Journal of Comparative Psychology 2001, Vol. 115, No. 3, 258-271 Copyright 2001 by the American Psychological Association, Inc. 0735-7036/01/$5.00 DOI: 10.1037//0735-7036.115.3.258 The Production and Perception of Long Calls by Cotton-Top Tamarins (Saguinus oedipus): Acoustic Analyses and Playback Experiments This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. Daniel J. Weiss, Brian T. Garibaldi, and Marc D. Mauser Harvard University The authors' goal was to provide a better understanding of the relationship between vocal production and perception in nonhuman primate communication. To this end, the authors examined the cotton-top tamarin's (Saguinus oedipus) combination long call (CLC). In Part 1 of this study, the authors carried out a series of acoustic analyses designed to determine the kind of information potentially encoded in the tamarin's CLC. Using factorial analyses of variance and multiple discriminant analyses, the authors explored whether the CLC encodes 3 types of identity information: individual, sex, and social group. Results revealed that exemplars could be reliably assigned to these 3 functional classes on the basis of a suite of spectrotemporal features. In Part 2 of this study, the authors used a series of habituationdishabituation playback experiments to test whether tamarins attend to the encoded information about individual identity. The authors 1st tested for individual discrimination when tamarins were habituated to a series of calls from 1 tamarin and then played back a test call from a novel tamarin; both oppositeand same-sex pairings were tested. Results showed that tamarins dishabituated when caller identity changed but transferred habituation when caller identity was held constant and a new exemplar was played (control condition). Follow-up playback experiments revealed an asymmetry between the authors' acoustic analyses of individual identity and the tamarins' capacity to discriminate among vocal signatures; whereas all colony members have distinctive vocal signatures, we found that not all tamarins were equally discriminable based on the habituation—dishabituation paradigm. Many species that dwell in dense environments develop a system of long-range vocalizations (Chapman & Weary, 1990; Cheney & Seyfarth, 1996; Marten, Quine, & Marler, 1977; Mitani & Nishida, 1993; Ryan & Sullivan, 1989; Wiley & Richards, 1978). In general, visually restrictive habitats promote greater vocal communication between individual animals (Altmann, 1967; Cheney & Seyfarth, 1996; Morton, 1986; Norcross & Newman, 1993; Romer & Bailey, 1986). Call morphology may become increasingly differentiated in such habitats so that subtle variations denote different contexts (Snowdon, Cleveland, & French, 1983). The acoustic variation within these calls may convey information about territorial status, group spacing, group location, or individual identity (e.g., Beecher, 1982; Chapman & Weary, 1990; Cleveland & Snowdon, 1982; Mitani, 1985). Therefore, to gain a better understanding of the design of communication signals, researchers must identify the sources of variation and attempt to discern the types of information conveyed. To that effect, this study focuses on the long call produced by cotton-top tamarins (Saguinus oedi- pus), a species that inhabits dense tropical rainforest and has a highly diversified vocal repertoire. Cleveland and Snowdon (1982) identified three classes of long calls produced by captive cotton-top tamarins. On the basis of their analyses, the long calls we examined in the present study may be classified as "combination long calls" (CLCs), vocalizations consisting of two syllable types: chirps and whistles. Cleveland and Snowdon noted that the CLC has the greatest variability across tamarins, perhaps due to individual tamarin differences. CLCs appear to occur in a variety of circumstances involving social excitement, disturbance, pairing with a new mate, play, and social isolation (Cleveland & Snowdon, 1982). The most common context appears to be social isolation (Cleveland & Snowdon, 1982; Thurwachter, 1980). Given the variability in acoustic parameters noted (e.g., overall number of syllables and call duration) between callers and the context in which the calls occur, the CLC provides an excellent opportunity for investigating the relative contribution of individual, sex, and group identity to call morphology. In fact, the closely related phee call in marmosets, which seems to occur in similar contexts, encodes information about both the sex and the identity of the caller (Jorgensen & French, 1998; Norcross & Newman, 1993). Our research program represents an attempt to explore the meaningful components of a communication system among tamarins by designing playback experiments that are informed by detailed acoustic analyses. This integration of information from studies of production and perception has been successfully carried out with a number of species (e.g., birds: Brown, Dooling, & O'Grady, 1988; Chaiken, 1992; Emlen, 1972; pikas: Conner, 1985; frogs: Burmeister, Wilczynski, & Ryan, 1999; Wilczynski, Rand, & Ryan, 1999; and primates: Hauser, 1998b; May, Moody, Daniel J. Weiss, Brian T. Garibaldi, and Marc D. Mauser, Department of Psychology, Harvard University. This work was supported by Harvard University's Mind Brain and Behavior Grant 368422. We thank Douwe Yntema, Kristen Vickers, and Eric Loken for their advice on analyzing the data; Kerry Jordan for her help in running the experiments and scoring trials; Fritz Tsao and Joe Swingle for their help in analyzing data; and Asif Ghazanfar for his comments on an earlier version of this article. Correspondence concerning this article should be addressed to Daniel J. Weiss, who is now at Brain and Cognitive Sciences Department, Meliora Hall, Office 494, University of Rochester, Rochester, New York 14627. Electronic mail may be sent to dweiss@bcs.rochester.edu. 258 This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. VOCAL PRODUCTION AND PERCEPTION BY TAMARINS & Stebbins, 1988; Snowdon & Cleveland, 1980; Snowdon et al., 1983). First, we report the findings from an acoustic analysis of the CLC designed to assess the types of information encoded in this call. Next, we report the results from a series of playback experiments designed to test whether tamarins can be discriminated by voice alone, and more specifically, by the acoustic morphology of their CLCs. The goal of the acoustic analyses was to determine whether the morphology of the tamarin's CLCs carries information about individual identity, sex, and group membership. Given previous work with other closely related species, we anticipated finding evidence of individual vocal signatures (e.g., common marmosets, Jones, Harris, & Catchpole, 1993; black-tufted-ear marmosets: Jorgensen & French, 1998) and sex-specific vocal signatures (e.g., red-chested moustached tamarins: Masataka, 1988; golden lion tamarins: Benz, French, & Leger, 1990). Evidence of group-level acoustic signatures has been weaker, coming primarily from chimpanzees (Mitani & Brandt, 1994) and the more closely related pygmy marmosets (Elowson & Snowdon, 1994). Such withingroup acoustic convergence has been taken as evidence of vocal learning. To date, there is no evidence showing that convergence in call morphology occurs when multiple groups are housed in close proximity, a situation that arose in our captive colony. The playback experiments were designed to complement the acoustic analysis of individual variation within CLCs. For a tamarin to be recognized by voice alone, two conditions must be satisfied. First, the sender must produce a signal that contains an individual signature. Second, the receiver must be able to detect differences between signals on the basis of the parameters that define the signature (Beecher, 1982). On the basis of these conditions, individual recognition has been successfully demonstrated in a variety of species, including pikas (Conner, 1985), tungara frogs (Wilczynski et al., 1999), budgerigars (Brown et al., 1988), Cory's shearwaters (Bretagnolle & Lequette, 1990), and European starlings (Chaiken, 1992). Because prior work (e.g., Snowdon et al., 1983) and the acoustic analyses presented in Part 1 suggest that individual identity is encoded in the tamarin's CLC, the playback experiments presented in Part 2 were designed to explore the perceptual component of individual recognition, assessing whether tamarins actually perceive differences in CLCs from different tamarins. Our study is structured as follows. In Part 1, we present the results of our acoustic analyses of individual, sex, and group identity and then discuss these results in light of current thinking about the sources of acoustic variation and their role in perceptual classification. In Part 2, we report findings from several experiments designed to test the extent of individual discrimination using a habituation-dishabituation paradigm; in the wild this technique has been successfully used with primates (e.g., Cheney & Seyfarth, 1988; Hauser, 1998a; Zuberbuehler, Cheney, & Seyfarth, 1999), but in captivity it has only been successfully used with speech stimuli (Ramus, Hauser, Miller, Morris, & Mehler, 2000) as opposed to species-specific signals. We conclude by discussing the benefits of integrating studies of production and perception as well as some of the limits of our experimental approach. Part 1: Call Production and Acoustic Morphology The ability to convey and receive information regarding individual identity represents an important adaptation for social com- 259 munication (Beecher, 1982; Cheney & Seyfarth, 1990). Thus, it is not surprising that this capacity is widespread across many species (e.g., frogs, Davis, 1987; dolphins: Sayigh et al., 1999; passerine birds: Stoddard, 1996). In nonhuman primates, the ability to recognize individuals on the basis of vocal signals alone has received much attention. The focus of many of these studies has been on calls that are able to convey identity over long distances (e.g., Chapman & Weary, 1990; Kajikawa & Hasegawa, 1996; Mitani, Gros-Louis, & Macendonia, 1997; Snowdon & Cleveland, 1980). The capacity of cotton-top tamarins to recognize other tamarins on the basis of the acoustics of the CLC seems likely because related species have demonstrated this capability (e.g., pygmy marmosets: Snowdon & Cleveland, 1980; common marmosets: Jones et al., 1993; black-tufted-ear marmosets: Jorgensen & French, 1998). Given these studies, it seems likely that most, if not all, Callitrichids will exhibit an individual signature system. This includes cotton-top tamarins. Sex differences in vocal morphology have also been reported in a number of primate species (e.g., vervet monkeys, Cercopithecus aethiops, Seyfarth, Cheney, & Marler, 1980), including several Callitrichids (e.g., moustached tamarins, Saguinus mystax, Heymann, 1987; red-chested moustached tamarins, Saguinus I. Labiatus, Masataka, 1988; golden lion tamarins, Leontopithecus rosalia, Benz et al., 1990; common marmosets, Callithrix