CLASSIFICATION OF DIVERSE CALL TYPES USING CLUSTER ANALYSIS TECHNIQUES

ABSTRACT

We investigated the problem of categorizing the repertoire of a group of highly varied vocalizations. A set of Weddell seal Leptonychotes weddellim-air calls recorded near Davis, Antarctica, were examined. The repertoire size was estimated by first subjectively assigning each call (based on auditory and spectrographic patterns) to one of a large number of provisional call types. A set of frequency, duration and waveform measurements were made on every call. For each provisional call type, the mean value of each of these measures was calculated. These mean values were used to perform Cluster Analyses of the provisional call types. Beginning with calls clustered closest together, the most similar provisional call types were successively amalgamated until further joining would link two clearly dissimilar vocalizations (rising versus falling frequency sweeps in this case). The Weddell seal repertoire contained 12 call types ranging from long sinusoidal upsweeps to growls. This procedure provides a method of estimating the minimum repertoire size of a sample of calls. Dichotomous Sorting using Principal Components Analysis can also be used to categorize calls but will probably be of greater value when applied to finding subdivisions within a single call type.     

Variations in underwater vocalizations of Weddell seals (Leptonychotes weddelli) at the Vestfold Hills as a measure of breeding population discreteness

Weddell seal vocalizations from Davis Station showed similarities to those from McMurdo Sound and Palmer Peninsula, but none were identical. One vocalization, DD1, was unique to Davis Station. At all sites trills, or territorial defense calls, were the most common and had more types than other calls. Chugs, an aggressive sound, were common at all sites. Weddell seal calls from Davis Station showed similarities to sounds from McMurdo by sharing the use of prefixes and suffixes. Sounds from Davis Station shared the use of both ascending and descending trills and whistles with calls from Palmer Peninsula.Commonalities in underwater vocalizations may indicate the degree of mixing between breeding populations, in which case, Weddell seals in the Davis area probably are from a breeding population distinct from those at either McMurdo Sound or Palmer Peninsula.     

Weddell seals do not lengthen calls in response to conspecific masking

Some mammalian and avian species alter their vocal communication signals to reduce masking by background noises (including conspecific calls). A preliminary study suggested that Weddell seals (Leptonychotes weddellii) increase the durations of some underwater call types when overlapped by another calling seal. The present study examined the durations and overlapping sequences of Weddell seal calls recorded in Eastern Antarctica. The calling rate, call type (13 major categories), total duration, numbers of elements per call and overlapping order of 100–200 consecutive calls per recording location were measured. In response to increased conspecific calling rates, the call durations and numbers of elements (within repeated-element call types) did not change or became shorter. Calls that were not overlapped were 3.8?±?6.1 s long, the first call in a series of overlapped calls was 14.4?±?15.7 s and subsequent calls in an overlapping series were 6.5?±?10.3 s. The mean durations of non-overlapped and overlapped calls matched random distributions. Weddell seals do not appear to be adjusting the durations or timing of their calls to purposefully avoid masking each other’s calls. The longer a call is, the more likely it is to overlap another call by chance. An implication of this is that Weddell seals may not have the behavioural flexibility to reduce masking by altering the temporal aspects of their calls or calling behaviours as background noises (natural and from shipping) increase.     

Weddell seal in-air call sequences made with closed mouths

Weddell seals (Leptonychotes weddelli) near Davis, Antarctica, produce a number of in-air vocalizations during the breeding season. With mouth and nostrils closed, pups and adults of both sexes produce at least 8 call types. Many of these are similar in nature to the sounds made underwater. Calls range from long, high frequency (>5 kHz) whistles to short, low frequency (<0.2 kHz) grunts. Individual call elements are often repeated and up to 6 call types are strung together in highly variable sequences.     

Examining harp seal behavioural patterns via their underwater vocalizations

Recordings of harp seal (Pagophilus groenlandicus) underwater vocalizations, obtained at 3-h intervals over 7 days during the breeding season, were examined with regard to the relative occurrence of high frequency calls. An increase of high frequency calls at mid-March is possibly associated with the onset of courtship and mating. A daily pattern indicated that the seals are vocally active all night and least active at daybreak and in the early afternoon. The inactive afternoon period coincided with the time that the greatest number of seals were resting on the ice. Remote and/or automated monitoring of vocalizations may provide a means of examining various underwater activities of marine mammals.     

