Mother–pup vocal recognition in harbour seals: influence of maternal behaviour, pup voice and habitat sound properties

Animal Behaviour 105 (2015) 109e120

Contents lists available at ScienceDirect

Animal Behaviour journal homepage: www.elsevier.com/locate/anbehav

Motherepup vocal recognition in harbour seals: influence of maternal behaviour, pup voice and habitat sound properties
a, b, *, Gwe
nae €l Beauplet a, b, 1, Mike O. Hammill b, c, 2, Caroline C. Sauve Isabelle Charrier d, e, 3 D
epartement de Biologie, Universit
e Laval, Qu
ebec, QC, Canada Qu
ebec-Oc
ean, Qu
ebec, QC, Canada c Maurice Lamontagne Institute, Fisheries and Oceans Canada, Mont-Joli, QC, Canada d Universit
e Paris-Sud, Centre Neurosciences Paris Sud, UMR 8195, Paris, France e Centre national de la recherche scientifique (CNRS), Paris, France a

b

a r t i c l e i n f o Article history: Received 14 November 2014 Initial acceptance 6 January 2015 Final acceptance 10 March 2015 Available online 16 May 2015 MS. number: A14-00926R Keywords: harbour seal individual signature Phoca vitulina playback experiment propagation test vocal recognition

Motherepup vocal recognition abilities in pinnipeds reflect maternal reproductive strategies. In otariids, motherepup pairs are frequently separated during lactation, pups are highly mobile at an early stage, and the high densities of colonies increase the risks of confusion between individuals. Accordingly, vocal recognition between mothers and pups is well developed in this group. In contrast, among phocids, the young are less mobile and mothers normally stay with their pups throughout lactation. Hence, the risks of confusion between individuals are relatively low and mothereyoung vocal recognition abilities are less developed in this family. Harbour seals, Phoca vitulina, are unique among phocids as females forage during lactation and pups are highly mobile throughout the 21e42-day nursing period. We performed playback experiments on 18 breeding female harbour seals to assess their abilities to recognize the calls of their pup and to evaluate the effect of maternal protectiveness and pup vocal stereotypy on the recognition process. In addition, we performed propagation tests to assess pup call propagation efficiency in the environment. As early as 3 days after birth, females were more responsive to calls of their own pup than to calls of nonfilial pups. Female responses also varied depending on their protective behaviour displayed towards their pup. Vocal discrimination abilities were positively correlated with the level of individuality conveyed in pup calls. Considerable differences in the propagation efficiency of pup calls were measured among sites used for motherepup reunions. This study provides experimental evidence that wild harbour seal females recognize the calls of their pup among others, and that they memorize previous versions of their pup's calls, taking into account age- and size-related effects on vocalizations. Propagation tests further suggest that some motherepup reunion sites are better suited for vocal recognition than others. © 2015 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.

Parenteoffspring recognition is common in species providing directed parental care. While visual and olfactory cues are used for egg recognition in the nest or at close range for the confirmation of offspring identity (e.g. Helantera, Martin, & Ratnieks, 2014; Pike, 2011; Pitcher, Harcourt, Schaal, & Charrier, 2011), vocal communication generally evolves in situations requiring identity signals that are effective over long distances. The complexity of


, CNPS-CNRS UMR 8195, Universite
Paris Sud, bat. * Correspondence: C. C. Sauve 446, 91405 Orsay cedex, Paris, France.
). E-mail address: [email protected] (C. C. Sauve 1 E-mail address: [email protected] (G. Beauplet). 2 E-mail address: [email protected] (M. O. Hammill). 3 E-mail address: [email protected] (I. Charrier).

parenteoffspring vocal recognition systems is generally associated with the difficulties encountered by the pair to maintain contact and to find each other during the rearing period. Accordingly, mutual parenteyoung recognition is demonstrated in species characterized by high population densities at breeding sites (Aubin & Jouventin, 2002; Trillmich, 1981), frequent parenteoffspring separations (Aubin & Jouventin, 2002; Briefer & McElligott, 2011) and mobile offspring (Jones, Falls, & Gaston, 1987; Sayigh et al., 1999). Comparative studies of parenteoffspring vocal recognition abilities thus provide valuable insights into the selective pressures shaping communication systems. Moreover, it allows for a greater understanding of the role of vocal recognition in growth and survival of the young as well as in adult reproductive success.

http://dx.doi.org/10.1016/j.anbehav.2015.04.011 0003-3472/© 2015 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.

110

C. C. Sauv
e et al. / Animal Behaviour 105 (2015) 109e120

Pinnipeds have a great potential for comparative studies of parenteoffspring recognition due to their established phylogeny (Nyakatura & Bininda-Emonds, 2012) and contrasting maternal strategies observed among families (Boness & Bowen, 1996). Otariids alternate between nursing episodes on land and foraging trips at sea, leading to frequent and extended maternal absences from the breeding colony (Schulz & Bowen, 2004). During maternal absences, otariid pups generally gather at the periphery of the colony to rest and play together (Bowen, 1991), and are thus relatively mobile and active in the course of the long lactation period (4e18 months; Schulz & Bowen, 2004). Walrus, Odobenus rosmarus (Odobenidae), females also forage during lactation, and calves are highly mobile, following their mother at sea during foraging trips (Kovacs & Lavigne, 1992; Stewart & Fay, 2001). Extended motherecalf separations are therefore infrequent throughout the long lactation period (18e27 months; Fisher & Stewart, 1997), although short separations occur during female feeding events and predator attacks. The mothereoffspring bond in this species is particularly strong, sometimes persisting several years after weaning (Knudtson, 1998; Nowak, 2003). As for phocids, females usually fast alongside their pup throughout the relatively short nursing period (4e60 days; Schulz & Bowen, 2004). They rely on fat reserves stored before parturition to produce the highly energetic milk necessary for successful rearing of their pup (Boness & Bowen, 1996). Phocid pups are mostly sedentary, rarely undertaking movements within the breeding colony (Bowen, 1991). These differences in maternal strategies and pup mobility reflect contrasting selective pressures for the evolution of motherepup vocal recognition systems among the families of pinnipeds. Accordingly, evidence suggests that mutual motherepup vocal recognition is the norm in otariids and walruses whereas recognition abilities in phocids are more variable (total absence or unidirectional; Charrier, Aubin, & Mathevon, 2010; Insley, Phillips, & Charrier, 2003). In spite of the phylogenetic dichotomy regarding pinniped maternal strategies, breeding females of a few phocid species forage during the lactation period (Boness, Bowen, & Oftedal, 1994; Hammill, Lydersen, Ryg, & Smith, 1991; Lydersen & Kovacs, 1993; Wheatley, Bradshaw, Harcourt, & Hindell, 2008). Harbour seal, Phoca vitulina, females perform short foraging trips after 1 week postpartum, and spend most of their time at sea during the rearing period (Boness et al., 1994; Bowen, Boness, & Iverson, 1999). Harbour seal pups are also exceptional among phocids because they are highly active and mobile in the water throughout lactation (Bigg, 1981; Bowen et al., 1999; Jorgensen, Lydersen, Brix, & Kovacs, 2001), which lasts between 21 and 42 days (Riedman, 1990). These traits imply greater selection pressures for an effective motherepup vocal recognition system in harbour seals compared to most phocids. Harbour seal pups frequently utter calls that disappear from their vocal repertoire upon weaning (Renouf, 1984), suggesting their essential role is in maintaining contact with the mother. Previous work has shown that harbour seal and northern elephant seal, Mirounga angustirostris, vocalizations convey individual stereotypy levels intermediate between those found in otariids and walruses compared with other phocids (Charrier et al., 2010; Insley,
, Beauplet, Hammill, & Charrier, in press). However, 1992; Sauve although vocal stereotypy is a prerequisite for individual recognition, it does not imply that females can perceive and use such information to identify their offspring (Beecher, 1982). A captive harbour seal female was successfully trained to discriminate between single calls from nonfilial pup pairs (Renouf, 1985), confirming that females have the sensory and neurological capacities to perceive differences between pup vocalizations. However, as the broadcast stimuli in this experiment consisted of a single repeated call from each pup, it could not be established whether the female