jacchus, Norcross & Newman, 1993). Indeed, there is some evidence suggesting that long calling in marmosets and some tamarin species may play a role in mating (e.g., Yoneda, 1981; Masataka, 1988; Norcross & Newman, 1993). Our final acoustic analyses focus on the possibility of group convergence in call morphology, and consequently, on the significant between-group differences in acoustic structure. Three factors make these analyses different from those reported to date. First, all of our tamarin groups were housed in the same room, and thus had continuous acoustic contact. Second, there were no obvious motivational differences that could explain acoustic convergence within groups. Third, there were no genetic or environmental factors that could account for within-group convergence. Thus, if cage group identity is encoded in the CLC, it may indicate that some level of vocal learning is taking place. In one of the few laboratory studies conducted to find evidence of vocal learning in nonhuman primates, Elowson and Snowdon (1994) discovered that pygmy marmosets (Cebuella pygmaea) can modify their trill vocalization in response to their social environment. In addition, experiments on an array of species have shown that call structure may be modified as a result of changes in social pairings (e.g., spear-nosed bats, Boughman, 1997; pygmy marmosets, Snowdon & Elowson, 1999; budgerigars: Hile, Plummer, & Striedter, 2000). Our analyses may be more limited with respect to vocal learning in that we did not manipulate the social context of the colony. Nevertheless, if group identity is encoded in the CLCs of cotton-top tamarins, then vocal learning is a possible mechanism, and future work can be designed to explore this possibility and others. General Method Subjects. The subjects consisted of a colony of 13 cotton-top tamarins (Saguinus oedipus). The members of the colony were born in captivity at the New England Regional Primate Center (Southborough, MA) and then This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 260 WEISS, GARIBALDI, AND HAUSER housed at the Primate Cognitive Neuroscience Laboratory at Harvard University. Tamarins were housed in groups of 2 or 3 animals in a 6.0 X 5.0 X 2.5 ft (1.8 X 1.5 x 0.76 m) home cage made of stainless steel wire and Plexiglas. The cage contained tree branches, perches, and wooden nest boxes. At the start of the experiment, there were four mated pairs: one family group consisting of a mated pair and a female offspring and one female pair. During the course of the experiment, 1 female tamarin (MR) died and was replaced by a new male tamarin (RJ). The female tamarin pair was divided and paired with males. Thus, the colony consisted of five mated pairs and one family group. Their diet consisted of Purina tamarin and marmoset chow, crickets, mealworms, supplemental vitamins, and sunflower seeds. This completely balanced diet was supplemented by food received during experiments (typically Noyes [Lancaster, NH] bananas and sweet pellets, fruit, occasionally Froot Loops, and marshmallows). They were fed once a day in the early evening. Tamarins had ad libitum access to water. The cage group membership was as follows (male-female): Cage 1, AC and JG; Cage 2, DD, ES, and RB; Cage 3, RW and SH; Cage 4, SP and EN; Cage 5, ID and EM; and Cage 6, RJ and UB. The females that were separated (as mentioned above) were SH and UB. There were a number of related tamarins within the colony: RB was the daughter of DD and ES, SP was the daughter of UB, and RW and ID were twin brothers. Call acquisition. The CLCs used as exemplars for the acoustic analyses were recorded from May 1998 through March 1999. The calls were recorded in two contexts: either in isolation in an acoustic chamber or in an experiment focusing on the social aspects of communication. To obtain calls from the acoustic chamber (Model 400-A, Industrial Acoustics, Lodi, NJ) we transported the tamarins to a test room that was both visually and acoustically isolated from all other members of the colony. Tamarins were placed in a cage (45 X 45 X 20 cm) within the larger sound chamber for a period of up to 15 min. Sessions were recorded onto a digital audiotape (DAT; DaPl digital audio tape recorder, Tascam, Tokyo, Japan; sampling rate: 48 kHz) using a Sennheiser ME66 directional microphone (Sennheiser, Old Lyme, CT; frequency response: 50-20,000 Hz) or a Sennheiser MKH60P48 microphone (frequency response: 5020,000 Hz). The tape was then acquired digitally onto a Power Macintosh 7100/80 (Apple Computer, Inc., Cupertino, CA) using Sound Designer II software (1996) and an Audiomedia II card. Some sessions were recorded directly into the computer by passing the signal through the DAT and acquiring with Sound Designer II. For the calls recorded in the social communication experiment, we transported the tamarins into a test room that was visually and acoustically isolated from the colony room. The tamarins were placed in the social communication apparatus consisting of two runways separated by a Plexiglas barrier. Food troughs were placed out of sight at the end of the runway. The experiment tested for audience effects (sensu, Marler, Dufty, & Pickert, 1986) in the presence of food or predators. The calls that were used from these experiments were recorded while the tamarin was alone in the apparatus with no predator or other tamarin present. Sessions were taped with a Sennheiser ME66 directional microphone and the signal was passed through a Tascam DAT (see above) into a Power Macintosh 7100/80. The Sound Designer II software was used to acquire the sound onto the Audiomedia U sound card. All calls (regardless of origin) were hi-pass filtered just below the lowest frequency (as determined by a 512 fft spectrogram generated in Canary Version 1.2, 1995) and then normalized (using Sound Designer II). After the sound files were edited, we listened to them in the acoustic chamber to check for audible background noise or any distortion in the playback. A total of 129 CLCs were recorded from 12 different tamarins over a period of 1 year. A range of 3 to 21 exemplars per tamarin was examined. The calls were selected at random from our catalog of high quality recordings (obtained through the methods described above). By selecting randomly, we ensured that these calls were representative of the natural variation within our population. It should be noted that the number of vocalizations selected for each tamarin differed because of the individual differences in rates of production. Of the 129 CLCs analyzed, 74 were five-syllable CLCs. In addition to the multiple discriminant analyses (MDAs) reported on the overall call set, separate sets of MDAs were reported for this subset of calls. It should also be noted that the analyses performed for this study were repeated to ensure that the results were replicable (though these analyses are not reported here). Acoustic analysis. All calls were analyzed using Canary (Version 1.2, 1995) for Macintosh computers. Table 1 outlines the parameters that were measured for each acquired CLC (see Figure 1 for an example of a spectrogram generated for the analysis). Two factors contributed to the selection of spectral and temporal characteristics chosen for measurement. The first was that previous studies of nonhuman primates have shown that both temporal and spectral characteristics of vocalizations may be used in order to process species-specific vocalizations (e.g., see Fischer, 1998; May et al., 1988, 1989; Owren & Bernacki, 1988; Seyfarth & Cheney, 1984). In addition, in a previous study of the vocal repertoire of the cotton-top tamarin, the authors used a number of these parameters in their analysis (Cleveland & Snowdon, 1982; Snowdon et al., 1983). The use of similar parameters may facilitate comparisons across these different populations. Bonferroni adjusted factorial analyses of variance (ANOVAs) and t tests were initially performed to identify significant parameters for discriminating between individual tamarins and groups of tamarins. MDA was then used to generate functions to reliably categorize calls across parameters of interest. However, because sex is a dichotomous variable, a binary logistic regression was used instead of the MDA. Because the data set consisted of calls ranging from three to seven syllables, we focused many analyses on the first syllable and the average of the final two syllables, as this statistical approach helped avoid overfilling Ihe functions lo the data. To determine whelher our resulls were statistically significant, we used a highly conservative measure. Because there were unequal samples across tamarins, we computed chance on the basis of the largest sample from any one tamarin. For example, in Ihe individual identity analysis, the greatesl contribution came from a tamarin lhal produced 21 oul of Ihe 129 calls lhal were analyzed. Therefore, for lhal condition, chance was computed as 21/129 or 16% for all tamarins. There were Iwo groups of vocalizations for each analysis. The first group consisted of all CLCs recorded, regardless of how many syllables were present To control for some of Ihe variability of the CLCs, we restticted Ihe second group lo only five-syllable long calls. These calls were stereo- Table 1 Parameters Measured Parameter Temporal Call duration Syllable duration No. of syllables Intersyllable pause Duration to peak frequency of syllable Spectral Syllable start frequency Syllable end frequency Frequency contour Syllable peak frequency Mean high frequency Mean low frequency a fft size = 512. b fft size = 2,048. Measuremenl Waveform or spectrogram Waveform or spectrogram Spedrogram" Waveform or spectrogram From speclrogram Power spectrum al start of syllable6 Power spectrum al end of syllable Difference between peak frequency al start and end of syllable Speclrogram Spectrum Lowesl frequency in the syllable15 c On the basis of power spectrum. 