Does reduced social contact affect discrimination of distance cues and individual vocalizations?

ANIMAL BEHAVIOUR, 2003, 65, 911–922 doi:10.1006/anbe.2003.2153

Does reduced social contact affect discrimination of distance cues and individual vocalizations? LESLIE S. PHILLMORE, CHRISTOPHER B. STURDY & RONALD G. WEISMAN

Department of Psychology, Queen’s University (Received 5 July 2001; initial acceptance 2 October 2001; final acceptance 13 August 2002; MS. number: A9107R2)

For songbirds, experience with conspecific vocalizations during development is important for the ontogeny of both production of vocalizations and perceptual abilities in young oscines. We examined the effects of reduced experience with conspecific vocalizations during development on two auditory perceptual tasks: discrimination of distance cues and discrimination between individual vocalizations in black-capped chickadees, Poecile atricapillus. Discrimination of distance cues was nearly identical in fieldand isolate-reared chickadees, but the capacity for memorization of numbers of individual vocalizations was lower in isolate-reared chickadees. We conclude that discrimination of distance cues, and thus distance perception, is probably not learned through experience but is an innate skill. We also conclude that discrimination between many individual vocalizations, a task demanding memorization, is aided by early experience with conspecific vocalizations. 

2003 Published by Elsevier Science Ltd on behalf of The Association for the Study of Animal Behaviour.

Many temperate-living songbirds establish and defend territories to obtain the resources necessary for rearing their young. To gain and hold such territories, male songbirds need highly developed auditory skills to use long-distance communication (McGregor 1991). One such skill is estimating the distance to (i.e. ranging) singing conspecifics using acoustic information alone. Another skill is discriminating between conspecifics, for example between neighbours or between neighbours and strangers. Female songbirds need similar auditory skills to find suitable males for mating (Ratcliffe & Otter 1996) and to maintain contact with mates over the territory. Are these perceptual abilities a natural talent that these songbirds can instinctively perform? Or must these skills be learned through experience, similar to the way song production must be learned from experience with adult male tutors (reviewed in Baptista 1996)? One way to answer these questions is to compare the performance of songbirds that have had experience using such skills (i.e. birds raised in the wild by natural parents) with the performance of songbirds that have never had these experiences (i.e. birds reared in the laboratory by humans). Correspondence and present address: L. S. Phillmore, Department of Psychology, University of Western Ontario, London, ON N6A 5C2, Canada (email: [email protected]). C. B. Sturdy is now at the Department of Psychology, P220 Biological Sciences Building, University of Alberta, Edmonton, AB T6G 2E9, Canada. R. G. Weisman is at the Department of Psychology, Queen’s University, Kingston ON K7L 3N6, Canada. 0003–3472/03/$30.00/0



Accurately estimating the distance to a singing conspecific is important for determining whether that conspecific is within territory boundaries, requiring an aggressive behavioural response, or outside that boundary, possibly requiring withholding that response. Field playback studies have shown that several species of songbirds can range singing conspecifics (e.g. western meadowlarks, Sturnella neglecta: McGregor & Falls 1984; Carolina wrens, Thryothorus ludovicianus: Naguib 1995). For example, male Carolina wrens overestimate the distance to a speaker playing back degraded song from within territory boundaries, frequently overflying the speaker (Naguib 1995). In the laboratory, Phillmore et al. (1998) used a go/no-go operant discrimination task (where responses to ‘go’ stimuli resulted in food reward and responses to ‘no-go’ stimuli resulted in lights out) to demonstrate that zebra finches, Taeniopygia guttata, and black-capped chickadees, Poecile atricapillus, can discriminate between vocalizations recorded over near (5-m) and far (75-m) distances. These field and laboratory tasks both require use of distance cues that are produced when sound is degraded by transmission through the environment. Degradation increases as distance increases, producing distance cues: decreased overall amplitude, attenuation of the high-frequency components of song, reverberation of the song by echo from the ground and vegetation, and decreased signal-to-noise ratio (Wiley & Richards 1982; Dabelsteen et al. 1993; see also Nelson & Stoddard 1998). Male songbirds use one or more of these cues (Naguib 1997b) to judge distance to the sender.

911 2003 Published by Elsevier Science Ltd on behalf of The Association for the Study of Animal Behaviour.

912

ANIMAL BEHAVIOUR, 65, 5

Laboratory operant research (Phillmore et al. 1998) and field playback research (Naguib et al. 2000) agree in suggesting that songbirds scale the perceptual distance cues in a vocalization on a continuum. For example, male chaffinches, Fringilla coelebs, have a systematic representation of different distances based on auditory cues that allows them to vary their responses according to the actual threat posed by a rival (Naguib et al. 2000). Discrimination between conspecific individuals is also a useful skill for territory defence. Withholding responses to the vocalizations of familiar, established territory neighbours reduces unnecessary energy expenditure, and responding aggressively to unfamiliar conspecific strangers reduces the threat to a territory and mate. Both field and laboratory studies of individual recognition demonstrate that many species of songbirds have the ability to discriminate among conspecifics based only on vocalizations. Field studies use a clever variant of the neighbour–stranger paradigm: neighbour and stranger vocalizations are played from the direction of the neighbour’s correct and incorrect territorial boundaries (Falls & Brooks 1975). White throated sparrows, Zonotrichia albicollis, responded more strongly to strangers’ vocalizations when the playback of neighbour and stranger was from the neighbour’s correct boundary, but responded equally as strongly to the songs of neighbour and stranger vocalizations when the playback was from an incorrect boundary. Other studies (e.g. with hooded warblers, Wilsonia citrina, and Kentucky warblers, Oporornis formosus, Godard & Wiley 1995) found similar results when the vocalizations of a territorial neighbour were played from the correct and incorrect boundary. Laboratory studies have used go/no-go operant procedures to show that different species of birds can discriminate between various individual vocalizations (e.g. barn and cliff swallows, Hirundo rustica and H. pyrrhonota, respectively, Loesche et al. 1991; song sparrows, Melospiza melodia, Beecher et al. 1994; European starlings, Sturnus vulgaris, Gentner & Hulse 1998). For example, Loesche et al. (1991) demonstrated that barn and cliff swallows could discriminate between at least 10 pairs of their own and each other’s vocalizations. In a subsequent study, Stoddard et al. (1992) showed that song sparrows could discriminate at least 32 pairs of conspecific songs. Phillmore et al. (2002) showed that black-capped chickadees could discriminate simultaneously between four pairs of male chickadee songs and between four pairs of female zebra finch distance calls. Furthermore, Phillmore et al. (2002) found that chickadees transferred their recognition of these individual vocalizations to exemplars of these same vocalizations that had been transformed by recording them at 25 m. Experience is necessary for the males of most songbird species to learn their songs: they must interact with their fathers and other nearby males (Baptista 1996). Isolating the young from adult males and their songs during development results in aberrant song production (Volman & Khanna 1995). Experience is also important for female songbirds, because it influences song preference. For example, females of various songbird species demonstrate preferences for familiar (home) dialect