relied on interindividual or intervocalization differences when discriminating the calls (pseudoreplication; Kroodsma, 1989) and, thus, whether interpup call variability is sufficient for harbour seal females to recognize the calls of their offspring. Moreover, the fact that a female can learn to discriminate between calls from nonfilial pups does not imply that such ability is effectively used in a natural context (notion of generalization; McGregor et al., 1992). Additional research is therefore needed to determine unambiguously whether harbour seal females are capable of vocal recognition of their offspring during lactation. Motherepup recognition processes can be affected by numerous maternal and offspring characteristics, as well as environmental factors. Interindividual differences in maternal behaviour (e.g. the extent to which females protect their young, by monitoring and limiting their movements) have been described in various mammals (e.g. Champagne, Francis, Mar, & Meaney, 2003; Dwyer & Lawrence, 2000; Fairbanks, 1996). These differences in maternal behaviour, associated with female experience and temperament (Fairbanks, 1996; Maestripieri, 1993; Plush, Hebart, Brien, & Hynd, 2011), are likely to influence female motivation to reunite with their offspring, and therefore the attention they pay to identity signals produced by young. The quality of recognition signals might also affect vocal recognition of pups by females. Pups with stable (i.e. low intraindividual variability) and distinctive (important deviation from interindividual mean) calls are more readily discriminated. In this
et al. (in press) found that individual vocal stereotypy regard, Sauve
of harbour seal pup calls differs greatly among individuals. Sauve et al. (in press) also provided evidence that pup calls are stereotyped from the first day after birth, supporting the idea that mothers can recognize their pups' calls within hours following parturition (Bartholomew, 1959). Yet, the onset of female abilities to recognize their pup's voice (if any) remains to be tested. It therefore appears that maternal protectiveness might influence female response to pup calls, and that pup vocal stereotypy might influence the vocal recognition process. These factors must thus be considered when testing female discrimination abilities. The environment within which vocal communication occurs might also affect the recognition process. Given the variable topography, proximity to anthropogenic activities and exposure to waves, wind and currents characteristic of harbour seal habitat (J. J. Burns, 2009; Robillard, Lesage, & Hammill, 2005; Thompson et al., 1997), the efficiency of pup call propagation might vary between and within colonies. In addition, acoustic characteristics of vocalizations are usually degraded differentially with distance (e.g. Wiley & Richard, 1982). Depending on the cues used in vocal recognition of offspring (if any), such asymmetrical propagation of call features may potentially influence female vocal recognition abilities. Our aim in this study was to determine whether harbour seal females are capable of recognizing their pup's calls, and if so, what maternal, offspring and environmental characteristics might affect the recognition process. We had three primary objectives. The first was to experimentally test the hypothesis that free-ranging harbour seal females can discriminate the calls of their pup from calls of age-matched, nonfilial pups. The second objective was to assess the effects of maternal and offspring traits on the vocal recognition process (if any). We hypothesized that the strength of the maternal response behaviour towards broadcast pup calls would be positively correlated with female motivational state prior to playback and the pup's individual vocal stereotypy. We also expected females to recognize the calls of their pup from the first few days after birth. Finally, the third objective was to investigate the differences in sound degradation between sites where mothers and pups reunited. We expected contrasting degradation levels among

C. C. Sauv
e et al. / Animal Behaviour 105 (2015) 109e120

acoustic parameters measured on the calls, as well as intra- and intercolony transmission efficiency differences. METHODS Study Site and Animal Handling This study was conducted at two harbour seal colonies on the south shore of the St Lawrence River estuary that were visited alternately (weather permitting). At the Bic colony (48
240 N, 68
510 W; N z 100 pups/year; Van de Walle, 2013), animals haul out on an island located about 4 km off the coast or utilize nearby rocky reefs that extend around the island. This colony encompasses a study area of approximately 40 km2. Located about 60 km
tis colony (48
410 N, 68
010 W; N z 30 pups/ downstream, the Me year; Van de Walle, 2013) encompasses a small bay and some small reefs 100 m off the coast. The Metis colony is sheltered from the prevailing southwesterly winds and covers a 10 km2 area along the costal shoreline. Throughout the 2013 breeding season (from mid-May to July), 77 harbour seal pups were captured in the water using a dip-net from a 5 m inflatable boat, then transferred to a stationary, 7 m hard-hulled motorboat for handling. Pups were marked with a pyramid tag (Seal Hat®, Dalton, U.K.) glued (Loctite no. 422 cyanoacrylate glue and no. 7452 Accelerator, Loctite Corp., Mississauga, ON, Canada) on the head and tagged with a uniquely numbered Jumbotag (Dalton, U.K.) in a hindflipper. Upon each capture, the pup was weighed (±0.5 kg; Salter spring scale, West Bromwich, U.K.) and a proxy for female protectiveness towards her pup (‘maternal protection level’) was assigned by two observers based on the following criteria: 0: female absent; 1: female swims away from pup as boat approaches; 2 and 3: female keeps the pup in sight at a long (>10 m) and close (<10 m) distance, respectively; 4: female attempts to jump into the boat to remain with the pup. Maternal protection level was determined prior to any playback test while the female was searching for her pup in the absence of vocal stimulus. Handling time rarely exceeded 10 min from capture to release. When the female was present, it was also noted whether the motherepup pair reunited within 5 min following pup release. As most pups were not captured immediately after parturition, pup age (in days) was determined using the degeneration of the um
, Hammill, & Barette, 2003). bilical cord (see Dube Recording Procedure Upon each capture (except captures when playback trials were performed), we recorded airborne mother attraction calls uttered by the pup for about 60 s following the pup's transfer to the bigger boat and prior to any handling. Maternal attraction calls have been defined as calls produced by pups during various motherepup interactions (i.e. during searching, reunion and nursing events; Stirling, 1970). Recordings were performed using a shotgun microphone (frequency response: 40 Hz e 20 kHz, ±2.5 dB; Sennheiser ME 67, Sennheiser electronic GmbH & Co. KG, Wedemark, Germany) connected to a Marantz PMD 661 digital recorder (frequency response: 20 Hz e 24 kHz, ±1 dB; Marantz Europe, Eindhoven, The Netherlands) at a 44.1 kHz sampling frequency. This procedure allowed the recording of pup calls in a standardized context where the motherepup pair was recently separated and both animals were attempting to find each other, which stimulated the pup to vocalize. Since calls did not exceed 10 kHz, recordings were resampled at 22.05 kHz. High-quality vocalizations (i.e. no background noise, no overlap with vocalizations from other animals) were selected for use as playback stimuli during the subsequent capture of the pup. In natural contexts, harbour seal pup

