Ethology 107, 961Ð974 (2001) Ó 2001 Blackwell Wissenschafts-Verlag, Berlin ISSN 0179±1613

RESEARCH PAPERS Laboratoire de Biologie Animale, Universite Jean Monnet, St-Etienne; CEFE±CNRS, UPR 9056, Montpellier; CNRS, NAMC MeÂcanismes de Communication, Orsay

Acoustic Communication in a Black-Headed Gull Colony: How Do Chicks Identify Their Parents? Isabelle Charrier, Nicolas Mathevon, Pierre Jouventin & Thierry Aubin Charrier, I., Mathevon, N., Jouventin, P. & Aubin, T. 2001: Acoustic communication in a blackheaded gull colony: how do chicks identify their parents? Ethology 107, 961Ð974.

Abstract Chicks perform conspicuous begging behaviour in response to the arrival of a parent. In seabirds colonies, as nests are close to each other, chicks are permanently surrounded by sound and visual stimuli produced by adult conspeci®cs approaching their nests. However, in spite of these conditions, black-headed gull chicks begin to vocalize as their parent approaches even before they can see it. In this paper, we report ®eld experiments testing soundbased discrimination of parents by black-headed gull chicks. Focusing on the `long call', i.e. the signal emitted by parents when coming back to the nest, we investigate here the acoustic parameters used for this recognition process. By playback experiments using modi®ed `long calls', we demonstrated that signals without amplitude modulation still elicit responses in chicks. In contrast, frequency modulation appears essential. In the frequency domain, experiments revealed that frequency analysis is precise. Chicks did not react when the frequency spectrum of parental call was shifted 20 Hz down or up. The totality of harmonics is not necessary: chicks require only two harmonics to discriminate between parents. Signal redundancy is of great signi®cance since a minimum of four successive syllables in parental `long call' are required to elicit reaction in the chick. Corresponding author: Nicolas Mathevon, Laboratoire de Biologie Animale, Universite Jean Monnet, 42023 St-Etienne cedex 2, France. E-mail: mathevon@ univ-st-etienne.fr

U. S. Copyright Clearance Center Code Statement: 0179-1613/2001/10711±0961$15.00/0 www.blackwell.de/synergy

962

I. Charrier, N. Mathevon, P. Jouventin & T. Aubin

Introduction In order to signal to parents their energy needs, and to elicit parental care, o€spring perform conspicuous begging behaviour (O'Connor 1984; Stamps et al. 1989). The evolution of begging may be driven by various constraints including the e€ectiveness of parental feeding elicitation, competition among siblings, and the eventual costs in terms of signal production and predation (Harper 1986; Godfray 1991, 1995; Leech & Leonard 1996). As parental resources may be limited, competition among a litter will select for intense begging (Leonard & Horn 1998; Leonard et al. 2000). However, begging behaviour can represent an energetic cost and, by its conspicuous aspect, may increase predation pressure (Krebs & Dawkins 1984; Harper 1986; Redondo & Castro 1992). Thus, o€spring should perform begging only when the probability of getting more food or care from their parents is suciently high to compensate for the possible costs. This situation is common within solitary nesting birds where any adult approaching a nest is likely to be the nest's owner bringing some food: nestlings may then initiate begging as soon as they see or hear any stimulus which may arise from a coming adult, e.g. a human voice elicits begging in chicks of magpie Pica pica (Redondo & Castro 1992). In seabird colonies such as those of the black-headed gull Larus ridibundus, nests are close to each other (2±5 nests per m2, Cramp & Simmons 1983) and many birds may ¯y over a nest at any given moment. The chicks are, therefore, permanently surrounded by sound and visual stimuli produced by adult conspeci®cs. A chick may consider each adult ¯ying over its nest to be a potential parent. Performing begging behaviour in response to any approaching ¯ying adult is likely to be maladaptive in terms of energy expenditure and risk of attack and injury from other adults (Prott & McLean 1991). To avoid inappropriate begging, o€spring might wait until the adult lands on their nest. However, ®eld observations show that chicks begin to vocalize as their parent approaches and before it lands. This performance requires that o€spring are able to distinguish their parents from all the other ¯ying adults approaching the nest. There are only two possible cues usable by chicks to perform this discrimination: visual patterns and acoustic features of the adult call. To be reliable, identi®cation by sight requires a short distance between chicks and their parents since the nest environment is often obstructed by vegetation. By contrast, sound identi®cation does not require that the receiving chick see the bird emitting the sound. It allows the chicks to initiate begging before they actually see their parent. Pioneering studies showed that sound-based individual recognition between parents and young is present in many Laridae (Beer 1970a, Beer 1970b; Evans 1970; Buckley & Buckley 1972; Miller & Emlen 1975), but the acoustic parameters used for this recognition process are not known. Here we present a ®eld experimental study which investigates how black-headed gull's o€spring identify the call of their potential feeders.

