 1988]. Captive and wild pygmy marmosets use Trills and J calls (Fig. 1) to maintain short-range contact and to mediate interactions between group members [Elowson & Snowdon, 1994; Elowson et al., 1992; Pola & Snowdon, 1975; Snowdon & Elowson, 1999; Snowdon & Hodun, 1981]. These contact calls are variants of a sinusoidally frequency modulated tone (Trills and J calls lower frequency: above 7 kHz; peak frequency: approx. 12 kHz) and have a pulsatile temporal structure (Trills: 32–38 cycles/sec, J calls: 17–23 cycles/sec). Trills are usually emitted when animals are no more than 10 m apart while they are feeding on exudates, foraging for insects, traveling or resting, whereas J calls are emitted at caller–receiver distances between 11 to 20 m. When J calls are produced at distances less that 5 m of a potential receiver, the calling animal either begins to travel within 5 min after emitting the call, or approaches an exudate source, or responds to a bout of babbling of an infant [de la Torre & Snowdon, 2002; Snowdon & Hodun, 1981]. The relationship between the use of these call types and the distance between the calling animal and the potential receivers suggests that pygmy marmosets vary the use of these calls in a way appropriate to the effects of habitat acoustics [de la Torre & Snowdon, 2002; Snowdon & Hodun, 1981]. Yépez et al.  found that each of four populations of pygmy marmosets in northeastern Ecuador had distinct preferences for tree species used for exudate feeding that were unrelated to the abundance of each species in that population. These results are an indication of population differences in behavior and, coupled with preliminary indications of differences in vocal structure, led us to hypothesize that each population would differ from the others in the structure of Trills and J calls. METHODS Study Area We gathered data in five areas in northeastern Ecuador in an east–west transect of approximately 300 km and a north–south transect of approximately 100 km. La Hormiga population is located in the margins of the Laguna Grande of the Cuyabeno hydrographic system. The Zancudococha population is located on the edges of the Zancudococha Lake. The San Pablo population is located in the margins of the Aguarico River. The Sacha population is located in the margins of the Napo River. The Amazoonico population is located in the margins of the Arajuno River (Fig. 2). Each population is separated from the others by at least 30 km; large rivers and large extensions of open areas or terra firme forests, not used by the marmosets, are natural boundaries for all populations. These areas have an altitude of 230–360 m above sea level. The habitats of San Pablo, Sacha and Amazoonico are varzea forests, seasonally flooded by white-water rivers, with different degrees of human alteration caused by agriculture and selective logging. The areas of Zancudococha and La Hormiga are edge habitats between terra firme and igapo forests seasonally flooded by black-water rivers. These habitats have not been affected by agricultural Fig. 1. Spectrograms of pygmy marmoset vocalizations. (a) Trill and (b) J call from La Hormiga (H) and Amazoonico (A) populations. FFT size: 256, Hanning window. Am. J. Primatol. Vocal Variability in Pygmy Marmosets / 335 activities but the area of La Hormiga had a high rate of tourism during the years when our research was carried out at that site (1996–1998) [de la Torre et al., 2000]. Fig. 2. Location of the studied populations of pygmy marmosets in Ecuadorian Amazonia (1, La Hormiga; 2, Zancudococha; 3, San Pablo; 4, Sacha; 5, Amazoonico). Data Collection We observed 14 groups of pygmy marmosets in the five populations for a total of 1,722 hr of direct observation. All observations were carried out in the rainy season (March through August) (Table I). Body lengths of adult and subadult individuals in all groups ranged from 13 to 15 cm and did not differ among populations. Group size varied from three to eight individuals, and the