over unfamiliar (alien) dialects (reviewed in Searcy & Yasukawa 1996). Studies of relative pitch perception in black-capped chickadees (Njegovan & Weisman 1997) and song note, relative pitch and frequency perception in zebra finches (Sturdy et al. 2001) have shown that songbirds reared in isolation from adult conspecific vocalizations show deficits in their ability to discriminate biologically relevant stimuli compared with normal adult birds. The discovery that experience with adult male conspecifics influences the development of songbirds’ absolute and relative perceptual abilities led us to speculate about the possible effect of early experience on the recognition of individual vocalizations and the ranging of these vocalizations. The ability to recognize neighbours’ songs requires an ability to monitor the locations of senders, so one might speculate that a bird deficient in individual recognition might be deficient in ranging as well. Although familiarity with vocalizations may improve ranging (Naguib 1997a), the idea that songbirds must have experience with a vocalization to range it accurately has received little support (reviewed in Wiley 2000). Thus, little is known about the effect of experience on individual recognition and ranging by adult songbirds. Phillmore et al. (1998) found that fieldreared black-capped chickadees learned an operantdiscrimination task based on distance cues faster than laboratory-reared zebra finches. Phillmore et al. (1998) speculated that the chickadees discriminated faster because they had extensive ranging experience, whereas the zebra finches had no ranging experience. Here, we evaluated the effects of experience (and its opposite, isolation) on the discrimination of distance cues and discrimination between individual vocalizations in blackcapped chickadees, which use individual vocalization cues and distance cues in laboratory discriminations and in the field (Fotheringham & Ratcliffe 1995; Phillmore et al. 1998). GENERAL METHOD

Animals Subjects were 16 experimentally naı¨ve adult blackcapped chickadees. All the chickadees were captured from the wild within 50 km of Kingston, Ontario, Canada (Environment Canada permit CA001). The field-reared group consisted of four males and four females, captured as adults (at least 15 months old) and housed in our laboratory for 2 months before testing. We sexed fieldreared birds after capture using a composite score of mass, wing chord and tail length (Desrochers 1990). We assigned the birds to age class by the shape and extent of the white margin of the outer rectrices (Pyle 1997). The isolate-reared group consisted of four males and four females, captured as nestlings. Two broods of nestlings were taken at about 5 days posthatch and a third brood was taken at about 15 days posthatch. The isolate-reared birds were hand-reared as nest groups until 35 days posthatch, sexed by laparotomy, then housed individually until they were about 15 months old. One female

PHILLMORE ET AL.: ISOLATE CHICKADEE PERCEPTION

isolate-reared chickadee became ill between testing in experiments 1 and 2, so data from this bird are available only for experiment 1.

Feeding and Housing Regimes We fed adult chickadees a ground mixture of three parts crushed unsalted sunflower seeds, two parts isolated soy protein (Tara Natural Foods, Kingston, Ontario) one part Purinature turkey starter (Purina, Woodstock, Ontario), and vitamin supplement (Theralin, Cranbury, New Jersey, U.S.A.). Before testing and between experiments 1 and 2, the food mixture was freely available. During experiments 1 and 2, the mixture was available only during feeder visits on rewarded (positive) trials. Water, grit and cuttlebone were freely available. On alternate days we added vitamin supplements to the water (Hagen Vitamin Supplement Conditioner for Birds, Montreal, Quebec, Canada). We provided a small amount of hard-boiled egg once a week as a further nutritional supplement. When they were nestlings, the isolate-reared group received the diet described by Lanyon (1979). Once they were eating independently, the isolate-reared group received the same food as the field-reared group described above. Isolate birds were reared together in one room (about 32 m), but were housed as separate nest groups in nest cups (500-ml plastic containers). These isolate birds could hear each other and other laboratory sounds, but could not hear adult chickadees. After the isolates were eating without assistance, we transferred them as nest groups to large cages (about 0.40.50.4 m deep). At 35 days posthatch, the isolates were transferred to individual cages (about 0.30.40.4 m deep) housed side by side on racks. The isolate birds were visually isolated from each other and in a separate room (about 43 m) from field-reared birds assigned for their housing. An isolate never heard vocalizations from any birds except other isolates. Captive field-reared birds were kept in individual cages (also about 0.30.40.4 m deep) in a room assigned for their housing. Except during experiments 1 and 2, when they were in experimental chambers, the birds remained in their assigned rooms. The birds were maintained on a light:dark cycle typical for the season in Kingston, Ontario. Temperature was maintained at about 20C.

Apparatus During experiments 1 and 2, each bird was housed in a standard budgerigar cage (0.30.40.4 m deep). A wire floor, attached near the bottom of the cage, ensured that spilled food was not eaten. Each cage was contained in a ventilated, sound-attenuating enclosure with a 9-W twin-tube fluorescent bulb. Each cage contained several perches, a water bottle and a grit container. An opening (0.110.16 m) in the cage allowed the bird access to the feeder (Njegovan et al. 1994). A perch in front of the feeder opening and the feeder itself were equipped with infrared sensors to monitor the position of the bird. A

microcomputer controlled the experiment and recorded the responses in each chamber. Song and call stimuli were stored in Macintosh Plus computers and were played using a NAD 310 integrated amplifier (NAD Electronics Ltd, London) through a Realistic Minimus-7 two-way speaker (cross-over at 2000 Hz, Radio Shack, Barrie, Ontario) beside the feeder.

Stimulus Preparation Original recordings Our goal was to produce recordings of the same vocalizations at various distances from the source to serve as discriminative stimuli. The prototypes for our simulations were the songs of 12 free-living male black-capped chickadees recorded in the wild at the Queen’s University Biology Station several years before the present experiments and the distance calls of 12 free-living female zebra finches recorded in the wild at Alice Springs and northern Victoria, Australia (R. Zann, personal communication). The stimulus vocalizations had not been heard previously by the subjects but were identical to those described and used by Phillmore et al. (1998).