Sound-Based Discrimination in Black-Headed Gull

963

Methods Biological Material

Recordings and playback experiments were performed during the breeding season of the black-headed gull from early May to mid-June 1999, at the Etang de la Ronze (Plaine du Forez, near Saint-Etienne, France). This pond accommodates the largest colony of black-headed gulls in southern Europe with about 4000 pairs (Rimbert 1999). The black-headed gull is a semialtricial bird. Chicks remain in the nest until the age of 3±4 weeks. Playback tests have been made when chicks were in the nest and dependent on their parents for food, i.e. between 3±4 days and 3±4 weeks. To facilitate chick identi®cation, nests were marked with a numbered label. Recording Procedure

To approach and observe birds closely without frightening them, we used a ¯oating observation blind camou¯aged with various pieces of typical vegetation taken from the shore. Recordings were taken with a Revox M 3500 microphone (frequency bandwidth: 150±18000 Hz, ‹ 1dB) connected to a Sony TC-D5M audiotape-recorder. During the recordings, the microphone was mounted on a horizontal boom extending approximately 1 m from the blind. The microphone was 80 cm above the water level. Adults emit `long calls' in ¯ight, from a height of 3±4 m above the nest until landing, when they come back to feed their chicks. During recording, the distance between the emitting bird and the microphone was approximately 2 m. The `long call' emitted by the adults when mates, or parents and chicks, meet each other (according to Tinbergen 1959; Cramp & Simmons 1983) is composed of several syllables. The syllable number varies between 4 and 12, with a median at 5. Each syllable is a complex sound, composed of a fundamental frequency and its harmonics (Fig. 1). Signal Acquisition and Analysis

Calls were digitized with a 16-bit acquisition card at a 22050-Hz sample rate with a 120-dB/octave antialiasing ®lter, using acquisition software (Cool Edit 1996; Syntrillium Software). Signals were then stored on the hard disk of a PC computer. To examine and modify sounds in the temporal and frequency domains, we used the SYNTANA analytical package (Aubin 1994). Playback Procedure

For playback experiments we used a Sony TC-D5M tape-recorder connected to an Audax unidirectional loudspeaker via a 10-W battery-powered ampli®er

Fig. 1: Structure of a natural `long call'. (a) Spectrogram of a `long call' composed of 5 syllables (window size: 1024, sliding: 10). (b) Frequency spectrum of a syllable. (c) Distribution of the number of syllables in natural long calls. There are always 4 syllables at least, with a peak at 5 (n ˆ 105 calls from 22 individuals)

964 I. Charrier, N. Mathevon, P. Jouventin & T. Aubin

Sound-Based Discrimination in Black-Headed Gull

965

constructed in the laboratory (frequency response 1±9 kHz, ‹ 4 dB). During playback, the loudspeaker was put on the front of the observation blind. The distance from the loudspeaker to the nest was 2 m. Signals were played at a natural sound pressure level (SPL ˆ 83±92 dB measured at 1 m with a BrueÈl & Kjaer Sound Level Meter type 2235). We tested the chicks while the parents were out of the nest, looking for food, and when the background noise in the neighbourhood was the lowest. Each nest may contain several chicks. To get independent samples and to avoid problems such as chicks not responding directly to the playback but to the reaction of their siblings, each of our tested chicks belonged to a di€erent nest. To prevent habituation, we tested the chicks with three di€erent experimental signals a day (with a minimum of 2 h between each). Each chick was tested every 2 or 3 days. Each experimental signal was broadcast once. Each chick was tested with the whole set of experimental signals. For each chick, the order of presentation of the signals was randomized. Hence, the observed responses for the whole group of tested chicks were neither a result of cumulative excitation nor dependent on playback order or on the age of chick. Criteria of Response

