- Email: [email protected]
Vocal matching by orange-fronted conures (Aratinga canicularis)
Behavioural Processes 82 (2009) 133–139
Contents lists available at ScienceDirect
Behavioural Processes journal homepage: www.elsevier.com/locate/behavproc
Vocal matching by orange-fronted conures (Aratinga canicularis) Thorsten J.S. Balsby ∗,1 , Jack W. Bradbury Cornell Laboratory of Ornithology, Cornell University, Ithaca, NY, Denmark
a r t i c l e
i n f o
Article history: Received 9 October 2008 Received in revised form 25 May 2009 Accepted 25 May 2009 Keywords: Contact call Fission–fusion Imitation Parrots
a b s t r a c t The functions of vocal matching have been clarified in territorial songbirds, compositionally stable groups of birds and mammals, and species with multiple alarm or assembly signals. The functions of vocal matching are less well understood in fission/fusion species that are non-territorial, live in groups with variable composition, and lack multiple alarm signals. Here we present the results of interactive playbacks in a fission/fusion parrot species, the orange-fronted conjure (Aratinga canicularis), that provide evidence of vocal matching. A randomly selected loud contact call (chee) per trial was played to passing wild flocks and short-term captives in Costa Rica. Of the trials where subjects interacted, 30% of wild flocks and 21% of captive trials showed significantly linear or curvilinear changes in similarity between the stimulus and response chees over the course of the trial. Surprisingly, both convergent and divergent sequences were observed, and many trials lacking a single trend showed disjunct changes in stimulus–response similarity. These results suggest that chee exchanges prior to flock fusions are not simply an exchange of greetings but are more likely some form of negotiation. This would explain the presence of convergent, divergent, and variable patterns of stimulus–response similarity seen in our experiments. © 2009 Elsevier B.V. All rights reserved.
1. Introduction Vocal matching (Catchpole and Slater, 1995; Tyack, 2008) occurs widely but for different functions. Some passerines with multiple song types use song type matching to mediate agonistic interactions during territorial defense (McGregor et al., 1992; Vehrencamp, 2000, 2001). Guenon monkeys with different alarm calls for different types of predators may echo a sentry’s alarm while adhering to call type (Cheney and Seyfarth, 1990; Zuberbuhler, 2000, 2001; Uster and Zuberbuhler, 2001). Animals that live in stable groups often converge on specific call variants that differentiate their group from others. Avian examples include cardueline finches (Marler and Mundinger, 1975; Mundinger, 1971, 1979; Groth, 1993); chickadees (Mammen and Nowicki, 1981; Nowicki, 1989); and parrots (Saunders, 1983; Farabaugh et al., 1994; Wright, 1996; Wanker et al., 1998; Bartlett and Slater, 1999; Hile and Striedter, 2000; Hile et al., 2000; Wanker and Fisher, 2001). There is a fourth and less intuitive context of vocal matching. Bottle-nosed dolphins (Tursiops truncatus) live in fission–fusion societies within insufficient compositional group stability to support group-specific calls (Connor et al., 2000). Instead, each individual regularly broadcasts its own distinctive “signature whis-
∗ Corresponding author. Tel.: +4551862515. E-mail address: [email protected] (T.J.S. Balsby). 1 Present address: Animal Behaviour Group, Department of Biology, University of Copenhagen, Universitetsparken 15, 2100 Copenhagen E, Denmark. 0376-6357/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.beproc.2009.05.005
tle” (Sayigh et al., 1990, 1999; Smolker et al., 1993; Mann and Smuts, 1998; Janik, 1999; Tyack, 2000, 2003; Watwood et al., 2005; Janik et al., 2006). Dolphins can mimic the signature whistles of other individuals (Sayigh et al., 1990; Janik, 2000; Tyack, 2003). It has been suggested that this form of vocal matching solicits the attention of and further interaction with the individual whose call has been copied, similar to humans calling out the name of an associate at a noisy social gathering (Tyack, 1993, 2003; Janik and Slater, 1998). Orange-fronted conures (Aratinga canicularis) live in a similar fission/fusion society (Bradbury et al., 2001; Bradbury, 2003). They differ from dolphins in that minimal social units are mated pairs (Hardy, 1963, 1965; Bradbury, 2003), whereas dolphins live in matrilineal female groups and separate all-male coalitions (Connor et al., 2000; Moller et al., 2001; Krutzen et al., 2003). Like dolphins, these conures exhibit loud contact calls (chees) that are individually distinctive even within mated pairs (Bradbury et al., 2001; Cortopassi and Bradbury, 2006). Each individual conure has a few chee types in their repertoire that can be distinguished on spectrograms (Cortopassi and Bradbury, 2006). Chees are broadcast throughout the day, both in flight and when perched, and are exchanged when flocks meet and often just prior to fissions or fusions of flocks (Bradbury, 2003). Field playbacks to assess the salience of chee geographical variation revealed partial matching of response chees to fixed stimuli (Vehrencamp et al., 2003). This suggested that wild conures, like dolphins, might use vocal matching to coordinate interactions with other birds. Unfortunately, the unpredictable movements of conure flocks (Bradbury et al., 2001) make it difficult to record natural
134
T.J.S. Balsby, J.W. Bradbury / Behavioural Processes 82 (2009) 133–139
transactions in the field. We, therefore, undertook interactive playbacks to see how often vocal matching of chees could be elicited experimentally, and to what degree a respondent might match the playback stimulus. One set of playbacks was performed to passing wild flocks, whereas the second utilized single birds caught in the wild and held for short periods in an aviary. While neither playback fully emulated natural flock interactions, both elicited response calling that could be quantified and examined for matching with stimuli. Both experiments allowed us to test if orange-fronted conures modify their chee during vocal interactions in any directed manner. If orange-fronted conures match our playback we expect to see increased similarity between the playback call and the bird responding to playback. In a controlled aviary experiment we also wanted to test if birds were more likely to respond to playback chees more similar to their own signature chee. 2. Materials and methods 2.1. Dates and locations All experiments were undertaken within the Area de Conservacion de Guanacaste, Costa Rica in zones exhibiting early successional stages of seasonal dry forest. Playbacks to wild flocks were undertaken on 26 days between April 1 and May 24, 2001 at four sites 2–7 km apart: Centeno (10◦ 52.69 N, 85◦ 34.37 W), Finca Jenny (10◦ 51.80 N, 85◦ 34.46 W), Pocosol (10◦ 53.55 N, 85◦ 36.02 W), and Santa Rosa (10◦ 50.35 N, 85◦ 37.07 W). Aviary experiments were undertaken at the Centeno site during May 23–August 5, 2003 (13 birds) and June 9–16, 2004 (3 birds). 2.2. Stimulus preparation Wild flock playback stimuli were single chees recorded in 2000 from naturally behaving wild birds within 1 km of each playback site using an 816T MKH Sennheiser microphone and a HHB Portadat (PDR1000) recorder (44.1 kHz/16 bit sampling). Aviary trial stimuli were recorded near the Centeno and Finca Jenny playback sites in 2001–2002. All stimuli were filtered to remove low frequency noise, and their amplitudes standardized using Syrinx (http://www.syrinxpc.com). Aviary stimuli were marked with a 35 ms 14 kHz tone (above this species’ range of hearing (Wright et al., 2003)) to distinguish them from responses during analyses. 2.3. Wild flock playback protocols Playbacks to wild flocks were rotated among the four sites so that three or more days (mean = 6.2 days) elapsed between return to the same site. This rotation rate relative to flock compositional turnover (Bradbury et al., 2001), the spacing between sites relative to range size (Bradbury et al., 2001), access to 24 different stimuli, and the high density of conures all reduced the chances that identical flocks would be exposed to the same stimulus during this series. Playback sessions were restricted to morning (5–9 a.m.) and afternoon (3–5 p.m.), which reflects the daily activity pattern of conures. Because radio-tracking showed birds were unlikely to return to the same site on the same day (Bradbury et al., 2001), morning and afternoon sessions on the same day used the same site. Flock visitation rates were similar for the 25 morning sessions (3.4 ± 0.4 flocks/h) and the 13 afternoon sessions (3.7 ± 0.5 flocks/h). Sixteen sessions were undertaken at the Centeno site, whereas seven to eight sessions were completed at each of the Santa Rosa, Pocosol, and Finca Jenny sites. Stimulus source and playback sites were identical in 70–80% of the sessions; the remaining stimuli were drawn from one of the other three sites. The average number of sessions throughout the series in which a given stimulus was used was 1.1 for