261 VOCAL PRODUCTION AND PERCEPTION BY TAMARINS Time(s) This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. Figure 1. A spectrogram from a five-syllable long call. Note that temporal information such as syllable duration, overall duration, and intersyllable intervals may be easily extracted from this representation. typed in that they always consisted of two chirps followed by three whistles. The five-syllable CLCs were also used as stimuli in our perceptual studies. Results Individual identity. A Bonferroni adjusted factorial ANOVA (p < .006) shows that there are a number of spectral and temporal parameters that significantly differ across tamarins (see Tables 2 and 3). A pairwise comparison across all tamarins shows that different pairs of tamarins vary with respect to the number of significant spectral and temporal differences. This suggests that individual tamarin callers differ from each other in a variety of ways and that some tamarins may be more closely related to others in their call morphology; this finding is clearly limited by the fact that it is restricted to the parameters that we selected to measure. A series of MDAs were used to explore whether we could reliably distinguish tamarins on the basis of the acoustic parameters of their CLCs. MDA: First syllable and average of two terminal syllables. Sixty-five of the 129 CLCs were used in a discriminant function analysis. A series of functions were generated that reliably predicted tamarin identity for 91% of the calls. When the 64 calls that we withheld were entered into the function, they were correctly classified 63% of the time. This result is significantly above chance (binomial test, chance = 16%, p < .001). Overall, this analysis was able to reliably predict individual identity 77% of the time. Figure 2 shows the plot of the centroids (function averages) of each tamarin for the first three canonical functions used in the analysis. This plot displays a three-dimensional representation of the relationship between the CLCs of each tamarin on the basis of the most important canonical functions. The associated eigenvalues were greater than 10% of the sum of the eigenvalues of all functions. This is a common criterion used in choosing which functions should be included in an analysis (see Yntema, 1997). Overall, several spectral and temporal parameters were important for classification. Spectral parameters included the mean low frequency of the last two syllables and the mean high and mean low frequencies of the first syllable. Temporal parameters included total duration of the call and average duration to the peak frequency of the two terminal syllables. MDA: First syllable and average of the last two syllables for five-syllable calls. Forty of the 74 five-syllable CLCs were used to create a MDA that was able to predict the identity of the caller with an accuracy of 98%. When the remaining 34 calls were entered into the analysis, we correctly classified them 56% of the time (binomial test, chance = 19%, p < .001). In total, the discriminant analysis was able to predict the identity of the caller with an accuracy of 78%. There were a number of spectral and temporal parameters that were important in classifying the data set. The temporal parameters included overall call duration, average syllable duration (for the two terminal whistles), and average duration to peak frequency (for the two terminal whistles). The most significant spectral parameters included the mean high frequency and the mean low frequency of the first syllable as well as the average mean low frequency of the two terminal syllables. Sex differences. We performed a series of / tests (Bonferroni adjusted, p < .005) across each parameter of the call to determine whether there were any significant differences based on the sex of the caller. Table 2 shows the results of these tests using data from Table 2 Significant Differences: All Calls Syllable order Parameter Temporal Duration No. of syllables Syllable duration Intersyllable pause Duration to peak frequency of syllable Spectral Syllable start frequency Syllable pause frequency Syllable end frequency Frequency contour Mean high frequency Mean low frequency Note. 1st syllable , S, C I, C ,C ,C ,C ,C 2nd to last syllable I, S, C I, C I, S, C I, S I, S, C Last syllable ,S, C ,c ,S, C ,S Syllable average I, S, C I, C I, S I I, S, C I C I, S I.C ,c c ,C I I I, C I, c I, c I = individual identity; S = sex identity; C = cage group membership. For all values, p < .005. 262 WEISS, GARIBALDI, AND HAUSER Table 3 Significant Differences: 5-Syllable Calls Syllable order Parameter Temporal Duration Syllable duration Intersyllable pause Duration to peak frequency of syllable Spectral This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. Syllable start frequency Syllable pause frequency Syllable end frequency Frequency contour Mean high frequency Mean low frequency Note. 1st syllable 2nd syllable 3rd syllable 4th syllable 5th syllable I, S, C I, S, C I, S, C I I, S, C I, S, C I, S, C I, S, C I, S I I I I, S C C C C C I, S I, S, C C C C I, S I I = individual identity; S = sexual identity; C = cage group membership. For all values, p < .005. the first syllable and the two terminal syllables in addition to the averages across all syllables. Temporal parameters accounted for most of the significant differences between sexes. Female calls tended to be longer than male calls by an average of 437 ms. The terminal syllables in female CLCs were also significantly longer in duration (M difference = 159 ms, SD = 23.0 ms). The intersyllable latency before the second-to-last syllable was longer for female calls (M difference = 10 ms, SD = 2.5 ms). The duration to the peak frequency of the next-to-last syllable was longer for 4' I EN4 JG1 u. U CF1 Figure 2. This figure contains a plot of each tamarin's centroids (across all long calls) for the first three canonical functions (CF) used in the analysis of individual identity. Note that this type of plot represents an acoustic space for the tamarin callers. The distances between the centroids indicate how the tamarins differ with respect to each of these dimensions. The superscript numbers show the cage group membership for each tamarin. Lowercase letters represent male tamarins, and uppercase letters represent female tamarins. females than for males (M difference in duration = 176 ms, SD = 26.0 ms). Binary logistic regression: First syllable and average of the last two syllables. Sixty-five calls were used in the binary logistic regression. A linear function was created that classified the sex of the caller with an accuracy of 85%. The generated function was able to classify the remainder of the call set (64 calls) with an accuracy of 70% for males (binomial test, chance = 43%, p = .004) and 76% for females (binomial test, chance = 57%, p = .017). Overall the classification average was 76% for males (binomial test, chance = 43%, p < .001) and 82% for females (binomial test, chance = 57%, p < .001) across all CLCs. Of further relevance, the parameters that were significant in the t tests were also the most important in defining the function. These included call duration, syllable duration, and duration to the peak frequency of the call. Sex differences in five-syllable calls. A total of 29 male and 45 female five-syllable CLCs were used in this analysis. Bonferroni adjusted t tests were used to compare parameters across all syllables. The results from these tests were similar to those observed with all CLCs included (see Table 3). Female calls were significantly longer than male calls (including longer whistles). Further, the duration to the peak frequency of the last three syllables was also significantly longer in female calls. Binary logistic regression: First syllable and average of the two terminal syllables in five-syllable long calls. There were not enough five-syllable long calls from both sexes to generate a reliable function with the first half of the calls and then check the reliability with the remaining calls. We, therefore, entered all of the calls into the regression and found that the calls could be assigned to the correct sex with an accuracy of 85% (76% of males, 91% of females). These results are significantly above chance (binomial test: males, chance = 39%, p = .0001; females, chance = 61%, p < .00001). Many of the parameters that were important in determining the function when all calls were analyzed were also important in this analysis. These included duration of the entire call, syllable duration, and duration to the peak frequency. This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. VOCAL PRODUCTION AND PERCEPTION BY TAMARINS Cage-group membership. In general, several temporal and spectral parameters appeared to differ between cage groups (see Table 2). Temporal parameters included duration of the entire call, duration of the last two syllables, number of syllables in the call, duration to the peak frequency of the fourth syllable, and the intersyllable latency between the third and fourth syllables. Important spectral parameters included the start, peak, and end frequency of the first syllable as well as the mean high and mean low frequencies of the first syllable. The temporal parameters that were significant included the overall duration, the duration of the last two syllables, duration to the peak frequency of the fourth syllable, and the intersyllable latency between the fourth and fifth syllables. MDA: First syllable and average of last two syllables. A series of functions were created that predicted cage-group membership with an accuracy of 77%. The remaining calls were classified by these functions with an accuracy of 52% (binomial test: chance = 25%, p < .001). Using all calls, we reliably predicted the analysis of cage-group membership with an accuracy of 64%. The results across individual cages varied, ranging from 30% to 80%. The first two canonical (discriminant) functions generated in the analysis accounted for 74% of the variance in the data. Figure 3 shows the plot of the centroids for each cage group on the first 2 canonical functions. The parameters that were important for generating the functions were similar to the parameters that were significant in the factorial ANOVAs. The important temporal parameters included total duration of the call, the duration of the first syllable, and the average duration of the last two syllables. Important spectral parameters included the mean high and mean low frequencies of the first syllable as well as the start and peak frequencies of the first syllable. Cage-group differences in five-syllable long calls. A series of factorial ANOVAs (Bonferroni adjusted, p < .006) revealed significant differences across a number of parameters (see Table 3). Significant temporal parameters included total call duration, duration of the two terminal syllables, and duration to the peak frequency of the fifth syllable. Significant spectral parameters included the mean high and mean low frequencies of the first three syllables and the start, peak, and end frequencies of the first syllable. MDA: First syllable and the average of the last two syllables. Forty of the 79 five-syllable CLCs were used in the initial analysis. A series of functions were generated that predicted cage-group membership with an accuracy of 80%. The remaining calls were then correctly classified with an accuracy of 56% (binomial test: 2 1 5: 0 -1 4 3 -2 6 - 1 0 1 2 3 CF1 Figure 3. This figure contains a plot of each cage's centroids (across all long calls) for the first two canonical functions (CF) generated in the analysis of cage group membership. 