home-range size varied from 0.15 to 1.2 ha (Table I). Distances between neighboring groups varied from 30 m (groups 1 and 2 at Zancudococha) to 1,000 m (groups 1 and 2 at La Hormiga). All the research methods were reviewed and approved by the Ecuadorian Ministry of the Environment, which gave us the legal permit to conduct the research. Vocalizations Two to three field workers observed the 14 groups of pygmy marmosets. We classified marmosets in each group by size and other morphological characters (e.g., presence of whitish nasal stripe) into approximate age classes [Soini, 1988]. Sex determination was possible in adults and in most of the subadult animals based on the scrotal pigmentation of males. The vocalizations of the marmosets in a group were recorded using focal samples, each sample lasting a total of 5 min. At least two focal samples of the adult (male and female) members of the groups were carried out per day of observation in TABLE I. Population Coordinates (UTM), Months and Hours of Observation, Group Size Range, and Home Range Size of all Studied Groups UTM coordinatesa Months of observation Group Hours of observationb Group size rangec Home range size (ha)d Amazoonico 219290E, 9883728N July 2003 June 2004 A1 A2 114 80 4–5 3–4 0.15 0.40 San Pablo 341767E, 9969737N July 2001 July 2002 June 2003 July 2004 P2 P4 P5 126 107 110 5–7 4–6 4–7 0.45 0.22 0.31 Sacha 337296E, 9946861N August 2001 June 2002 August 2003 S1 S2 S3 153 143 159 4–6 5–8 5–7 0.37 1.20 0.40 La Hormiga 368315E, 3685N April, May 1997 G1 G2 G3 105 100 105 6 4 6 0.90 1.09 0.53 Zancudococha 445459E, 9933749N June, July, August 1997 Z1 Z2 Z3 140 140 140 6–7 7 5–6 0.73 0.40 0.78 Population a Zone 18, Datum PSAD56. Hours of observation rounded to the nearest hour, rainy season only. Variance in group size is the result of differences in group sizes among field seasons. d Home range size was estimated by connecting the extreme location points of group members during the study period; the periphery that enclosed all points was considered as the home range perimeter and the area inside the perimeter was calculated [Yépez et al., 2005]. b c Am. J. Primatol. 336 / de la Torre and Snowdon each group. These samples were evenly distributed throughout the day. The vocalizations of the focal individuals were recorded with a high fidelity stereo recorder (Marantz PMD 222, Marantz, frequency response: 40–14,000 Hz) and a unidirectional microphone (Sennheiser ME66, Sennheiser, Germany, frequency response: 40–20,000 Hz72.5 dB) at a recording volume that optimized the recording of the calls. In all groups the recording distances ranged from 1 to 10 m. We annotated the vocalizations emitted by the focal animals on the tape. Focal recordings were done only when animals were visible and could be clearly identified. Acoustic analysis We used Signal 2.29 (Engineering Design) and SoundEdit (Macromedia) to construct spectrograms of calls (FFT size: 256, Hanning window). We selected 15 Trills and 15 J calls with high sound clarity, emitted by the adult male and the adult female in each of the groups. These vocalizations were selected from at least five focal samples carried out on different days throughout a sampling period and were evenly distributed throughout the day (in those groups that were observed in different years we used the vocalizations recorded in only 1 year for each focal animal to avoid the possibility of changes in group composition between years). We analyzed a total of 420 Trills and 420 J calls. We measured the acoustical variables with the cursors of the display screen and recorded four variables: (1) total call duration (in ms); (2) number of cycles (for Trills) or notes (for J calls)/sec, obtained by dividing the number of complete cycles or notes by the total duration of the call (in sec); (3) minimum frequency (in kHz); (4) upper or maximum frequency (in kHz). We also calculated the frequency bandwidth (maximum minus minimum frequency in kHz) of each call. Data Analyses We used