Centeno, 1.0 for Pocosol and Santa Rosa, and 2.2 for Finca Jenny. Within a session, we used a single stimulus per trial and 1–4 stimuli/session. Wild flock playbacks were controlled and documented with Syrinx (http://SyrinxPC.com) on an Acer computer (Travelmate 610). When conures were seen or heard, we began playing the stimulus at 2–4 s intervals until they either began replying with lags similar to normal exchanges between flocks or left the area. If they appeared to be responding, we played the current stimulus interactively in response to any utterance by them (keeping response lags at 2–4 s) until they stopped calling and left. Audio output from the laptop was supplied to a Harman Kardon CA212 amplifier and 30 W omni-directional speaker (RadioShack Model 40-1352) hidden in a tree 25–40 m away and 3–5 m from the ground. The amplifier was adjusted to a natural output sound level of 90–95 dB at 1 m (as measured with Radio Shack Sound Level Meter Model 33-2055). While one experimenter oversaw the playbacks, the other recorded the size of target flocks, whether a call was a stimulus or response, and if possible, which respondent gave the call. They also monitored proximity sufficiently to define each trial as close (flock < 10 m from the microphone and speaker at all times), moderate (10–50 m), far (> 50 m), or variable (both moderate and close proximities noted). A Sennheiser MKH 816T directional microphone was mounted on a tripod near the experimenters and kept pointed at the respondents while recording all of their calls on the HHB Portadat. 2.4. Aviary playback protocol Sixteen wild conures were netted within 0.5 km of the Centeno playback site for the aviary playbacks. Only a single bird was held in an aviary at a time. The two hardware cloth aviaries (3.5 m × 1.8 m × 1.8 m) were stationed in the seasonal forest about 300 m apart to minimize interactions. Each aviary included a covered shelter at one end, perches, feeding tables, and a blind at the opposite end where an observer could run the playbacks and record all responses without disturbing the subject. Conditions for capture, husbandry, and release of the birds followed an IACUC-approved protocol. Aviary playbacks began an average 2.5 days after capture (range 1–6). Ten stimuli were used in 2003 and a different ten used in 2004. Both sets of stimuli were recorded within the typical 6 km home-range (Bradbury et al., 2001). Because conures sometimes emit chees in doublets, a subset of aviary stimuli consisted of the same chee repeated twice at the typical spacing of 50–80 ms. Stimulus playback used a Yamaha SU200 sampler, a Harman Kardon CA212 amplifier, and a Q6 Sound Sphere loudspeaker (35 W, 4 ). We adjusted output pressure levels to 57 dB at 5 m distance and 1.5 m above the ground using a RadioShack sound level meter (Model 33-2055). The speaker was hidden 7–9 m from the aviary 2–4 m above the ground and near perches used by visiting wild conures. Each aviary bird received five successive trials between 6:35 a.m. and 12:30 p.m. A trial consisted of a pre-playback period (5 min), a playback/treatment period (average duration 96.5 ± 2.97 (SE) s, range 24–154 s), and a post-playback period (5 min). Successive trials were separated by at least 15 min and a minimum of 5 min since the last passage of wild conures. Trials were abandoned if wild birds arrived during the pre- or playback periods; post-playback periods ended early if an aviary bird started responding to outside birds. Playback began with a single chee. If the subject bird responded we continued with a stimulus playback 2–4 s after each response by the subject until we had played the single chee eight times and the double chee twice. If a bird did not respond, we played back chees at roughly 15 s intervals. If the aviary bird gave double chees, we matched the first two with our double chees. If the aviary bird never produced double chees during a playback, we played the stimulus double chees as the last two stimuli in
T.J.S. Balsby, J.W. Bradbury / Behavioural Processes 82 (2009) 133–139
135
Fig. 1. Plot showing first two principal coordinate values (PCO) from the SPCC similarity matrix for stimulus and successive responses in playback trial to six wild birds recorded at less than 10 m. Two response call types were identified by visual inspection. Open circles indicate positions of successive responses with arrow showing temporal direction of sequence; dark circle shows location of stimulus. Spectrograms of sample calls are provided and their positions on trajectory labeled accordingly. Note sudden jump from initial call type on right to final call type more similar to stimulus on left.
the sequence. Subject vocalizations were recorded in all periods for 13 birds and during the playback and post-playback periods only for 3 birds using a Sennheiser MKH 816T microphone and Marantz PMD 670 or 690 recorders. We also recorded “baseline chees” from each subject on at least two occasions. Baseline chees are defined as spontaneous calls emitted when the bird was not interacting or had not interacted recently with other birds. Chees were considered baseline if two or more minutes had elapsed since the last playback or interaction with a wild bird. As a control for propagation and equipment distortions, we re-recorded all stimuli three times with the speaker and microphone at usual playback positions.
2.5. Spectrographic analyses Syrinx was used to batch-extract all chees and emission times from trial recordings. Extracted chees were assigned to stimulus or response categories using the log for the playback annotations and/or high frequency markers. Call rates for aviary trials were computed at this time. Spectrograms were generated in Syrinx (0.5–10 kHz band limit, Hann window, 90% overlap and 256 (flock playbacks) or 512 (aviary) point FFTs), and visually assigned to within-trial chee types (Cortopassi and Bradbury, 2006). We noted whether the stimulus could be visually matched to one of the respondent chee types. Several clean examples of each aviary bird’s
136
T.J.S. Balsby, J.W. Bradbury / Behavioural Processes 82 (2009) 133–139 Table 1 Numbers of visually identified chee types in responses to playbacks by wild flocks of different sizes. Weighted call types were computed from the fractional contribution of each call type to a flock’s total sample by summing the products of that fraction and its logarithm to the base 2, and then raising 2 to the power of that sum. This index corrects for situations in which multiple call types were present but were not used equally often (Cortopassi and Bradbury, 2006). Flock size
1 2 3–4 5–6 7–28
Number of flocks
8 18 13 11 21
Call types (flock)
Weighted call types (flock)
Mean ± SD
Range
Mean ± SD
1–3 1–6 1–4 1–6 1–6
1.33 1.81 1.66 2.62 2.57
1.50 2.17 2.15 3.00 3.20
± ± ± ± ±
0.93 1.29 1.14 1.55 1.28
± ± ± ± ±
0.67 1.18 0.73 1.50 1.20
Range 1.00–2.85 1.00–5.59 1.00–3.03 1.00–5.99 1.00–6.00
used nonparametric methods (Spearman). Given low signal–noise ratios, wild flock trials assigned to the “far” category were excluded from SPCC-PCO analyses. We assumed that repeat playbacks of the same stimulus to the same wild flocks were rare enough that trials were independent. 3. Results 3.1. Wild flock playback trials
Fig. 2. Another trial to wild birds plotted in same manner as in Fig. 1. This trial involved four birds recorded at 10 m or less which produced a single call type throughout the trial. Unlike the prior figure, this plot showed a gradual approach to the stimulus coordinates, in part affected by increasing call duration, and then a sudden return back to a short duration call just before the flock left the site.
baseline chee types were set aside for later comparisons. Chee types were distinguished by the contour and length of the different parts (Bradbury et al., 2001) of the chee. During interactions the chee types may vary gradually resulting in several variants of the chee type, these changes however were much smaller than the differences between types. Similarities between stimulus, response, and baseline chees (aviary playbacks) within a trial were quantified by assembling all pair-wise spectrographic cross-correlations (SPCC) (wild bird: bandwidth 0.5–12 kHz, FFT = 512 samples/frame, 50% overlap, and Hann window; aviary: bandwidth 0.5–10 kHz, FFT = 512 samples, 90% overlap, and Hann window) into a single matrix (Cortopassi and Bradbury, 2000). In Figs. 1 and 2, we use principal coordinate ordination (see Cortopassi and Bradbury, 2000) on the matrix to illustrate how the chees change. 2.6. Statistics Trials in both experiments were considered successful if subjects produced six or more chees in response to playback stimuli. Statistical testing used JMP, or the PROC GLM or PROC MIXED procedures in SAS ver 9.12 (SAS Institute Inc., Cary, NC, USA). Mixed models accounted for non-independence of chees recorded in the same trial by blocking for trial. Where necessary, continuous variables were transformed to meet conditions of residual normality, and parametric methods applied. Some regressions on aviary birds
Overall, 169 (41%) of the 409 playbacks to wild flocks resulted in vocal interactions where the flock perched in the area. Successful exchanges lasted an average 6.6 min ± 5.4 (SD) (range 1–27) and led to an average 28.7 respondent chees ± 25.3 (range 6–166). Success rate was unaffected by playback site (23 = 0.37, p = 0.946), stimulus source site (24 = 2.89, p = 0.577), time of day: (21 = 1.23, p = 0.889), or whether stimulus source and playback site were the same (21 = 1.79, p = 0.116). In the 71 successful trials where flocks perched less than 50 m from the microphone, the SPCC similarity between stimulus and responses averaged 0.670 ± 0.088 (range 0.192–0.835) with an average per trial coefficient of variation of 7% ± 4% (range 2–21%). Within-trial mean SPCC values were unrelated to playback site (F4,66 = 2.02, p = 0.103), stimulus source (F4,66 = 1.96, p = 0.111), time of day (F1,69 = 0.93, p = 0.339), duration of the exchange (r = 0.070, F1,69 = 0.343, p = 0.560), flock size (r = −0.136, F1,66 = 1.237, p = 0.270), or combinations of these variables with or without interaction terms. For the 71 successful trials in which flocks responded, we identified an average 0.74 chee-types/bird in the flock/trial ± 0.64 (range: 0.11–3.00; Table 1). Five of seven solitary birds that responded emitted only one type, a sixth produced three types with one being used 80% of the time, and the seventh also produced three types with two of them accounting for 47% and 32% of the responses. Although total and weighted numbers of types increased significantly with flock size (total types: r = 0.451, F1,65 = 16.58, p < 0.0001; weighted types: r = 0.403, F1,65 = 12.61, p = 0.0007), the rate of increase was extremely low (0.1 additional types/additional bird). Despite the low number of response call types produced, 75% of the trials produced a type that appeared to be a match to the stimulus using visual inspection. 3.1.1. Matching of playback stimuli in wild flocks Many responding parrots changed their chee’s structure monotonically in relation to the playback stimulus, as shown by a persistent directional change in the SPCC between the stimulus and successive response chees (see Fig. 1). Other trials appeared to show alternating intervals of convergence and divergence relative to the playback stimuli (see Fig. 2). A general linear model relating consecutive SPCC values within successful trials, relative order of response nested within trial, and relative order squared nested within trial was highly significant overall (F212, 1823 = 17.97, p = 0.0001), as were each of the three effects (trial: F1,70 = 45.73, p < 0.0001; relative
T.J.S. Balsby, J.W. Bradbury / Behavioural Processes 82 (2009) 133–139
137
Fig. 4. Spearman correlation coefficients of SPCC similarity between stimulus and subject response chees versus order of response for each aviary trial of 14 different subjects. Points for multiple trials on the same individual, each involving a different stimulus, are connected by a vertical line.
