13

It is widely known that most of the lemur vocal repertoires are poorly studied or even completely undescribed. Recent analysis of Eulemur vocal repertoires from the phonetic point of view allow the categorisation of tonal calls, noises and composite emissions of different subspecies and species. These developments suggest that, for understanding the evolution of the Eulemur vocal repertoire, comparative analysis of the vocal repertoires of captive specimens provides a viable method to study species relationships and therefore to identify the developmental processes that underlie evolutionary change. In this work we discuss results from a comparative acoustic analysis of vocalisations in Eulemur rubriventer, Eulemur macaco (E. m. maccco, E. m. Flavifrons), Eulemur coronatus, Eulemur mongoz and Eulemur fulvus. Detailed acoustical analysis was necessary to generate data for the classification of the naturally occurring vocalisations. This classification is essential to figure out which vocal types could be compared and to make a preliminary description of the vocal repertoires. Digitised calls were categorised and then statistically analysed. We measured 40 different acoustic parameters per vocalisation, including dominant frequency, four spectrum peaks, index of formant dispersion, duration and harmonic composition. We found three main categories: at single unit, simple structure (featuring Tonal calls, Snorts, Gulps);  b) multiple unit, simple structure (featuring Grunts, Long Grunts, Hoots); c) multiple unit, complex structure (Snort- Grunt- Clear Calls, Grunt -Clear Calls, Snort-Grunt, Long Grunt -Clear calls, Rising Grunts). Moreover, the vocal repertoires are quite divergent among species of lemurs that are closely related; these results suggest that differences in the phonatory apparatus and in their social behaviour have been underestimated.

Many insects and other arthropods use stridulatory organs without resonant structures. In such systems the resulting signal (sound pressure versus time) contains a few to several hundred small spikes. The repetition period of the spikes is 0.5 to 2 ms in many species while the spikes are sometimes as short as 0.1 ms. In time-frequency representations (TFRs) like sonograms, the problem is to find the optimal time window to analyse such signals. With a short-time Fourier transform it is not possible to get a sharp 2-dimensional image with high time and high frequency resolution simultaneously. For nonresonant stridulatory signals it is better to use TFRs based on the Wigner distribution. Unfortunately, this produces cross components that complicate a more detailed analysis. To avoid these problems, special methods based on general Cohen's class TFRs have been developed, for example the Adaptive Optimal-Kernel TFR by Baraniuk and Jones. The use of new and established methods of digital signal processing will be discussed.

The main sound producing apparatus in singing cicadas is a pair of tymbals, positioned in the lateral parts of the first abdominal segments. In addition to this, some cicadas produce sounds also by other means, often associated with wing movements. To get a better insight into the wing clicking mechanism in Cicadatra persica and Pagiphora anullata we used a commercial high-speed digital video camera JVC GR-DVL 9800E. With this camera one can fill each frame on the tape with 2 or 4 smaller subframes and after deinterlacing extract two half-frames from each. This corresponds to an 8x final recording speed and is adequate for a recording frequency of 200 frames per second. For such analyses we used a Digital Origin EditDV computer program and Adobe Photoshop 5 software. To compare sound tracks with corresponding image sequences we exported sound clips from EditDV to Canary 1.2. Oscillograms were subsequently transferred to Photoshop 5 for comparison. Selected video clips were exported from EditDV to Photoshop as image sequences, which were subsequently deinterlaced, sorted and inserted into the same picture for comparison of sound and video tracks. Since there exists a systematic delay between both tracks we had to shift the video track backwards for 21-23 ms. After this correction we analysed our video recordings of Cicadatra and Pagiphora specimens to associate sounds with specific wing movements. From this information we made conclusions about the sound producing mechanism. Nevertheless, some additional experiments with other methods are needed to prove the conclusions.