Logic and preparation of stimuli When recording free-living birds, it is often difficult to minimize and standardize the distance between the bird and the microphone. This produces variability in the distance cues present in undegraded vocalizations, which is compounded by re-recording meant to degrade the signal. We began with simulated (rather than only digitized) vocalizations because we could not eliminate distance cues present in recordings of natural song. Specifically, we eliminated background noise and reverberation visible on the sound spectrogram or Fast Fourier Transforms (FFTs). These simulations were used as the source for producing the degraded signals. We digitized the original recordings into a Power Macintosh 7100 using Sound Designer II software and an Audiomedia Digital Signal Processing Board (Digidesign, Menlo Park, California, U.S.A.) at 44-k 16-bit samples/s. We sampled frequency and amplitude measures from each natural song or call (SoundScope 8, GW Instruments, Somerville, Massachusetts, U.S.A.) every 12.5 ms in prototype zebra finch calls and every 25 ms in prototype chickadee songs, because zebra finch distance calls have more rapid frequency and amplitude modulation than chickadee songs. Sample frequency and amplitude measures were input into TurboSynth SC software (Digidesign) to create synthetic copies of the prototypes. To ensure accurate synthesis, sound spectrograms, FFTs and the human ear were used to compare simulations with their prototype. We verified the validity of our simulations in informal playback experiments. We played our simulated male chickadee songs to free-living male chickadees and we played female zebra finches distance calls to male zebra finches. Each type of synthesized stimulus produced species-typical responses in male conspecifics similar to those elicited by natural vocalizations (Phillmore et al. 1998).

913

914

ANIMAL BEHAVIOUR, 65, 5

Figure 1. Sound spectrograms (FFT settings: Hamming window, 1024 FFT points, 184-Hz filter) of representative male black-capped chickadee songs and female zebra finch distance calls used as stimuli in the distance cue discrimination task (experiment 1, stimulus set A) and the individual vocalization discrimination task (experiment 2, stimulus set B). Stimulus set A: one vocalization from each species recorded at four distances (75, 50, 25 and 5 m). Stimulus set B: two additional individual vocalizations recorded at 5 m.

Distance playback We made playback recordings of the simulated vocalizations in an area known to be inhabited by black-capped chickadees (Pangman Conservation Reserve, Queen’s University Biology Station, about 50 km north of Kingston, Ontario). The terrain is relatively flat and well forested (coniferous and deciduous trees). We recorded between 0700 and 1000 hours Eastern Standard Time on 29 October and 1 November 1996. Weather conditions were similar on all recording days. A Sony TCD-D7 DAT player and a Sony SRS-77G amplified speaker were used to play simulated calls and a Sony Walkman Pro and a Sennheiser MKH-816 shotgun microphone were used to record them. The playback speaker was mounted on a pole with the centre of the woofer 1.5 m from the ground, a height at which both chickadees and zebra finches are known to sing. The microphone was mounted on a tripod so that the centre of the end of the baffle was 1.5 m from the ground and aimed directly at the speaker. The amplitude of the stimuli was set at 80 dB when measured from 1 m away from the speaker and 1.5 m from the ground (Realistic Sound Level Meter, Radio Shack, Barrie, Ontario). The amplitude of the stimuli and the recording level on the Walkman Pro were not adjusted during recording. Simulated vocalizations were played back 5, 25, 50 and 75 m from the microphone, yielding four sets of stimuli (the 5-, 25-, 50- and 75-m sets). Each set of vocalizations consisted of 12 simulated chickadee songs and 12 simulated zebra finch calls (e.g. Fig. 1). Vocalizations recorded at 5 m included some background sounds and were sub-

ject to mild degradation (compared with our original almost noise-free simulations); vocalizations recorded at farther distances were more degraded, that is, they were decreased in amplitude (and therefore, signal-to-noise ratio), and had increased reverberation and highfrequency attenuation compared with the vocalizations recorded at 5 m.

Microphone selection In conducting playback research with songs and calls, one has the choice of using a directional or an omnidirectional microphone. Both kinds of microphones have advantages and individual researchers have their preferences. Many researchers that study distance perception and sound transmission over distance (e.g. Lemon et al. 1980; Naguib 1995, 1997a, b) use directional microphones. When recorded on-axis (directly aimed at the sound source), directional microphones provide a better signal-to-noise ratio in recordings, retain more of the higher frequencies in the signal and preserve most of the reverberation in the signal. Other researchers (e.g. Nelson & Stoddard 1998) favour omnidirectional microphones, which record more reverberation from behind the microphone and, therefore, provide somewhat more accurate cues to the actual distance to the speaker. In this study, a directional microphone provided an audible signal at 75 m but at some loss in the absolute level of reverberation. However, our interest was in the discrimination of relative distance cues rather than the estimation of actual distances. Hence, a directional

PHILLMORE ET AL.: ISOLATE CHICKADEE PERCEPTION

microphone provided clean, focused recordings useful for our purposes.

Final preparation and amplitude control We digitized the vocalization sets at 44-k 16-bit samples/s and all at the same input level with Sound Designer II software. The files were converted to 22-k 8-bit samples/s using WaveConvert software (Waves, Tel-Aviv, Israel) and the L1 Ultramaximizer Plug-In (Waves) for Sound Designer II, respectively. To eliminate aliasing, before conversion, the files were digitally low-passfiltered (<10 007 Hz) with Q10 Paragraphic Equalizer Plug-In (Waves) for Sound Designer. Songs and calls were all played back in the field at a single amplifier setting, rerecorded in the field at a single recorder setting, and played back in the chambers at a single amplifier setting. These settings were standardized to produce an 80-dB amplitude at 1 m in the field and a 75-dB amplitude in the chamber for vocalizations recorded at 5 m in the field. In effect, this procedure holds the amplitude of the background stimuli constant while the amplitude of the foreground song or call decreases with distance in the field.

Training Procedures and Experimental Design Each chickadee in the field-reared and isolate-reared groups was trained in both the distance cue and the individual vocalization discrimination tasks. We counterbalanced the order of training in the two tasks across birds separately in each group. That is, two males and two females in each group learned the distance cue task first then learned the individual vocalization task, and the remaining birds in each group learned the individual vocalization task first, then the distance cue task. We designated the distance cue task as experiment 1 and the individual vocalization task as experiment 2. Between experiments, the birds were removed from the experimental chambers and allowed free feeding in their housing rooms during a minimum rest period of 24 days. In each task, we gave nondifferential followed by discrimination training. The purpose of nondifferential training was to ensure high and uniform percentages of response across the stimuli that during subsequent discrimination training served as rewarded (standard) and unrewarded (comparison) songs and calls.

Nondifferential training Nondifferential training in the first task began after a bird had learned to use the perch and feeder. Nondifferential training in the second task began immediately after the rest period between tasks. The within-trial sequence began when a bird landed on the perch, breaking an infrared beam. If the bird remained on the perch for 1 s, a single vocalization was selected randomly and without replacement and played once. If the bird entered the feeder after the stimulus was played, breaking an infrared beam, it was rewarded with 1-s access to food. A 30-s intertrial interval (ITI) followed. If the bird left the perch without entering the feeder, the trial ended after 1 s. If the bird remained on the perch, the

trial ended after 1 s, a 60-s ITI followed. The 60-s ITI was used to increase the probability of the bird leaving the perch on all trials. If the bird left the perch before the vocalization had finished playing, the trial ended, and the chamber lights turned off during a 5-s ITI. This procedure helped to ensure that the birds listened to each vocalization in its entirety. During each of the two nondifferential training periods, the birds heard each of the vocalizations played during later discrimination training. Nondifferential training lasted until the bird responded consistently to all stimuli and did not leave the request perch before stimuli were finished playing on more than 10% of trials.