In natural conditions, when a chick is alone, or with its siblings, waiting in the nest to be fed by its parents, it is usually silent and lying quietly. When returning to the nest, a parent emits its `long call'. At this moment, the chick holds up its head, often moves around the nest and cries. This reaction constitutes the beginning of the begging behaviour which continues with several calls and di€erent movements if the parent lands on the nest (Cramp & Simmons 1983). This reaction is observed only in response to the arrival of the chick's own parents, i.e. when a neighbouring adult is returning to its nest, the chick does not react. To assess the nature of response during playback experiments, the behaviour of chicks was observed for 5 min before and after signal emission. We established a two-point behavioural scale: a positive response and a negative one. A response was considered positive when the emission of an experimental signal induced a similar reaction to that in natural conditions: the chicks held up the head, moved in the nest and sometimes cried. The chick's reaction was often less strong than in natural interactions with parents (certainly due to the lack of visual signal in the experimental condition), but the observed behaviours were unambiguous. If the emitted signal induced no change in the behaviour of chicks, we considered it a negative response. Two observers were present to discuss the chick's response and approve a positive or negative response. As the chick's responses were easy to interpret, no confusion between behavioural responses arose. Experimental Signals

We tested 19 experimental signals using natural long calls. In a ®rst experiment, we tested to see if black-headed gull chicks respond selectively to their

966

I. Charrier, N. Mathevon, P. Jouventin & T. Aubin

parents' calls. We then investigated the acoustic features required for the elicitation of response in chicks. Taken into account the temporal and frequential domains of the sound, we modi®ed temporal parameters such as the number of syllables, the frequency and the amplitude modulations, and frequency parameters such as the frequency bandwidth and the pitch values of harmonics. Experiment 1: can chicks discriminate parental from nonparental calls? To test the ability of chicks to discriminate between their parents and other birds, we played back a natural `long call' of one of the parents and a `long call' of a nonparental bird. To rule out e€ects of particular individuals, each chick was tested with a call coming from di€erent nonparental individuals. We then used 16 di€erent nonparental calls for this experiment. A nonparental bird was either a neighbouring bird (familiar), i.e. a bird which has its nest in a radius of 5 m from the tested chick's nest, or a non-neighbouring bird (nonfamiliar), i.e. a bird whose nest was situated at a minimum of 20 m from the tested chick's nest. Eight chicks were tested with neighbouring birds' calls while the other 8 chicks were tested with non-neighbouring birds' calls. In both these chicks' series, the presentation of parental and stranger signals were randomized between chicks: 8 chicks were tested ®rst with the natural signal, and 8 with the signal of a stranger bird ®rst. Experiment 2: does any call's syllable bring the required information for call discrimination by chicks? The `long call' is composed of several syllables. For each tested chick, we built an experimental long call by repeating 8 times one syllable taken at random among the successive syllables of one chick's parental call. These calls were then composed of 8 identical parental syllables. The silence duration between syllables was representative of intersyllable silences present in the natural call. To prevent spectral artefacts arising from an abrupt gap in amplitude, `cut and paste' was carried out at the minimum (near zero) amplitude values of the signal. Experiment 3: is syllabic redundancy an important feature for parental call discrimination? The number of syllables di€ers among calls of an individual and among calls of di€erent individuals. Most long calls contain 4±6 syllables (Fig. 1). To test if this syllable number is important for the elicitation of chick's response, we presented experimental `long calls' composed of 1, 2, 3 or 4 syllables to the chicks. For each tested chick, four di€erent natural parental calls were used to build the experimental `long calls'. From these natural calls, we select at random either 1, 2, 3 or 4 successive syllables to make the experimental signals. Each experimental `long call' was presented only once to each chick.