each of the 15 calls from the adult male and female in a group to evaluate inter-individual variability with unpaired t-tests for each acoustic variable within each group. For analyzing population differences, we averaged all the acoustic measurements for a given call type to obtain a single value for each variable per subject; thus, degrees of freedom in our interpopulation analyses were based on the number of focal animals (N 5 28 animals) and not the number of individual calls. We carried out two complementary statistical analyses of the acoustic variables of Trills and J calls, after confirming that the data set met the conditions of normality, independence and homoscedasticity. Nested ANOVAs (SuperAnova 1.11, Abacus Concepts for MacIntosh) were used to determine interpopulation differences Am. J. Primatol. after taking into account group and individual variability, with population as the main factor, group as the factor nested within population and individuals nested within group. Significant differences between pairs of populations were determined with the Fisher’s PLSD test (P 5 0.05). We also used discriminant function analyses (SPSS version16.0) to complement ANOVA results for individual acoustic variables to determine which variables contributed most to differentiation of populations with Wilks’ lambda. The resulting discriminant functions were used to classify Trills and J calls by population and the proportion of correctly classified calls was tested against the expected values (20% for each population for each call type), using a w2 test. Sex differences were evaluated with unpaired t-tests (StatView SE; Abacus Concepts for MacIntosh). RESULTS Individual and Sex Differences Within each group the Trills and J calls of the individual animals differed significantly in at least one of the variables that we measured, but across the entire sample (14 males and 14 females) we found no consistent or significant pattern of differences between sexes in either Trills and J calls (Appendix A). Interpopulation Differences After controlling for group and individual variability, we found significant interpopulation differences in all of the acoustic variables of Trills and J calls. Trills The results of the nested ANOVAs showed that Trill structure differed significantly among all populations (Table II). Marmosets at Amazoonico emitted the shortest Trills with the narrowest bandwidth and the highest number of cycles/sec whereas marmosets at Sacha emitted the longest Trills. Marmosets at La Hormiga and Zancudococha emitted Trills with the lowest number of cycles/sec. The lowest minimum frequency was recorded in the Trills of La Hormiga and Zancudococha. The highest maximum frequency was recorded in the Trills of San Pablo. In the discriminant function analysis, cycles per second, minimum frequency and duration were the main variables in statistically differentiating Trills from the five populations. Trills were correctly classified to the population of the caller in 71.4% of the cases, compared with the expected 20% for each population (w216 ¼ 59:1, Po0.0001). Using the cross-validation, leave-one-out method, classification was correct in 50.0% of the cases (w216 ¼ 47:5, Po0.0001). Vocal Variability in Pygmy Marmosets / 337 TABLE II. Interpopulation Differences in Trill Acoustic Parameters (Mean 7 SE; Populations With Different Letters Differ Significantly—Fisher’s PSLD Tests—on That Variable; F and P Values From Nested ANOVAs) Population Amazoonico Sacha San Pablo Zancudococha Hormiga F4,11 P Duration (ms) Cycles/sec 211.25710a 348.17717b 277.33723c 283.57718c 310.3776c,b 5.86 0.009 38.5270.6a 33.9370.6b 33.2770.6b 31.1070.7c 31.1770.7c 30.57 0.0001 Max. freq (kHz) 10.5370.3a 11.9470.3b,c 12.2370.4b 11.2270,1a,c 11.4070.2a,b,c 4.53 0.02 Min. freq (kHz) 8.0270.2b,c 8.7170.2a,c 8.7270.2a 7.8170.1b 7.7370.1b 5.88 0.009 Bandwidth (kHz) 2.5170.2a 3.2370.2b 3.5170.2b 3.4170.1b 3.6670.3b 4.92 0.02 TABLE III. Interpopulation Differences in J Call Acoustic