Discrimination training We defined rewarded songs and calls as standard vocalizations and unrewarded songs and calls as comparison vocalizations. During discrimination training, visits to the feeder after standard (S+) songs and calls were still rewarded with access to food, but visits to the feeder after comparison (S
) songs and calls resulted in a 30-s ITI with the chamber lights out. As during nondifferential training, on each trial, a single song or call was selected randomly and without replacement from the stimulus pool and played. The birds were tested in day-long sessions of about 1000 trials each.

Response Measures As one measure of discrimination during training, we calculated a discrimination ratio for performance on standard stimuli separately for each experiment. Within each experiment, we calculated a discrimination ratio separately for songs and calls. The discrimination ratio equalled the mean percentage of response to all the standard stimuli divided by the mean percentage of response to all the stimuli. Discrimination was at chance when the ratio was 0.50 (responding to all stimuli equally) and perfect when the ratio was 1.00 (responding to only the standards). The discrimination ratio provides an overall measure of discrimination between many standards and comparison stimuli, but it does not tell how many standards were well discriminated. For example, a bird could achieve a high discrimination ratio without mastering the perceptual task by responding to only a few of the many standard stimuli and responding little to other standard and comparison stimuli. To provide a measure of how many standards were well discriminated (i.e. how a bird performed in the perceptual task), we followed Weisman et al. (1994) in adapting the two-tailed 95% confidence interval (CI) from sampling statistics (Sokal & Rohlf 1980). In this adaptation, the 95% CI was calculated from the XSD percentages of response for the comparison stimuli (separately for songs and calls) on each day of training. In the absence of a true difference between an individual standard and the distribution of comparison stimuli, the response to about 19 of 20 comparisons should be included in the 95% CI (= X1.96 SD) (Sokal & Rohlf 1980). In other words, when an individual

915

916

ANIMAL BEHAVIOUR, 65, 5

standard exceeds the 95% CI for the comparison stimuli, there is statistical evidence that the standard was well discriminated. Because the actual number of standards varied between the discrimination tasks in experiments 1 and 2, we analysed the percentage of standards above the 95% CI in each task.

Statistical Analyses When the discrimination ratio is near 0.50 or 1.00, sample values may not be normally distributed. Unfortunately, tests for normality can be more sensitive to deviations from normality than the analysis of variance (ANOVA) itself (Hays 1994). As a running check on whether our analyses of the original data were yielding inflated P values, we conducted parallel ANOVAs of the original data and arcsine square-root transformations of the discrimination ratios and rank transformations of the percentage of standards above the 95% CI (see Conover & Iman 1981 for a more detailed rationale for parallel analyses). In each instance, analysis of transformed data yielded virtually the same levels of significance as analysis of the untransformed data; we report the results for ANOVAs of the untransformed data here. EXPERIMENT 1: DISCRIMINATION OF DISTANCE CUES We adapted our laboratory conditioning assay of distance cue perception from Phillmore et al. (1998). Briefly, chickadees were trained to discriminate between degraded (recorded at 75 m) and undegraded (recorded at 5 m) versions of vocalizations from two species: male black-capped chickadee songs (Weisman et al. 1990) and female zebra finch distance calls (Zann 1996). We tested transfer of learning to vocalizations recorded at the same distance as the training vocalizations, and also tested transfer to vocalizations recorded at intermediate distances (25 and 50 m) to the training vocalizations. Phillmore et al. (1998) did not include females in their study so experiment 1 will provide the first evidence for females in the operant distance cue assay. Njegovan & Weisman (1997) found that chickadees isolated from contact with conspecific song performed more poorly in discriminating the frequency ratio between notes in the conspecific ‘fee bee’ song than chickadees that were not isolated. The isolate-reared birds in the present experiment had no contact with adult conspecific vocalizations or with the normal localization cues emanating from conspecific vocalizations produced at a distance from the receiver in a natural habitat. Isolates were brought into the laboratory from the field as 5- to 15-day-old nestlings, which eliminated experience with flock and mate location and with territorial defence in three-dimensional space. In the laboratory, the farthest apart any isolate-reared chickadee was from another isolate was about 3 m, minimizing the amount of degradation in any vocalizations that the isolates heard from each other. Furthermore, isolates were visually isolated from each other, so forming any association between

auditory distance cues and actual visual distance in space would be, at best, difficult for these birds. One might hypothesize that experience with ranging vocalizations is necessary to discriminate based on distance cues and, more specifically, that isolates should fail at putting distance-degraded vocalizations on to a continuum from near to far. An alternative hypothesis is that ranging is innate and, therefore, experience is unnecessary to discriminate based on distance cues as adults. In experiment 1, we tested isolate-and field-reared chickadees in distance cue discriminations (Phillmore et al. 1998) for evidence to support one of these hypotheses.

Methods Discrimination training To avoid possible confounding effects of pseudoreplication (Kroodsma 1990), we prepared two separate sets of training vocalizations; half the birds heard one set and half the birds heard the other set. The first set consisted of nine of the 12 chickadee songs and nine of the 12 zebra finch calls recorded at 75 m, randomly selected to serve as standards. The same nine songs and calls recorded at 5 m were used as comparisons. This left three songs and three calls (recorded at both distances) to test transfer of training. The random selection process was repeated with the same 12 songs and 12 calls to generate a second set of training and test stimuli; the test stimuli were not repeated between sets. This procedure allowed us to use a total of six songs and six calls (recorded at both distances) to test transfer of training, reducing the chance that test stimuli were special in some way and perhaps identifiable by that characteristic rather than as new exemplars of the learned category. Training continued until performance on discrimination ratios was at least 0.80 for three consecutive days.

Transfer to 75- and 5-m vocalizations not heard in training To prevent birds from learning to discriminate between the test stimuli, we introduced the recordings into the stimulus pool as probe trials reinforced at 15%, rather than as additional ‘training’ stimuli reinforced at the same levels as original training stimuli (100% and 0%). To reduce the possibility that birds would use the reduced probability of reinforcement (15%) to identify test stimuli as probes, we reduced the percentage reinforcement for each 75-m (standard) training stimulus to 85%. In other words, during the test phase, responses on some trials resulted in neither reward (food availability) nor punishment (lights out), but that occurrence could not be used to single out transfer test stimuli from original training stimuli. Therefore, in the beginning of the test phase we reduced the percentage reinforcement for each 75-m (standard) training stimulus to 85% for two sessions. Then, we alternated between transfer sessions and discrimination training. During transfer sessions, one simulated chickadee song and one simulated zebra finch call,

PHILLMORE ET AL.: ISOLATE CHICKADEE PERCEPTION

each recorded at 75 and 5 m, were added to the stimulus pool. Thus, over three transfer sessions all the birds in each group heard all 12 75-m and 5-m test vocalizations.