263 chance = 34%, p = .004). The combined results yielded a classification average of 69%. Accuracy varied across cage group from 29% to 100%. The first two functions accounted for 95% of the variance in the data. The parameters that were most important for these functions resembled the list of parameters that were significant in the factorial ANOVAs. Important temporal parameters included total duration of the call, the average duration of the fourth and fifth syllables, and the duration to the peak frequency in the first and last two syllables. Important spectral parameters included the mean high and mean low frequencies of the first syllable and the start, peak, and end frequencies of the first syllable. Discussion The main findings from our analyses are that the CLCs of the cotton-top tamarin can be categorized by individual, sex, and group identity. For individual identity, both spectral and temporal parameters were important for classification. For sexual identity, temporal parameters seemed most important. For group identity, temporal parameters were important as were some spectral parameters related to the first syllable. Although these findings do not allow us to conclude that tamarins use the selected features to identify the identity, sex, or group of the caller, or that they are able to do so, they do reveal that within this call type, there is the potential for such recognition. It is not surprising that we found acoustic information about individual tamarin identity encoded in the CLCs. As mentioned earlier, individual recognition is widespread among many different species (see above) because it plays an important role in social communication. The finding that sexual identity may be contained within the acoustics of the call suggests that long calling may play a role in mating or group composition. This is consistent with findings from closely related species (e.g., common marmosets: Norcross & Newman, 1993). In fact, the importance of temporal factors in this analysis resembles findings in red-chested moustached tamarins (Masataka, 1988) as well as in the common marmosets' phee calls (Norcross & Newman, 1993). It is also possible, of course, that sex differences in call morphology are a by-product of differences in vocal tract morphology that were selected for reasons other than calling and mate choice. These findings need to be augmented with both field studies as well as perceptual experiments in order to verify that sexual identity can be discriminated and used in mate choice. We may look to the closely related phee call in marmosets to gain some insight as to why sex differences in CLCs may be important. It has been hypothesized that phee calls may function to advertise reproductive status to conspecifics (Moynihan, 1970). In addition, field studies have shown that solitary tamarins may use long calls as a mechanism for vocally searching for mates. Young female tamarins have been observed to approach unfamiliar male callers (see Masataka, 1988; Norcross & Newman, 1993; Yoneda, 1981). In contrast, preliminary results from phonotaxis experiments, in which calls were played from known and unknown individual tamarins, suggest that female tamarins approach known males, whereas males approach unknown females (Miller, Hauser, Miller, & Gilda Costa, 2001). More research is necessary to This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 264 WEISS, GARIBALDI, AND HAUSER determine whether and how the CLC plays a role in attracting potential mates. Our finding that information about group identity is encoded in the acoustics of the CLC suggests that some degree of vocal learning may occur within each cage group. This finding is of particular interest because there have been few examples of vocal learning in primates (see Janik & Slater, 1997). In addition, although previous studies of vocal learning have found differences between colonies (e.g., Elowson & Snowdon, 1994; Marshall, Wrangham, & Arcadi, 1999), the tamarins used in our analyses were housed within the same room, allowing for visual, olfactory, and acoustic contact between members of different cages. However, further research is necessary in order to establish that vocal learning occurs. This includes experiments designed to measure the extent to which call matching occurs (during calling bouts) as well as recordings from newly mated pairs. The extent to which our findings will generalize to other species or other call types within species must be tempered by a number of factors. Norcross and Newman (1993) discovered differences in call structure that depended on the context of the recordings. When vocalizations were recorded from an isolated tamarin as opposed to a group-housed tamarin, there were more syllables per call and an overall increase in call duration. These modifications may have made the callers easier to localize. Likewise, in our study, it is important to bear in mind the context in which these calls were recorded. The calls were taken from tamarins that were isolated from their groups. Thus, these findings may not extend to CLCs that are produced by tamarins that are not isolated. Similarly, one must be careful in comparing the acoustics from this set of recorded tamarin calls with other call sets. Given the findings of Jorgensen and French (1998) that phee calls in marmosets may not be stable over time despite the persistence of individual characteristics, it would be interesting to investigate the stability of the tamarin CLC. It should be noted that the entire tamarin call sample used in this study was recorded within a 12-month period to diminish the effects of longitudinal changes. As mentioned earlier, in the study of communication it is important to consider aspects of production and perception. On the basis of the fact that the cotton-top tamarin's CLC encodes acoustic information about tamarin caller identity, sex, and group membership, in the following playback experiments, we investigated whether such acoustic differences are perceptually salient. Specifically, we asked whether cotton-top tamarins can discriminate individual tamarins by their calls alone. If tamarins are able to perceive differences between callers, then future experiments may investigate which features are perceptually salient and underlie this ability. Part 2: Playback Experiments Several studies, spanning a wide variety of primate species, have shown that individual identity may be encoded in contact calls (e.g., chimpanzees, Marler & Hobbett, 1975; squirrel monkeys, Symmes, Newman, Talmage-Riggs, & Lieblich, 1979; ringtailed macaques, Macedonia, 1986; common marmosets, Jones et al., 1993; blue monkeys, Butynski, Chapman, & Chapman, 1992; black-tufted-ear marmosets, Jorgensen & French, 1998). However, these studies do not address whether individual animals use the information encoded in the signal to identify other individual animals by voice. Playback experiments are essential for addressing the problem of individual recognition from the perceptual perspective. Studies of several primate species have used playback experiments to explore the question of individual recognition (e.g., vervet monkeys: Cheney & Seyfarth, 1980, 1982; rhesus macaques: Hansen, 1976; Kendall, Rodman, & Edmond, 1996; barbary macaques: Hammerschmidt & Fischer, 1998; pygmy marmosets: Snowdon & Cleveland, 1980; chimpanzees: Kajikawa & Hasegawa, 1996). Of these studies, few have been performed with captive populations (e.g., Kajikawa & Hasegawa, 1996; Snowdon & Cleveland, 1980) and none have made use of the habituationdishabituation paradigm that has proved effective in field studies (e.g., Cheney & Seyfarth, 1988; Kendall et al., 1996; Seyfarth & Cheney, 1990; Hauser, 1998a; Zuberbuehler et al., 1999). This paradigm is powerful because it allows researchers to investigate perceptual classification in animals without training. There are several reasons to expect that tamarins can recognize individuals on the basis of their CLCs. First, results from our acoustic analyses (reported earlier) suggest that information about individual tamarin identity is contained within the CLC. Second, previous studies with related species (e.g., pygmy marmosets, Snowdon & Cleveland, 1980), as well as one study with cotton-top tamarins (Snowdon et al., 1983), provide evidence of individual recognition; the latter study used a different kind of long call and a different method. Finally, the function of the CLC has been described as an affiliative signal (e.g., Cleveland & Snowdon, 1982), sometimes occurring when an individual animal is separated from its group. Therefore, one would expect CLCs to provide sufficient information for individual recognition. These experiments represent only an initial step in our attempt to identify the critical features mediating perceptual classification of vocalizations in cotton-top tamarins. Because we anticipated finding evidence of individual recognition for the cotton-top tamarin's CLC, we generated the following prediction: If we habituate our test subject to several CLC exemplars from one individual and then present this subject with a CLC from a different individual, the test subject should dishabituate (i.e., respond) to the new individual's call. However, if the subject is habituated to CLCs from one individual and then presented with a novel CLC from the same individual, the subject should transfer habituation (i.e., not respond). Therefore, our first series of playbacks used this procedure to test for responses to changes in tamarin caller identity for both opposite and same-sex pairings. General Method Apparatus. The test cage used for the playback experiments was housed in an acoustic chamber. During experiments, we placed the tamarins in a wire and cloth test cage (45 X 45 X 20 cm) with a wire floor. A thin black cotton sheet hung behind the cage. Behind the sheet, an Alesis Monitor One speaker (Alesis Corporation, Santa Monica, CA; frequency range, 45-18,000 Hz ± 3 dB) was mounted on a shelf above the box, either directly behind, or to the left or right of center (location was changed between experiments to prevent habituation to sounds emitted from one location). A video camera (Videolabs Flexcam Videolabs, Minneapolis, MN) was used to record and monitor the sessions. An Alesis RA-100 amplifier drove the speaker. Hie experimenters watched the session on a monitor outside of the acoustic chamber. The experiment was run using a HyperCard (1996) This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. VOCAL PRODUCTION AND PERCEPTION BY TAMARINS program on a Power Macintosh 7100/80 AV. We played back calls using an Audiomedia n sound card (sampled at 48 kHz) outputting to an Alesis Monitor 1 speaker. A log of each session was recorded in HyperCard. This allowed the experimenters to enter data on-line and keep track of time elapsed between trials. We recorded vocal responses during trials using a Sennheiser MKH60P48 directional microphone (frequency response, 5020,000 Hz). Stimuli. Calls used in this experiment were recorded during sessions of another experiment. We recorded additional calls while the animals were in the playback chamber. In both cases, the calls were recorded on a Tascam DAT using a Sennheiser ME-60 microphone. We hi-pass filtered the calls using Sound Designer n software. To determine the frequency of the filter, we calculated a spectrogram (1024 fft) and measured the lowest frequency of the call. After filtering the call at that frequency, all of the calls were normalized to 100% for amplitude. Only high quality calls were selected for use in this experiment. We screened each call by examining the spectrograms (1024 fft) as well as listening to each exemplar to ensure that they were free of any artifacts (e.g., cage noise, clipping, etc.). Procedure. We placed a tamarin into the playback chamber and allowed it to acclimate for 1 min before we played back any recorded sounds. After 1 min elapsed, we waited for the tamarin to face away from the speaker before playing the first call of the habituation set. Habituation stimuli were always randomly presented. If more than one cycle of the habituation set was required during a session, we presented the second set of stimuli in a different, randomized order than the first set. For each of the experiments presented in this study, we played a series of calls until the tamarin failed to respond on three consecutive trials. After three no-response trials, we played the test call followed by a posttest if necessary (see below). Responses were measured using the following parameters: head turning toward speaker, body orienting toward speaker, movement toward speaker, freezing response (if animal was moving and then froze when the recorded call was played), and vocal responses (producing vocalizations within 5 s of hearing the recorded call). A "no response" was recorded in the absence of these responses. During sessions, trials with ambiguous responses (as determined by the experimenters while running a session) were treated functionally as "yes" responses. Ambiguous responses included trials in which the tamarin was facing the speaker before any sound was played (because of either an errant press by the experimenter or a slight latency between button press and sound) or when experimenters were unable to determine the direction of the tamarin's gaze. This conservative approach was adopted to minimize the number of trials that needed to be rerun (and thereby minimize exposure to the test stimuli). Intertrial interval was set at a minimum of 10 s and at a maximum of 30 s. The session was aborted if we were unable to play back a call within this 30-s window. However, this did not occur during our experiments. Trials started once the tamarin's face was visible and turned away from the speaker. In addition, if the tamarin spontaneously produced a long call, we waited at least 5 s before playing back a call. When tamarins failed to respond to the test call, we played back a posttest stimulus. The posttest stimulus, a tamarin scream, represented a call that was both acoustically and functionally different from the habituation and test calls. Scream exemplars were recorded while the medical staff caught the tamarins and administered tuberculosis shots and then normalized for amplitude. The duration of the posttest screams was within the range of the habituation stimuli. The purpose of the posttest was to ensure that the tamarin had not habituated to the playback setup in general. Failure to respond to the test stimulus could indicate that the tamarin had either habituated to the playback setup in general or had perceptually clustered the habituation and test stimuli into one category. If the tamarin had habituated to the playback setup, then it should ignore the posttest stimulus. In contrast, if failure to respond to the test call was due to the tamarin's perception of similarity between the habituation series and the test stimulus, then it should respond to the posttest stimulus. Sessions in 265 which the tamarins failed to respond to the posttest were rerun at a later date. Analyses. Because responses to playbacks can be difficult to score in real time, we digitized (Adobe Premiere version 4.2; Adobe Systems Inc., San Jose, CA) the last 13 habituation trials (including the three no response habituation trials), the test, and the posttest. This ensured that there would be a distribution of both yes and no response trials, thereby reducing the opportunity for biased scoring. These trials were assigned code file names and then randomized to ensure that the scorers were naive to the condition. Once scores were obtained, we consulted the master list to determine the condition. Two scorers then analyzed each of the trials and recorded a yes, no, or ambiguous response; the type of response (e.g., head turn, countercall, etc.) and the duration of response were recorded as well. Trials in which it was not clear whether a tamarin had looked at the speaker were deemed ambiguous. These responses were treated as yes responses if they occurred in the habituation series. If a test trial was determined to be ambiguous, then the trial was rerun. This occurred in less than 8% of all sessions. Sessions were discarded if scorers assessed any of the three habituation trials (the three trials preceding the test trial) as a yes response. Likewise, sessions in which there was disagreement between scorers on any of the three no response habituation trials, the test, or the posttest stimuli were also discarded. Finally, sessions in which the tamarin did not respond to both the test and the posttest were also discarded. Condition 1: Individual recognition. In the first condition, we presented tamarins with a habituation set composed of eight different fivesyllable CLCs from male RW. The test stimulus was a five-syllable CLC from female ES. RW and ES were from different groups. The posttest stimulus was a scream from male SP matched in length and filtered in accordance with the habituation and test stimuli. As a control for this condition, we presented tamarins with the same habituation series from RW and a novel five-syllable CLC from RW. An additional test was conducted in which the habituation series and test stimuli were produced by tamarins of the same sex. In this condition, the habituation set consisted of eight different five-syllable CLCs from female ES. The test stimulus was a five-syllable CLC from female JG. The posttest stimulus consisted of a scream from either male SP or male ID. Condition 2: Individual recognition of same sex cage mates. Because the acoustic analyses revealed that cage mates sounded more similar to each other in some dimensions than to noncage mates, it is possible that tamarins tested in Condition 1 used those features that are salient for distinguishing among cage groups rather than individuals. Therefore in Condition 2 we also set out to test whether tamarins could discriminate callers regardless of sex and cage group membership. Because of the configuration of our colony, we had access to only a single group with two tamarins of the same sex: the mother-daughter pair, ES and RB, respectively. If our colony of tamarins could discriminate between these callers, then we could be fairly confident that they are able to recognize the differences between any two tamarins on the basis of their calls. If the tamarins failed to discriminate between these callers, however, then dishabituation in Condition 1 could be explained by the perceptual salience of both sex and group membership. In other words, although tamarins may fail to recognize individual tamarins on the basis of their five-syllable CLC, they can recognize the sex of the caller or its group membership. The habituation set for Condition 2 consisted of 8 five-syllable CLCs produced by female ES. The test stimulus was a five-syllable CLC produced by female RB. The posttest stimulus was a scream from male ID. A second habituation set was created in which RB exemplars were used for habituation and an ES exemplar was used as the test stimulus. A number of tamarins were inaccessible during the ES-RB condition, and thus, we ran only 10 tamarins. Condition 3: Discrimination of acoustically similar callers. One potential confound in Condition 2 was that the callers were more acoustically similar than those used in Condition 1 (see Discussion). To begin exploring this possibility, we designed a final series of playbacks to test whether This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 266 WEISS, GARIBALDI, AND HAUSER tamarins would discriminate between two unrelated callers with acoustically similar CLCs. On the basis of the results from the MDA (reported earlier), we chose CLCs from tamarins ID and RB as stimuli; the coordinates for ID and ES (the mother from the previous condition) were similar in the three most important acoustical dimensions and differed from the calls of RB along similar dimensions. Likewise, pairwise ANOVAs comparing the acoustics of CLCs from RB and ID revealed no significant differences (Garibaldi, 1999). The habituation set consisted of 8 five-syllable CLCs produced by female RB. The test stimulus was a five-syllable CLC produced by male ID. The posttest stimulus was a scream from ID. A second habituation set was created in which five ID exemplars were used for habituation and an RB exemplar was used as the test stimulus. The IDhab-RBtest condition was run 12 months after the other conditions (see General Discussion) when the colony expanded, allowing for 14 tamarins to be tested. Results Condition 1: Individual recognition. In the RWhab-EStes, condition, 12 out of 13 tamarins responded to the test call following habituation (see Figure 4). Tamarins' responses for this condition included 9 head turns, 1 approach, 5 long calls, and 2 chirps. Note that four test trials had multiple responses (e.g., one trial contained a head turn and a long call, etc.). The average number of calls played back until habituation was 22.8 (SD = 15.60, range = 4-58). In the control condition involving RWhab-RWtest, 3 out of 13 tamarins dishabituated to the novel RW call (see Figure 4). All three tamarins responded with head turns. The average number of calls played back until habituation was 19.8 (SD = 12.20, range = 5-49). There was a statistically significant difference in the number of tamarins responding to the test trial for these two conditions (paired sign test, p < .02). There was, however, no difference between conditions in the number of trials to habituation, paired t test, r(12) = 0.77, p < .46. For the EShab-JGtest condition, 11 out of 13 tamarins responded to the test call following habituation (see Figure 4). The responses for this condition included 7 head turns, 2 approaches, 5 long calls, and 1 trilled vocalization (consisting of multiple chirps). The average number of calls played back until habituation was 25.18 RW-ES RW-RW ES-JG ES-ES • Response Q No Response Figure 4. Results from both the initial change of identity condition, RWhab-ESt<:st including the control RWhab-RWtest, and from the same-sex change of identity condition, ES^-JG,^, including the control EShab- (SD = 27.59, range = 5-99). In the control condition involving EShab-ES,,.st, 2 out 13 tamarins dishabituated to the novel ES call (see Figure 4). One tamarin responded with a trilled vocalization and the other responded with an antiphonal call. The average number of trials until habituation was 12.84 (SD = 7.02, range = 4-27). There was a statistically significant difference in the number of individual tamarins responding to the test trial for these two conditions (paired sign test, p < .004). There was no significant difference between conditions in