Parameters (Mean 7 SE; Populations With Different Letters Differ Significantly—Fisher’s PSLD Tests—on That Variable; F and P Values From Nested ANOVAs) Population Amazoonico Sacha San Pablo Zancudococha Hormiga F4,11 P Duration (ms) 590.50728a 582.50719a 730.67730b 751.83720b 705.17727b 7.32 0.004 Notes/sec 17.3370.8a 18.4070.7a,b 18.5871.4a,b 17.7171.0a 20.4270.8b 3.96 0.03 J calls The results of the nested ANOVAs showed that there was significant variation among populations in all the acoustic parameters of this call type (Table III). Marmosets at Sacha and Amazoonico emitted the shortest J calls. The highest number of notes/sec and the lowest minimum frequency were in J calls at La Hormiga. The highest maximum frequency was in J calls at San Pablo. The narrowest bandwidth was in J calls of Amazoonico. In the discriminant function analysis, maximum frequency and duration were the main variables in statistically differentiating J calls from the five populations. J calls were correctly classified to the population of the caller in 78.6% of the cases, compared with the expected 20% for each population (w216 ¼ 70:4, Po0.0001). Using the cross-validation, leave-oneout method, classification was correct in 57.1% of the cases (w216 ¼ 42:6, P 5 0.0003). DISCUSSION The structure of the Trills and J calls recorded from the wild groups of pygmy marmosets is within the range of variation recorded in captive colonies [Elowson & Snowdon, 1994; Pola & Snowdon, 1975; Snowdon, 1993; Snowdon & Cleveland, 1980]. Significant differences in at least one acoustic variable of Trills and of J calls make possible individual recognition between the focal animals in each of the study groups in a similar manner as was Max. freq (kHz) Min. freq (kHz) Bandwidth (kHz) 10.4270.2a 12.3670.3b 13.5870.4c 12.3970.2b 11.8270.1b 11.36 0.0007 7.970.2b,c 8.3370.1b 8.8970.3a 7.9270.1c 7.6270.1c 15.8 0.0002 2.4870.1a 4.0370.3b 4.6970.4b 4.4670.1b 4.2070.2b 6.16 0.008 reported in captive pygmy marmosets by Snowdon and Cleveland . However, we did not find any consistent differences in the whole data set that could be attributable to sex only. Trills and J calls had significantly different structures in different populations. These differences exist despite the presence of some intrapopulation variability, as shown by the results of the nested ANOVAs and in the discriminant function analyses. Pair-wise comparisons for variables of the two call types indicated that only La Hormiga and Zancudococha populations did not differ from each other except for notes/sec in J calls, but they differed from all other populations (Tables II and III). Based on the similar body sizes of the marmosets in all the groups and populations, we believe that the potential effect of body size on the frequency differences found in the two call types among populations is negligible. Recording distances were similar for all groups and were within the range of minimum distortion for Trills and J calls [de la Torre & Snowdon, 2002] so we also feel confident that the observed differences were not an artifact of sampling acoustic differences among populations. Some of the differences we found in Trills and J calls could be related to differences in the acoustic characteristics of the habitats. The ‘‘local adaptation hypothesis’’ states that the acoustic characteristics of the environment through which a signal is normally transmitted have selectively influenced the form of a signal to reduce distortion [Gish & Morton, 1981]. Am. J. Primatol. 