1 (a) Field-reared 0.9

Transfer to stimuli re-recorded at intermediate distances 0.8

0.7 Discrimination ratio

After testing of transfer to untrained stimuli was complete, we gave four sessions of discrimination training (standards reinforced at 85%). Then, we alternated between intermediate distance transfer sessions and discrimination training. Test stimuli were recordings at 25 and 50 m of the same nine randomly selected songs and calls used in the training set, each reinforced at 15%. The percentages of reward to standard and the test stimuli were 85 and 15%, respectively, to make it less likely that birds could learn to discriminate the test stimuli during transfer. During each transfer session, three simulated songs and three simulated calls, each recorded at 25 and 50 m, were added to the stimulus pool. Therefore 12 different intermediate distance stimuli were added to the stimulus pool for each transfer session, for a total of 36 50-m and 25-m test vocalizations over the three transfer sessions.

Chickadee songs Zebra finch calls

0.6

0.5 1 (b) Isolate-reared 0.9

0.8

0.7

Results Gender effects In the acquisition of discrimination between 75- and 5-m vocalizations, we found no significant differences between males and females in the discrimination ratios (ANOVA: F1,12 =0.001, P=0.98) or the percentage of standards above the 95% CI (F1,12 =0.045, P=0.835), nor did males and females differ significantly on the final day of acquisition in the discrimination ratio (F1,12 =0.081, P=0.781) or the percentage of standards above the 95% CI (all the birds discriminated all 18 standard vocalizations). Therefore, we pooled acquisition performance across gender (Figs 2, 3). We also found no significant differences between males and females in transfer to test vocalizations across distance cues (F1,11 =0.335, P=0.575 and F1,11 =0.663, P=0.433, respectively). Therefore, we pooled transfer performance (Fig. 4).

Acquisition and final-day performance There were no significant differences between the discrimination ratios of field- and isolate-reared chickadees in acquisition rates over the first 4 training days (F1,14 =0.604, P=0.450) or final-day performance (F1,14 = 1.016, P=0.331; Fig. 2). Both groups learned to discriminate chickadee songs faster than zebra finch calls (F1,14 =5.724, P=0.031). By the final day of training, however, there was no significant difference in discrimination ratios between the songs and calls (F1,14 =3.981, P=0.066). There were no significant differences between field- and isolate-reared chickadees in acquisition rates (percentage of standards >95% CI; F1,14 =0.549, P=0.471). Both groups learned to discriminate more chickadee songs than zebra finch calls during acquisition (F1,14 =5.862, P=0.030; Fig. 3). By the final training day, all birds discriminated all nine standard songs and all nine standard calls. The rearing groups did not differ

Chickadee songs Zebra finch calls

0.6

0.5

1

2 3 Training day

4

Final day

Figure 2. Acquisition and final-day performance plotted as mean discrimination ratios across training days for field-reared (a) and isolate-reared (b) chickadees in the distance cue discrimination task (experiment 1).

significantly in days to attain the training criterion (F1,14 =1.016, P=0.331; Fig. 3).

Transfer across distance cues Responses of field- and isolate-reared chickadees to the trained 75-m standards and 5-m comparison vocalizations appeared to form a continuous linear generalization function with their responses to the distance-transformed 25- and 50-m test versions of the same vocalizations used in acquisition training (Pearson correlation: field-reared: r5 =0.985, P<0.01; Fig. 4a; isolate-reared: r6 =0.965, P<0.05; Fig. 4b). There were no significant differences between the generalization functions for field- and isolate-reared birds (F1,13 =0.141, P=0.713) or between the generalization functions for chickadee songs and zebra finch calls (F1,13 =0.112, P=0.743). These results suggest that, compared with field-reared birds, isolate rearing has little effect on chickadees’ ability to scale the distance cues in vocalizations. In addition to examining transfer to distancetransformed versions of the training vocalizations, we also examined transfer to untrained test chickadee songs and zebra finch calls recorded at the same distances as the training vocalizations. Both field-reared and isolatereared chickadees responded less to test versions than to

917

ANIMAL BEHAVIOUR, 65, 5

100

(a) Field-reared

(a) Field-reared

100

Y = 1.043X – 1.817 75

75

50

25

Chickadee songs Zebra finch calls

0

100

(b) Isolate-reared

Percentage of response

50

% Standards > 95% Cl

918

25

0 100 (b) Isolate-reared Y = 0.949X + 3.298

75

75

50

50

25

0

Chickadee songs Zebra finch calls

1

2 3 Training day

4

Final day

Figure 3. Acquisition and final-day performance plotted as percentage of standards above the 95% CI across training days for fieldreared (a) and isolate-reared (b) chickadees in the distance cue discrimination task (experiment 1).

training versions of the 75-m standard vocalizations (F1,13 =16.520, P=0.001), but isolates responded less to the test vocalizations than did field-reared chickadees (F1,13 =5.939, P=0.03).

Power analyses In these tests, the discrimination of distance cues was little affected by rearing conditions. We found no significant differences between field- and isolate-reared chickadees in the acquisition or generalization of discrimination based on distance cues. To determine whether these null results represented no real effect or a type II error, we compared our results to minimum detectable isolation effect sizes for a power of 80% using our sample sizes and variances (Cohen 1988). In our power analyses, we focused on the results for S+s >95% CI, because in discriminations of multiple standard and comparison stimuli, the discrimination ratio is often insensitive to experimental manipulations (e.g. Njegovan & Weisman 1997; Sturdy et al. 2001). We compared our minimum detectable effect sizes to previously published results and discuss them in terms of their biological relevance. Specifically, we looked to Njegovan & Weisman’s (1997) comparisons of isolate and

25

0

0

25

50 Distance (m)

75

Figure 4. Responses of field-reared (a) and isolate-reared (b) chickadees to training and test vocalizations in the distance cue discrimination task (experiment 1). Triangles: percentage of response to standard (75 m) and comparison (5 m) training vocalizations. Circles: percentage of response to test vocalizations (5, 25, 50 and 75 m). Filled symbols: response to chickadee songs. Open symbols: response to zebra finch calls.