Gut length, food transit time and diving habit in phocid seals

The southern elephant seal (Mirounga leonina) has the ability to dive for 2 h and reach depths of 1200 m. This creature is also exceptional in having a small intestine that is 25 times body length. Krockenberger and Bryden advanced the hypothesis that the long small intestine has developed to compensate for the extended periods with reduced or even abolished intestinal blood perfusion during diving. To test this hypothesis we have measured small-intestinal lengths in crabeater (Lobodon carcinophagus), Weddell (Leptonychotes weddellii), Ross (Ommatophoca rossi), leopard (Hydrurga leptonyx), harp (Phoca groenlandica), ringed (Phoca hispida) and hooded (Cystophora cristata) seals and related them to available data on their maximal dive duration. We found no significant correlation (P > 0.05) between intestinal length relative to body length and diving ability, but we found that small-intestinal internal area was significantly (P < 0.05) related to body length. A crude scanning electron microscopical examination of the small intestines of Weddell, crabeater, hooded and harp seals failed to reveal any gross anatomical differences between small-intestinal surfaces. This suggests that gut dimension in this variety of phocid species with widely differing diving ability is not related to diving habit, but is instead related to body size. The transit time of digesta was determined in two 1-year-old harp seals by use of radiopaque polyethylene rings of 4-mm diameter followed by X-ray examination, as markers for the solid phase passage, and chromium ethylene-diaminetetra acetic acid (Cr-EDTA) as a marker for the liquid phase. The transit time for the Cr-EDTA marker was 6.9 h ± 0.5 SE (range 4.5–8 h, n= 7), while 80% of the polyethylene markers appeared in the colon after 17.6 h ± 1.0 SE (range 14–21.5 h, n= 6) and were sometimes retained in the colon for several hours before defecation. These transit times did not change significantly (P > 0.05) in response to repetitive diving over a period of 8 h. This indicates that the often-used Cr-EDTA is not a good measure for digesta passage time when used alone in seals, and that the hypothesis of Krockenberger and Bryden is most likely wrong. Received: 17 December 1997 / Accepted: 4 May 1998     

Geographic variation of Weddell seal (<Emphasis Type="Italic">Leptonychotes weddellii</Emphasis>) airborne mother–pup vocalisations

Geographic variation of vocal behaviour may be reflective of genetic, cultural and ecological differences between locations. Such differences have been previously reported in six pinnipeds, including Weddell seals (Leptonychotes weddellii). This study investigated whether the acoustic characteristics of airborne Weddell seal mother and pup ‘contact’ calls vary geographically between the Vestfold Hills and McMurdo Sound. From each location, 72 mother and 120 pup calls were analysed, with 11 acoustic characteristics measured. MANOVA and Discriminant function analysis (DFA) identified significant differences in several fundamental frequency characteristics. The cross-validated DFA classified greater than 60% of mother and pup vocalisations to the correct location. For novel calls, the probabilities of correctly classifying these percentages by chance are low. Additionally, ordinal logistic regression identified significant differences in the maximum energy distribution of mother and pup vocalisations. For pups, this variation is likely the result of genetic differences between the two populations. Nevertheless, results support suggestions that, when compared with underwater repertoire differences, airborne within-call differences are less efficient at identifying Weddell seal breeding populations.     

Through-ice communication by Weddell seals may not be practicable

Possible communication between territorial male Weddell seals (Leptonychotes weddellii) under the ice with females on the ice was investigated. In-air and underwater recordings of underwater calls were made at three locations near Davis, Antarctica. Most underwater calls were not detectable in air, often because of wind noise. In-air call amplitudes of detectable calls ranged from 32–74 dB re. 20 Pa at 86 Hz down to 4–38 dB re. 20 Pa at 3.6 kHz. Most of these would be audible to humans. Only 26 of 582 amplitude measurements (from 230 calls) ranged from 5 dB to a maximum of 15 dB above the minimum harbour-seal (Phoca vitulina) in-air detection threshold. Seals on the ice could likely hear a few very loud underwater calls but only if the caller was nearby and there were no wind noises. The low detectability of underwater calls in air likely precludes effective communication between underwater seals and those on the ice.     