338 / de la Torre and Snowdon When we quantitatively evaluated the ambient noise levels and the reverberation properties of the habitats of the groups [de la Torre & Snowdon, 2002, de la Torre & Snowdon, in preparation] we found that La Hormiga had the greatest ambient noise and Amazoonico the least. Longer duration and wider bandwidth may increase the detectability of the calls because of temporal summation [Aubin & Mathevon, 1995; Zwicker et al., 1957] and are expected in calls from noisier habitats. Additionally, the frequency of a call may be adapted to be above the upper range of ambient noise [Snowdon & Hodun, 1981]. Thus, calls from populations with a lower range of ambient noise should be lower in frequency than those from populations with a broader range of ambient noise. However, although the calls from Amazoonico did have the shortest duration and smallest bandwidth as predicted, the calls from La Hormiga were not the longest calls; Trills, but not J calls, had the widest bandwidth and there was no relationship across populations between ambient noise range and lower minimum frequency of calls. Slower repetition rates (fewer number of cycles or notes per second) and lower frequencies reduce the distorting effects that sound reverberation has on the transmission of signals with fast repetition rates and high frequencies [Brenowitz, 1986; Richards & Wiley, 1980; Wiley, 1991; Wiley & Richards, 1978]. Zancudococha, Amazoonico and Sacha habitats had the greatest reverberation compared with the other two sites; however, although the calls from Zancudococha had lowest repetition rates as predicted, Trills from Amazoonico and Sacha had the highest repetition rates and none of the calls from these three populations had the lowest minimum frequency. Thus, although some predictions about call structure based on habitat acoustics were supported, several others were not. It is possible that other habitat acoustic variables, such as amplitude fluctuations specific to one population or another, may have influenced call structure [Brenowitz, 1986; Richards & Wiley, 1980; Wiley & Richards, 1978]. However, we currently have no data to evaluate this. The presence of different predator assemblages in different populations may also drive changes in vocal structure [Marler, 1955; Zimmermann et al., 2000] but our observations do not suggest variation in predators among populations. Social influences may play a role in the differences in vocal structure. Studies with captive pygmy marmosets and Wied’s black tufted-ear marmosets C. kuhlii have provided evidence of plasticity in acoustic structure that allow them to adjust their vocalizations through subtle changes in frequency and temporal parameters, in response to changes in their social environment (e.g., the presence of a novel companion) [Elowson & Snowdon, 1994; Rukstalis et al., 2003; Snowdon & Elowson, 1999]. As Trills Am. J. Primatol. and J calls are emitted during interactions between individuals under fluid environmental conditions, we view vocal plasticity as a functional adaptation of the marmosets to cope with their changing environment where individuals may adapt their call structure to match that of a new mate or social group [de la Torre & Snowdon, 2002; Elowson & Snowdon, 1994; Elowson et al., 1992]. Natural changes in the social environment of wild pygmy marmoset groups include changes in group size and composition owing to migrations (e.g., presence of novel companions) [de la Torre et al., 2000; Soini, 1988]. These environmental changes are not unique for pygmy marmosets so we expect vocal plasticity to be reported in other primates as more detailed studies are carried out on multiple populations of the same species. Vocal plasticity is considered a basis for learning processes [Egnor & Hauser, 2004; Snowdon et al., 1997] known to be related to dialects in some bird species [e.g., Marler, 1970; Nottebohm, 1972], whales [Noad et al., 2000], bats [Boughman, 1998] and primates [Crockford et al., 2004], and might account for some of the population differences we have reported, but this hypothesis is difficult to test in the field. Interpopulation differences in the acoustic structure of vocalizations in the five studied marmoset populations are consistent with independent interpopulation differences found in exudate feeding preferences in the same populations [Yépez et al., 2005]. Genetic variation owing to river barriers promoting separation between populations and the