normal chickadees in another perceptual operant task (relative pitch discrimination, using the same apparatus as we did in the present work). For example, over acquisition, our minimum detectable effect for the percentage of S+s >95% CI was 3.91, or 22% of 18 S+s. Is this a biologically significant effect size? Njegovan & Weisman (1997) reported that their field-reared chickadees discriminated 24% more rewarded standards on day 4 and 44% more standards on day 10 of acquisition than did isolate-reared chickadees. Therefore, if the effect sizes in this experiment were as large as those found by Njegovan & Weisman (1997), we would have detected them at 80% power. We also calculated the minimum detectable effect size at 80% power for the days to criterion measure as 3.13 days. Njegovan & Weisman’s (1997) isolates required 10 and 26 days more than did field-reared control groups to reach criterion on tasks with 4 and 9 S+s, respectively. Again, if the effect sizes in the present experiment were as large as those found by Njegovan & Weisman (1997), we would have detected them at 80% power, especially

PHILLMORE ET AL.: ISOLATE CHICKADEE PERCEPTION

considering that our sample size was eight isolates versus their three isolates. It is therefore likely that the effect of isolation on days to criterion in the distance cue discrimination was very small or zero. Finally, we found a minimum detectable effect size of 22% response between field- and isolate-reared chickadees for transfer to novel distances (at 25 and 50 m). Njegovan & Weisman (1997) did not conduct transfer tests in their research, so no comparison with published results is available. However, we do have evidence that an effect of that size is biologically relevant: the graded effect of increased distance between responding to 25- and 50-m vocalizations is, on average, in both groups, about 22%. In other words, we could detect a different effect of this size using the same data. In contrast, the difference between the field- and isolate-reared groups was much smaller. Therefore, it is likely that our results were not due to type II error, but rather that there was little or no effect of isolation on responding to these distances. In summary, although we cannot rule out the possibility that larger samples might lead to significant differences where none now exist, our power analyses suggest that any differences revealed by larger sample sizes are likely to be small and of little biological importance. In summary, field- and isolate-reared chickadees were nearly identical in discriminations based on distance cues: they acquired the discrimination at the same rate and to the same level of accuracy, and they generalized the discrimination to cues for intermediate distances equally well. We observed only one significant difference between field- and isolate-reared chickadees: isolates transferred less well to vocalizations recorded from different birds during test sessions than to those heard during training. EXPERIMENT 2: DISCRIMINATION OF INDIVIDUAL VOCALIZATIONS We adapted our assay of individual vocalization discrimination from that of Phillmore et al. (2002). Chickadees were trained to discriminate between individual conspecific vocalizations (4 male black-capped chickadee songs) and between individual heterospecific vocalizations (4 female zebra finch distance calls). Phillmore et al. (2002) found that field-reared chickadees quickly and accurately discriminated between these individual vocalizations. Some evidence suggests that isolate songbirds might be able to discriminate well between individual vocalizations. Loesche et al. (1991) compared barn and cliff swallows discriminating between individual vocalizations (begging calls) from both species. These swallows were reared in isolation from conspecific vocalizations; however, they were still able to learn five pairs of vocalizations from each species, for a total of 10 pairs. These results appear to support the hypothesis that juvenile experience is unnecessary for adult recognition of individual vocalizations. However, the alternative hypothesis, that experience greatly improves the recognition of individual vocalizations, is not necessarily supported or

tested. Loesche et al. (1991), who were not interested in the role of experience during development, did not compare their isolates to any field-reared swallows. Loesche et al.’s (1991) isolate swallows required hundreds of trials to acquire their individual discriminations. Therefore it is possible that although these isolate swallows could perform this task, it is not known whether their performance, relative to field-reared birds, was impaired. Here we tested both field- and isolate-reared chickadees in discriminations between individual vocalizations to provide unequivocal evidence about the role of experience during development.

Methods We presented eight different individuals’ vocalizations from the 5-m vocalization set: four songs and four calls. Two of the songs and two of the calls were standards and the other two songs and two calls were comparisons. Half of the birds received one set of individuals’ vocalizations, and the other half received a different set of different individuals’ vocalizations. Therefore, we trained the birds to discriminate the songs of two individual conspecifics from two other individual conspecifics, and, simultaneously, the calls of two individual heterospecifics from the calls of two other individual heterospecifics. Training continued until the birds reached a discrimination ratio of 0.80 on both songs and calls and discriminated more than 50% of the four standards above the 95% CI (i.e. 3 of the 4 individual vocalizations). If the bird did not reach criterion, training was discontinued at a maximum of 21 days. This learning criterion differs from that used in experiment 1 by including the number of standards above the 95% CI and by allowing slow-learning birds more days of training.

Results Gender effects We performed analyses of the discrimination ratio, percentage of standards above the 95% CI, and the number of sessions to criterion using rearing group (field versus isolate rearing) by gender (male versus female) by vocalization (songs versus calls) ANOVAs, yielding 10 possible main effects and interactions with gender. We found significant effects of gender only in the final percentage of standards above the CI (F1,10 =6.435, P=0.029; interaction with rearing: F1,10 =6.435, P=0.029). Table 1 shows the final percentage of standards above the 95% CI. Field-reared males and females discriminated all of the standard vocalizations; isolates (males and females) discriminated a lower percentage of standards above the 95% CI than did field-reared birds (F1,10 =8.000, P=0.015). Isolate males and females discriminated significantly fewer standards, but isolate females discriminated even fewer standards than did isolate males. In the other eight analyses, gender had small and inconsistent nonsignificant effects, with higher discrimination ratios for females than males but slower acquisition (measured by the percentage of standards >95% CI) for females than for

919

ANIMAL BEHAVIOUR, 65, 5

Table 1. Means±SEs for the final percentages of standards above the 95% CI in experiment 2

(a) Field-reared chickadees 100

Rearing group

75 Gender

Field

Isolate

Males Females

100±1% 100±1%

94±9% 75±8%

1

0.8

% Standards > 95% Cl

50

Discrimination ratio

920

0 100

50

0.4

25 0

0

Field-reared

Isolate-reared

Figure 5. Final day performance in the individual vocalization discrimination task (experiment 2; ■: black-capped chickadee songs; [: zebra finch calls) plotted as the mean±SE discrimination ratios for field-reared and isolate-reared chickadees.

males. Hence, except for the gender effect on the final percentage of standards above the 95% already presented (Table 1), we pooled performance for males and females in subsequent analyses of the results of experiment 2.

Stimulus-vocalization effects We plotted the discrimination of chickadee songs and zebra finch calls separately to maintain consistency with reporting in experiment 1. There were no significant differences between songs and calls in discrimination ratio (F1,12 =0.029, P=0.868) or percentage of standards above the 95% CI (F1,12 =0.001, P=0.999). Chickadees did not have a large or sustained advantage discriminating between individual chickadee songs over zebra finch calls (Figs 5, 6).

Final-day performance and sessions to criterion Discrimination ratios on the final day of training were high for both groups, and there was no significant difference between field- and isolate-rearing groups (F1,12 = 2.675, P=0.128). The nonsignificant difference between the field- and isolate-reared groups was small and in favour of the isolate-reared group (Fig. 5). We calculated the minimum detectable effect size as 0.137 with 80% power. As in other work from this laboratory (Sturdy et al. 2001), the discrimination ratio was relatively insensitive for detecting the effects of isolate rearing.