Sound-Based Discrimination in Black-Headed Gull

967

Experiment 4: are temporal parameters used for discrimination? A natural `long call' presents amplitude modulation and frequency modulation. To test the importance of both these parameters for individual recognition, we constructed two types of signals: a) signal without amplitude modulation but with natural frequency modulation (Fig. 2). To build this signal, we used the analytic signal concept which allows demodulation of amplitude modulated signal using Hilbert transform (Seggie 1987; Mbu-Nyamsi et al. 1994). (b) Signal without frequency modulation but with natural amplitude modulation (Fig. 2). To make this signal, we synthesized a carrier frequency consisting of a harmonic series with the mean frequency value of each harmonic of the parental call. Thus the frequency values of the harmonic series were kept constant. Moreover, the natural repartition of energy between harmonics was respected. A natural amplitude modulation was extracted from the call of the parent using the analytic signal calculation (BreÂmond & Aubin 1992; MbuNyamsi et al. 1994) and then applied to this carrier frequency without frequency modulation. Each chick was tested with one parental call without amplitude modulation, and with one other parental call without frequency modulation. Experiment 5: is the spectral pro®le important for parental call's discrimination? Two types of experimentation were carried out: to assess the accuracy of pitch discrimination, natural long calls were shifted up or down. Eight linear shifts were applied to 8 di€erent parental `long calls', for each chick. This was done by picking a data record through a square window, applying short-term overlapping (50%) fast Fourier transform (FFT), followed by a linear shift (+ or )) of each spectrum and by a short-term inverse fast Fourier transform (FFT±1, Randall & Tech 1987). The window size was 4096 points (DF ˆ 5Hz). The values were ‹ 50, ‹ 30, ‹ 20 and ‹ 10 Hz. In these experimental signals the natural temporal structure (frequency and amplitude modulations) was kept. To test if all harmonics are essential for individual recognition, we performed ®ltration of the natural signal. Signals were ®ltered by band-pass digital ®lters. This was done by applying optimal ®ltering with an FFT (Press et al. 1988). The window size was 4096 points (DF ˆ 5Hz).We constructed two experimental signals: one with only the two most powerful harmonics and the other one with the most powerful harmonic (Fig. 2). For each chick, two di€erent parental `long calls' were used to make these experimental signals. Statistics

To assess the signi®cance of behavioural changes obtained with experimental signals we used the McNemar Test (Sokal & Rohlf 1995).

968

I. Charrier, N. Mathevon, P. Jouventin & T. Aubin

969

Sound-Based Discrimination in Black-Headed Gull

Fig. 2: Spectrograms of modi®ed signals in the time (a, b) and frequency (c, d) domains. (a) Signal without amplitude modulation. (b) Signal without frequency modulation. (c) Filtered signal composed of the 2 most powerful harmonics. (d) Filtered signal composed of the most powerful harmonic

Results Table 1 presents the results of experiments. Experiment 1: chicks respond speci®cally to parental calls Sixteen chicks were tested with the natural signal of one of the parent and with the one of a nonparental bird. Chicks never respond to familiar nonparental calls nor to nonfamiliar calls. Only the parental signals elicited positive responses (Table 1). This experience con®rms that chicks are able to discriminate parental calls among calls of other individuals. Table 1: Responses of chicks to experimental signals. Modi®cations were made in both temporal and frequency domains (McNemar Test, *p < 0.01) Experiment

Number of tested birds

Number of responding birds

Parental signal Non-parental signal

16 16

16 0*

Long call composed by a repetition of 8 syllables

15

15

Long call composed by: 1 syllable 2 syllables 3 syllables 4 syllables

11 11 11 11

1* 4* 6* 11

Long call without AM Long call without FM

11 6

11 0*

6 6

6 1*

16 16 16 16 12 12 12 12

0* 1* 5* 12 11 2* 2* 0*

Filtrations 2 harmonics 1 harmonic Linear shift (Hz) +50 +30 +20 +10 )10 )20 )30 )50

970

I. Charrier, N. Mathevon, P. Jouventin & T. Aubin

Experiment 2: each call's syllable contains the totality of required information for call discrimination The experimental signal composed of 8 identical parental syllables elicited positive responses for each of the chicks tested (Table 1). Experiment 3: redundancy increases the probability of chick's response The percentage of positive responses increased with the syllable number. However, even `long calls' with 1 or 2 syllables managed to elicit some positive responses. 100% positive responses were obtained with `long calls' composed of 4 syllables (Table 1). Experiment 4: amplitude modulations are not used for call discrimination. Frequency modulations are essential for sound elicited behaviour The signal without amplitude modulation elicited a positive response for each trial (Table 1). The signal without frequency modulation did not induce any positive responses from the chicks (Table 1). Experiment 5: the frequency analysis is precise and 2 harmonics alone are sucient to elicit chick's response The higher the absolute value of shifts, the more numerous were the negative responses. Results obtained with positive shifts were symmetric to those obtained with negative shifts (Table 1). Only the + and ±10 Hz signals elicited the more numerous positive responses (91.6 and 75% respectively). The signal with 2 harmonics elicited 100% positive responses. On the contrary, the signal with only 1 harmonic elicits 16.7% positive responses (Table 1). Discussion This paper reports ®eld experiments investigating how black-headed gull's chicks may discriminate between parental calls and nonparental ones. The emphasis is on which call parameters are required to elicit the chick's reaction. Discrimination Between Parental and Nonparental Acoustic Signals