low mobility of marmosets could account for some of the observed differences but at present we have no data to test this hypothesis. This study provides the first evidence for natural population variation in vocal structure in a neotropical primate and one of the few demonstrations in the whole order. These variations can be partially explained by differences in habitat acoustics. Mechanisms of social learning and genetic isolation of different populations may account for the remaining differences. Given the documented interpopulation variability in vocal production in this study coupled with the variability in exudate feeding in pygmy marmosets [Yépez et al., 2005], the loss of even one population may imply the loss of a unique behavioral variation. This consideration needs to be taken into account while planning future conservation actions in Ecuadorian Amazon and in other areas of the distribution of pygmy marmosets. ACKNOWLEDGMENTS All the research reported in this manuscript adhered to the American Society of Primatologists Principles for the ethical treatment of nonhuman primates; all the research objectives and protocols reported in the manuscript were reviewed and Vocal Variability in Pygmy Marmosets / 339 approved by the Ecuadorian Ministry of the Environment, which gave us the legal permit to conduct the research. We are grateful for the support from the following institutions: Wisconsin Regional Primate Center (NCCR Grant P51 RR000167), Ecolap - USFQ, Amazoonico - Selva Viva, Transturi and Sacha Lodge. We thank Pablo Yépez, Delfı́n Payaguaje, Alfredo Payaguaje, Monserrat Bejarano, Lucı́a de la Torre, Daniel Payaguaje, Fernanda Tomaselli, Carolina Proaño, Hernán Castañeda, Santiago Molina, Beatriz Romero and Margarita Brandt for their help in the field work and Carolina Proaño for the sound analyses of some vocalizations. We thank Karen Strier for her suggestions on the data analyses and Rosamunde E. A. Almond, Katherine A. Cronin, Tatiana Humle and two anonymous reviewers for critical feedback. Appendix A Individual differences in acoustic parameters (mean7SE—on second line in each cell) of Trills and J calls of individuals (adult male, M; adult female, F) in the five populations (P, San Pablo; S, Sacha; A, Amazoonico; H, Hormiga; Z, Zancudococha; numbers following the letters indicate group number within each population). Variables with significant differences between individuals in a group (unpaired ttests, Po0.05) are marked with a star (). Trills Group Individual Duration (sec) Cycles/sec Peak freq (kHz) P1 M P1 F P2 M P2 F P3 M P3 F S1 M S1 F S2 M S2 F S3 M S3 F A1 M A1 F A2 M A2 F H1 M H1 F 251 1 249 3 301 38 275 34 214 11 374 39 328 53 373 65 359 37 379 28 271 2 379 2 194 11 195 13 218 34 238 15 311 24 309 30 34.1 0.6 33.6 0.7 35.6 0.4 33 0.8 31.7 0.8 31.6 1.1 34 0.7 32.4 0.8 35.5 0.7 34.7 0.7 32.2 0.7 34.8 0.7 38.1 0.8 37.9 1.0 40.2 0.9 37.9 1.0 33.6 0.5 30.5 0.7 11.88 0.30 11.54 0.19 11.71 0.17 13.41 0.21 13.57 0.09 11.28 0.18 11.67 0.12 12.12 0.15 10.86 0.32 12.59 0.17 11.88 0.15 12.52 0.15 11.30 0.11 10.45 0.18 9.74 0.09 10.62 0.12 11.78 0.2 12.21 0.15 Min freq (kHz) 7.97 0.19 8.52 0.18 8.62 0.20 9.51 0.22 9.43 0.23 8.29 0.16 8.35 0.1 8.41 0.17 8.10 0.16 8.98 0.29 9.52 0.09 8.91 0.26 8.50 0.18 7.77 0.08 7.82 0.14 8.01 0.18 7.32 0.12 7.7 0.09 Bandwidth (kHz) 3.90 0.22 3.02 0.19 3.08 0.2 3.90 0.29 4.15 0.25 2.99 0.23 3.32 0.16 3.72 0.18 2.77 0.24 3.61 0.24 2.35 0.18 3.61 0.27 2.81 0.19 2.69 0.19 1.92 0.12 2.61 0.17 4.46 0.18 4.52 0.17 Am. J. Primatol. 