the number of trials until habituation, paired t test, f(12) = 1.69, p < .12. Condition 2: Individual recognition of same-sex cage mates. In the EShab-RBtest condition, 5 out of 10 tamarins dishabituated to the RB test call. The observed responses included 4 head turns, 1 trilled response (chirps), and 1 antiphonal call. The average number of calls played back until habituation was 26.3 (SD = 20.60, range = 5-57). The proportion of those responding was not significant when compared with baseline results from EShab-ES,est (paired sign test, p < .38). In the RBhab-ESt(,st condition, 8 out of 13 tamarins dishabituated to the ES test call (see Figure 5). The observed responses included 4 head turns, 2 approaches, and 4 antiphonal long calls (see Figure 6). The average number of calls played back until habituation was 28.7 (SD = 21.00, range = 5-66). The proportion of those responding was not significant when compared with the baseline results (paired sign test, p < .13). There was no difference in the number of trials until habituation in the EShab-RBtest and in the RBhab-ESMS, conditions, paired t test, t(9) = 0.16, p < .88. Condition 3: Discrimination of acoustically similar callers. In the RBhab-IDKst condition, 9 out of 13 tamarins dishabituated to the ID test call (see Figure 7). The responses included 4 head turns, 2 approaches, 1 chirp, and 2 antiphonal long calls. The average number of calls played back until habituation was 18.6 (SD = 14.30, range = 5-58). The proportion of those responding was significant when compared with baseline results (paired sign test, p < .04). In the IDhab-RBtest condition, 11 out of 14 tamarins dishabituated to the RB test call. The responses included 5 head turns and 10 antiphonal calls. The average number of calls played back until habituation was 17.5 (SD = 15.30, range = 5-50). The proportion of those responding was significant when compared with the baseline results (paired sign test, p < .02). There was no difference in the number of trials until habituation in the RBhab-IDtest and in the IDhab-RBtcst conditions, paired t test, f(12) = 0.66, p < .53. Playbacks of own calls. Because of the design of our experiments, a number of tamarins heard their own calls played back either during the habituation series or as the test call. This presented an interesting situation because the tamarins had never heard their own calls broadcast from an external source. One possible response is that tamarins perceived these calls as coming from an unfamiliar caller. A second possibility is that tamarins perceive such calls as familiar but not as their own call (e.g., from a cage mate). Finally, tamarins may recognize the call as their own and respond differently from how they respond to unfamiliar and familiar callers. Research conducted with birds has shown that when an individual bird's own song is played back, it responds with an intermediate response between that of a familiar neighbor and that of a stranger (e.g., Brooks & Falls, 1975; Searcy, McArthur, Peters, & Marler, 1981; Yasukawa, Bick, Wagman, & Mailer, 1982; 267 VOCAL PRODUCTION AND PERCEPTION BY TAMARINS This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. General Discussion Figure 5. Results from the related female tamarins' change of identity condition, RBhab-ESMst as well as the reverse EShab-RB,cst. McArthur, 1986). Snowdon et al. (1983) studied quiet and normal long calls in cotton-top tamarins and reported that tamarins responded to their own calls in the same way that they responded to a cage mate's calls. We did not run experiments involving unknown tamarins. Consequently, we cannot contrast the tamarin's response to its own call as opposed to calls produced by unfamiliar tamarins. We can, however, explore differences in response to self-call trials and other call categories. Summary. Across all conditions, tamarins that heard their own calls during the habituation series took an average of 11.9 trials (SD = 7.04) to habituate. This is not significantly different from the number of trials those same tamarins required to habituate to other call sets (M = 9.9, SD = 5.30), paired t test, f(5) = 0.52, p < .62. For conditions involving a change in caller identity, tamarins whose calls were played as the habituation series responded to the test call in three out of six trials. This is not significantly different from how the rest of the sample performed on change of identity conditions, unpaired t test, f(57) = 0.976, p < .34. For tamarins whose calls were played as the test call, 5 out of 6 tamarins responded. This is not significantly different from how the rest of the sample performed on change of identity conditions, unpaired t test, ?(51) = 0.66, p < .52. In control conditions in which tamarin caller identity remained constant (i.e., RWhab-RWtest, ES^-ES^, both tamarins transferred habituation to the novel exemplar from their own repertoire. Results from Condition 1 provide support for the hypothesis that cotton-top tamarins can discriminate between individual tamarins on the basis of their five-syllable CLC. Of the 26 trials involving a change in caller identity, tamarins responded 23 times. In addition, the results from the control conditions indicate that the methodology is well suited for identifying which aspects of the CLC may be meaningful to the perceiver. Of the 26 control trials involving novel calls from the same caller, only 5 tamarins dishabituated. This finding is critical in that it indicates that the tamarins were not simply responding to a novel acoustic change in the test condition. Rather, tamarins responded to changes that were meaningful—in this case, a change in caller identity. These results also show that the habituation-dishabituation procedure can be successfully used with captive nonhuman primates to test for perceptual classification of species-specific calls. Additional evidence for the effectiveness of this procedure comes from studies testing speech perception with cotton-top tamarins (Ramus et al., 2000). The results from this first condition suggest that antiphonal calling may be another assay that indicates whether tamarins find the change in caller identity meaningful. This assay has been used before in studies with pygmy marmosets (Snowdon & Cleveland, 1980) and cotton-top tamarins (Snowdon et al., 1983). Overall, in 11 out of 26 trials, tamarins called back to the test call when identity changed, whereas only 1 out of 26 control trials yielded a similar response (see Figure 6 for a diagram of antiphonal calling responses by condition). It should be noted that in tests of speech perception, tamarins never respond with antiphonal calling (Ramus et al., 2000), suggesting that this behavior may be restricted to conspecific vocalizations. The implications of the antiphonal calling data are discussed below. Results from Condition 2 suggest that tamarins fail to discriminate between two same-sex, related tamarins from the same group. The results of this condition underscore the importance of considering aspects of both production and perception in the study of communication. On the basis of the acoustic analyses alone, one might have predicted that these 2 tamarins could be discriminated. Both ES and RB have discrete points on the plot of individual centroids on the canonical functions generated by the MDA (see Figure 2). By performing perceptual experiments in conjunction with these analyses, we obtained a more complete understanding of how tamarins recognize individual tamarins on the basis of acoustics alone. 12 •s 6 I. I RW-ES RW-RW ES-JG ES-ES RB-ES I ES-RB Figure 6. Antiphonal calling data across all conditions. ID-RB RB-ID 268 WEISS, GARIBALDI, AND HAUSER 12 RB-ID ID-RB H Response This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 171 No Response Figure 7. Results from the acoustically similar tamarin callers' change of identity condition, RBhab-IDttst as well as the reverse This conclusion, however, is limited to our methodological assay. In contrast to a psychophysical task that asks whether Signal A is discriminable from Signal B, the habituation-dishabituation procedure asks whether Signal A is meaningfully different from Signal B (see Nelson & Marler, 1990). Thus, although the tamarins apparently did not perceive a meaningful difference between the CLCs of ES and RB, a psychophysical task may well show that their calls are discriminable. An open question is whether tamarins can discriminate individual tamarins of the same sex and the same group. Recall that the relationship between the 2 tamarins whose calls were played in Condition 2 was that of parent-offspring. It may be that unrelated tamarins within a group can be discriminated. Although our laboratory is not currently set up to allow for us to answer this question, tamarin groups in the wild often contain unrelated tamarins of the same sex (Neyman, 1977). In addition, it would also be interesting to test whether members of the same group have an advantage in discriminating related tamarins from their own group. One limitation to Condition 2 was that when the motherdaughter pair were plotted using MDA, the centroids were closer than those of the other tamarins used in Condition 1 (see Figure 2). In addition, pairwise ANOVAs revealed that there was only one parameter that significantly differed between these 2 tamarins (the starting frequency for some of the syllables in their calls). In contrast, the tamarins from the first condition differed along more dimensions. In particular, the CLCs of RW and ES differed in call duration, in duration of the individual syllables, and in duration to the peak frequency of the call. The CLCs of ES and JG differed in the duration of the syllables and the intersyllable interval (Garibaldi, 1999). Overall, this suggests that the calls produced by these 2 tamarins have greater overlap in call morphology than the exemplars used for previous test conditions. Therefore, there are multiple interpretations for the playback results. Condition 3 examined whether results from the first two conditions could be explained solely on the basis of acoustic similarity and, thus, degree of discriminability. Specifically, if ES and RB could not be discriminated because of acoustic similarity, then ID and ES should also prove difficult to discriminate. Results from Condition 3 indicate that tamarins can discriminate the CLCs of two unrelated tamarins even though the degree of acoustic similarity is high. The findings from this condition again underscore the importance of considering both production and perception when studying communication. On the basis of the relative acoustic proximity of ID and ES (used in Condition 2), one could not have predicted the results from these experiments— specifically, that RB and ES were difficult to discriminate between whereas ID and RB were not. Although studies of production help establish the kinds of information encoded in the signal (e.g., individual identity, sexual identity, etc.), it is only through perceptual experiments that we can identify which acoustic features