As in many other anurans, breeding males of the European green toad (Bufo viridis complex) produce two different types of acoustic signals: the advertisement call and the release call. The advertisement call is a long-range vocalisation that elicits positive phonotactic responses in receptive females and either aggressive or negative phonotactic responses in conspecific males. The release call is a short-range vocalisation and it is produced in two different contexts: by a paired male during his attempts to kick away a competitor; and by a single male when he is clasped by a second individual as a consequence of a fight or of an erroneous mating attempt. In the present work we analyse the release and advertisement calls of 222 individuals recorded from 22 populations distributed over a large portion of the green toad range (from the Italian peninsula, Corsica and Sardinia southward to Morocco and Israel, and eastward to Kyrgyzstan and Kazakhstan). We address the question of the relationships between function and structure of signals and we ask to what extent differences in function might explain the differences in signal structure and their pattern of variation. We compare the two types of calls on the basis of their acoustic properties and on the basis of the variation that these properties show at individual and populational levels. The advertisement and release calls differ markedly with respect to most of their acoustic properties (call duration, intensity, number and duration of pulses, fundamental frequency) and these differences might be often explained as the effect of their different functions. Despite these differences, however, the morpho-physiological constraints (temperature and body size) affect the signal temporal and spectral properties in a similar way, producing qualitatively consistent patterns of variation.

Food begging calls are highly stereotyped innate vocalisations whose structure is very similar in quite different species. It usually consists of a series of homologous harmonic pulses. To understand species-specificity and relationship to the adult sounds, using Canary ,software we recorded and processed food begging calls from five different species of African Poicephalus parrots pertaining to two different superspecies (P.  gulielmi, supersp. robustus), and P.  senegalus, P. meyeri, P. rufiventris, P. cryptoxanthus, supersp. meyeri) from birth to 60 days of age. As the food begging call is an innate vocalisation type, it also qualifies as a useful taxonomic character. In fact, it is evolutionarily slow to change, apparently because it is unlikely to be affected by those external pressures that more quickly shape morphological characters. The general principle is that in all Poicephalus species studied, the call is structured in a series of harmonic pulses. These increase in number in the first four weeks of age, then decrease. However, at all times, a clear-cut difference among all the species is maintained up to the tenth week of age, when the differences disappear and the pulse number is two for all species. For all species studied, the main specific character appears to be the utterance rate, that is the pause duration between the different pulses. By applying a correlator analysis routine to the different species' spectrograms and subsequently a multidimensional scaling to correlation values, it clearly appears that all species' parameters segregate neatly. Within the superspecies meyeri, it also appears that P. meyeri and P. senegalus are phylogenetically the most apart, bioacoustically confirming recent molecular results.

Male territorial songbirds interact with each other via full song over long distances. The relative timing of strophes in these singing interactions conveys information about male dominance with overlapping as a dominant signal indicating willingness to escalate. Other birds may eavesdrop on such interactions to assess future opponents. This requires that individuals at a distance perceive the intended time pattern of the interaction despite attenuation and masking effects. Particularly during the dawn chorus, song of individual males is masked by vocalisations of other birds of the same and other species. We here report on a dual-speaker playback experiment in which we simulate alternating, overlapping and random patterns of singing interactions between male blue tits in a natural habitat. We record these 'interactions' and the vocalisations masking them with an array of four microphones (Acoustic Location System, ALS) representing eavesdroppers. The ALS allows the location of the masking birds. Masking in time and frequency range as well as signal-to-noise ratio between 'interactants' and masking birds are taken into consideration in the analysis. We expect the 'eavesdroppers' to perceive the temporal pattern of the interaction differently, depending on their location relative to the 'interactants' and the masking they experience. This study gives insight into how masking may constrain the perception of time patterns in singing interactions that vary considerably and therefore cannot be reconstructed from any internal representation.

Flocking is the natural social structure of the budgerigar Melopsittacus undulatus. Its repertoire is well known and includes a number of acoustically and functionally distinct calls (contact call, alarm call, etc.). The social context plays a large role in the learning of these calls. Our aim in this study was to determine if the social context (presence or lack of a partner) may modulate the significance attached to a signal and thus modify the behavioural responses to a conspecific call. To investigate this, unmated paired birds and temporarily isolated ones were submitted to the playback of a specific alarm call. The tested sessions were videotaped and then analysed frame by frame. Multiple qualitative response measures were used and quantified. They included both vocalisations of subjects in response to the sound playback and a graded series of postures, ranging from a variation of orientation relative to the sound speaker location, up to alert behaviour and flight. Our results clearly indicate a behavioural plasticity of budgerigars according to the social context. The most prominent difference between isolated and paired birds is that the former are significantly more likely to give an acoustic response while the latter show an increased motricity and even an escape flight. Isolated birds give repeated calls in response to the playback signal, as they would maintain contact with the flock. At this level, paired birds are quite unresponsive. The modulation of the behavioural responses observed in this species indicates that the meaning of a signal is related to the social context.