Comparing marine mammal acoustic habitats in Atlantic and Pacific sectors of the High Arctic: year-long records from Fram Strait and the Chukchi Plateau

During the International Polar Year (IPY), acoustic recorders were deployed on oceanographic moorings in Fram Strait and on the Chukchi Plateau, representing the first coordinated year-round sampling of underwater acoustic habitats at two sites in the High Arctic. Examination of species-specific marine mammal calls recorded from autumn 2008–2009 revealed distinctly different acoustic habitats at each site. Overall, the Fram Strait site was acoustically complex compared with the Chukchi Plateau site. In Fram Strait, calls from bowhead whales (Balaena mysticetus) and a variety of toothed whales (odontocetes) were recorded year-round, as were airgun pulses from seismic surveys. In addition, calls from blue whales (Balaenoptera musculus) and fin whales (B. physalus) were recorded from June to October and August to March, respectively. Conversely, at the Chukchi Plateau site, beluga (Delphinapterus leucas) and bowhead whale calls were recorded primarily from May to August, with airgun signals detected only in September–October. Ribbon seal (Phoca fasciata) calls were detected in October–November, with no marine mammals calls at all recorded from December to February. Of note, ice-adapted bearded seals (Erignathus barbatus) were recorded at both sites, primarily in spring and summer, corresponding with the mating season for that species. Differences in acoustic habitats between the two sites were related to contrasts in sea ice cover, temperature, patterns of ocean circulation and contributions from anthropogenic noise sources. These data provide a provisional baseline for the comparison of underwater acoustic habitats between Pacific and Atlantic sectors of the High Arctic.     

Individual variation in pup vocalizations and absence of behavioral signs of maternal vocal recognition in Weddell seals (Leptonychotes weddellii)

Individually stereotyped vocalizations often play an important role in relocation of offspring in gregarious breeders. In phocids, mothers often alternate between foraging at sea and attending their pup. Pup calls are individually distinctive in various phocid species. However, experimental evidence for maternal recognition is rare. In this study, we recorded Weddell seal (Leptonychotes weddellii) pup vocalizations at two whelping patches in Atka Bay, Antarctica, and explored individual vocal variation based on eight vocal parameters. Overall, 58% of calls were correctly classified according to individual. For males (n= 12) and females (n= 9), respectively, nine and seven individuals were correctly identified based on vocal parameters. To investigate whether mothers respond differently to calls of familiar vs. unfamiliar pups, we conducted playback experiments with 21 mothers. Maternal responses did not differ between playbacks of own, familiar, and unfamiliar pup calls. We suggest that Weddell seal pup calls may need to contain only a critical amount of individually distinct information because mothers and pups use a combination of sensory modalities for identification. However, it cannot be excluded that pup developmental factors and differing environmental factors between colonies affect pup acoustic behavior and the role of acoustic cues in the relocation process.     

Geographic variations in underwater male Weddell seal Trills suggest breeding area fidelity

Adult Weddell seals (Leptonychotes weddellii) exhibit site fidelity to where they first breed but juveniles, and perhaps transient adult males, may disperse from their natal location. If there is mixing between adjacent breeding groups, we would expect that common vocalizations would exhibit clinal patterns. Underwater Trill vocalizations of male Weddell seals at Mawson, Davis, Casey, McMurdo Sound, Neumayer and Drescher Inlet separated by ca. 500 to >9,000 km, were examined for evidence of clinal variation. Trills are only emitted by males and have a known territorial defense function. Trills from Davis and Mawson, ca. 630 km apart, were distinct from each other and exhibited the greatest number of unique frequency contour patterns. The acoustic features (duration, waveform, frequency contour) of Trills from Neumayer and Drescher Inlet, ca. 500 km apart, were more distinct from each other than they were from the other four locations. General Discriminant Analysis and Classification Tree Analysis correctly classified 65.8 and 76.9% of the Trills to the correct location. The classification errors assigned more locations to sites >630 km away than to nearest neighbours. Weddell seal Trills exhibit geographic variation but there is no evidence of a clinal pattern. This suggests that males remain close to single breeding areas throughout their lifetime.     