Fig. 3. Plots of spectrographic cross-correlation (SPCC) similarity between stimulus and successive responses versus order of response for three successful wild flock playback trials. All three examples showed significant linear and/or curvilinear relationships, and all were recorded with the birds 10 m or less from the recording microphone. Trial 25: responses from a single pair emitting two visually identified call types. Trial 48: responses from a flock of 13 birds emitting four visually identified call types. Trial 86: responses of a single bird producing a single call type throughout the trial.
chees that diverged from, rather than matched, the stimuli (Fig. 4). This was verified statistically with a mixed model relating SPCC values to trial and order of response in which order of response showed no consistent linear pattern (F1,358 = 0.03, p = 0.862), but the interaction term between trial and order was highly significant (F33,358 = 2.6, p < 0.0001), suggesting that birds in some trials converged and in other trials diverged their call relative to the playback stimuli. Aviary birds matched stimuli with subtle changes of the chee and never changed chee type during interactions. Despite the short duration of trials, 8% (6 of 74) of all trials and 17.6% (6 of 34) of successful trials showed a significantly positive Spearman correlation between SPCC and order of response, and one trial showed a significant negative correlation. Bird identity was not a predictor of pattern and 5 of the 14 subjects showed both convergent and divergent trends in different trials (mean number of trials per subject = 2.4, range 1–5) (Fig. 4).
order: F1,71 = 3.44, p < 0.0001; relative order squared: F1,71 = 3.33, p < 0.0001). Twenty-one (30%) of these trials, which is substantially more than could be expected by random (5%), showed significant post hoc evidence of linear and/or curvilinear patterns of SPCC value versus relative order of emission (Fig. 3). Of these, 12 showed concave curvilinearity (lower initial and final values relative to middle values), 7 showed convex patterns (higher initial and final values), and the remaining two showed a positive linear slope and negative linear slope respectively. The duration of a trial, determined by the birds, varied between trials. Duration was significantly positively related to respondent flock size (r = 0.508, F1,69 = 22.96, p < 0.0001). Proximity between flock and microphone, number of call types, and presence of linear or curvilinear patterns did not influence trial duration. Trials that resulted in significant convergent interactions did not result in significantly longer interaction than significant divergent interactions (t-test t15.8 = 0.3, p = 0.77).
3.2.2. Chee similarity and response to playback Chee rates during and post-playback decreased through successive trials in a session, suggesting an effect of order or time of day (playback: intercept = −0.27, F1,55 = 27.6, p < 0.001; post-playback: intercept = −0.28, F1,54 = 8.3, p = 0.006). Similarity between stimulus and subject baseline chees did not affect the chee rate during playback (intercept = 1.70, F1,55 = 0.7, p = 0.407), but it did do so in the post-playback period with higher similarity leading to higher rates (intercept = 7.96, F1,54 = 8.1, p = 0.006). In all but three trials, birds only vocalized in the post-playback period if they had previously vocalized during playback. The number of chees given during and post-playback did not vary according to whether birds converged (rs > 0.21), diverged (rs < −0.21), or made non-directional changes (−0.21 < rs < 0.21 relative to the stimulus (mixed model, playback: F2,18 = 2.4, p = 0.115; post-playback: F2,18 = 1.5, p = 0.249).
3.2. Aviary playback trials
The wild flock playbacks sought to replicate natural interactions in which passing groups respond to chees from a static group by landing nearby and exchanging calls for several minutes. Such exchanges can end either with merging of the two groups or independent departures of the original flocks. Wild flocks responded about 40% of the time to similarly staged playbacks of recorded chees, and 30% of responding flocks showed significant linear or
3.2.1. Matching of stimuli calls Playbacks were successful for 34 of the 74 aviary trials and 14 of the 16 test subjects. As with the wild flocks, only a subset of the successful playbacks provided evidence of matching. However, inspection indicated that 42% of subjects more often produced
4. Discussion
138
T.J.S. Balsby, J.W. Bradbury / Behavioural Processes 82 (2009) 133–139
curvilinear relationships between the order of a response in an exchange and the degree of spectrographic similarity between it and the stimulus. Some exchanges showed a convergent match with the stimulus, some showed divergence away from the stimulus, and others displayed alternating patterns of convergence and divergence. Wild flock playbacks demonstrated significant vocal matching in near natural contexts, while aviary experiments allowed examination of variables such as baseline and signature chee types that were impossible to assess in the context of wild flock playbacks. It was extremely difficult to know by direct observation of the flocks how many birds responded in a subject flock, and thus whether changes in call type were due to one bird shifting between alternatives, or different birds contributing at different times. Although one might have predicted that solitary aviary birds would respond to any conspecific sound, captives were as selective to playbacks as were wild flocks. Of those that did respond, many showed convergent responses to some stimuli and divergent or highly variable responses to others. This selectivity was not significantly correlated with the SPCC similarity between the bird’s signature chee and the trial stimulus. However, that similarity did significantly predict how much the bird called in the post-playback period. Both wild and short-term captive conures favor an individuallyspecific “signature” chee when calling (Cortopassi and Bradbury, 2006). The same pattern was seen here in seven solitary birds responding to field playbacks, and all 16 of the aviary subjects. At least within a trial, each of these individuals modified an existing chee type relative to the stimulus rather than switch to some alternative type. Playbacks to wild flocks also elicited a surprisingly low number of chee types. It is not known whether multiple call types within a wild flock trial were emitted by different birds, or by a single bird that changed call types. In either case, the results suggest that even in large flocks, only a few birds exchanged chees with the playback. This suggests that there may be asymmetries in leadership roles within flocks, a notion reinforced by the rapid and directed manner with which conures move between feeding sites without prior reconnaissance (Bradbury, 2003). The existence of individual signatures may also explain the low (41–46%) rates at which either playback series elicited responses. We frequently observe over-flying flocks that do not respond to chees of roosting flocks, or roosting flocks that do not reply to the calls of over-flying flocks. While there may be situations where flock fusion is not advantageous, it could also be the case that conures discriminate between the chees of individual callers and respond selectively. In both sets of experiments, a substantial number of trials showed alternating intervals of stimulus–response convergence and divergence. This alternating variation, the curvilinear patterns seen in many of the field trials, the small number of individuals responding even in large flocks, and the long duration of many field trials suggest that there is more to these vocal exchanges than simple announcements of identity. The ubiquitous change of contact calls that we see during interactions in orange-fronted conures makes it unlikely that birds are “naming” each other using vocal labeling as suggested by Wanker et al. (2005) for stable captive groups of spectacled parrotlets (Forpus conspicillatus). In fact, it is unclear how such a system would be beneficial in the fission–fusion society of these conures. Instead, it seems likely that the exchange is some form of negotiation between a few individuals. Likely functions include the determination of subsequent leadership should the two groups fuse or a sorting out of dominance. Either alternative might explain our discovery here of divergent as well as convergent call modification during exchanges. The ability to imitate call variants of other individuals offers the possibility to interact with all individuals in the population; species with fixed vocal types can only match some of the song types in the population and interactions would thus be constrained to only