340 / de la Torre and Snowdon H2 M H2 F H3 M H3 F Z1 M Z1 F Z2 M Z2 F Z3 M Z3 F 293 29 304 34 339 20 307 17 342 38 280 20 251 25 323 23 286 17 220 33.1 0.4 30.1 0.4 28.9 1.0 30.8 0.9 28.6 0.9 31.1 0.4 32.1 0.9 30.6 0.8 30.1 0.4 33.9 11.53 0.23 11.01 0.24 10.85 0.18 11.01 0.12 11.42 0.16 11.16 0.05 11.23 0.21 11.53 0.18 10.74 0.17 11.22 8.07 0.10 7.59 0.11 7.6 0.13 8.12 0.70 8.09 0.12 7.74 0.13 8.02 0.1 7.74 0.14 7.26 0.14 7.99 3.46 0.26 3.42 0.20 3.25 0.22 2.89 0.15 3.33 0.23 3.43 0.18 3.20 0.24 3.79 0.17 3.48 0.19 3.23 J Calls Group Individual Duration (sec) Notes/sec Peak freq (kHz) Min freq (kHz) Bandwidth (kHz) P1 M P1 F P2 M P2 F P3 M SP3 F S1 M S1 F S2 M S2 F S3 M S3 F A1 M A1 F A2 M 797 140 644 100 639 41 790 68 781 8 733 11 570 43 608 4 568 4 605 4 504 4 640 5 516 4 611 3 586 38 17.7 0.5 18.3 0.4 22.5 0.4 22.8 0.3 14.1 0.6 16.1 0.8 16.7 0.9 15.9 0.7 19.7 0.8 18.5 0.9 20 0.9 19.6 0.9 18 0.6 19.2 0.8 16.7 0.4 13.35 0.16 12.19 0.24 14.95 0.11 14.40 0.22 14.02 0.26 12.63 0.29 11.89 0.2 12.85 0.21 11.67 0.2 13.05 0.26 11.56 0.2 13.16 0.25 10.45 0.21 9.85 0.16 10.40 0.10 7.67 0.17 8.72 0.17 9.98 0.08 9.72 0.09 8.54 0.12 8.72 0.17 8.15 0.06 8.22 0.18 8.15 0.06 8.08 0.17 8.45 0.57 8.93 0.18 7.75 0.09 7.51 0.14 8.24 0.04 5.68 0.19 3.47 0.26 4.97 0.14 4.68 0.25 5.48 0.28 3.92 0.29 3.73 0.22 4.63 0.31 3.51 0.18 4.96 0.16 3.12 0.67 4.23 0.30 2.708 0.28 2.33 0.19 2.16 0.10 Am. J. Primatol. Vocal Variability in Pygmy Marmosets / 341 A2 F H1 M H1 F H2 M H2 F H3 M H3 F Z1 M Z1 F Z2 M Z2 F Z3 M Z3 F 650 37 789 33 679 39 723 55 599 42 695 47 748 53 727 47 712 37 836 45 741 46 714 60 785 65 15.4 0.7 20.7 0.2 22 0.2 20.4 0.9 22.8 0.2 18.5 0.7 18.1 0.5 20.5 0.4 16.1 0.4 16.6 0.5 18.6 0.3 20 0.4 14.5 0.7 REFERENCES Aubin T, Mathevon N. 1995. Adaptation to severe conditions of propagation: long-distance distress calls of a colonial seabird. Bioacoust 6:153–161. Boinski S, Newman JD. 1988. Preliminary observations on squirrel monkeys (Saimiri oerstedi) vocalizations in Costa Rica. Am J Primatol 14:329–343. Boughman JW. 1998. Vocal learning by greater spear-nosed bats. Proc R Soc Lond B Biol Sci 265:227–233. Brenowitz EA. 1986. Environmental influences on acoustic and electric animal communication. Brain Behav Evol 28:32–42. Catchpole CK, Slater PJB. 1995. Bird song: biological themes and variations. Cambridge: Cambridge University Press. 248p. Crockford C, Herbinger I, Vigilant L, Boesch C. 2004. Wild chimpanzees produce group-specific calls: a case for vocal learning? Ethology 110:221–243. de la Torre S. 2000. Primates de la Amazonı́a Ecuatoriana (primates of Amazonian Ecuador). Quito: Proyecto PETRAMAZ/SIMBIOE. 60p. de la Torre S, Snowdon CT. 2002. Environmental correlates of vocal communication of wild pygmy marmosets. Anim Behav 63:847–856. de la Torre S, Snowdon CT, Bejarano M. 2000. Effects of human activities on pygmy marmosets in Ecuadorian Amazon. Biol Cons 94:153–163. Egnor SE, Hauser MD. 2004. A paradox in the evolution of primate vocal learning. 2004. Trends Neurosci 27: 649–654. Elowson AM, Snowdon CT. 1994. Pygmy marmosets, Cebuella pygmaea, modify vocal structure in response to changed social environment. Anim Behav 47:1267–1277. Elowson AM, Snowdon CT, Sweet CJ. 1992. Ontogeny of trill and J call vocalizations in the pygmy marmoset, Cebuella pygmaea. Anim Behav 43:703–715. 10.99 0.08 12.01 0.15 11.98 0.14 11.51 0.16 11.50 0.13 12.16 0.12 11.74 0.20 11.98 0.10 12.62 0.13 13.19 0.27 12.03 0.16 12.07 0.12 12.42 0.21 8.26 0.05 7.48 0.12 7.3 0.06 7.8 0.10 7.88 0.06 7.96 0.07 7.3 0.12 7.78 0.11 7.95 0.08 8.43 0.16 7.92 0.09 7.57 0.15 7.89 0.08 2.73 0.08 4.53 0.18 4.69 0.18 3.71 0.16 3.62 0.13 4.20 0.14 4.43 0.22 4.20 0.09 4.67 0.15 4.76 0.28 4.11 0.18 4.50 0.15 4.54 0.21 Ferrari SF, Lopes Ferrari MA. 1989. A re-evaluation of the social organisation of the Callitrichidae, with reference to the ecological differences between genera. Folia Primatol 52:132–147. Fischer J, Hammerschmidt K, Todt D. 1998. Local variation in Barbary macaque shrill barks. Anim Behav 56:623–629. Gish SL, Morton ES. 1981. Structural adaptations to local habitat acoustics in Carolina wren songs. Z Tierpsychol 56:74–84. Green S. 1975. Dialects in Japanese monkeys? Vocal learning and cultural transmission of locale specific behavior? Z Tierpsychol 38:304–314. Hershkovitz P. 1977. Living new world monkeys (Platyrrhini) with an introduction to primates. Vol I. Chicago: The University of Chicago Press. 