111

vocalizations differ according to the level of stress faced by the pup (Perry & Renouf, 1988). It is likely that calls from the present study were typical of stressed pups since they were recorded while animals were captured. Nevertheless, calls recorded in this context were found to carry an individual signature having the potential to be discriminated by females through a vocal recognition process
et al., in press). (Sauve To characterize pup voice consistency and distinctiveness, we calculated a proxy for pup vocal stereotypy as individual percentage correct classification scores generated by a discriminant func
et al., in tion analysis on six acoustic parameters (detailed in Sauve press). Playback Experiment To assess whether breeding harbour seal mothers discriminate their offspring's calls from calls of other pups, we tested 21 different females with two experimental stimuli: (1) maternal attraction calls from their own pup (control) and (2) maternal attraction calls from a nonfilial pup. Since maternal attraction call
et al., in characteristics change as pups age and grow (Sauve press), stimuli were always created using calls from the most recent recording available for each pup. This minimized the differences between the broadcast stimuli and the calls that the pup would utter when playback trials were conducted. Sound files containing vocalizations from pups of similar ages were also used for both treatments (own and nonfilial pup) within a given playback session. Each experimental stimulus consisted of two series of five airborne maternal attraction calls separated by 3 s of silence. Within each series, calls were interspersed by intervals of natural silence (0.3e0.7 s). Each series contained at least two different maternal attraction calls since collecting 10 high-quality calls for each pup was not always possible. Playback stimuli were created using Avisoft SASLab Pro (Avisoft Bioacoustics, Berlin, Germany; v5.2.07) and Goldwave (Goldwave Inc., St John's, Canada; v5.67). A playback session started immediately after the capture of a pup for which a stimulus sound file was available, while the mother was still in the area, but not looking in the direction of the boat. During each session, a stimulus was broadcast from the boat and the female's behavioural response was noted during the following 90 s. An additional 90 s interval allowed the female to return to her initial motivational state before the second stimulus was broadcast and the behavioural response noted for another 90 s. Treatment order (calls from own or nonfilial pup broadcast first) was randomized. Stimuli were broadcast simultaneously in air and underwater to mimic natural pup calling behaviour since calls produced when pups are at the water's surface propagate in both
et al., in press). media (Perry & Renouf, 1988; Renouf, 1984; Sauve Stimuli were broadcast using a MP3 player (Edirol R-09 WAVE/MP3 Recorder, Roland Corporation U.S., Los Angeles, CA, U.S.A.) connected to both a portable loudspeaker (frequency response: 70 Hz
das, France) held above e 15 kHz; Permio 8, TAG, Saint Jean de Ve the surface of the water (<50 cm) and an underwater speaker (frequency response: 200 Hz e 20 kHz; LL916, Lubell Labs, Columbus, OH, U.S.A.) placed at a minimum depth of 1 m. Amplitudes of playback stimuli were adjusted to 92e94 dB SPL (re. 20 mPa) and 153e155 dB SPL (re. 1 mPa) at 1 m in air and underwater, respectively. The aerial amplitude range corresponds to the mean amplitude of pups' calls measured with a sound level meter at 1 m (Brüel & Kjaer 2235, Brüel & Kjaer, Nærum, Denmark), while the underwater amplitude range approached the mean amplitude (148 ± 3 dB SPL re. 1 mPa) estimated from underwater calls recorded using a calibrated hydrophone of known effective sensitivity (165 dB SPL re. 1 V/mPa, Cetacean Research C54XRS, Cetacean

112

C. C. Sauv
e et al. / Animal Behaviour 105 (2015) 109e120

Research Technology, Seattle, WA, U.S.A.; frequency response: 16 Hz e 44 kHz, ±3 dB). When the motherepup pair is separated, a female hearing her pup calling usually approaches the pup (Evans & Bastian, 1969) and attempts to locate it with her head out of the water, while looking
, personal in the direction of the pup (I. Charrier & C. C. Sauve observation). Therefore, we predicted that a female hearing her own pup call would rapidly swim towards the loudspeakers, and then surface several times looking in the direction of the broadcast calls, while actively trying to locate her pup. If females discriminate calls uttered by their offspring from calls of other pups, females should stay at a greater distance from the loudspeaker and be less interested in looking towards it after hearing calls from nonfilial pups. Female behavioural responses to playback were monitored by an observer located in the boat and a digital video camera (GoPro Hero 3 Silver Edition, Woodman Labs, San Mateo, CA, U.S.A.) attached to the underwater speaker. This allowed the measure of three variables depicting a female's response to the playback: (1) the number of times the female looked in the direction of a speaker (‘number of looks’), either by surfacing while looking towards the aerial loudspeaker and/or by passing in front of the video camera; (2) the female's closest distance of approach to the aerial speaker estimated by the observer (‘closest approach’; ±1 m and ±5 m for short (<15 m) and greater distances, respectively); and (3) the time lag from stimulus broadcast to the female's first appearance at the water's surface or in front of the underwater camera (‘appearance latency’, s). The observer estimating each female's closest approach was the same throughout the study to ensure consistency of the results. This observer was familiar with the study sites, and distance estimations were facilitated by the presence of numerous surrounding reefs and coasts that served as visual cues. Most playback tests were performed near haul out sites, where different females swam nearby as the observer monitored the response of the tested female. Pelage spotting patterns are unique in individual adult harbour seals and are thought to remain the same over the animal's life span, thus photo-identification can be used to identify individuals (Cunningham et al., 2009; Hastings, Small, & Pendleton, 2012). To ensure unequivocal characterization of the female's response to both stimuli, we photographed the faces of surfacing females during both the playback experiment and the final motherepup reunion to compare the fur patterns with pictures taken previously while the motherepup pair was together. The three behavioural response variables (number of looks, closest approach and appearance latency) were introduced into a principal component analysis (PCA) to generate independent variables describing the female's strength of response to the playback stimuli (McGregor et al., 1992). Principal component (PC) scores were computed for PCs with eigenvalues >1. Pairwise Tukey and t tests were performed on these PC scores to assess differences in response to playback according to maternal protection level and treatment (own versus nonfilial pup), respectively. To investigate the effect of female and pup characteristics, as well as experimental settings on the behavioural responses to playback stimuli, PC scores were introduced as response variables in linear mixed models (LMM, package nmle; Pinheiro, Bates, DebRoy, Sarkar, & R Development Core Team, 2012) with treatment, pup age at playback (days), Dage (pup age at playback  pup age when calls used as stimuli were recorded, in days), maternal protection level and treatment order (own versus nonfilial pup) as fixed effects and female ID as a random effect. To obtain the best model explaining a female's response to playback, we performed stepwise model selection based on the likelihood ratio test (package ‘lmtest’; Zeilis & Hothorn, 2002) until eliminating fixed effects no longer generated better models than the previous one.

We first tested the relationship between a pup's individual vocal stereotypy level and the female's response to the playback stimuli using a linear model on pup vocal stereotypy level and the difference in PC scores obtained from both playback treatments for a given female (PCown  PCnonfilial). Second, we performed a onetailed t test to assess whether individual vocal stereotypy was greater for pups that reunited with their mother within 5 min following their release. To provide insight into the magnitude of both of these relationships, we computed a Cohen's d using the pffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffi formula d ¼ 2t= df ¼ 2r= 1  r 2 , where t is the Student's statistic, df is the degrees of freedom from the t test and r is the Pearson correlation coefficient (Borenstein, Hedges, Higgins, & Rothstein, 2009).

Propagation Tests The two colonies differed in their exposure to waves, wind and streams. Indeed, motherepup reunion sites (i.e. aquatic areas, often near haul out sites, where mothers and pups reunite after a separation) at the Bic colony are distributed around Bic Island, which acts as a wave- and wind-shield protecting different sites
tis colony is depending on the direction of the winds. The costal Me protected from prevailing winds (SW), but is more exposed during storms (usually NE winds). Sites within each colony also vary in their level of exposure to the elements. As waves, winds and streams mask and induce pattern loss in sounds, they are likely to influence the distance to which pups' maternal attraction calls can be detected by females at each reunion site. To investigate the propagation efficiency of maternal attraction calls at different reunion sites, were carried out underwater and aerial propagation tests at both colonies. Propagation tests consisted of 10 repetitions of a single pup call (maternal attraction call), played with a MP3 player (Edirol R-09 WAVE/MP3 Recorder) either in air (Permio 8 loudspeaker) or underwater (LL916 underwater speaker) at amplitude levels consistent with that of harbour seal pups calling in a natural context (92e94 dB SPL (re. 20 mPa) and 153e155 dB SPL (re. 1 mPa) at 1 m in air and underwater, respectively). Broadcast signals were re-recorded at increasing distances from the loudspeaker. In-air broadcast signals were re-recorded (Edirol R-09 WAVE/MP3 Recorder) at 1 m (control), 4 m, 8 m, 16 m, and so forth, until calls could not be heard by the experimenters (maximal distance range 256e2048 m). Similarly, underwater signals were re-recorded (C54XRS hydrophone connected to a Marantz PMD 661) at 1 m (control), 25 m, 50 m, 100 m, etc. Distances between the speaker and the microphone were measured using a range finder (Bushnell ScoutArc® 1000, Bushnell Corp., Overland Park, NS, Canada) for distances less than 100 m and a hand-held GPS (Garmin GPSMAP 60Cx, Garmin, Olathe, KS, U.S.A.) otherwise. This procedure was performed at seven different
tis, four at Bic; see interactive map) under reunion sites (three at Me similar weather conditions (light wind, low waves). Propagation tests were conducted in directions from which females were more likely to swim back to the colony and search for their pup after a feeding trip at sea. To quantify the degradations of the signals during propagation, we measured three acoustic parameters on propagated calls for which the quality of the vocalization was sufficient to allow reliable visual identification on the spectrograms. (1) We assessed the modification of the FM pattern using a spectrogram crosscorrelation function, which simultaneously analyses temporal, amplitude and frequency components (Clark, Marler, & Beeman, 1987; Khanna, Gaunt, & McCallum, 1997). Single spectrograms were created for each propagated call (Hamming window, FFT length ¼ 1024, frame size ¼ 100%, overlap ¼ 96.87%) and pairwise