Grouping in auditory temporal perception and vocal production is mutually adapted: the case of wriggling calls of mice

Auditory Gestalt perception by grouping of species-specific vocalizations to a perceptual stream with a defined meaning is typical for human speech perception but has not been studied in non-human mammals so far. Here we use synthesized models of vocalizations (series of wriggling calls) of mouse pups (Mus domesticus) and show that their mothers perceive the call series as a meaningful Gestalt for the release of instinctive maternal behavior, if the inter-call intervals have durations of 100–400 ms. Shorter or longer inter-call intervals significantly reduce the maternal responsiveness. We also show that series of natural wriggling calls have inter-call intervals mainly in the range of 100–400 ms. Thus, series of natural wriggling calls of pups match the time-domain auditory filters of their mothers in order to be optimally perceived and recognized. A similar time window exists for the production of human speech and the perception of series of sounds by humans. Neural mechanisms for setting the boundaries of the time window are discussed.     

Dual Function of Chip Calls Depending on Changing Call Rate Related to Risk Level in Territorial Pairs of White‐Eared Ground‐Sparrows

Different mechanisms have been proposed for encoding information into vocalizations: variation of frequency or temporal characteristics, variation in the rate of vocalization production, and use of different vocalization types. We analyze the effect of rate variation on the dual function of chip calls (contact and alarm) produced by White‐eared Ground‐sparrows (Melozone leucotis). We conducted an acoustic playback experiment where we played back 1 min of four chip call rates (12, 36, 60, 84 calls/min). We measured the response of territorial pairs using behavioral responses, and fine structural features of calls produced in response to those playbacks. White‐eared Ground‐sparrows showed more intense behavioral responses to higher than lower call rate playbacks. Both individuals of the pair approached the source of the playback stimulus faster, produced the first vocalization faster, produced more vocalizations, and spent more time close to the stimulus in higher call rate than in lower call rate playbacks. Frequency and duration characteristics of calls (chip and tseet) were similar in response to all call rate playbacks. Our playback experiment elicited different intensity of behavioral responses, suggesting that risk‐based information is encoded in call rate. Our results suggest that variation in the rate of chip call production serves a dual function in this species; calls are used at lower rates for pair contact and at higher rates for alarm/mobbing signals.     

Cardiorespiratory pattern of rest-associated apnea in a Weddell seal: a case study at an ice hole in Antarctica

Seals are known to be periodic breathers with eupneic and apneic phases at rest. We video-recorded a resting Weddell seal (Leptonychotes weddellii) that appeared head up from an ice hole in McMurdo Sound, Antarctica. From the video recorded, the duration of each eupneic and apneic phase was extracted, and heart rate in apneic phase was obtained by counting ripples of the sea surface around the neck of a seal. The results indicated that the distribution of instantaneous heart rate was bimodal. In addition, the higher the apneic heart rate, the shorter the apneic duration as well as the combination of apneic duration and successive eupneic duration (A–E duration). From these analyses, we characterized the periodic breathing of a resting Weddell seal.     

Annual and diurnal variations in the underwater vocalizations of Weddell seals

Summary Weddell seal vocalizations were recorded once a fortnight throughout 1984 at a major pupping site near Davis, Antarctica. Few vocalizations were heard between January and June. The number of vocalizations recorded increased from July to a peak in November and early December during the breeding period and then fell off rapidly through December. The increase in the number of seal vocalizations was correlated with a build up in seal numbers but some of the increase may be explained by an increased rate of vocalization by males.     

Advertisement calls of two anuran amphibians,Rana tigrina andTomopterna breviceps

Rana tigrina andTomopterna breviceps occur as sympatric species at Dharwad, India. Sexually mature males produce advertisement calls. The advertisement call of both the species consist of a number of calls produced in series forming a call group. Each call group ofRana tigrina comprises 10–40 calls, whereas that ofTomopterna breviceps consists 13–141 calls. Each call consists of a pulse group with variable number of pulses which lack pulse interval. Calls of both the species exhibit similarities in (i) call consisting of series of calls with a pulse group in each call, (ii) absence of pulse interval within the pulse group, (iii) the amplitude of the first pulse being always small, and (iv) the frequency spectrum beginning from 200 Hz. Based on the similarities in the spectral features of the calls, it is suggested that the two species may be closely related to each other.     