are actually meaningful to the tamarins. It should be noted that our acoustic analyses created only one of many possible renditions of the acoustic space for tamarin CLCs. Even the results from pairwise ANOVAs may be limited by sample size. It is possible that there is one parameter, or even a constellation of parameters, that distinguishes the ID calls from the ES calls and that underlies the perceptual playback results. One approach to solving this issue is to run playback experiments with calls that have individual parameters manipulated. For example, since ID and RB are of opposite sex, whereas ES and RB are of the same sex, we could manipulate temporal parameters of the call because these have been shown to be important in our acoustic analyses of sexual identity. This kind of manipulation would determine which parameters meaningfully contribute to individual tamarin recognition and, in turn, would help explain why the related females were more difficult to discriminate between than ID and RB were. Thus, part of our overall research program focuses on manipulating individual features and determining which parameters are meaningful in the perception of the CLC. The antiphonal calling data from this condition were somewhat surprising. In the RBhab-IDtest condition, only 2 out of 9 responding tamarins produced antiphonal calls. It is unclear why this test exemplar elicited few antiphonal responses and still was detected by most tamarins as a change in speaker identity (see Discussion). In contrast, 10 out of 11 responding tamarins called antiphonally to the test stimulus in the IDhab-RBtest condition. This represents a higher level of response than in other conditions (see Figure 6). One factor that may have contributed to this effect is that this condition was run 12 months after the initial individual tamarin recognition conditions. It is possible that because the test call had been recorded 12 months earlier, it sounded less familiar. Previous research has shown that the phee call in marmosets changes over a 12-month period (Jorgensen & French, 1998), and it is unknown whether this occurs in tamarin contact calls as well. One possibility not addressed in our study is that tamarins cannot distinguish between the calls of their cage mates, regardless of sex. However, this does not seem likely because Condition 3 showed that the tamarins are able to discriminate between acoustically similar callers of the opposite sex. It is likely that they would respond to the acoustic differences known to exist between the sexes. Future studies may address this issue further. One finding from the analysis of self-call perception is that tamarins are likely to dishabituate when their own calls are used as the test stimulus in conditions involving change of identity. However, because most of our tamarins responded to changes in speaker identity, further research is necessary to determine whether tamarins are particularly sensitive to the acoustics of their own call when it is played as a test. For tamarins that heard their own calls in the habituation series, half did not respond to the change of identity. However, because of the small sample, this finding is not statistically significant. In 269 This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. VOCAL PRODUCTION AND PERCEPTION BY TAMARINS addition, the number of exemplars prior to habituation did not change for tamarins that habituated to a series of their own calls. This result supports Snowdon et al.'s (1983) position that self-calls may be perceived as comparable with calls produced by known tamarins. Of further relevance, results from experiments designed to elicit antiphonal calling have found that there is no significant difference between the tamarins' rate of antiphonal calling to their own calls and the tamarins' rate of antiphonal calling to those of other tamarins (Ghazanfar, Flombaum, Miller, & Hauser, 2001); in these experiments as well, however, tamarins have not been tested in a condition involving their own calls versus foreign tamarins' calls. As mentioned earlier, results from studies on bird songs have found that birds respond to self-song with an intermediate response between that of a familiar neighbor and that of a stranger (e.g., Brooks & Falls, 1975; McArthur, 1986; Searcy et al., 1981; Yasukawa et al., 1982). One interpretation of these results is that self-song production provides a template to which other songs are compared as a criterion for species recognition (see McArthur, 1986; Searcy et al., 1981). Our data suggest that there is no distinction between hearing one's own calls and hearing a familiar individual's calls. Additional research is necessary to determine whether an individual primate's own calls play a special role in primates' vocal perception. When hearing a long call, conspecifics often respond with their own antiphonal long calls (Cleveland & Snowdon, 1982). Thus, antiphonal calling may provide the most natural assay for gauging responses. In our experiments, the first and third test conditions had more antiphonal calling than the second condition (in which discrimination was difficult; see Figure 6) or the control conditions. However, there were still some anomalous findings (see Condition 3) with respect to the number of antiphonal calls elicited by the different test stimuli. Further research is necessary to determine which components of the call are critical to eliciting these responses. For example, research on gorilla double grunts has found that there are specific acoustic cues that tend to elicit antiphonal calling (Seyfarth, Cheney, Harcourt, & Stewart, 1994). Studies by Ghazanfar et al. (2001) suggest that tamarins are more likely to produce antiphonal calls in response to complete CLCs when contrasted with playbacks of chirps alone, whistles alone, or white-noise bursts that mimic the temporal patterning of the CLC without the species-typical amplitude envelope. One goal of these experiments was to determine whether the habituation-dishabituation paradigm could be effectively implemented in captivity to address issues pertaining to meaningful differences in vocalizations. The results from these experiments suggest that this paradigm is effective. In fact, it should be noted that the IDhab-RBtest condition was run more than 12 months after the first condition of this study. The responses in this condition (which included a high proportion of antiphonal calling) suggest that the effect is robust and that the paradigm is consistent across time. The overall goals for the analyses and experiments reported in this study were met. Future research may determine which features of the CLC are considered meaningful by perceivers. This can be accomplished by habituating tamarins to a series of CLCs and then presenting a test call in which one acoustic parameter has been manipulated. The acoustic analyses reported in this study can be used to guide the creation of these test calls. References Adobe premiere [computer software]. (1998). San Jose, CA: Adobe Systems Inc. Altmann, S. A. (1967). The structure of primate social communication. In S. A. Altmann (Ed.), Social communication among primates (pp. 325362). Chicago: University of Chicago Press. Beecher, M. D. (1982). Signature systems and kin recognition. American Zoologist, 22, 477-490. Benz, J. J., French, J. A., & Leger, D. W. (1990). Sex differences in vocal structure in a callitrichid primate (Leontopithecus rosalia). American Journal of Primatology, 21, 257-264. Boughman, J. W. (1997). Greater spear-nosed bats give group-distinctive calls. Behavioral Ecology and Sociobiology, 40, 61—70. Bretagnolle, V., & Lequette, B. (1990). Structural variation in the call of the Cory's shearwater (Calonectris diomedea, aves, Procellariidae). Ethology, 85, 313-323. Brooks, R. J., & Falls, J. B. (1975). Individual recognition in whitethroated sparrows: I. Discrimination of songs of neighbors and strangers. Canadian Journal of Zoology, 53, 879-888. Brown, S. D., Dooling, R. J., & O'Grady, K. E. (1988). Perceptual organization of acoustic stimuli by budgerigars (Melopsittacus undulatus): in. Contact calls. Journal of Comparative Psychology, 102, 236247. Burmeister, S., Wilczynski, W., & Ryan, M. J. (1999). Temporal call changes and prior experience affect graded signalling in the cricket frog. Animal Behaviour, 57, 611-618. Butynski, T. M., Chapman, C. A., & Chapman, L. J. (1992). Use of male blue monkey "pyow" calls for long-term individual identification. American Journal of Primatology, 28, 279-284. Canary (Version 1.2) [computer software]. (1995). Ithaca, NY: Cornell Bioacoustics Workstation. Chaiken, M. (1992). Individual recognition of nestling distress screams by European starlings (Stumus vulgaris). Behaviour, 120, 139-150. Chapman, C. A., & Weary, D. M. (1990). Variability in spider monkeys' vocalizations may provide basis for individual recognition. American Journal of Primatology, 22, 279-284. Cheney, D. L., & Seyfarth, R. M. (1980). Vocal recognition in free-ranging vervet monkeys. Animal Behaviour, 28, 362-367. Cheney, D. L., & Seyfarth, R. M. (1982). How vervet monkeys perceive their grunts: Field playback experiments. Animal Behaviour, 30, 739751. Cheney, D. L., & Seyfarth, R. M. (1988). Assessment of meaning and the detection of unreliable signals by vervet monkeys. Animal Behaviour, 36, 477-486. Cheney, D. L., & Seyfarth, R. M. (1990). How monkeys see the world: Inside the mind of another species. Chicago: University of Chicago Press. Cheney, D. L., & Seyfarth, R. M. (1996). The function and mechanisms underlying baboon contact barks. Animal Behaviour, 52, 507-518. Cleveland, J., & Snowdon, C. T. (1982). The complex vocal repertoire of the adult cotton-top tamarin, Saguinus oedipus oedipus. Zeitschrift fur Tierpsychologie, 58, 231-270. Conner, D. A. (1985). The function of the pika short call in individual recognition. Zeitschrift fur Tierpsychologie, 67, 131-143. Davis, M. S. (1987). Acoustically mediated neighbor recognition hi the North American bullfrog, Rana catesbeiana. Behavioral Ecology and Sociobiology, 21, 185-190. Elowson, A. M., & Snowdon, C. T. (1994). Pygmy marmosets, Cebuella pygmaea, modify vocal structure in response to changed social environment. Animal Behaviour, 47, 1267-1277. Emlen, S. T. (1972). An experimental analysis of the parameters of bud song eliciting species recognition. Behaviour, 41, 130-171. Fischer, J. (1998). Barbary macaques categorize shrill barks into two call types. Animal Behaviour, 55, 799-807. This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 270 WEISS, GARIBALDI, AND HAUSER Garibaldi, B. T. (1999). Lang calling in cotton-top tamarins. Unpublished bachelor's thesis, Harvard University. Ghazanfar, A. A., Flombaum, J. I., Miller, C. T., & Hauser, M. D. (2001). The units of perception in the antiphonal calling behavior of cotton-top tamarins (Saguinus oedipus): Playback experiments with long calls. Journal of Comparative Physiology A., 187, 27-35. Hammerschmidt, K., & Fischer, J. (1998). Maternal discrimination of offspring vocalizations