1

2

3

4

5

6

Final day

(b) Isolate-reared chickadees

75

0.6

0.2

Chickadee songs Zebra finch calls

25

1

2

3

4 5 6 7 Training day

8

9

10 Final day

Figure 6. Acquisition and final-day performance in the individual vocalization discrimination task (experiment 2) plotted as the percentage of standards above the 95% CI across training days for field-reared (a) and isolate-reared (b) chickadees. The number of training days shown for field-reared chickadees is less than that of isolate-reared chickadees because 6 days of training was the minimum necessary for some field-reared birds to meet criterion for the discrimination (more than 50% of all 4 standards >95% CI).

The percentages of standards above the 95% CI across several days of acquisition suggest that isolates learned to discriminate individual vocalizations more slowly than did field-reared chickadees. An ANOVA (F1,12 =8.378, P=0.01) confirmed that isolates required almost twice as many days to learn the discrimination to criterion as the field-reared group (XSE: isolates: 17.31.8 days, N=7; field-reared: 9.32.1 days, N=7; Fig. 6). Although the isolates had much more training, they discriminated fewer standards than did the field-reared group on the final (criterion) day (F1,12 =8.000, P=0.0152; Fig. 6). In summary, isolate-reared chickadees learned to discriminate individual vocalizations more slowly than did fieldreared chickadees, and isolates learned to discriminate slightly fewer standards over this much longer training period.

DISCUSSION We tested whether rearing black-capped chickadees without experience with adult conspecific vocalizations impairs their auditory perceptual abilities and specifically their ability to use distance cues and to discriminate between individual vocalizations. In experiment 1, isolate rearing had no significant consistent or strong effects on the chickadees’ ability to use distance cues in an operant

PHILLMORE ET AL.: ISOLATE CHICKADEE PERCEPTION

discrimination. Most important, isolates and field-reared birds both used distance cues to respond to vocalizations recorded at different distances as though they were on the same linear continuum. The sole significant difference between isolates and field-reared chickadees in experiment 1 was in transferring the distance cue discrimination to vocalizations from individuals not heard during training. This result may be related to the finding in experiment 2 that isolate rearing significantly slowed the acquisition of an operant discrimination between individual vocalizations and mildly reduced the total number of individual vocalizations that isolates were able to discriminate. Our results suggest a disassociation between these two perceptual abilities, the discrimination of distance cues and the discrimination between individual vocalizations. The ability to perceive distance cues may be innate, but the recognition of individual vocalizations appears to require auditory contact with adult conspecifics. Both the failure of isolation to affect distance cue perception and its significant effect on individual vocalization perception were general to conspecific and heterospecific vocalizations. This result suggests that distance cue perception and the effect of experience on auditory recognition may be general across complex vocalizations.

Distance Cue Discrimination That isolation affected the discrimination of individual vocalizations, but had little or no effect on the discrimination of distance cues, suggests that a reduction in experience with adult conspecific vocalizations during development is detrimental to songbirds’ abilities to perceive conspecific vocalizations. Isolation had effects on discrimination of individual vocalizations in the present study and on the discrimination of relative pitch in chickadees elsewhere (Njegovan & Weisman 1997). We suggest that this evidence supports the idea that perception of distance cues is so fundamental a skill that experience does not contribute to its development. Animals must estimate distance based on acoustic information from various sources (e.g. mates, offspring, predators). A male chickadee that cannot estimate the distance to a conspecific rival, a mate in need of guarding, or a predator lacks survival skills essential to propagation of his genes. This is not to say that experience does not affect the perception of distance cues; experience probably contributes to the refinement of distance perception and ranging, especially in the field. Although isolate-reared chickadees discriminated distance cues about as competently as did field-reared chickadees, it is unknown whether isolates could translate distance cues into accurate ranging of conspecifics in the field. Moreover, more difficult distance cue discriminations, for example, between vocalizations recorded at 75 and 100 m, might reveal differences between isolates and field-reared birds. Testing isolates with more distances or with a task requiring actual estimation of distance may show that isolates are deficient compared to field-reared birds. Even more stringent conditions of isolation, such as rearing chickadees entirely out of hearing

of other isolates, might affect the discrimination of distance cues. We do not rule out the effect of experience with conspecific vocalizations on distance cue perception, but our evidence suggests that the effect may be subtle and limited.

Individual Vocalization Discrimination The individual vocalization task was based on the natural task of distinguishing between vocalizations of familiar neighbours and unfamiliar strangers, and of territorial neighbours. Both of these tasks, discrimination between individual vocalizations in the laboratory and the field, require memorization of conspecific vocalizations. We found that isolates took significantly more trials than field-reared birds to learn almost the same number of individual vocalizations. These results are similar to those of Sturdy et al. (2001), who found that rearing isolate zebra finches under similar conditions slowed the rate at which they learned a song-note memorization task. Although isolates in both studies eventually learned almost the same number of exemplars as did field-reared birds, they learned tasks requiring memorization of conspecific vocalizations at a significantly lower rate. These results demonstrate that experience affects the rate of learning, and could be evidence that isolate rearing impairs memory. An experiment that tested whether specific tutoring with conspecific vocalizations aided isolates’ performance on memorization tasks would clarify the source of these deficits.

Conclusion Isolating birds during development is proving to be a useful tool to study the importance of experience with conspecific vocalizations for the development of several different perceptual abilities. Sturdy et al. (2001) found that isolate rearing impaired relative pitch perception in zebra finches, which replicated an earlier finding in chickadees (Njegovan & Weisman 1997). Together with results from the present study, these studies provide strong support that experience with conspecific vocalizations is important for the development of auditory perception in songbirds. To extend this line of research, it would be useful to determine how and why an impoverished developmental situation causes deficits in auditory perception.

Acknowledgments This research was supported by a grant from the Natural Science and Engineering Research Council (Canada). We thank Laurene Ratcliffe for help obtaining the nestling chickadees and Jen Sartor, Alex Hernandez, Angel Ip and Shauna Novak for help in hand-rearing them. Laurene Ratcliffe and Barrie Frost provided helpful comments on the manuscript. This research was approved by the University Animal Care Committee of Queen’s University (Protocol No. 99-085).