C. C. Sauv
e et al. / Animal Behaviour 105 (2015) 109e120

correlation coefficients were calculated using Avisoft Correlator (Avisoft Bioaccoustics, Berlin, Germany; high pass cutoff ¼ 0 Hz, tolerate frequency deviation ¼ 1 Hz). (2) We measured the AM degradation by calculating the smoothed envelope using a Hilbert transformation (Mbu Nyamsi, Aubin, & Bremond, 1994). Maximal cross-correlation coefficients accounting for a time lag of ±10 ms were computed using R v.3.1.0 (R Development Core Team, 2011) between AM patterns from calls recorded at various distances and those recorded at 1 m (controls). Finally, (3) we measured the degradation of the frequency spectrum by calculating the averaged frequency spectrum on the whole length of the call (Hamming window; frequency resolution ¼ 0.336 Hz), and obtained Pearson correlation coefficients between propagated and control calls. For all three acoustic parameters, we performed correlation analyses separately for calls propagated in air and underwater. We computed average correlation coefficients for FM, AM and the frequency spectrum from calls recorded at a given distance and site and in a given medium, and extracted the furthest distance before which each value dropped below 0.5 (d0.5; m). This d0.5 variable was hereafter used as an index of sound propagation efficiency. Statistical Analyses We tested normality and homoscedasticity criteria with ShapiroeWilk and Levene (package car; Fox & Weisberg, 2011) tests, respectively. When the normality criterion was not respected, we used randomization tests (rand-test; modified from Mazerolle, 2013) with 500 permutations rather than t tests to compare statistics between groups. Results were considered significant at P < 0.05 and are presented as means ± SE, unless stated otherwise. Ethical Note All applicable international, national and institutional guidelines for the care and use of animals were followed. All procedures involving animals performed in this study were in accordance with
Laval (permit CPAUL 2011020) the ethical standards of Universite and Fisheries and Oceans Canada (permit IML 2013-001). As capture and tagging may induce some stress to the animals, the reduced handling time (<15 min from capture to release) limited the duration of this stress and avoided any risks of abandonment of the pups by the mother. All females tested in playback were seen to reunite with their pup following its release. RESULTS Female Vocal Recognition Abilities A total of 21 different females were submitted to playback tests. The results from three of these females were discarded because it was not possible to identify the pup's mother with certainty. This resulted in 18 high-confidence playback sessions, all performed on different females showing maternal protection levels of 2 (N ¼ 16) or 3 (N ¼ 2) since it was not possible to test females that swam away (level 1) or jumped into the boat (level 4). Tested females with maternal protection levels of 2 and 3 are hereafter referred to as displaying low and high protectiveness, respectively. The standardized loadings from the PCA were >j0.50j on at least one PC for the three behavioural variables, indicating that all of them were relevant in the PCA (Table 1; Hair, 1995). Only the first PC (PC1) had an eigenvalue >1 and was subsequently retained. All original behavioural measures contributed equally to PC1 and loaded significantly (loading >j0.50j) on this factor. Moreover, the signs of the loadings were consistent with PC1 representing a strength of response to playback variable (McGregor et al., 1992). A

113

Table 1 Harbour seal female responses to playback (Ntests ¼ 36) were described by a single composite variable (i.e. PC1) generated by the PCA on the three measured behavioural variables Standardized factor loadings

Number of looks Closest approach Latency to first appearance Eigenvalue Proportion of variance explained Cumulative proportion of variance explained

PC1

PC2

PC3

0.58 ¡0.58 ¡0.57 2.10 0.70 0.70

0.47 0.34 0.82 0.47 0.15 0.85

0.67 0.74 0.07 0.44 0.15 1.00

All original behavioural measures contributed equally to PC1. Components with an eigenvalue >1 are shown in bold.

shorter latency, closer approach towards the speakers and more looks towards the speakers all indicate a strong response; consistently appearance latency and closest approach showed negative loadings on PC1 whereas the loading for number of looks was positive. Model selection for the best modelling of PC1 scores led to the elimination of the treatment order from the fixed effects (see Table 2 for details). The only significant effects on female responses to the playback were (1) maternal protection level (bhigh ¼ 1.048, P ¼ 0.03; Table 3) and (2) the treatment (bown ¼ 1.067, P < 0.01). Females showed a greater overall response to calls from their own pup than to calls of nonfilial pups (one-tailed paired t test: t17 ¼ 3.36, P < 0.01; Fig. 1b). Specifically, females with low protection levels displayed a greater response to calls from their own pups than to calls of nonfilial young (one-tailed paired t test: t15 ¼ 3.78, P < 0.01) whereas highly protective females did not show a differential response to playbacks according to the treatment (t1 ¼ 0.08, P ¼ 0.53). Effects of Maternal and Offspring Traits on Vocal Recognition Regardless of the treatment (own versus nonfilial pup), females with low maternal protection levels (level 2) displayed weaker responses to the playback stimuli (i.e. PC1 scores) than highly protective females (level 3; one-tailed t test: t5.1 ¼ 3.90, P ¼ 0.01; Fig. 1). Pup age when a playback was conducted and pup age difference between call recording and playback session had no effect on a female's response to playback (Table 3). The strength of a female's preferential response to her pup during playback (i.e. PC1own  PC1nonfilial) and pup individual vocal stereotypy exhibited a positive correlation that was marginally significant (b ¼ 2.5 ± 1.3, R2 ¼ 0.22, N ¼ 15, P ¼ 0.08, d ¼ 1.06; Fig. 2a). Similarly, pups that reunited with their mother within 5 min after their release in the water had on average a greater vocal stereotypy than those that did not (one-tailed t test: t16.75 ¼ 1.54, P ¼ 0.07, d ¼ 0.56; Fig. 2b). Effects of Environmental Characteristics on Vocal Recognition The patterns of both AM and FM, as well as the frequency spectrum of propagated calls were differentially degraded with increasing distance from the source (Figs 3, 4). In underwater trials, the frequency spectrum was the least degraded acoustic feature with distance at every site tested, followed by FM and then AM (Fig. 5). Similarly, frequency spectra showed the weakest degradation with distance in 57% (4/7) of in-air propagation tests, while FM was most efficiently transmitted in 29% (2/7) of the trials. There was no difference between colonies in mean d0.5 in air and under water, respectively, for FM (rand-test; P ¼ 0.48 and 0.37), AM (rand-test; P ¼ 0.28 and 0.56) or frequency spectra (rand-test,

C. C. Sauv
e et al. / Animal Behaviour 105 (2015) 109e120

114

Table 2 The best model (bold type) of the behavioural responses to playback stimuli (PC1 scores; Table 1) determined by model selection included the treatment (own versus nonfilial pup), pup age, pup age difference between call recording and playback session and maternal protection level (MPL) as fixed effects, and female ID as a random effect Model

Coefficient P values

PC1 score~treatmentþageþDageþMPLþt.orderþjIDpup PC1 score~treatmentþageþDageþMPLþjIDpup PC1 score~treatmentþageþMPLþjIDpup

Treatment

Age

DAge

MPL

t.order

0.004 0.004 0.004

0.376 0.430 0.5681

0.399 0.510

0.026 0.029 0.033

0.482

Log likelihood

P likelihood ratio test

60.80 60.85 58.53

0.74 0.03

Reference levels for each factor were as follows: treatment: ‘nonfilial’; MPL: ‘low’; treatment order (t.order): ‘own ¼ first treatment broadcast’. The best modelling of PC1 scores identified by the model selection procedure is shown in bold.