In our experiments, nonparental `long calls' never elicited positive responses whereas chicks always reacted to parental `long calls'. This reaction indicates that black-headed gull chicks are able to discriminate the calls of their parents among the calls of any other bird, the latter being either a neighbouring (familiar) or a non-neighbouring (nonfamiliar) bird. Chicks are then able to categorize acoustic signals into two categories: parental or nonparental. Our experiments did not

Sound-Based Discrimination in Black-Headed Gull

971

investigate if this categorization is more accurate and allows chicks to di€erentiate between nonparental familiar vs. nonfamiliar birds' calls. Our results simply show that chicks do not respond to such calls. On the contrary and in spite of the confusing environment of the colony, chicks perform perception and identi®cation of parental calls. We then assess that recognition of parents' calls by blackheaded gull chicks is accurate and reliable. Acoustic Features Required to Elicit Chicks' Response

Experiments in temporal domain showed that all signals without amplitude modulation still elicited positive responses. Amplitude modulations are not used for parental call recognition by chicks. In contrast, amplitude modulation alone is not sucient to elicit a response in chicks. Frequency modulation appears to be essential: signals without frequency modulation never elicited positive responses. Of course, the lack of response to calls when frequency modulation was removed does not necessarily imply that frequency modulations are essential for parental recognition per se, and may be based on a failure in species recognition. A similar study in king penguins Aptenodytes patagonicus has brought the same kind of results (Jouventin et al. 1999). Frequency modulation appears to be essential for parental calls recognition in king penguin species, as it is for the black-headed gull. Maybe this is a common feature among numerous colonial seabirds. Experiments in the frequency domain revealed that the black-headed gull chick is particularly sensitive to di€erences in frequency values. Signals with a linear shift of 20±30 Hz up or down elicited a majority of negative responses. The decline of positive responses was still obvious with a shift of 10 Hz up or down. Young black-headed gulls can detect as little as a 1% shift in the frequency of their parents calls. It is known that birds can discriminate a 1% change in frequency (Dooling 1982). It is a particularly accurate and outstanding frequency analysis considering that this species live in a noisy environment. Colonial birds develop high capacities of signal detection in the noise. Previous studies in king penguins have shown that adults and chicks can detect their mate or parents calls with a signal-to-noise ratio of )4 to )6 dB (Aubin & Jouventin 1998). Filtered signals showed that a single harmonic is not sucient to elicit a response. Chicks pay attention to the width of their parental calls since two harmonics are at least necessary for chicks to recognize the parental call. The Importance of Redundancy

`Long calls' are made by a succession of a number of syllables. The experimental signal composed of the repetition of the same parental syllable elicited a positive response for all chicks tested. Each syllable contains the same information with regard to parental calls vs. nonparental calls discrimination by the chicks. However, a call composed of an isolated parental syllable elicits only

972

I. Charrier, N. Mathevon, P. Jouventin & T. Aubin

few responses in chicks. In contrast, 4 successive parental syllables elicit responses for all the tested chicks. This importance of redundancy could be due to several factors. First, the repetition may be a part of the information coding, e.g. calls with less than 4 syllables could be considered by the chick as a nonspecies speci®c signal. Such modi®ed calls may not constitute a sucient stimulus for species discrimination, and then be outside the range that chicks will accept as a potential parent regardless of the `parental identity' of syllables per se. This explanation seems to be dubious since calls composed of 2 or 3 syllables still elicit responses in a number of tested chicks. The second possible explanation is that redundancy allows chicks to check several successive times the information contained in the syllable. Chick's response occurs only when the chick is sure of the parental identity of the emitting bird. The information coding is then secured by redundancy. This system may be especially useful in the jamming environment of the colony where the background noise impairs the communication process (Aubin & Jouventin 1998). According to the mathematical theory of communication (Shannon & Weaver 1949), the volume of information is expressed by the following equation: V ˆ FcT with c ˆ log…1 ‡ S/N†