1117p. Hodun A, Snowdon CT, Soini P. 1981. Subspecific variation in the long calls of the tamarin, Saguinus fuscicollis. Z Tierpsychol 57:97–110. Janik V, Slater PBS. 1997. Vocal learning in mammals. Adv Stud Behav 26:59–99. Marler P. 1955. Characteristics of some animal calls. Nature 176:6–8. Marler P. 1965. Communication in monkeys and apes. In: de Vore I, editor. Primate behavior. Field studies of monkeys and apes. New York: Holt, Rinehart & Winston. p 544–584. Marler P. 1970. Birdsong and speech development: could there be parallels? Am Sci 58:669–673. Marshall AJ, Wrangham RW, Arcadi AC. 1999. Does learning affect the structure of vocalizations in chimpanzees? Anim Behav 58:825–830. Masataka N. 1992. Attempts by animal caretakers to condition Japanese macaque vocalizations result inadvertently in individual specific calls. In: Nishida T, McGrew WC, Marler P, Pickford M, deWaal FBM, editors. Topics in primatology. Vol. 1. Tokyo: University of Tokyo Press. p 271–278. Am. J. Primatol. 342 / de la Torre and Snowdon Mitani JC, Brandt KL. 1994. Social factors influence acoustic variability in the long distance calls of male chimpanzees. Ethology 96:233–252. Mitani JC, Hunley KL, Murdoch ME. 1999. Geographic variation in the calls of wild chimpanzees: a reassessment. Am J Primatol 47:133–151. Noad MJ, Cato DH, Bryden MM, Jenner M-N, Jenner KCS. 2000. Cultural revolution in whale songs. Nature 408: 537. Nottebohm F. 1972. The origins of vocal learning. Am Nat 106:116–140. Pola YV, Snowdon CT. 1975. The vocalizations of pygmy marmosets (Cebuella pygmaea). Anim Behav 23:826–842. Richards DG, Wiley RH. 1980. Reverberations and amplitude fluctuations in the propagation of sound in a forest: implications for animal communication. Am Nat 115:381–399. Rukstalis M, Fite JE, French JA. 2003. Social changes affect vocal structure in a Callitrichid primate (Callithrix kuhlii). Ethology 109:327–340. Seyfarth RM. 1987. Vocal communication and its relation to language. In: Smuts BB, Cheney DL, Seyfarth RM, Wrangham RW, Struhsaker W, editors. Primate societies. Chicago: The University of Chicago Press. p 440–451. Snowdon CT. 1993. A vocal taxonomy of the callitrichids. In: Rylands AB, editor. Marmosets and tamarins, systematics, behaviour and ecology. Oxford: Oxford Scientific Publications. p 78–94. Snowdon CT, Cleveland J. 1980. Individual recognition of contact calls in pygmy marmosets. Anim Behav 28:717–727. Snowdon CT, Cleveland J. 1984. ‘‘Conversations’’ among pygmy marmosets. Am J Primatol 7:15–20. Am. J. Primatol. Snowdon CT, Elowson AM. 1999. Pygmy marmosets modify call structure when paired. Ethology 105:893–908. Snowdon CT, Hodun A. 1981. Acoustic adaptations in pygmy marmoset contact calls: locational cues vary with distances between conspecifics. Behav Ecol Sociobiol 9:295–300. Snowdon CT, Elowson AM, Roush RS. 1997. Social influences on vocal development in New World primates. In: Snowdon CT, Hausberger M, editors. Social influences on vocal development. Cambridge: Cambridge University Press. p 234–248. Soini P. 1988. The pygmy marmoset, genus Cebuella. In: Mittermeier RA, Rylands AB, Coimbra-Filho AF, da Fonseca GAB, editors. Ecology and behavior of neotropical primates. Vol. 2. Washington, DC: World Wildlife Fund. p 79–129. Wiley RH. 1991. Associations of song properties with habitats for territorial oscine birds of eastern North America. Am Nat 138:973–993. Wiley RH, Richards DG. 1978. Physical constraints on the acoustic communication in the atmosphere: implications for the evolution of animal vocalizations. Behav Ecol Sociobiol 3:69–94. Yépez P, de la Torre S, Snowdon CT. 2005. Interpopulation differences in exudate feeding of pygmy marmosets in Ecuadorian Amazonia. Am J Primatol 66:145–158. Zimmermann E, Vorobieva E, Wrogemann D, Hafen T. 2000. Use of vocal fingerprinting for specific discrimination of gray (Microcebus murinus) and rufous mouse lemurs (Microcebus rufus). Int J Primatol 21:837–852 Zwicker E, Flottorp G, Stevens SS. 1957. Critical bandwidth in loudness summation. J Acoust Soc Am 29:548–557.