Table 3 The best LMM of PC1 scores (see Table 2) indicated that female's strength of response to playback stimuli (N\ ¼ 18) was greater following the broadcast of calls from their own pup than following the broadcast of calls of nonfilial pups (reference level for treatment effect) Effect

Coefficient

df

t

P

Treatment (own) Pup age (days) Age difference (days) Protection level (high)

1.067 0.024 0.021 1.048

15 15 15 16

3.35 0.81 0.67 2.64

0.00 0.43 0.51 0.03

Highly protective females responded more intensely to playback than less protective females (reference level for protection-level effect). Significant outcomes are shown in bold.

P ¼ 0.55 and 0.37). However, sound propagation efficiency varied considerably within colonies (Fig. 5). In-air propagation efficiency (i.e. d0.5 and coefficients of correlation in decreasing order) at sites
tis were 1 > 2 > 3 for FM patterns and 2 > 1 > 3 for both from Me AM patterns and frequency spectra (Fig. 5). For underwater prop
tis site ranks were 3 > 1 > 2 for both FM agation efficiency, the Me patterns and frequency spectra, while ranks for AM patterns were 3 > 2 > 1. When sorted similarly, Bic site ranks for in-air propagation were 6 > 4 > 7 > 5 for FM patterns, 6 ¼ 7 > 5 > 4 for AM patterns and 5 > 7 > 6 > 4 for the frequency spectra. Rankings for underwater propagation were 4 > 5 > 7 > 6 for both FM and AM patterns while they were 4 > 7 > 5 > 6 for the frequency spectrum (Fig. 5).

*

Response to playback (PC1)

5 ** 3

1

−1

−3

DISCUSSION Female Vocal Recognition Abilities Maternal responses to playback stimuli differed depending on whether the broadcast calls were from their own pup or from an unrelated pup, indicating that harbour seal females have the ability to recognize their pup's call. Females therefore perceive and use individual vocal markers previously identified in harbour seal pup calls (Khan, Markowitz, & McCowan, 2006; Perry & Renouf, 1988;
et al., in press) for individual vocal recognition. Renouf, 1984; Sauve High population density, land breeding and maternal absences from the breeding site represent biological traits that seem to favour maternal vocal recognition abilities in phocids (Table 4). Although harbour seals breed in relatively low population density areas, most breeding sites from the St Lawrence estuary (rocky reefs) are generally flooded at high tide (Boulva & McLaren, 1979; Lesage, Hammill, & Kovacs, 1995), forcing pups to take to the water within hours after birth (Bigg, 1981). Female foraging trips at sea during the lactation period (Boness et al., 1994) imply an effective need to re-locate their pup when they return to the breeding colony. In instances where pups accompany their mother at sea, a recognition system effective over long distances is also required for the pair to reunite after female diving bouts, as pups perform shorter dives than females (Bowen et al., 1999). Therefore, the ability of a female harbour seal to recognize her pup's calls is in accordance with general environmental and ecological constraints favouring the evolution of motherepup recognition. Some harbour seal populations breed on sand beaches and therefore face less landscape variation with tide levels. These more stable meeting sites could ease the use of spatial cues by females searching for their pup after a foraging trip (e.g. Trimble & Insley, 2010) and could thus influence the reliance on vocal communication. Vocal recognition has been shown to vary among populations of given species. For instance, grey seals, Halichoerus grypus, from two distinct colonies exhibited contrasting vocal recognition abilities (McCulloch & Boness, 2000; McCulloch, Pomeroy, & Slater, 1999). As harbour seals are widespread along North Atlantic and Pacific coasts (J. J. Burns, 2009; Thompson et al., 1997), it would be interesting to perform playback experiments on harbour seal females from other populations to assess divergences in vocal recognition abilities. Effects of Maternal and Offspring Traits on Vocal Recognition

−5

Nonfilial own Low MPL

Nonfilial own High MPL

Figure 1. Responses of female harbour seals to playback of calls from nonfilial pups and from their own pups (PC1 scores, see Table 1) relative to the females' maternal protection level (MPL; low MPL ¼ 2, N ¼ 16; high MPL ¼ 3, N ¼ 2). Boxes extend from the first to the third quartile, with a thick line indicating the median. Whiskers extend to the most extreme data point, which was no more than 1.5 times the interquartile range. *P < 0.05; **P < 0.01.

Maternal behaviour has been shown to be influenced by female temperament (Fairbanks, 1996; Maestripieri, 1993; Plush et al., 2011). The maternal protection level variable was an approach we developed to combine female motivational state at the time of the playback trial and personality (boldness personality trait; Wilson, Clark, Coleman, & Dearstyne, 1994). The relationship between maternal protection level and female temperament was supported by its repeatability within individuals (Weinstein, Capitanio, &

C. C. Sauv
e et al. / Animal Behaviour 105 (2015) 109e120

80

4 (a)

115

(b)

70 3 60 50

∇

ISL

PB

2 1

40 30

0

20 −1

10

−2 0

10

20

30

40 ISL

50

60

70

0

80

No reunion

Reunion

Figure 2. (a) Relation between female harbour seals' differential response to playback (PC1own  PC1nonfilial ¼ DPB) and pup individual vocal stereotypy level (ISL). (b) Pup ISL for motherepup pairs that did not reunite (no reunion) and that reunited (reunion) within 5 min following handling. Boxes extend from the first to the third quartile, with a thick line indicating the median. Whiskers extend to the most extreme data point, which was no more than 1.5 times the interquartile range.

4

(a)

1m

4m

8m

16 m

32 m

1m

4m

8m

16 m

32 m

3

2

Frequency (kHz)

1

4 (b)

3

2

1

Time Figure 3. Spectrograms illustrating attenuation of aerial harbour seal pup calls recorded at increasing distances at (a) site 5 and (b) site 6. Figure generated using Seewave R (Sueur, Aubin, & Simonis, 2008).

C. C. Sauv
e et al. / Animal Behaviour 105 (2015) 109e120

116

1

800

(a)

(a)

600 *

0.8 400 *

*

200 0

0.4 300 0

64

128

256 Distance (m)

1

1

2

3

(b)

4

*

512

5

6

7

*

*

225

(b)

d0.5 (m)

Coefficient of correlation

0.6

150 75

0.8

0 0.6

1

2

3

4

5

6

7

1600 (c) 0.4 1200 0.2 0

800 200

400

800 Distance (m)

1600

Figure 4. Attenuation of the three measured acoustic parameters of harbour seal pup calls when propagated (a) in air and (b) underwater, at site 4. Symbols denote mean correlation coefficients and error bars denote the SEs for the FM pattern (circles), the AM pattern (triangles) and the frequency spectrum (squares). The dotted line indicates the 0.5 correlation threshold used to compare sites.