part of the population. The ability to imitate contact calls would therefore be especially beneficial in species that live in dynamic social flocks such as the fission–fusion system. The social complexity could therefore favor the evolution of a flexible communication system. If natural interactions involve progressive and coordinated changes in chee structure by both parties, our experimental design clearly made this type of adjustment impossible. Responding birds may have modified their chees a certain amount and then waited for the targeted recipient, in this case the playback, to modify their chee in reply. This modification never occurred. Instead of a steady increase (or decrease) in similarity as the interaction progressed, our failure to change chees could easily have generated the oscillating and suddenly shifting response patterns seen in many trials. The production of agonistic call types just prior to the departure of many wild flocks may also reflect the failure of our playbacks to “behave normally”. Current studies in which short-term captives in the aviary are recorded during interactions with visiting wild flocks will hopefully clarify this issue. Acknowledgements This research was funded by a NSF grant to Jack Bradbury (IBN 02-29271). Special thanks go to Roger Blanco, Maria Marta Chavez, and the staff at the Area Conservacion Guanacaste for support in this study. We thank Marisa Adler, Erin Bohman, Susannah BuhrmanDeever, Ben McCue, Angelika Poesel, Judith Scarl, Jacob Slominski, and Jennifer Webb for field and analytical assistance. Francoise Vermeleyen provided statistical advice on the use of mixed models. References Bartlett, P., Slater, P.J.B., 1999. The effect of new recruits on the flock specific call of budgerigars (Melopsittacus undulatus). Ethol. Ecol. Evol. 11, 139–147. Bradbury, J.W., 2003. Vocal communication in wild parrots. In: DeWaal, F.B.M., Tyack, P.L. (Eds.), Animal Social Complexity: Intelligence, Culture and Individualized Societies. Harvard University Press, Cambridge, MA, pp. 293–316. Bradbury, J.W., Cortopassi, K.A., Clemmons, J.R., 2001. Geographical variation in the contact calls of orange-fronted parakeets. Auk 118, 958–972. Catchpole, C.K., Slater, P.J.B., 1995. Bird Song, Biological Themes and Variation. Cambridge University Press, Cambridge, p. 248. Cheney, D.L., Seyfarth, R.M., 1990. How Monkeys See the World. Chicago University Press, Chicago, IL, p. 388. Connor, R.C., Wells, R., Mann, J., Read, A., 2000. The bottlenose dolphin—social relationships in a fission: fusion society. In: Mann, J., Connor, R., Tyack, P.L., Whitehead, H. (Eds.), Cetacean Societies: Field studies of Whales and Dolphins. University of Chicago Press, Chicago, IL, pp. 91–126. Cortopassi, K.A., Bradbury, J.W., 2000. The comparison of harmonically rich sounds using spectrographic cross-correlation and principal coordinates analysis. Bioacoustics 11, 89–127. Cortopassi, K.A., Bradbury, J.W., 2006. Contact call diversity in wild orange-fronted parakeet pairs, Aratinga canicularis. Anim. Behav. 71, 1141–1154. Farabaugh, S.M., Linzenbold, A., Dooling, R.J., 1994. Vocal plasticity in buderigars (Melopsittacus undulatus): evidence for social factors in the learning of contact calls. J. Comp. Psychol. 108, 81–92. Groth, J.G., 1993. Call matching and positive assortative matching in Red crossbills. Auk 110, 398–491. Hardy, J.W., 1963. Epigamic and reproductive behavior of the orange-fronted parakeet. Condor 65, 169–199. Hardy, J.W., 1965. Flock social behavior of the orange-fronted parakeet. Condor 67, 140–156. Hile, A.G., Plummer, T.K., Striedter, G.F., 2000. Male vocal imitation produces call convergence during pair bonding in budgerigars, Melopsittacus undulatus. Anim. Behav. 59, 1209–1218. Hile, A.G., Striedter, G.F., 2000. Call convergence within groups of female budgerigars (Melopsittacus undulatus). Ethology 106, 1105–1114. Janik, V.M., 1999. Origins and implications of vocal learning in bottlenose dolphins. In: Box, H.O., Gibson, K.R. (Eds.), Mammalian Social Learning: Comparative and Ecological Perspectives. Symposia of the Zoological Society of London, No. 72. Cambridge University Press, Cambridge, UK, pp. 308–332. Janik, V.M., 2000. Whistle matching in wild dolphins (Tursiops truncatus). Science 289, 1355–1357. Janik, V.M., Slater, P.J.B., 1998. Context-specific use suggests that bottlenose dolphin signature whistles are cohesion calls. Anim. Behav. 56, 829–838. Janik, V.M., Sayigh, L.S., Wells, R.S., 2006. Signature whistle shape conveys identity information to bottlenose dolphins. Proc. Natl. Acad. Sci. U.S.A. 103, 8293–8297.
T.J.S. Balsby, J.W. Bradbury / Behavioural Processes 82 (2009) 133–139 Krutzen, M., Sherwin, W.B., Connor, R.C., Barre, L.M., Van de Casteele, T., Mann, J., Brooks, R., 2003. Contrasting relatedness patterns in bottlenose dolphins (Tursiops sp.) with different alliance strategies. Proc. R. Soc. Lond. B 270, 497–502. Mammen, D.L., Nowicki, S., 1981. Individual differences and within flock convergence in chickadee calls. Behav. Ecol. Sociobiol. 9, 179–186. Mann, J., Smuts, B.B., 1998. Natal attraction: allomaternal care and mother-infant separations in wild bottlenose dolphins. Anim. Behav. 55, 1097–1113. Marler, P., Mundinger, P.C., 1975. Vocalizations, social organization, and breeding biology of the twite, Acanthus flavirostris. Ibis 117, 1–17. McGregor, P.K., Dabelsteen, T., Shepherd, M., Pedersen, S.B., 1992. The signal value of matched singing in great tits: evidence from interactive playback experiments. Anim. Behav. 43, 987–998. Moller, L.M., Beheregaray, L.B., Harcourt, R.G., Krutzen, M., 2001. Alliance membership and kinship in wild male bottlenose dolphins (Tursiops aduncus) of southeastern Australia. Proc. R. Soc. Lond. B 268, 1941–1947. Mundinger, P.C., 1971. Vocal imitation and individual recognition of finch calls. Science 168, 480–482. Mundinger, P.C., 1979. Call learning in the carduelinae: ethological and systematic considerations. Syst. Zool. 28, 279–283. Nowicki, S., 1989. Vocal plasticity in captive black-capped chickadees: the acoustic basis and rate of call convergence. Anim. Behav. 37, 64–73. Saunders, D.A., 1983. Vocal repertoire and individual vocal recognition in the shortbilled white-tailed black cockatoo, Calyptorhynchus funereus latirostris Carnaby. Aust. Wildl. Res. 10, 527–536. Sayigh, L.S., Tyack, P.L., Wells, R.S., Scott, M.D., 1990. Signature whistles of freeranging bottlenose dolphins, Tursiops truncatus: stability and mother-offspring comparisons. Behav. Ecol. Sociobiol. 26, 247–260. Sayigh, L.S., Tyack, P.L., Wells, R.S., Solow, A., Scott, M., Irvine, A.B., 1999. Individual recognition in wild bottlenose dolphins: a field test using playback experiments. Anim. Behav. 57, 41–50. Smolker, R.A., Mann, J., Smuts, B.B., 1993. The use of signature whistles during separations and reunions among wild bottlenose dolphin mothers and calves. Behav. Ecol. Sociobiol. 33, 393–402. Tyack, P.L., 1993. Animal language research needs a broader comparative and evolutionary framework. In: Roitblat, H.L., Herman, L.M., Nachtigall, P.E. (Eds.), Language and Communication: Comparative Perspectives. Lawrence Erlbaum Associates Publishers, London, UK, pp. 115–152. Tyack, P.L., 2000. Functional aspects of cetacean communication. In: Mann, J., Connor, R., Tyack, P.L., Whitehead, H. (Eds.), Cetacean Societies: Field studies of Whales and Dolphins. University of Chicago Press, Chicago, IL, pp. 270–307.
139
Tyack, P.L., 2003. Dolphins communicate about individual-specific social relationships. In: DeWaal, F.B.M., Tyack, P.L. (Eds.), Animal Social Complexity: Intelligence, Culture and Individualized Societies. Harvard University Press, Cambridge, MA, pp. 342–361. Tyack, P.L., 2008. Convergence of calls as animals form social bonds, active compensation for noisy communication channels, and the evolution of vocal learning in mammals. J. Comp. Psychol. 122, 319–331. Uster, D., Zuberbuhler, K., 2001. The functional significance of Diana monkey ‘clear’ calls. Behaviour 138, 741–756. Vehrencamp, S.L., 2000. Handicap, index, and conventional signal elements of bird song. In: Amundsen, T., Rosenqvist, G., Espmark, Y. (Eds.), Animal Signals: Signalling and Signal Design in Animal Communication. Tapir Publishers, Trondheim, Norway, pp. 277–300. Vehrencamp, S.L., 2001. Is song-type matching a conventional signal of aggressive intentions? Proc. R. Soc. Lond. B 268, 1637–1642. Vehrencamp, S.L., Ritter, A.F., Keever, M., Bradbury, J.W., 2003. Responses to playback of local versus distant contact calls in the orange-fronted conure, Aratinga canicularis. Ethology 109, 37–54. Wanker, R., Fisher, J., 2001. Intra- and Interindividual variation in the contact calls of spectacled parrotlets (Forpus conspicillatus). Behaviour 138, 709– 726. Wanker, R., Apcin, J., Jennerjahn, B., Waibel, B., 1998. Discrimination of different social companions in spectacled Parrotlets (Forpus conspicillatus): evidence for individual recognition. Behav. Ecol. Sociobiol. 43, 197–202. Wanker, R., Sugama, Y., Prinage, S., 2005. Vocal labelling of familiy members in spectacled parrotlets, Forpus conspicillatus. Anim. Behav. 70, 111– 118. Watwood, S.L., Owen, E.C.G., Tyack, P.L., Wells, R.S., 2005. Signature whistle use by temporarily restrained and free-swimming bottlenose dolphins, Tursiops truncatus. Anim. Behav. 69, 1373–1386. Wright, T.F., 1996. Regional dialects in the contact call of a parrot. Proc. R. Soc. Lond. B 263, 867–872. Wright, T.F., Cortopassi, K.A., Bradbury, J.W., Dooling, R.J., 2003. Hearing and vocalizations in the conure Aratinga canicularis. J. Comp. Psychol. 117, 87–95. Zuberbuhler, K., 2000. Referential labelling in Diana monkeys. Anim. Behav. 59, 917–927. Zuberbuhler, K., 2001. Predator-specific alarm calls in Campbell’s monkeys, Cercipithecus campbelli. Behav. Ecol. Sociobiol. 50, 414–422.