…1†

where F represents the width of the frequency band (Hz), c the intensity of the signal, T the duration of the signal and S/N the signal/noise ratio. In the colony, because of the loud background noise, the S/N ratio is low. Background noise is the summation of conspeci®c vocalizations. If birds of the colony increase c, they increase the level of the noise too. By repeating syllables, i.e. by repeating several times the information needed by the chick to decide whether or not to respond, black-headed gull adults use temporal redundancy to increase the factor T in equation 1 and so to keep the volume of information constant. Our experiments showed that a long call must contain a minimum of 4 parental syllables to elicit a chick's reaction. This result is linked to the fact that natural long calls are composed of at least 4 syllables. Most of the time, natural long calls present more than 5 syllables, giving a `safety margin' for the recognition. Moreover, the spectrum of natural long call is composed of more than 10 harmonics. This frequency redundancy increases the factor F in equation 1. On the other hand, the long call of the blackheaded gull presents some features allowing the emitter to be located. The gaps of amplitude between syllables and silences, and amplitude modulation are known to represent acoustical spatial cues (Wiley & Richards 1978; Knudsen 1999). Moreover, a recent study showed that a broadband signal exceeding 3 kHz, as the black-headed gull, is necessary for unambiguous localization (Saberi et al. 1999). With these strategies a black-headed gull adult enhances its chances to be heard, localized and recognized in spite of the noise of the colony.

Sound-Based Discrimination in Black-Headed Gull

973

Adaptive Signi®cance of Chicks' Response to Parental Call

Our work shows that black-headed gull chicks react selectively to the calls of their parents. One can not exclude the possibility that chicks' reaction to their parents' call could be no more than a `Pavlovian salivation' in response to a stimulus associated with the presentation of food. Nevertheless, chicks are able to distinguish accurately parental calls from nonparental calls using particular sound features. What could be, if any, the biological signi®cance of this ability? We suggest that there are two alternative, although not exclusive, hypotheses. First, by enabling the chick to react to its parent as soon as possible, before the parent lands and is visible to the chick, this ability may enhance the impact of the begging behaviour on the parent. In a litter, chicks carrying out good discrimination of parental calls, and consequently reacting earlier than others, may gain more food or care than others. A second hypothesis is that chicks' reaction may facilitate the ®nding of the nest by the adult and/or may incite the adult to land. A previous study has shown that laughing gull Larus atricilla parents react by approaching the voices of chicks by orienting towards the direction from which the sounds come (Beer 1979). If chicks responded to any arriving adult, the latter could be confused by multiple signals and could land on a neighbouring nest: this would be maladaptive since black-headed gull adults vigorously defend their territory (Cramp & Simmons 1983). Acknowledgements The study was supported by the University Jean Monnet and the CNRS (CEPE and NAMC). We thank Jean-Dominique Lebreton and his team especially, for allowing us to work at La Ronze colony and for their kind advice in the ®eld. Prof. Peter Bowden helped with the English translation.

Literature Cited Aubin, T. 1994: SYNTANA: a software for the synthesis and analysis of animal sounds. Bioacoustics 6, 80Ð81. Aubin, T. & Jouventin, P. 1998: The cocktail party e€ect in king penguin colonies. Proc. R. Soc. London 265, 1665Ð1673. Beer, C. G. 1970a: On the responses of laughing gull chicks, Larus atricilla, to the calls of adults. I. Recognition of voice of the parents. Anim. Behav. 18, 652Ð660. Beer, C. G. 1970b: On the responses of laughing gull chicks, Larus atricilla, to the calls of adults. II. Age changes and responses to di€erent types of call. Anim. Behav. 18, 661Ð667. Beer, C. G. 1979: Vocal communication between laughing gull parents and chicks. Behaviour 70, 118Ð146. BreÂmond, J. C. & Aubin, T. 1992: The role of amplitude modulation in distress-call recognition by the black-headed gull (Larus ridibundus). Ethol. Ecol. Evol. 4, 187Ð191. Buckley, P. A. & Buckley, F. G. 1972: Individual egg and chick recognition by adult royal terns (Sterna maxima maxima). Anim. Behav. 20, 457Ð462. Cool Edit 1996: Syntrillium software. Syntrillium Software Corp., Phoenix, AZ. Cramp, S. & Simmons, K. E. L. 1983: Handbook of birds of Europe the Middle East and North Africa. The Birds of the Western Palearctic, Vol. III. Oxford Univ. Press, Oxford, London, New York. Dooling, R. J. 1982: Auditory perception in birds. In: Acoustic Communication in Birds, Vol. 1. (Kroodsman, D. E. & Miller, E. H., eds). Academic Press, New York, pp. 95Ð130.