The aim of this paper is to describe the seasonal variation in responses by Green Woodpecker Picus viridis, Great Spotted Woodpecker Dendrocopos major, Nuthatch Sitta europaea and Short-toed Treecreeper Certhia brachydactyla, to playback. The study was carried out in the province of Bologna, at altitudes ranging between 100 and 1,100 m a.s.l.. Between February and May 1996 through 1998 1,299 field playback experiments were carried out in 419 series of 1-4 species, the order of which was chosen at random. As regards D. major, we stimulated it first with drumming and then with vocalisations. Drumming and calls were played for two minutes from a tape recorder (100, with a control period of three minutes. Three out of four species showed a marked seasonality in their response to playback: P. viridis Chi-squared = 39.07, df = 9, P < 0.0001; D. major Chi-squared = 42.34, df = 9, P < 0.00001; S. europaea Chi-squared = 33.41, df = 9, P < O.OOI. An analysis of the individual components of Chi-squared shows that: in March P. viridis responds more frequently than expected (highest value in the second 10-day period), while in May it responds less frequently than expected. D. major responds more than expected in the third 10-day period of February, in March and in the first and second 10-day periods of April (highest value in the second 10-day period of March) whereas less than expected in May. S. europaea responds more often than expected in the third 10-day period of February and in the first 10-day period of March and less in May. As far as C. brachydactyla is concerned, there is no significant link between response and season, probably due to the longer breeding period (2 clutches per year).

According to the signature whistle hypothesis, bottlenose dolphins Tursiops truncatus emit individually sequential stereotyped whistles, which may function to broadcast the whistler's identity and location. Such whistles have been found in other dolphin species repertories. 12 hours of sound recordings of tucuxi dolphins Sotalia fluviatilis from Sepetiba Bay were analysed with the aim to detect such signature whistles. All of the whistle sonograms in which contours were stereotyped and sequential were selected and classified in "Types'' by means of two different methods: Visual Observation (VO) and a quantitative method named Contour Similarity technique (CS). 202 whistle sequences were chosen and classified in 27 Types by means of the OV method. Although the individuality of the Types was not demonstrated, the possibility that they might really represent signature whistles is reinforced by the following factors: the majority of the Types were formed of whistle sequences quite close to each other; they were mainly emitted during feeding and travelling behaviours; and were in accord with the general species repertoire. The CS technique did not correlate with the VO method in the whistles classification and it does not seem to be useful for the S. fluviatilis repertoire. I also selected 19 whistles sequences formed from two simultaneous components which resemble a kind of vocal interaction between dolphins known in the literature as "duet''

Acoustical communication plays one of the crucial roles in functioning of white whale reproductive gathering. However its mechanisms have not been clearly understood so far. This study was conducted during the summer season in 1999 at a beluga reproductive gathering off cape Beluzjy, Great Solovetsky Island, the White Sea. The sound recordings were made using a Sony MZ-R55 MD-recorder with frequency response from 20 to 20,000 Hz ± 3 dB. Processing and analysis of the acoustical stream were conducted using Cool Edit Pro 1.0 software. Two behavioural situations were analysed (83 minutes, 2771 signals): group-resting (24 min., 958 signals) and sexual behaviour (59 min., 1813 signals). The average vocalisation rate during group-resting was 39.4 signals/min and 5.3 signals per whale/min. The following proportions of the main types of vocalisations were: "squeak'' (24%), "U'' (19%), "whistles'' (11%), "vowels'' (7%), "chirrup'' (6.6%), "scream'' (6.2%), "bleating'' (6%), "creak'' (5.9). During sexual behaviour, the mean vocalisation rate was 30.7 signals/min and 1.7 signals per whale/min; these values are significantly lower than the statistical means 53.7 signals/min and 3.6 signals per whale/ min. The following types of signals predominated during sexual behaviour: "squeak'' (24.4%), "vowels'' (10.8%), "chirrup'' (10%), "bleating" (8.6%), "whistles'' (6.5%), "creak" (5.6%). Thus it was found that during sexual behaviour the average number of signal/min decreased and group composition and structure of the main types of vocalisation were changed due to a decrease in proportions of "U", "scream'' and "whistles'' types of vocalisations. Time analysis showed that an intensification of the sexual interactions was accompanied by decreases in total signal production. The bursts (peaks) of vocal activity preceded behavioural peaks rather than accompanied them. Thus, our results confirmed the important role of the sound communication to maintain group unity and co-ordinate activities of beluga whales.