Contents lists available at ScienceDirect
Behavioural Processes journal homepage: www.elsevier.com/locate/behavproc
Vocal matching by orange-fronted conures (Aratinga canicularis) Thorsten J.S. Balsby ∗,1 , Jack W. Bradbury Cornell Laboratory of Ornithology, Cornell University, Ithaca, NY, Denmark
a r t i c l e
i n f o
Article history: Received 9 October 2008 Received in revised form 25 May 2009 Accepted 25 May 2009 Keywords: Contact call Fission–fusion Imitation Parrots
a b s t r a c t The functions of vocal matching have been clarified in territorial songbirds, compositionally stable groups of birds and mammals, and species with multiple alarm or assembly signals. The functions of vocal matching are less well understood in fission/fusion species that are non-territorial, live in groups with variable composition, and lack multiple alarm signals. Here we present the results of interactive playbacks in a fission/fusion parrot species, the orange-fronted conjure (Aratinga canicularis), that provide evidence of vocal matching. A randomly selected loud contact call (chee) per trial was played to passing wild flocks and short-term captives in Costa Rica. Of the trials where subjects interacted, 30% of wild flocks and 21% of captive trials showed significantly linear or curvilinear changes in similarity between the stimulus and response chees over the course of the trial. Surprisingly, both convergent and divergent sequences were observed, and many trials lacking a single trend showed disjunct changes in stimulus–response similarity. These results suggest that chee exchanges prior to flock fusions are not simply an exchange of greetings but are more likely some form of negotiation. This would explain the presence of convergent, divergent, and variable patterns of stimulus–response similarity seen in our experiments. © 2009 Elsevier B.V. All rights reserved.
1. Introduction Vocal matching (Catchpole and Slater, 1995; Tyack, 2008) occurs widely but for different functions. Some passerines with multiple song types use song type matching to mediate agonistic interactions during territorial defense (McGregor et al., 1992; Vehrencamp, 2000, 2001). Guenon monkeys with different alarm calls for different types of predators may echo a sentry’s alarm while adhering to call type (Cheney and Seyfarth, 1990; Zuberbuhler, 2000, 2001; Uster and Zuberbuhler, 2001). Animals that live in stable groups often converge on specific call variants that differentiate their group from others. Avian examples include cardueline finches (Marler and Mundinger, 1975; Mundinger, 1971, 1979; Groth, 1993); chickadees (Mammen and Nowicki, 1981; Nowicki, 1989); and parrots (Saunders, 1983; Farabaugh et al., 1994; Wright, 1996; Wanker et al., 1998; Bartlett and Slater, 1999; Hile and Striedter, 2000; Hile et al., 2000; Wanker and Fisher, 2001). There is a fourth and less intuitive context of vocal matching. Bottle-nosed dolphins (Tursiops truncatus) live in fission–fusion societies within insufficient compositional group stability to support group-specific calls (Connor et al., 2000). Instead, each individual regularly broadcasts its own distinctive “signature whis-
∗ Corresponding author. Tel.: +4551862515. E-mail address: [email protected] (T.J.S. Balsby). 1 Present address: Animal Behaviour Group, Department of Biology, University of Copenhagen, Universitetsparken 15, 2100 Copenhagen E, Denmark. 0376-6357/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.beproc.2009.05.005
tle” (Sayigh et al., 1990, 1999; Smolker et al., 1993; Mann and Smuts, 1998; Janik, 1999; Tyack, 2000, 2003; Watwood et al., 2005; Janik et al., 2006). Dolphins can mimic the signature whistles of other individuals (Sayigh et al., 1990; Janik, 2000; Tyack, 2003). It has been suggested that this form of vocal matching solicits the attention of and further interaction with the individual whose call has been copied, similar to humans calling out the name of an associate at a noisy social gathering (Tyack, 1993, 2003; Janik and Slater, 1998). Orange-fronted conures (Aratinga canicularis) live in a similar fission/fusion society (Bradbury et al., 2001; Bradbury, 2003). They differ from dolphins in that minimal social units are mated pairs (Hardy, 1963, 1965; Bradbury, 2003), whereas dolphins live in matrilineal female groups and separate all-male coalitions (Connor et al., 2000; Moller et al., 2001; Krutzen et al., 2003). Like dolphins, these conures exhibit loud contact calls (chees) that are individually distinctive even within mated pairs (Bradbury et al., 2001; Cortopassi and Bradbury, 2006). Each individual conure has a few chee types in their repertoire that can be distinguished on spectrograms (Cortopassi and Bradbury, 2006). Chees are broadcast throughout the day, both in flight and when perched, and are exchanged when flocks meet and often just prior to fissions or fusions of flocks (Bradbury, 2003). Field playbacks to assess the salience of chee geographical variation revealed partial matching of response chees to fixed stimuli (Vehrencamp et al., 2003). This suggested that wild conures, like dolphins, might use vocal matching to coordinate interactions with other birds. Unfortunately, the unpredictable movements of conure flocks (Bradbury et al., 2001) make it difficult to record natural
134
T.J.S. Balsby, J.W. Bradbury / Behavioural Processes 82 (2009) 133–139
transactions in the field. We, therefore, undertook interactive playbacks to see how often vocal matching of chees could be elicited experimentally, and to what degree a respondent might match the playback stimulus. One set of playbacks was performed to passing wild flocks, whereas the second utilized single birds caught in the wild and held for short periods in an aviary. While neither playback fully emulated natural flock interactions, both elicited response calling that could be quantified and examined for matching with stimuli. Both experiments allowed us to test if orange-fronted conures modify their chee during vocal interactions in any directed manner. If orange-fronted conures match our playback we expect to see increased similarity between the playback call and the bird responding to playback. In a controlled aviary experiment we also wanted to test if birds were more likely to respond to playback chees more similar to their own signature chee. 2. Materials and methods 2.1. Dates and locations All experiments were undertaken within the Area de Conservacion de Guanacaste, Costa Rica in zones exhibiting early successional stages of seasonal dry forest. Playbacks to wild flocks were undertaken on 26 days between April 1 and May 24, 2001 at four sites 2–7 km apart: Centeno (10◦ 52.69 N, 85◦ 34.37 W), Finca Jenny (10◦ 51.80 N, 85◦ 34.46 W), Pocosol (10◦ 53.55 N, 85◦ 36.02 W), and Santa Rosa (10◦ 50.35 N, 85◦ 37.07 W). Aviary experiments were undertaken at the Centeno site during May 23–August 5, 2003 (13 birds) and June 9–16, 2004 (3 birds). 2.2. Stimulus preparation Wild flock playback stimuli were single chees recorded in 2000 from naturally behaving wild birds within 1 km of each playback site using an 816T MKH Sennheiser microphone and a HHB Portadat (PDR1000) recorder (44.1 kHz/16 bit sampling). Aviary trial stimuli were recorded near the Centeno and Finca Jenny playback sites in 2001–2002. All stimuli were filtered to remove low frequency noise, and their amplitudes standardized using Syrinx (http://www.syrinxpc.com). Aviary stimuli were marked with a 35 ms 14 kHz tone (above this species’ range of hearing (Wright et al., 2003)) to distinguish them from responses during analyses. 2.3. Wild flock playback protocols Playbacks to wild flocks were rotated among the four sites so that three or more days (mean = 6.2 days) elapsed between return to the same site. This rotation rate relative to flock compositional turnover (Bradbury et al., 2001), the spacing between sites relative to range size (Bradbury et al., 2001), access to 24 different stimuli, and the high density of conures all reduced the chances that identical flocks would be exposed to the same stimulus during this series. Playback sessions were restricted to morning (5–9 a.m.) and afternoon (3–5 p.m.), which reflects the daily activity pattern of conures. Because radio-tracking showed birds were unlikely to return to the same site on the same day (Bradbury et al., 2001), morning and afternoon sessions on the same day used the same site. Flock visitation rates were similar for the 25 morning sessions (3.4 ± 0.4 flocks/h) and the 13 afternoon sessions (3.7 ± 0.5 flocks/h). Sixteen sessions were undertaken at the Centeno site, whereas seven to eight sessions were completed at each of the Santa Rosa, Pocosol, and Finca Jenny sites. Stimulus source and playback sites were identical in 70–80% of the sessions; the remaining stimuli were drawn from one of the other three sites. The average number of sessions throughout the series in which a given stimulus was used was 1.1 for