Gosling, 2008): 88.9% (N ¼ 16/18) of tested females in this study were attributed either the same maternal protection level or one of two consecutive maternal protection levels throughout the lactation period. Highly protective females displayed stronger responses to the playback regardless of the treatment (own versus nonfilial pup calls) and responded similarly to calls of their own pups and to calls of nonfilial young. It is unlikely that some harbour seal females in this study lacked recognition abilities. The absence of preferential response to calls of their own pup by highly protective females might be attributable to their hormonal status (lactogenic and sex steroid hormones; Numan, 1994), bold personality and related strong motivation to reunite with their pup, which may incite them to pay less attention to vocalization characteristics and to react strongly to any pup vocalization heard when searching for their young. Animal boldness has been associated with a greater amount of time spent near offspring, greater parental care and higher weaning success (Budaev, Zworykin, & Mochek, 1999; Reale, Gallant, Leblanc, & Festa-Bianchet, 2000). In the present study, bolder or more protective harbour seal females (i.e. that responded more intensely to pup calls) could maintain better contact with their pup during the nursing period. However, in situations where risks of confusion are high (e.g. high population density areas), the behaviour displayed by highly protective females could result in confusion over offspring identity, which would have important outcomes on pup growth and survival, and ultimately on female

400 0

*

* * 1

2

3

4

5

* 6

7

Figure 5. Furthest distances before which the mean correlation coefficients of the measured acoustic parameters of harbour seal pup calls fell below 0.5 (d0.5) when propagated in air (black bars) and underwater (white bars) at the seven sites testedfrequency spectra: (a) FM, (b) AM and (c) frequency spectra. Refer to interactive map for identification of site numbers. Asterisks indicate propagation trials where correlation coefficients remained 0.5 at the maximal distance at which propagated calls were analysed.

reproductive success unless pup identity is confirmed upon reunion by a nonacoustic recognition mode (e.g. olfaction; Dehnhardt, 2002; Miller, 1991). Pup age did not affect female response to the playback stimuli, suggesting that the onset of maternal vocal recognition takes place within the first few days after birth and that female discrimination abilities remain constant throughout the nursing period (tested females' pup age range 3e37 days). The absence of Dage effect on the playback responses indicate that females retain older versions of their offspring vocalizations despite modifications throughout
et al., in press). Long-term vocal recognition was rearing (Sauve demonstrated at higher levels in otariids, as female sub-Antarctic fur seals, Arctocephalus tropicalis, and northern fur seals, Callorhinus ursinus, recognize the calls of their pups several months/years after birth and weaning (Charrier, Mathevon, & Jouventin, 2003; Insley, 2000). Since maternal attraction calls disappear from harbour seal pups' repertoire upon weaning (Renouf, 1984), it is unlikely that long-term vocal recognition abilities provide any adaptive benefit in this species (e.g. postweaning cooperation allowed by kin recognition). As Charrier et al. (2003) suggested, memorization of older versions of pups' calls is likely to represent a

C. C. Sauv
e et al. / Animal Behaviour 105 (2015) 109e120

117

Table 4 Maternal vocal recognition abilities of the five phocid species studied to date Pup mobility

Breeding substrateb

Lactation lengthb,c (days)

Maternal vocal recognition

0

Lowd

Land

22e29

Yesj

HeM

0

Nullehighe,f

Fast ice or land

15e21

Population dependentk,l

MeL

40

Fast ice

45e50

Nom

L

55

Moderate ehighg Highh

Land

21e42

Yesn

L

0

Lowi

Land

35e42

Noo,p

Species

Population densitya

Northern elephant seal, Mirounga angustirostris Grey seal, Halichoerus grypus Weddell seal, Leptonychotes weddellii Harbour seal, Phoca vitulina Hawaiian monk seal, Neomonachus schauinslandi

H

Maternal foraging (% time at sea)b

Population density: low (L), medium (M), high (H). High population density, frequent maternal absences and high offspring mobility are biological traits generally favouring the evolution of parenteoffspring vocal recognition systems. a Riedman (1990). b Schulz and Bowen (2004). c Riedman (1990). d Reiter, Stinson, and Leboeuf (1978). e Haller, Kovacs, and Hammill (1996). f Jenssen, Asmul, Ekker, and Vongraven (2010). g JM Burns (1999). h Bowen et al. (1999). i Boness, Craig, Honigman, and Austin (1998). j Petrinovich (1974). k McCulloch and Boness (2000). l McCulloch et al. (1999). m Van Opzeeland, Van Parijs, Frickenhaus, Kreiss, and Boebel (2012). n Present study. o Job, Boness, and Francis (1995). p Result inferred from observations (no experimental study).

by-product of strong learning experienced by females during the course of the rearing period. Females whose pup calls were more stereotyped displayed greater selectivity for calls from their own pup during playback. Moreover, pups that reunited with their mothers within 5 min following handling displayed greater vocal stereotypy than those that did not. This suggests that females would be more efficient in recognizing their pup's voice when it is highly stable and distinctive. Pup vocal stereotypy rather than female discrimination abilities would therefore be the limiting factor in the harbour seal motherepup vocal recognition system. Females rearing pups that produce weakly stereotyped calls are thus likely to depend more on spatial cues to locate their offspring and on nonacoustic recognition modes to confirm pup identity upon reunion. Effects of Environmental Characteristics on Vocal Recognition The three acoustic features used to describe maternal attraction call propagation efficiency showed considerable differences in distance-related degradation. Spectral features of pup maternal attraction calls propagated with a high reliability, which suggests that some of these highly individually stereotyped characteristics of
et al., in press) could be used by females for vocal pup calls (Sauve recognition over long distances. Because our underwater speaker was less sensitive to frequencies >800 Hz, the degradation of the frequency spectrum measured on calls propagated underwater might have been overestimated. Spectral features of underwater maternal attraction calls might thus propagate at greater distances than those reported in this study. Both AM and FM patterns propagated with similar efficiency during in-air propagation tests. However, in underwater propagation tests, AM was degraded more rapidly than FM. Under calm weather conditions, noise and medium absorption are likely to be more important in shallow coastal waters where currents and physical barriers (e.g. seaweed, rocks)

are present than over the surface. This could explain the relatively poor propagation of AM patterns obtained in underwater trials compared to aerial propagation tests. Frequency modulation slopes represent other call characteristics that have been shown to be
et al., in individually stereotyped in harbour seal pup calls (Sauve press) and have also been shown to be (1) highly resistant to degradation with distance and (2) involved in recognition processes in other colonial species (Aubin & Jouventin, 2002; Charrier et al., 2010; Charrier, Pitcher, & Harcourt, 2009; Pitcher, Harcourt, & Charrier, 2012). By providing evidence that the most individually stereotyped acoustical parameters in harbour seal pup calls are also the ones that propagate most efficiently in the habitat, propagation tests further suggest that spectral characteristics and FM pattern are likely to be used by females for vocal recognition of their offspring. However, playback tests on females using modified pup calls are required to validate this hypothesis. Throughout the nursing period, 38.9% (7/18) of pups were consistently captured at the same reunion site. Other pups (11/18) moved around the Bic Island, with consecutive captures being separated by up to 6.5 km (linear distance between sites 4 and 6). Although it was not possible to determine whether these movements were undertaken by the pups alone or while accompanied by their mother, this demonstrates that given motherepup pairs used various reunion sites. Tides, winds and wave conditions are likely to induce pups to move from one site to another while females are foraging at sea. Acoustic communication over long distances (>100 m) is therefore expected to be particularly important for the maintenance of the motherepup contact during nursing in harbour seals. The propagation tests performed in this study demonstrate that individually stereotyped acoustic features of both airborne and underwater harbour seal pup maternal attraction calls can propagate at such a range of distances. Overall, both colonies had similar sound propagation characteristics for all three acoustic features measured. Nevertheless,