in barbary macaques (Macaco sylvanus). Primates, 39, 231-236. Hansen, E. W. (1976). Selective responding by recently separated juvenile rhesus monkeys to the calls of their mother. Developmental Psychobiology, 9, 83-88. Hauser, M. D. (1998a). Functional referents and acoustic similarity: Field playback experiments with rhesus monkeys. Animal Behaviour, 55, 1647-1658. Hauser, M. D. (1998b). Orienting asymmetries in rhesus monkeys: The effect of time-domain changes on acoustic perception. Animal Behaviour, 56, 41-47. Heymann, E. W. (1987). Behavior and communication of moustached tamarins, Saguinus mystax mystax (Primates: Calitrichidae), in an outdoor enclosure: I. The physical structure of long calls of the moustached tamarin. Primate Report, 17, 45-52. Hile, A. G., Plummer, T. K., & Striedter, G. F. (2000). Male vocal imitation produces call convergence during pair bonding in budgerigars, Melopsittacus undulatus. Animal Behaviour, 59, 1209-1218. HyperCard Player (Version 2.3.5) [Computer software]. (1996). Cupertino, CA: Apple Computer, Inc. Janik, V. M., & Slater, P. J. B. (1997). Vocal learning in mammals. Advances in the Study of Behavior, 26, 59—99. Jones, B. S., Harris, D. H., & Catchpole, C. K. (1993). The stability of the vocal signature in phee calls of the common marmoset, Callithrix jacchus. American Journal of Primatology, 31, 67-75. Jorgensen, D. D., & French, J. A. (1998). Individuality but not stability in marmoset long calls. Ethology, 104, 729-742. Kajikawa, S., & Hasegawa, T. (1996). How chimpanzees exchange information by pant-hoots: A playback experiment. The Emergence of Human Cognition and Language, 3, 123—128. Macedonia, J. M. (1986). Individuality in the contact call of the ring-tailed lemur (Lemur catta). American Journal of Primatology, 11, 163-179. Marler, P., Dufty, A., & Pickert, R. (1986). Vocal communication in the domestic chicken: II. Is a sender sensitive to the presence and nature of a receiver? Animal Behaviour, 34, 194-198. Marler, P., & Hobbett, L. (1975). Individuality in a long-range vocalization of wild chimpanzees. Zeitschrift fur Tierpsychologie, 38, 97-109. Marshall, A. J., Wrangham, R. W., & Arcadi, A. C. (1999). Does learning affect the structure of vocalizations in chimpanzees? Animal Behaviour, 58, 825-830. Marten, K., Quine, D. B., & Marler, P. (1977). Sound transmission and its significance for animal vocalization: n. Tropical habitats. Behavioral Ecology and Sociobiology, 2, 291-302. Masataka, N. (1988). The response of red-chested moustached tamarins to long calls from their natal and alien populations. Animal Behaviour, 36, 55-61. May, B. J., Moody, D. B., & Stebbins, W. C. (1988). The significant features of Japanese monkey coo sounds: A psychophysical study. Animal Behaviour, 36, 1432-1444. May, B. J., Moody, D. B., & Stebbins, W. C. (1989). Categorical perception of conspecific communication sounds by Japanese macaques, Macacafuscata. Journal of the Acoustical Society of America, 85, 837-847. McArthur, P. (1986). Similarity of playback songs to self song as a determinant of response strength in song sparrows (Melospiz/a melodia). Animal Behaviour, 34, 199-207. Mitfer, C. T., Hauser, M. D., Miller, J., & Gilda Costa, R. (2001). Patterns of selective phonotaxis in cotton-top tamarins. Manuscript in preparation. Mitani, J. C. (1985). Responses of gibbons (Hylobates muelleri) to self, neighbor, and stranger duets. International Journal of Primatology, 6, 193-200. Mitani, J. C., & Brandt, K. L. (1994). Social factors influence the acoustic variability in the long-distance calls of male chimpanzees. Ethology, 96, 233-252. Mitani, J. C., Gros-Louis, J., & Macendonia, J. M. (1997). Selection for acoustic individuality within the vocal repertoire of wild chimpanzees. International Journal of Primatology, 17, 569-583. Mitani, J. C., & Nishida, T. (1993). Contexts and social correlates of long-distance calling by male chimpanzees. Animal Behaviour, 45, 735746. Morton, E. S. (1986). Predictions from the ranging hypothesis for the evolution of long-distance signals in birds. Behaviour, 83, 66-86. Moynihan, M. (1970). The control, suppression, decay, disappearance, and replacement of displays. Journal of Theoretical Biology, 29, 85-112. Nelson, D. A., & Marler, P. (1990). The perception of birdsong and an ecological concept of signal space. In W. C. Stebbins & M. A. Berkley (Eds.), Comparative perception: Vol. II. Complex signals (pp. 443-478). New York: Wiley. Neyman, P. F. (1977). Aspects of the ecology and social organization of free-ranging cotton-top tamarins (Saguinus oedipus) and the conservation status of the species. In D. G. KJeiman (Ed.), The biology and conservation of the Callitrichidae (pp. 39-71). Washington, DC: Smithsonian Institution Press. Norcross, J. L., & Newman, J. D. (1993). Context and gender-specific differences in the acoustic structure of common marmoset (Callithrix jacchus) phee calls. American Journal of Primatology, 30, 37-54. Owren, M. J., & Bernacki, R. (1988). The acoustic features of vervet monkey (Cercopithecus aethiops) alarm calls. Journal of the Acoustical Society of America, 83, 1927-1935. Ramus, F., Hauser, M. D., Miller, C. T., Morris, D., & Mehler, J. (2000, April 14). Language discrimination by human newboms and cotton-top tamarin monkeys. Science, 288, 349-351. Rendall, D., Rodman, P. S., & Edmond, R. E. (1996). Vocal recognition of individuals and kin in free-ranging rhesus monkeys. Animal Behaviour, 51, 1007-1015. Romer, H., & Bailey, W. J. (1986). Insect hearing in the field: H. Male spacing behaviour and correlated acoustic cues in the bushcricket Mygalopsis marki. Journal of Comparative Physiology, 159, 627-638. Ryan, M. J., & Sullivan, B. K. (1989). Transmission effects on temporal structure in the advertisement calls of two toads, Bufo woodhousii and Bufo valliceps. Ethology, 80, 182-189. 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. Animal Behaviour, 57, 41-50. Searcy, W. A., McArthur, P., Peters, S. S., & Marler, P. (1981). Response of male song and swamp sparrows to neighbour, stranger, and self songs. Behaviour, 77, 152-163. Seyfarth, R. M., & Cheney, D. L. (1984). The acoustic features of vervet monkey grunts. Journal of the Acoustical Society of America, 75, 129134. Seyfarth, R. M., & Cheney, D. L. (1990). The assessment by vervet monkeys of their own and another species' alarm calls. Animal Behaviour, 40, 754-764. Seyfarth, R. M., Cheney, D. L., Harcourt, A. H., & Stewart, K. J. (1994). The acoustic features of gorilla double grunts and their relation to behavior. American Journal of Primatology, 33, 31—50. Seyfarth, R. M., Cheney, D. L., & Marler, P. (1980). Vervet monkey alarm calls: Semantic communication in a free-ranging primate. Animal Behaviour, 28, 1070-1094. This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. VOCAL PRODUCTION AND PERCEPTION BY TAMARINS Snowdon, C. T., & Cleveland, J. (1980). Individual recognition of contact calls by pygmy marmosets. Animal Behaviour, 28, 717-727. Snowdon, C. T., Cleveland, J., & French, J. A. (1983). Responses to context- and individual-specific cues in cotton-top tamarin long calls. Animal Behaviour, 31, 91-101. Snowdon, C. T., & Elowson, M. A. (1999). Pygmy marmosets modify call structure when paired. Ethology, 105, 893-908. Sound Designer II [Computer software]. (1996). Palo Alto, CA: Digidesign. Stoddard, P. K. (1996). Vocal recognition of neighbors by territorial passerines. In D. E. Kroodsma & E. H. Miller (Eds.), Ecology and evolution of acoustic communication in birds (pp. 356-374). Ithaca, NY: Cornell University Press. Symmes, D., Newman, J. D., Talmage-Riggs, G., & Lieblich, A. K. (1979). Individuality and stability of isolation peepes in squirrel monkeys. Animal Behaviour, 27, 1142-1152. Thurwachter, W. (1980). Einzel- und gruppenverhalten des Krallenaffchens Saguinus oedipus oedipus (Wollkopftamarin). Gefangenschaft unter besonderer Berucksichtigung der LautauBerung "Long Call." Zulassungsarbeit: Univ. Freiburg i. Br. Wilczynski, W., Rand, A. S., & Ryan, M. J. (1999). Female preferences for temporal order of call components in the tungara frong: A Bayesian analysis. Animal Behaviour, 58, 841-851. Wiley, R. H., & Richards, D. G. (1978). Physical constraints on acoustic communication in the atmosphere: Implications for the evolution of animal vocalizations. Behavioral Ecology and Sociobiology, 3, 69-94. Yasukawa, K., Bick, E. I., Wagman, D. W., & Marler, P. (1982). Playback and speaker-replacement experiments on song-based neighbor, stranger, and self-discrimination in male red-winged blackbirds. Behavioral Ecology and Sociobiology, 10, 211-215. Yntema, D. (1997). Psychology 1952 course materials. Unpublished course notes, Harvard University. Yoneda, M. (1981). Ecological studies of Saguinus fuscicollis and Saguinus labiatus with reference to habitat segregation and height preference. Kyoto University Overseas Research Report New World Monkeys, 2, 43-50. Zuberbuehler, K., Cheney, D. L., & Seyfarth, R. M. (1999). Conceptual semantics in a nonhuman primate. Journal of Comparative Psychology, 113, 33-42. Received July 21, 2000 Revision received February 22, 2001 Accepted February 24, 2001 Members of Underrepresented Groups: Reviewers for Journal Manuscripts Wanted If you are interested in reviewing manuscripts for APA journals, the APA Publications and Communications Board would like to invite your participation. Manuscript reviewers are vital to the publications process. As a reviewer, you would gain valuable experience in publishing. The P&C Board is particularly interested in encouraging members of underrepresented groups to participate more in this process. If you are interested in reviewing manuscripts, please write to Demarie Jackson at the address below. Please note the following important points: • • • • 271 To be selected as a reviewer, you must have published articles in peer-reviewed journals. The experience of publishing provides a reviewer with the basis for preparing a thorough, objective review. To be selected, it is critical to be a regular reader of the five to six empirical journals that are most central to the area or journal for which you would like to review. Current knowledge of recently published research provides a reviewer with the knowledge base to evaluate a new submission within the context of existing research. To select the appropriate reviewers for each manuscript, the editor needs detailed information. Please include with your letter your vita. In your letter, please identify which APA journal(s) you are interested in, and describe your area of expertise. Be as specific as possible. For example, "social psychology" is not sufficient—you would need to specify "social cognition" or "attitude change" as well. Reviewing a manuscript takes time (1-4 hours per manuscript reviewed). If you are selected to review a manuscript, be prepared to invest the necessary time to evaluate the manuscript thoroughly. Write to Demarie Jackson, Journals Office, American Psychological Association, 750 First Street, NE, Washington, DC 20002-4242.