Centeno, 1.0 for Pocosol and Santa Rosa, and 2.2 for Finca Jenny. Within a session, we used a single stimulus per trial and 1–4 stimuli/session. Wild flock playbacks were controlled and documented with Syrinx (http://SyrinxPC.com) on an Acer computer (Travelmate 610). When conures were seen or heard, we began playing the stimulus at 2–4 s intervals until they either began replying with lags similar to normal exchanges between flocks or left the area. If they appeared to be responding, we played the current stimulus interactively in response to any utterance by them (keeping response lags at 2–4 s) until they stopped calling and left. Audio output from the laptop was supplied to a Harman Kardon CA212 amplifier and 30 W omni-directional speaker (RadioShack Model 40-1352) hidden in a tree 25–40 m away and 3–5 m from the ground. The amplifier was adjusted to a natural output sound level of 90–95 dB at 1 m (as measured with Radio Shack Sound Level Meter Model 33-2055). While one experimenter oversaw the playbacks, the other recorded the size of target flocks, whether a call was a stimulus or response, and if possible, which respondent gave the call. They also monitored proximity sufficiently to define each trial as close (flock < 10 m from the microphone and speaker at all times), moderate (10–50 m), far (> 50 m), or variable (both moderate and close proximities noted). A Sennheiser MKH 816T directional microphone was mounted on a tripod near the experimenters and kept pointed at the respondents while recording all of their calls on the HHB Portadat. 2.4. Aviary playback protocol Sixteen wild conures were netted within 0.5 km of the Centeno playback site for the aviary playbacks. Only a single bird was held in an aviary at a time. The two hardware cloth aviaries (3.5 m × 1.8 m × 1.8 m) were stationed in the seasonal forest about 300 m apart to minimize interactions. Each aviary included a covered shelter at one end, perches, feeding tables, and a blind at the opposite end where an observer could run the playbacks and record all responses without disturbing the subject. Conditions for capture, husbandry, and release of the birds followed an IACUC-approved protocol. Aviary playbacks began an average 2.5 days after capture (range 1–6). Ten stimuli were used in 2003 and a different ten used in 2004. Both sets of stimuli were recorded within the typical 6 km home-range (Bradbury et al., 2001). Because conures sometimes emit chees in doublets, a subset of aviary stimuli consisted of the same chee repeated twice at the typical spacing of 50–80 ms. Stimulus playback used a Yamaha SU200 sampler, a Harman Kardon CA212 amplifier, and a Q6 Sound Sphere loudspeaker (35 W, 4 ). We adjusted output pressure levels to 57 dB at 5 m distance and 1.5 m above the ground using a RadioShack sound level meter (Model 33-2055). The speaker was hidden 7–9 m from the aviary 2–4 m above the ground and near perches used by visiting wild conures. Each aviary bird received five successive trials between 6:35 a.m. and 12:30 p.m. A trial consisted of a pre-playback period (5 min), a playback/treatment period (average duration 96.5 ± 2.97 (SE) s, range 24–154 s), and a post-playback period (5 min). Successive trials were separated by at least 15 min and a minimum of 5 min since the last passage of wild conures. Trials were abandoned if wild birds arrived during the pre- or playback periods; post-playback periods ended early if an aviary bird started responding to outside birds. Playback began with a single chee. If the subject bird responded we continued with a stimulus playback 2–4 s after each response by the subject until we had played the single chee eight times and the double chee twice. If a bird did not respond, we played back chees at roughly 15 s intervals. If the aviary bird gave double chees, we matched the first two with our double chees. If the aviary bird never produced double chees during a playback, we played the stimulus double chees as the last two stimuli in
T.J.S. Balsby, J.W. Bradbury / Behavioural Processes 82 (2009) 133–139
135
Fig. 1. Plot showing first two principal coordinate values (PCO) from the SPCC similarity matrix for stimulus and successive responses in playback trial to six wild birds recorded at less than 10 m. Two response call types were identified by visual inspection. Open circles indicate positions of successive responses with arrow showing temporal direction of sequence; dark circle shows location of stimulus. Spectrograms of sample calls are provided and their positions on trajectory labeled accordingly. Note sudden jump from initial call type on right to final call type more similar to stimulus on left.
the sequence. Subject vocalizations were recorded in all periods for 13 birds and during the playback and post-playback periods only for 3 birds using a Sennheiser MKH 816T microphone and Marantz PMD 670 or 690 recorders. We also recorded “baseline chees” from each subject on at least two occasions. Baseline chees are defined as spontaneous calls emitted when the bird was not interacting or had not interacted recently with other birds. Chees were considered baseline if two or more minutes had elapsed since the last playback or interaction with a wild bird. As a control for propagation and equipment distortions, we re-recorded all stimuli three times with the speaker and microphone at usual playback positions.
2.5. Spectrographic analyses Syrinx was used to batch-extract all chees and emission times from trial recordings. Extracted chees were assigned to stimulus or response categories using the log for the playback annotations and/or high frequency markers. Call rates for aviary trials were computed at this time. Spectrograms were generated in Syrinx (0.5–10 kHz band limit, Hann window, 90% overlap and 256 (flock playbacks) or 512 (aviary) point FFTs), and visually assigned to within-trial chee types (Cortopassi and Bradbury, 2006). We noted whether the stimulus could be visually matched to one of the respondent chee types. Several clean examples of each aviary bird’s
136
T.J.S. Balsby, J.W. Bradbury / Behavioural Processes 82 (2009) 133–139 Table 1 Numbers of visually identified chee types in responses to playbacks by wild flocks of different sizes. Weighted call types were computed from the fractional contribution of each call type to a flock’s total sample by summing the products of that fraction and its logarithm to the base 2, and then raising 2 to the power of that sum. This index corrects for situations in which multiple call types were present but were not used equally often (Cortopassi and Bradbury, 2006). Flock size
1 2 3–4 5–6 7–28
Number of flocks
8 18 13 11 21
Call types (flock)
Weighted call types (flock)
Mean ± SD
Range
Mean ± SD
1–3 1–6 1–4 1–6 1–6
1.33 1.81 1.66 2.62 2.57
1.50 2.17 2.15 3.00 3.20
± ± ± ± ±
0.93 1.29 1.14 1.55 1.28
± ± ± ± ±
0.67 1.18 0.73 1.50 1.20
Range 1.00–2.85 1.00–5.59 1.00–3.03 1.00–5.99 1.00–6.00
used nonparametric methods (Spearman). Given low signal–noise ratios, wild flock trials assigned to the “far” category were excluded from SPCC-PCO analyses. We assumed that repeat playbacks of the same stimulus to the same wild flocks were rare enough that trials were independent. 3. Results 3.1. Wild flock playback trials
Fig. 2. Another trial to wild birds plotted in same manner as in Fig. 1. This trial involved four birds recorded at 10 m or less which produced a single call type throughout the trial. Unlike the prior figure, this plot showed a gradual approach to the stimulus coordinates, in part affected by increasing call duration, and then a sudden return back to a short duration call just before the flock left the site.
baseline chee types were set aside for later comparisons. Chee types were distinguished by the contour and length of the different parts (Bradbury et al., 2001) of the chee. During interactions the chee types may vary gradually resulting in several variants of the chee type, these changes however were much smaller than the differences between types. Similarities between stimulus, response, and baseline chees (aviary playbacks) within a trial were quantified by assembling all pair-wise spectrographic cross-correlations (SPCC) (wild bird: bandwidth 0.5–12 kHz, FFT = 512 samples/frame, 50% overlap, and Hann window; aviary: bandwidth 0.5–10 kHz, FFT = 512 samples, 90% overlap, and Hann window) into a single matrix (Cortopassi and Bradbury, 2000). In Figs. 1 and 2, we use principal coordinate ordination (see Cortopassi and Bradbury, 2000) on the matrix to illustrate how the chees change. 2.6. Statistics Trials in both experiments were considered successful if subjects produced six or more chees in response to playback stimuli. Statistical testing used JMP, or the PROC GLM or PROC MIXED procedures in SAS ver 9.12 (SAS Institute Inc., Cary, NC, USA). Mixed models accounted for non-independence of chees recorded in the same trial by blocking for trial. Where necessary, continuous variables were transformed to meet conditions of residual normality, and parametric methods applied. Some regressions on aviary birds
Overall, 169 (41%) of the 409 playbacks to wild flocks resulted in vocal interactions where the flock perched in the area. Successful exchanges lasted an average 6.6 min ± 5.4 (SD) (range 1–27) and led to an average 28.7 respondent chees ± 25.3 (range 6–166). Success rate was unaffected by playback site (23 = 0.37, p = 0.946), stimulus source site (24 = 2.89, p = 0.577), time of day: (21 = 1.23, p = 0.889), or whether stimulus source and playback site were the same (21 = 1.79, p = 0.116). In the 71 successful trials where flocks perched less than 50 m from the microphone, the SPCC similarity between stimulus and responses averaged 0.670 ± 0.088 (range 0.192–0.835) with an average per trial coefficient of variation of 7% ± 4% (range 2–21%). Within-trial mean SPCC values were unrelated to playback site (F4,66 = 2.02, p = 0.103), stimulus source (F4,66 = 1.96, p = 0.111), time of day (F1,69 = 0.93, p = 0.339), duration of the exchange (r = 0.070, F1,69 = 0.343, p = 0.560), flock size (r = −0.136, F1,66 = 1.237, p = 0.270), or combinations of these variables with or without interaction terms. For the 71 successful trials in which flocks responded, we identified an average 0.74 chee-types/bird in the flock/trial ± 0.64 (range: 0.11–3.00; Table 1). Five of seven solitary birds that responded emitted only one type, a sixth produced three types with one being used 80% of the time, and the seventh also produced three types with two of them accounting for 47% and 32% of the responses. Although total and weighted numbers of types increased significantly with flock size (total types: r = 0.451, F1,65 = 16.58, p < 0.0001; weighted types: r = 0.403, F1,65 = 12.61, p = 0.0007), the rate of increase was extremely low (0.1 additional types/additional bird). Despite the low number of response call types produced, 75% of the trials produced a type that appeared to be a match to the stimulus using visual inspection. 3.1.1. Matching of playback stimuli in wild flocks Many responding parrots changed their chee’s structure monotonically in relation to the playback stimulus, as shown by a persistent directional change in the SPCC between the stimulus and successive response chees (see Fig. 1). Other trials appeared to show alternating intervals of convergence and divergence relative to the playback stimuli (see Fig. 2). A general linear model relating consecutive SPCC values within successful trials, relative order of response nested within trial, and relative order squared nested within trial was highly significant overall (F212, 1823 = 17.97, p = 0.0001), as were each of the three effects (trial: F1,70 = 45.73, p < 0.0001; relative
T.J.S. Balsby, J.W. Bradbury / Behavioural Processes 82 (2009) 133–139
137
Fig. 4. Spearman correlation coefficients of SPCC similarity between stimulus and subject response chees versus order of response for each aviary trial of 14 different subjects. Points for multiple trials on the same individual, each involving a different stimulus, are connected by a vertical line.