118

C. C. Sauv
e et al. / Animal Behaviour 105 (2015) 109e120

propagation tests identified important within-colony heterogeneity of sound transmission efficiency, indicating that reunion sites are associated with differential constraints on signal propagation efficiency. Overall, locations best suited for aerial sound
tis colony were sites 1 and 2, whereas site propagation at the Me 3 showed the worst aerial sound propagation. Sites 1 and 2 represent a single motherepup reunion area tested in both directions. Propagation trajectories at these sites were parallel to the rocky coast, which might channel sound waves, as opposed to the open and sandy bay at site 3. By contrast, the most efficient
tis was site 3. This could underwater sound propagation site at Me be explained by the protection from dominant currents and waves provided by the bay, and by the absence of underwater reefs impeding sound propagation as found in sites 1 and 2. The important aerial degradation of acoustic signals at site 3 could thus be compensated by its high underwater transmission efficiency. At the Bic colony, the three acoustic features measured led to different rankings for in-air propagation. Nevertheless, the best correlation coefficients were obtained at site 6 for both AM and FM patterns. Although no fine-scale site use data were collected in the study area, the elongated rocky reef and surrounding aquatic area at site 6 were observed to be used extensively as haul out sites and motherepup reunion sites, respectively. Underwater sound propagation at Bic was the most efficient at site 4, which seemed infrequently used by motherepup pairs to haul out or reunite, while the highest sound degradation was found at site 6. Site 4 is a particularly deep-water area compared to the other sites tested. There was therefore less turbulence and no obstacles to sound propagation at the location of the hydrophone (depth z 2 m) at this site, which might explain the better underwater propagation. Considering these results, it is difficult to determine which tested sites are best suited for motherepup vocal recognition and successful reunions at the Bic colony. For instance, sites 6 and 7 might represent interesting trade-offs between acoustic properties, protection from the elements and haul out opportunities. In both colonies, the distribution of hauled out motherepup pairs was not proportional to the space available on each haul out site (Beauplet
, n.d.), suggesting that site characteristics might influence & Sauve spatial use by breeding seals (Ban & Trites, 2007). Exposure to waves and substrate as well as anthropogenic disturbance and noise levels are known to influence pinniped haul out site choice (e.g. Acevedo-Gutierrez & Cendejas-Zarelli, 2011; Ban & Trites, 2007; Cordes, Duck, Mackey, Hall, & Thompson, 2011; Suryan & Harvey, 1999). As they are likely to affect motherepup recognition efficiency, acoustic properties could represent additional characteristics influencing harbour seal haul out and reunion site selection during the rearing season. The potential impact of site acoustic properties on the efficiency of motherepup vocal communication could be assessed by additional playback experiments at the different sites coupled with distribution surveys of motherepup haul out sites and reunion sites. Conclusions This experimental study provides evidence that harbour seal females from the St Lawrence estuary are able to recognize and distinguish the calls of their own pups from the calls of other pups during the nursing period. This ability was affected by the quality of individual pup call signature, which could further affect the successfulness of motherepup reunions following separations. Propagation tests also revealed important differences in acoustic propagation efficiency between reunion sites at the two breeding colonies. Physical characteristics of these sites are thus likely to affect vocal communication.

Acknowledgments
becThis study is a contribution to the research program of Que
an, and was supported by Natural Sciences and Engineering Oce Research Council of Canada (grant number 390930-2010 to G.B. and 431802-2012 to C.C.S.), Fonds de Recherche Nature et Technologies
bec (grant number 2010-NC-132906 to G.B. and Master du Que Research scholarship to C.C.S.), Centre National de la Recherche Scientifique, France (to I.C.) and Department of Fisheries and Oceans (to M.O.H.). We gratefully acknowledge H. Allegue, P. Carter, J.-F. Gosselin, B. Paradis, S. Turgeon and J. Van de Walle for their precious contribution to fieldwork. Supplementary Material Supplementary material for this article is available, in the online version, at http://dx.doi.org/10.1016/j.anbehav.2015.04.011. References Acevedo-Gutierrez, A., & Cendejas-Zarelli, S. (2011). Nocturnal haul-out patterns of harbor seals (Phoca vitulina) related to airborne noise levels in Bellingham, Washington, USA. Aquatic Mammals, 37, 167e174. http://dx.doi.org/10.1578/ am.37.2.2011.167. Aubin, T., & Jouventin, P. (2002). How to vocally identify kin in a crowd: the penguin model. Advances in the Study of Behvavior, 31, 243e277. http://dx.doi.org/ 10.1016/s0065-3454(02)80010-9. Ban, S., & Trites, A. W. (2007). Quantification of terrestrial haul-out and rookery characteristics of Steller sea lions. Marine Mammal Science, 23, 496e507. http:// dx.doi.org/10.1111/j.1748-7692.2007.00130.x. Bartholomew, G. A. (1959). Mothereyoung relations and the maturation of pup behaviour in the Alaska fur seal. Animal Behaviour, 7, 163e171.
, C. C. (n.d.). [Distribution of captured harbour seal pups at Bic Beauplet, G., & Sauve
tis, St. Lawrence estuary, 2008e2013]. Unpublished raw Island and Pointe Me data. Beecher, M. D. (1982). Signature systems and kin recognition. American Zoologist, 22, 477e490. Bigg, M. A. (1981). Harbour seals. In S. H. Ridgway, & R. J. Harrison (Eds.), Handbook of marine mammals (Vol. 2, pp. 1e27). London, U.K.: Academic Press. Boness, D. J., & Bowen, W. D. (1996). The evolution of maternal care in pinnipeds. Bioscience, 46, 645e654. Boness, D. J., Bowen, W. D., & Oftedal, O. T. (1994). Evidence of a maternal foraging cycle resembling that of otariid seals in a small phocid, the harbor seal. Behavioral Ecology and Sociobiology, 34, 95e104. http://dx.doi.org/10.1007/ bf00164180. Boness, D. J., Craig, M. P., Honigman, L., & Austin, S. (1998). Fostering behavior and the effect of female density in Hawaiian monk seals, Monachus schauinslandi. Journal of Mammalogy, 79, 1060e1069. http://dx.doi.org/10.2307/1383115. Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H. R. (2009). Introduction to meta-analysis. Hoboken, NJ: J. Wiley. Boulva, J., & McLaren, I. A. (1979). Biology of the harbor seal, Phoca vitulina, in eastern Canada. Bulletin of the Fisheries Research Board of Canada, 200, 1e24. Bowen, W. D. (1991). Behavioural ecology of pinniped neonates. In D. Renouf (Ed.), The behaviour of pinnipeds (pp. 66e127). London, U.K.: Chapman & Hall. Bowen, W. D., Boness, D. J., & Iverson, S. J. (1999). Diving behaviour of lactating harbour seals and their pups during maternal foraging trips. Canadian Journal of Zoology, 77, 978e988. http://dx.doi.org/10.1139/cjz-77-6-978. Briefer, E., & McElligott, A. G. (2011). Mutual mothereoffspring vocal recognition in an ungulate hider species (Capra hircus). Animal Cognition, 14, 585e598. http:// dx.doi.org/10.1007/s10071-011-0396-3. Budaev, S. V., Zworykin, D. D., & Mochek, A. D. (1999). Individual differences in parental care and behaviour profile in the convict cichlid: a correlation study. Animal Behaviour, 58, 195e202. http://dx.doi.org/10.1006/anbe.1999.1124. Burns, J. M. (1999). The development of diving behavior in juvenile Weddell seals: pushing physiological limits in order to survive. Canadian Journal of Zoology, 77, 737e747. Burns, J. J. (2009). Harbor seal and spoted seal, Phoca vitulina and P. largha. In B. Würsig, & J. G. M. Thewissen (Eds.), Encyclopedia of marine mammals. New York, NY: Academic Press. Champagne, F. A., Francis, D. D., Mar, A., & Meaney, M. J. (2003). Variations in maternal care in the rat as a mediating influence for the effects of environment on development. Physiology & Behavior, 79, 359e371. Charrier, I., Aubin, T., & Mathevon, N. (2010). Motherecalf vocal communication in Atlantic walrus: a first field experimental study. Animal Cognition, 13, 471e482. http://dx.doi.org/10.1007/s10071-009-0298-9. Charrier, I., Mathevon, N., & Jouventin, P. (2003). Fur seal mothers memorize subsequent versions of developing pups' calls: adaptation to long-term recognition or evolutionary by-product? Biological Journal of the Linnean Society, 80, 305e312. http://dx.doi.org/10.1046/j.1095-8312.2003.00239.x.