921

922

ANIMAL BEHAVIOUR, 65, 5

References Baptista, L. F. 1996. Nature and its nurturing in avian vocal development. In: Ecology and Evolution of Acoustic Communication in Birds (Ed. by D. Kroodsma & E. Miller), pp. 39–60. Ithaca, New York: Cornell University Press. Beecher, M. D., Campbell, S. E. & Burt, J. M. 1994. Song perception in the song sparrow: birds classify by song type but not by singer. Animal Behaviour, 47, 1343–1351. Cohen, J. 1988. Statistical Power Analysis. 2nd edn. Hillsdale, New Jersey: L. Erlbaum. Conover, W. J. & Iman, R. L. 1981. Rank transformations as a bridge between parametric and nonparametric statistics. American Statistician, 35, 124–129. Dabelsteen, T., Larson, O. N. & Pedersen, S. B. 1993. Habitatinduced degradation of sound signals: quantifying the effects of communication sounds and bird location on blur ratio, excess attenuation, and signal-to-noise ratio in blackbirds song. Journal of the Acoustical Society of America, 93, 2206–2220. Desrochers, A. 1990. Sex determination of black-capped chickadees with a discriminant analysis. Journal of Field Ornithology, 61, 79–84. Falls, J. B. & Brooks, R. J. 1975. Individual recognition by song in white-throated sparrows II. Effects of location. Canadian Journal of Zoology, 53, 1412–1420. Fotheringham, J. R. & Ratcliffe, L. 1995. Song degradation and estimation of acoustic distance in black-capped chickadees (Parus atricapillus). Canadian Journal of Zoology, 73, 858–868. Gentner, T. Q. & Hulse, S. M. 1998. Perceptual mechanisms for individual vocal recognition in European starlings, Sturnus vulgaris. Animal Behaviour, 56, 579–594. Godard, R. & Wiley, R. H. 1995. Individual recognition of song repertoires in two wood warblers. Behavioral Ecology and Sociobiology, 37, 119–123. Hays, W. L. 1994. Statistics. 5th edn. Austin, Texas: Harcourt Brace. Kroodsma, D. E. 1990. Using appropriate experimental designs for intended hypotheses in ‘song’ playbacks, with examples for testing effects of song repertoire sizes. Animal Behaviour, 40, 1138– 1150. Lanyon, W. E. 1979. Development of song in the wood thrush (Hylocichla mustelina), with notes on a technique for hand-rearing passerines from the egg. American Museum Novitates, 2666, 1–27. Lemon, R. E., Struger, J., Lechowicz, M. J. & Norman, R. F. 1980. Song features and singing heights of American warblers: maximization or optimization of distance? Journal of the Acoustical Society of America, 69, 1169–1176. Loesche, P., Stoddard, P. K., Higgins, B. J. & Beecher, M. D. 1991. Signature versus perceptual adaptations for individual vocal recognition in swallows. Behaviour, 118, 15–25. McGregor, P. K. 1991. The singer and the song: on the receiving end of bird song. Biological Review, 66, 57–81. McGregor, P. K. & Falls, J. B. 1984. The response of western meadowlarks (Sturnella neglecta) to the playback of undegraded and degraded songs. Canadian Journal of Zoology, 62, 2125–2128. Naguib, M. 1995. Auditory distance assessment of singing conspecifics in Carolina wrens: the role of reverberation and frequency attenuation. Animal Behaviour, 50, 1297–1307. Naguib, M. 1997a. Ranging of songs in Carolina wrens: effects of familiarity with the song type on use of different cues. Behavioral Ecology and Sociobiology, 40, 385–393.

Naguib, M. 1997b. Use of song amplitude for ranging in Carolina wrens, Thryothorus ludovicianus. Ethology, 103, 723–731. Naguib, M., Klump, G. M., Hillman, E., Griessmann, B. & Teige, T. 2000. Assessment of auditory distance in a territorial songbird: accurate feat or rule of thumb? Animal Behaviour, 59, 715–721. Nelson, B. & Stoddard, P. 1998. Accuracy of auditory distance and azimuth perception by a passerine bird in natural habitat. Animal Behaviour, 56, 467–477. Njegovan, M. & Weisman, R. 1997. Pitch discrimination in fieldand isolation-reared black-capped chickadees (Parus atricapillus). Journal of Comparative Psychology, 111, 294–301. Njegovan, M., Hilhorst, B., Ferguson, S. & Weisman, R. 1994. A motor driven feeder for operant training in song birds. Behavior Research Methods, Instruments, and Computers, 26, 26–27. Phillmore, L. S., Sturdy, C. B., Ramsay, S. M. & Weisman, R. G. 1998. Discrimination of auditory distance cues by black-capped chickadees (Poecile atricapillus) and zebra finches (Taeniopygia guttata). Journal of Comparative Psychology, 112, 282–291. Phillmore, L. S., Sturdy, C. B., Turyk, M. & Weisman, R. G. 2002. Discrimination of individual vocalizations by black-capped chickadees (Poecile atricapilla). Animal Learning and Behavior, 30, 43–52. Pyle, P. 1997. Identification Guide to North American Birds. Part I. Bolinas, California: Slate Creek Press. Ratcliffe, L. M. & Otter, K. 1996. Sex differences in song recognition. In: Ecology and Evolution of Acoustic Communication in Birds (Ed. by D. Kroodsma & E. Miller), pp. 339–355. Ithaca, New York: Cornell University Press. Searcy, W. A. & Yasukawa, K. 1996. Song and female choice. In: Ecology and Evolution of Acoustic Communication in Birds (Ed. by D. Kroodsma & E. Miller), pp. 454–473. Ithaca, New York: Cornell University Press. Sokal, R. R. & Rohlf, F. J. 1980. Biometry. New York: W. H. Freeman. Stoddard, P. K., Beecher, M. D., Loesche, P. & Campbell, S. E. 1992. Memory does not constrain individual recognition in a bird with song repertoires. Behaviour, 122, 274–287. Sturdy, C. B., Phillmore, L. S., Sartor, J. J. & Weisman, R. G. 2001. Reduced social contact causes auditory perceptual deficits in zebra finches, Taeniopygia guttata. Animal Behaviour, 62, 1207– 1218. Volman, S. & Khanna, H. 1995. Convergence of untutored song in group-reared zebra finches Taeniopygia guttata. Journal of Comparative Psychology, 109, 211–221. Weisman, R., Ratcliffe, L., Johnsrude, I. & Hurley, T. A. 1990. Absolute and relative pitch production in the song of the blackcapped chickadee. Condor, 92, 118–124. Weisman, R., Njegovan, M. & Ito, S. 1994. Frequency-ratio discrimination by zebra finches (Taeniopygia guttata) and humans (Homo sapiens). Journal of Comparative Psychology, 108, 363– 372. Wiley, R. H. 2000. A new sense of the complexities of bird song. Auk, 117, 861–868. Wiley, R. H. & Richards, D. G. 1982. Adaptions for acoustic communication in birds: sound transmission and signal detection. In: Acoustic Communication in Birds. Vol. 1 (Ed. by D. Kroodsma & D. Miller), pp. 131–181. New York: Academic Press. Zann, R. 1996. The Zebra Finch: a Synthesis of Field and Laboratory Studies. Oxford: Oxford University Press.