Fig. 3. Plots of spectrographic cross-correlation (SPCC) similarity between stimulus and successive responses versus order of response for three successful wild flock playback trials. All three examples showed significant linear and/or curvilinear relationships, and all were recorded with the birds 10 m or less from the recording microphone. Trial 25: responses from a single pair emitting two visually identified call types. Trial 48: responses from a flock of 13 birds emitting four visually identified call types. Trial 86: responses of a single bird producing a single call type throughout the trial.
chees that diverged from, rather than matched, the stimuli (Fig. 4). This was verified statistically with a mixed model relating SPCC values to trial and order of response in which order of response showed no consistent linear pattern (F1,358 = 0.03, p = 0.862), but the interaction term between trial and order was highly significant (F33,358 = 2.6, p < 0.0001), suggesting that birds in some trials converged and in other trials diverged their call relative to the playback stimuli. Aviary birds matched stimuli with subtle changes of the chee and never changed chee type during interactions. Despite the short duration of trials, 8% (6 of 74) of all trials and 17.6% (6 of 34) of successful trials showed a significantly positive Spearman correlation between SPCC and order of response, and one trial showed a significant negative correlation. Bird identity was not a predictor of pattern and 5 of the 14 subjects showed both convergent and divergent trends in different trials (mean number of trials per subject = 2.4, range 1–5) (Fig. 4).
order: F1,71 = 3.44, p < 0.0001; relative order squared: F1,71 = 3.33, p < 0.0001). Twenty-one (30%) of these trials, which is substantially more than could be expected by random (5%), showed significant post hoc evidence of linear and/or curvilinear patterns of SPCC value versus relative order of emission (Fig. 3). Of these, 12 showed concave curvilinearity (lower initial and final values relative to middle values), 7 showed convex patterns (higher initial and final values), and the remaining two showed a positive linear slope and negative linear slope respectively. The duration of a trial, determined by the birds, varied between trials. Duration was significantly positively related to respondent flock size (r = 0.508, F1,69 = 22.96, p < 0.0001). Proximity between flock and microphone, number of call types, and presence of linear or curvilinear patterns did not influence trial duration. Trials that resulted in significant convergent interactions did not result in significantly longer interaction than significant divergent interactions (t-test t15.8 = 0.3, p = 0.77).
3.2.2. Chee similarity and response to playback Chee rates during and post-playback decreased through successive trials in a session, suggesting an effect of order or time of day (playback: intercept = −0.27, F1,55 = 27.6, p < 0.001; post-playback: intercept = −0.28, F1,54 = 8.3, p = 0.006). Similarity between stimulus and subject baseline chees did not affect the chee rate during playback (intercept = 1.70, F1,55 = 0.7, p = 0.407), but it did do so in the post-playback period with higher similarity leading to higher rates (intercept = 7.96, F1,54 = 8.1, p = 0.006). In all but three trials, birds only vocalized in the post-playback period if they had previously vocalized during playback. The number of chees given during and post-playback did not vary according to whether birds converged (rs > 0.21), diverged (rs < −0.21), or made non-directional changes (−0.21 < rs < 0.21 relative to the stimulus (mixed model, playback: F2,18 = 2.4, p = 0.115; post-playback: F2,18 = 1.5, p = 0.249).
3.2. Aviary playback trials
The wild flock playbacks sought to replicate natural interactions in which passing groups respond to chees from a static group by landing nearby and exchanging calls for several minutes. Such exchanges can end either with merging of the two groups or independent departures of the original flocks. Wild flocks responded about 40% of the time to similarly staged playbacks of recorded chees, and 30% of responding flocks showed significant linear or
3.2.1. Matching of stimuli calls Playbacks were successful for 34 of the 74 aviary trials and 14 of the 16 test subjects. As with the wild flocks, only a subset of the successful playbacks provided evidence of matching. However, inspection indicated that 42% of subjects more often produced
4. Discussion
138
T.J.S. Balsby, J.W. Bradbury / Behavioural Processes 82 (2009) 133–139
curvilinear relationships between the order of a response in an exchange and the degree of spectrographic similarity between it and the stimulus. Some exchanges showed a convergent match with the stimulus, some showed divergence away from the stimulus, and others displayed alternating patterns of convergence and divergence. Wild flock playbacks demonstrated significant vocal matching in near natural contexts, while aviary experiments allowed examination of variables such as baseline and signature chee types that were impossible to assess in the context of wild flock playbacks. It was extremely difficult to know by direct observation of the flocks how many birds responded in a subject flock, and thus whether changes in call type were due to one bird shifting between alternatives, or different birds contributing at different times. Although one might have predicted that solitary aviary birds would respond to any conspecific sound, captives were as selective to playbacks as were wild flocks. Of those that did respond, many showed convergent responses to some stimuli and divergent or highly variable responses to others. This selectivity was not significantly correlated with the SPCC similarity between the bird’s signature chee and the trial stimulus. However, that similarity did significantly predict how much the bird called in the post-playback period. Both wild and short-term captive conures favor an individuallyspecific “signature” chee when calling (Cortopassi and Bradbury, 2006). The same pattern was seen here in seven solitary birds responding to field playbacks, and all 16 of the aviary subjects. At least within a trial, each of these individuals modified an existing chee type relative to the stimulus rather than switch to some alternative type. Playbacks to wild flocks also elicited a surprisingly low number of chee types. It is not known whether multiple call types within a wild flock trial were emitted by different birds, or by a single bird that changed call types. In either case, the results suggest that even in large flocks, only a few birds exchanged chees with the playback. This suggests that there may be asymmetries in leadership roles within flocks, a notion reinforced by the rapid and directed manner with which conures move between feeding sites without prior reconnaissance (Bradbury, 2003). The existence of individual signatures may also explain the low (41–46%) rates at which either playback series elicited responses. We frequently observe over-flying flocks that do not respond to chees of roosting flocks, or roosting flocks that do not reply to the calls of over-flying flocks. While there may be situations where flock fusion is not advantageous, it could also be the case that conures discriminate between the chees of individual callers and respond selectively. In both sets of experiments, a substantial number of trials showed alternating intervals of stimulus–response convergence and divergence. This alternating variation, the curvilinear patterns seen in many of the field trials, the small number of individuals responding even in large flocks, and the long duration of many field trials suggest that there is more to these vocal exchanges than simple announcements of identity. The ubiquitous change of contact calls that we see during interactions in orange-fronted conures makes it unlikely that birds are “naming” each other using vocal labeling as suggested by Wanker et al. (2005) for stable captive groups of spectacled parrotlets (Forpus conspicillatus). In fact, it is unclear how such a system would be beneficial in the fission–fusion society of these conures. Instead, it seems likely that the exchange is some form of negotiation between a few individuals. Likely functions include the determination of subsequent leadership should the two groups fuse or a sorting out of dominance. Either alternative might explain our discovery here of divergent as well as convergent call modification during exchanges. The ability to imitate call variants of other individuals offers the possibility to interact with all individuals in the population; species with fixed vocal types can only match some of the song types in the population and interactions would thus be constrained to only