974

I. Charrier, N. Mathevon, P. Jouventin & T. Aubin

Evans, R. M. 1970: Parental recognition and the mew call in Larus bulleri. The Auk 87, 503Ð513. Godfray, H. C. J. 1991: Signalling of need by o€spring to their parents. Nature 352, 328Ð330. Godfray, H. C. J. 1995: Evolutionary theory of parent-o€spring con¯ict. Nature 376, 133Ð138. Harper, A. B. 1986: The evolution of begging: sibling competition and parent-o€spring con¯ict. Am. Nat. 128, 99Ð114. Jouventin, P., Aubin, T. & Lengagne, T. 1999: Finding a parent in a king penguin colony: the acoustic system of individual recognition. Anim. Behav. 57, 1175Ð1183. Knudsen, E. I. 1999: Early experience and critical periods. In: Fundamental Neuroscience (Zigmond, M. J., Bloom, F. E., Landis, S. C., Roberts, J. L. & Squire, L. R., eds). Academic Press, pp. 637Ð654. Krebs, J. R. & Dawkins, R. 1984: Animal signals: mind reading and manipulation. In: Behavioural Ecology An Evolutionary Approach (Krebs & Davies, ed.). Blackwell, Oxford, pp. 380Ð402. Leech, S. M. & Leonard, M. L. 1996: Is there an energetic cost to begging in nestling tree swallows (Tachycineta bicolor) ? Proc. R. Soc. London 263, 983Ð987. Leonard, M. L. & Horn, A. G. 1998: Need and nestmates a€ect begging in tree swallows. Behav. Ecol. Sociobiol. 18, 199Ð206. Leonard, M. L., Horn, A. G., Gozna, A. & Ramen, S. 2000: Brood size and begging intensity in nestling birds. Behav. Ecol. 11, 196Ð201. Mbu-Nyamsi, R. G., Aubin, T. & BreÂmond, J. C. 1994: On the extraction of some time dependent parameters of an acoustic signal by means of the analytical signal concept. Its application to animal sound study. Bioacoustics 5, 187Ð203. Miller, D. E. & Emlen, S. T. 1975: Individual chick recognition and family integrity in the Ring-billed Gull (Larus delawarensis). Behaviour 52, 124Ð144. O'Connor, R. J. 1984: The Growth and Development of Birds. Wiley, New York. Press, W. H., Flannery, B. P., Teukolsky, S. A. & Vetterling, W. T. 1988: Numerical Recipes in C. The Art of Scienti®c Computing. Cambridge Univ. Press, New York. Prott, F. M. & McLean, I. G. 1991: Recognition of parents' calls by chicks of the snares crested penguin. Bird Behav. 9, 103Ð113. Randal, R. B. & Tech, B. 1987: Frequency analysis. Naerum: BruÈel & Kjaer. Redondo, T. & Castro, F. 1992: Signalling of nutritional need by magpie nestlings. Ethology 92, 193Ð204. Rimbert, P. 1999: Les oiseaux de la Loire.- LPO- Loire. Saberi, K., Takahashi, Y., Farahbod, H. & Konishi, M. 1999: Neural bases of an auditory illusion and its elimination in owls. Nature Neuroscience 7, 581Ð681. Seggie, D. 1987: The application of analytic signal analysis in speech processing. Processing Inst. Acoustics 8, 82Ð85. Shannon, C. E. & Weaver, W. 1949: The mathematical theory of communication. Univ. of Illinois Press, Urbana. Sokal, R. R. & Rohlf, F. J. 1995: Biometry, 3rd edn. Freeman, New York. Stamps, J. A., Clark, A., Arrowood, P. & Kus, B. 1989: Begging behaviour in budgerigars. Ethology 81, 177Ð192. Tinbergen, N. 1959: Comparative studies of behaviour of Gulls (Laridae): a progress report. Behaviour 15, 1Ð70. Wiley, R. H. & Richards, D. G. 1978: Physical constraints on acoustic communication in the atmosphere: implication for the evolution of animal vocalisations. Behav. Ecol. Sociobiol. 3, 69Ð94. Received: July 2, 2000 Initial acceptance: November 7, 2000 Final acceptance: February 5, 2001 (A. Kacelnik)