Experimental evidence of seawater drinking in juvenile hooded (Cystophora cristata) and harp seals (Phoca groenlandica)

This study was undertaken to measure whether young harp seals (Phoca groenlandica) and hooded seals (Cystophora cristata) drink seawater and, if so, to investigate how the excess salt load is handled. Blood and urine samples were collected from hooded seal pups (n=3) and harp seal pups (n=3) after 2 weeks of freshwater exposure, at intervals during 3 weeks of seawater exposure and, finally, after 2 weeks of re-exposure to fresh water. Total water turnover, as measured by injection of tritiated water, was 2200 ml · day⁻¹ and 3300 ml · day⁻¹ in hooded seals and harp seals, respectively. The extent of mariposia was taken as the difference between total water turnover and influx of water through food (free and metabolic water) and respiratory water exchange. Seawater drinking amounted to 14% and 27% of total water turnover (r_H2O) for the hooded seals and harp seals, respectively. Further evidence of mariposia was obtained from an increase in the excretion rate of the urine osmolytes Na⁺, Cl⁻ and Mg²⁺, during the period of seawater exposure. It is concluded that water influx due to seawater drinking can not be excluded as a source of error when estimating food consumption of free-ranging harp seals and hooded seals, by use of labeled water techniques. Accepted: 11 May 2000     

Communication inGalago crassicaudatus

The olfactory, visual and, auditory communicative behaviors of a colony of feralGalago crassicaudatus (four males ad three females) observed over a two year period are documented. The enclosure was constructed to simulate a natural environment, but all stimulus possibilities available in the wild were not present and several new ones were introduced. It is not known how many of such differences may have affected the animals' behavior, however not all behaviors observed in the wild appear in the lab. The characteristics of several calls are described in detail and illustrated with sonagrams. These include an agonistic call, a sexual call, a possible alarm or mobbing call, a possible territorial or spacing call and a “low grunting sound” which appears in contexts of sexual and territorial calls. Compared to “higher” primates these galagos exhibit a small number of vocalizations. The communicative system appears directly correlated to social organization.     

Variation in and Meaning of Alarm Calls in a Social Desert Rodent Rhombomys opimus

The great gerbil (Rhombomys opimus), a social rodent that lives in family groups, emits three different alarm vocalizations in the presence of predators: a rhythmic call; a faster more intense call; and a single whistle. We tested the hypothesis that the alarm calls communicate risk of predation. We quantified the relationship between predator distance and type of alarm call via human approaches to gerbils. We also tested responses of focal adults in family groups to playback broadcasts of the different calls and controls of bird song and tape noise. Results showed that alarm calls were related to distance from a predator. Gerbils gave the rhythmic call when the predator was farthest away, the more intense call as the predator moved closer; and a short whistle when startled by a close approach of the predator. Gerbils stopped feeding and stood vigilant in a frozen alert posture in response to playbacks of all three alarm calls. They decreased non‐vigilant behavior to the alarm vocalizations more than to the controls and decreased non‐vigilant behavior significantly more in response to the intense alarm and whistle compared with the rhythmic alarm. We conclude that one function of gerbil alarm calls is to communicate response urgency to family members. The rhythmic alarm communicates danger at a distance, whereas the intense alarm and whistle signal the close approach of a predator.     

Satellite imagery can be used to detect variation in abundance of Weddell seals (Leptonychotes weddellii) in Erebus Bay, Antarctica

The Weddell seal population in Erebus Bay, Antarctica, represents one of the best-studied marine mammal populations in the world, providing an ideal test for the efficacy of satellite imagery to inform about seal abundance and population trends. Using high-resolution (0.6 m) satellite imagery, we compared counts from imagery to ground counts of adult Weddell seals and determined temporal trends in Erebus Bay during November 2004–2006 and 2009, and December 2007. Seals were counted from QuickBird-2 and WorldView-1 images, and these counts were compared with ground counts at overlapping locations within Erebus Bay during the same time. Counts were compared across years and within individual haul-out locations. We counted a total of 1,000 adult Weddell seals from five images across all years (for a total of 21 satellite-to-ground count comparisons), approximately 72% of the total counted on the ground at overlapping locations. We accurately detected an increase in abundance during 2004–2009. There was a strong, positive correlation (r = 0.98, df = 3, P < 0.003) between ground counts and counts derived from the imagery. The correlation between counts at individual haul-out locations was also strong (r = 0.80, df = 19, P < 0.001). Detection rates ranged from 30 to 88%. Overall, our results showed the utility of high-resolution imagery to provide an accurate way to detect the presence and variation in abundance of Weddell seals. Our methods may be applied to other species in polar regions, such as walruses or polar bears, particularly in areas where little is known about population status.