part of the population. The ability to imitate contact calls would therefore be especially beneficial in species that live in dynamic social flocks such as the fission–fusion system. The social complexity could therefore favor the evolution of a flexible communication system. If natural interactions involve progressive and coordinated changes in chee structure by both parties, our experimental design clearly made this type of adjustment impossible. Responding birds may have modified their chees a certain amount and then waited for the targeted recipient, in this case the playback, to modify their chee in reply. This modification never occurred. Instead of a steady increase (or decrease) in similarity as the interaction progressed, our failure to change chees could easily have generated the oscillating and suddenly shifting response patterns seen in many trials. The production of agonistic call types just prior to the departure of many wild flocks may also reflect the failure of our playbacks to “behave normally”. Current studies in which short-term captives in the aviary are recorded during interactions with visiting wild flocks will hopefully clarify this issue. Acknowledgements This research was funded by a NSF grant to Jack Bradbury (IBN 02-29271). Special thanks go to Roger Blanco, Maria Marta Chavez, and the staff at the Area Conservacion Guanacaste for support in this study. We thank Marisa Adler, Erin Bohman, Susannah BuhrmanDeever, Ben McCue, Angelika Poesel, Judith Scarl, Jacob Slominski, and Jennifer Webb for field and analytical assistance. Francoise Vermeleyen provided statistical advice on the use of mixed models. References Bartlett, P., Slater, P.J.B., 1999. The effect of new recruits on the flock specific call of budgerigars (Melopsittacus undulatus). Ethol. Ecol. Evol. 11, 139–147. Bradbury, J.W., 2003. Vocal communication in wild parrots. In: DeWaal, F.B.M., Tyack, P.L. (Eds.), Animal Social Complexity: Intelligence, Culture and Individualized Societies. Harvard University Press, Cambridge, MA, pp. 293–316. Bradbury, J.W., Cortopassi, K.A., Clemmons, J.R., 2001. Geographical variation in the contact calls of orange-fronted parakeets. Auk 118, 958–972. Catchpole, C.K., Slater, P.J.B., 1995. Bird Song, Biological Themes and Variation. Cambridge University Press, Cambridge, p. 248. Cheney, D.L., Seyfarth, R.M., 1990. How Monkeys See the World. Chicago University Press, Chicago, IL, p. 388. Connor, R.C., Wells, R., Mann, J., Read, A., 2000. The bottlenose dolphin—social relationships in a fission: fusion society. In: Mann, J., Connor, R., Tyack, P.L., Whitehead, H. (Eds.), Cetacean Societies: Field studies of Whales and Dolphins. University of Chicago Press, Chicago, IL, pp. 91–126. Cortopassi, K.A., Bradbury, J.W., 2000. The comparison of harmonically rich sounds using spectrographic cross-correlation and principal coordinates analysis. Bioacoustics 11, 89–127. Cortopassi, K.A., Bradbury, J.W., 2006. Contact call diversity in wild orange-fronted parakeet pairs, Aratinga canicularis. Anim. Behav. 71, 1141–1154. Farabaugh, S.M., Linzenbold, A., Dooling, R.J., 1994. Vocal plasticity in buderigars (Melopsittacus undulatus): evidence for social factors in the learning of contact calls. J. Comp. Psychol. 108, 81–92. Groth, J.G., 1993. Call matching and positive assortative matching in Red crossbills. Auk 110, 398–491. Hardy, J.W., 1963. Epigamic and reproductive behavior of the orange-fronted parakeet. Condor 65, 169–199. Hardy, J.W., 1965. Flock social behavior of the orange-fronted parakeet. Condor 67, 140–156. Hile, A.G., Plummer, T.K., Striedter, G.F., 2000. Male vocal imitation produces call convergence during pair bonding in budgerigars, Melopsittacus undulatus. Anim. Behav. 59, 1209–1218. Hile, A.G., Striedter, G.F., 2000. Call convergence within groups of female budgerigars (Melopsittacus undulatus). Ethology 106, 1105–1114. Janik, V.M., 1999. Origins and implications of vocal learning in bottlenose dolphins. In: Box, H.O., Gibson, K.R. (Eds.), Mammalian Social Learning: Comparative and Ecological Perspectives. Symposia of the Zoological Society of London, No. 72. Cambridge University Press, Cambridge, UK, pp. 308–332. Janik, V.M., 2000. Whistle matching in wild dolphins (Tursiops truncatus). Science 289, 1355–1357. Janik, V.M., Slater, P.J.B., 1998. Context-specific use suggests that bottlenose dolphin signature whistles are cohesion calls. Anim. Behav. 56, 829–838. Janik, V.M., Sayigh, L.S., Wells, R.S., 2006. Signature whistle shape conveys identity information to bottlenose dolphins. Proc. Natl. Acad. Sci. U.S.A. 103, 8293–8297.
T.J.S. Balsby, J.W. Bradbury / Behavioural Processes 82 (2009) 133–139 Krutzen, M., Sherwin, W.B., Connor, R.C., Barre, L.M., Van de Casteele, T., Mann, J., Brooks, R., 2003. Contrasting relatedness patterns in bottlenose dolphins (Tursiops sp.) with different alliance strategies. Proc. R. Soc. Lond. B 270, 497–502. Mammen, D.L., Nowicki, S., 1981. Individual differences and within flock convergence in chickadee calls. Behav. Ecol. Sociobiol. 9, 179–186. Mann, J., Smuts, B.B., 1998. Natal attraction: allomaternal care and mother-infant separations in wild bottlenose dolphins. Anim. Behav. 55, 1097–1113. Marler, P., Mundinger, P.C., 1975. Vocalizations, social organization, and breeding biology of the twite, Acanthus flavirostris. Ibis 117, 1–17. McGregor, P.K., Dabelsteen, T., Shepherd, M., Pedersen, S.B., 1992. The signal value of matched singing in great tits: evidence from interactive playback experiments. Anim. Behav. 43, 987–998. Moller, L.M., Beheregaray, L.B., Harcourt, R.G., Krutzen, M., 2001. Alliance membership and kinship in wild male bottlenose dolphins (Tursiops aduncus) of southeastern Australia. Proc. R. Soc. Lond. B 268, 1941–1947. Mundinger, P.C., 1971. Vocal imitation and individual recognition of finch calls. Science 168, 480–482. Mundinger, P.C., 1979. Call learning in the carduelinae: ethological and systematic considerations. Syst. Zool. 28, 279–283. Nowicki, S., 1989. Vocal plasticity in captive black-capped chickadees: the acoustic basis and rate of call convergence. Anim. Behav. 37, 64–73. Saunders, D.A., 1983. Vocal repertoire and individual vocal recognition in the shortbilled white-tailed black cockatoo, Calyptorhynchus funereus latirostris Carnaby. Aust. Wildl. Res. 10, 527–536. Sayigh, L.S., Tyack, P.L., Wells, R.S., Scott, M.D., 1990. Signature whistles of freeranging bottlenose dolphins, Tursiops truncatus: stability and mother-offspring comparisons. Behav. Ecol. Sociobiol. 26, 247–260. Sayigh, L.S., Tyack, P.L., Wells, R.S., Solow, A., Scott, M., Irvine, A.B., 1999. Individual recognition in wild bottlenose dolphins: a field test using playback experiments. Anim. Behav. 57, 41–50. Smolker, R.A., Mann, J., Smuts, B.B., 1993. The use of signature whistles during separations and reunions among wild bottlenose dolphin mothers and calves. Behav. Ecol. Sociobiol. 33, 393–402. Tyack, P.L., 1993. Animal language research needs a broader comparative and evolutionary framework. In: Roitblat, H.L., Herman, L.M., Nachtigall, P.E. (Eds.), Language and Communication: Comparative Perspectives. Lawrence Erlbaum Associates Publishers, London, UK, pp. 115–152. Tyack, P.L., 2000. Functional aspects of cetacean communication. In: Mann, J., Connor, R., Tyack, P.L., Whitehead, H. (Eds.), Cetacean Societies: Field studies of Whales and Dolphins. University of Chicago Press, Chicago, IL, pp. 270–307.
139
Tyack, P.L., 2003. Dolphins communicate about individual-specific social relationships. In: DeWaal, F.B.M., Tyack, P.L. (Eds.), Animal Social Complexity: Intelligence, Culture and Individualized Societies. Harvard University Press, Cambridge, MA, pp. 342–361. Tyack, P.L., 2008. Convergence of calls as animals form social bonds, active compensation for noisy communication channels, and the evolution of vocal learning in mammals. J. Comp. Psychol. 122, 319–331. Uster, D., Zuberbuhler, K., 2001. The functional significance of Diana monkey ‘clear’ calls. Behaviour 138, 741–756. Vehrencamp, S.L., 2000. Handicap, index, and conventional signal elements of bird song. In: Amundsen, T., Rosenqvist, G., Espmark, Y. (Eds.), Animal Signals: Signalling and Signal Design in Animal Communication. Tapir Publishers, Trondheim, Norway, pp. 277–300. Vehrencamp, S.L., 2001. Is song-type matching a conventional signal of aggressive intentions? Proc. R. Soc. Lond. B 268, 1637–1642. Vehrencamp, S.L., Ritter, A.F., Keever, M., Bradbury, J.W., 2003. Responses to playback of local versus distant contact calls in the orange-fronted conure, Aratinga canicularis. Ethology 109, 37–54. Wanker, R., Fisher, J., 2001. Intra- and Interindividual variation in the contact calls of spectacled parrotlets (Forpus conspicillatus). Behaviour 138, 709– 726. Wanker, R., Apcin, J., Jennerjahn, B., Waibel, B., 1998. Discrimination of different social companions in spectacled Parrotlets (Forpus conspicillatus): evidence for individual recognition. Behav. Ecol. Sociobiol. 43, 197–202. Wanker, R., Sugama, Y., Prinage, S., 2005. Vocal labelling of familiy members in spectacled parrotlets, Forpus conspicillatus. Anim. Behav. 70, 111– 118. Watwood, S.L., Owen, E.C.G., Tyack, P.L., Wells, R.S., 2005. Signature whistle use by temporarily restrained and free-swimming bottlenose dolphins, Tursiops truncatus. Anim. Behav. 69, 1373–1386. Wright, T.F., 1996. Regional dialects in the contact call of a parrot. Proc. R. Soc. Lond. B 263, 867–872. Wright, T.F., Cortopassi, K.A., Bradbury, J.W., Dooling, R.J., 2003. Hearing and vocalizations in the conure Aratinga canicularis. J. Comp. Psychol. 117, 87–95. Zuberbuhler, K., 2000. Referential labelling in Diana monkeys. Anim. Behav. 59, 917–927. Zuberbuhler, K., 2001. Predator-specific alarm calls in Campbell’s monkeys, Cercipithecus campbelli. Behav. Ecol. Sociobiol. 50, 414–422.