C. C. Sauv
e et al. / Animal Behaviour 105 (2015) 109e120 Charrier, I., Pitcher, B. J., & Harcourt, R. G. (2009). Vocal recognition of mothers by Australian sea lion pups: individual signature and environmental constraints. Animal Behaviour, 78, 1127e1134. http://dx.doi.org/10.1016/ j.anbehav.2009.07.032. Clark, C. W., Marler, P., & Beeman, K. (1987). Quantitative-analysis of animal vocal phonology: an application to swamp sparrow song. Ethology, 76, 101e115. Cordes, L. S., Duck, C. D., Mackey, B. L., Hall, A. J., & Thompson, P. M. (2011). Longterm patterns in harbour seal site-use and the consequences for managing protected areas. Animal Conservation, 14, 430e438. http://dx.doi.org/10.1111/ j.1469-1795.2011.00445.x. Cunningham, L., Baxter, J. M., Boyd, I. L., Duck, C. D., Lonergan, M., Moss, S. E., et al. (2009). Harbour seal movements and haul-out patterns: implications for monitoring and management. Aquatic Conservation: Marine and Freshwater Ecosystems, 19, 398e407. Dehnhardt, G. (2002). Sensory systems. In R. Hoelzel (Ed.), Marine mammal biology. An evolutionary approach (pp. 116e141). Oxford, U.K.: Blackwell.
, Y., Hammill, M. O., & Barette, C. (2003). Pup development and timing of Dube pupping in harbour seals (Phoca vitulina) in the St Lawrence River estuary, Canada. Canadian Journal of Zoology, 81, 188e194. Dwyer, C. M., & Lawrence, A. B. (2000). Maternal behaviour in domestic sheep (Ovis aries): constancy and change with maternal experience. Behaviour, 137, 1391e1413. Evans, W. E., & Bastian, J. (1969). Marine mammal communication: social and ecological factors. In H. T. Anderson (Ed.), The biology of marine mammals (pp. 425e476). New York, NY: Academic Press. Fairbanks, L. A. (1996). Individual differences in maternal style: causes and consequences for mothers and offspring. Advances in the Study of Behavior, 25, 579e611. Fisher, K. I., & Stewart, R. E. A. (1997). Summer foods of Atlantic walrus Odobenus rosmarus rosmarus in Foxe Bassin, Northwest Territories. Canadian Journal of Zoology, 75, 1166e1175. Fox, J., & Weisberg, S. (2011). car: An R companion to applied regression (2nd ed.). Thousand Oaks, CA: Sage. Hair, J. F. (1995). Multivariate data analysis with readings (4th ed.). Upper Saddle River, NJ: Prentice Hall. Haller, M. A., Kovacs, K. M., & Hammill, M. O. (1996). Maternal behaviour and energy investment by grey seals (Halichoerus grypus) breeding on land-fast ice. Canadian Journal of Zoology, 74, 1531e1541. http://dx.doi.org/10.1139/z96-167. Hammill, M. O., Lydersen, C., Ryg, M., & Smith, T. G. (1991). Lactation in the ringed seal Phoca hispida. Canadian Journal of Fisheries and Aquatic Sciences, 48, 2471e2476. http://dx.doi.org/10.1139/f91-288. Hastings, K. K., Small, R. J., & Pendleton, G. W. (2012). Sex- and age-specific survival of harbor seals (Phoca vitulina) from Tugidak Island, Alaska. Journal of Mammalogy, 93, 1368e1379. Helantera, H., Martin, S. J., & Ratnieks, F. L. W. (2014). Recognition of nestmate eggs in the ant Formica fusca is based on queen derived cues. Current Zoology, 60, 131e136. Insley, S. J. (1992). Mothereoffspring separation and acoustic stereotypy: a comparison of call morphology in 2 species of pinnipeds. Behaviour, 120, 103e122. http://dx.doi.org/10.1163/156853992x00237. Insley, S. J. (2000). Long-term vocal recognition in the northern fur seal. Nature, 406, 404e405. http://dx.doi.org/10.1038/35019064. Insley, S. J., Phillips, A. V., & Charrier, I. (2003). A review of social recognition in pinnipeds. Aquatic Mammals, 29, 181e201. http://dx.doi.org/10.1578/ 016754203101024149. Jenssen, B. M., Asmul, J. I., Ekker, M., & Vongraven, D. (2010). To go for a swim or not? Consequences of neonatal aquatic dispersal behaviour for growth in grey seal pups. Animal Behaviour, 80, 667e673. http://dx.doi.org/10.1016/ j.anbehav.2010.06.028. Job, D. A., Boness, D. J., & Francis, J. M. (1995). Individual variation in nursing vocalizations of Hawaiian monk seal pups, Monachus schauinslandi (Phocidae, Pinnipedia), and lack of maternal recognition. Canadian Journal of Zoology, 73, 975e983. http://dx.doi.org/10.1139/z95-114. Jones, I. L., Falls, J. B., & Gaston, A. J. (1987). Vocal recognition between parents and young of ancient murrelets, Synthliboramphus antiquus (Aves, Alcidae). Animal Behaviour, 35, 1405e1415. http://dx.doi.org/10.1016/s0003-3472(87)80013-1. Jorgensen, C., Lydersen, C., Brix, O., & Kovacs, K. M. (2001). Diving development in nursing harbour seal pups. Journal of Experimental Biology, 204, 3993e4004. Khan, C. B., Markowitz, H., & McCowan, B. (2006). Vocal development in captive harbor seal pups, Phoca vitulina richardii: age, sex, and individual differences. Journal of the Acoustical Society of America, 120, 1684e1694. http://dx.doi.org/ 10.1121/1.2226530. Khanna, H., Gaunt, L. L., & McCallum, D. A. (1997). Digital spectrographic crosscorrelation: tests of sensitivity. Bioacoustics, 76, 101e115. Knudtson, P. (1998). The world of the walrus. San Francisco, CA: Sierra Club Books. Kovacs, K. M., & Lavigne, D. M. (1992). Maternal investment in otariid seals and walruses. Canadian Journal of Zoology, 70, 1953e1964. Kroodsma, D. E. (1989). Inappropriate experimental designs impede progress in bioacoustic research: a reply. Animal Behaviour, 38, 717e719. Lesage, V., Hammill, M. O., & Kovacs, K. M. (1995). Harbour seal (Phoca vitulina) and grey seal (Halichoerus grypus) abundance in the St Lawrence estuary. Canadian Techincal Report of Fisheries and Aquatic Sciences, 2613. iiie19. Lydersen, C., & Kovacs, K. M. (1993). Diving behaviour of lactating harp seal, Phoca groenlandica, females from the Gulf of St Lawrence, Canada. Animal Behaviour, 46, 1213e1221. http://dx.doi.org/10.1006/anbe.1993.1312.

119

Maestripieri, D. (1993). Maternal anxiety in rhesus macaques (Macaca mulatta): II. Emotional beases of individual differences in mothering style. Ethology, 95, 32e42.
Mazerolle, M. (2013). I. Estim
es, param etres et r
e
echantillonnage [R script]. Universite
bec en Abitibi Te
miscamingue, Institut de Recherche sur les Fore ^ts. du Que Mbu Nyamsi, R. G., Aubin, T., & Bremond, J. C. (1994). On the extraction of some time dependent parameters of an acoustic signal by means of the analytic signal concept. Its application to animal sound study. Bioacoustics, 5, 187e203. http:// dx.doi.org/10.1080/09524622.1994.9753244. McCulloch, S., & Boness, D. J. (2000). Motherepup vocal recognition in the grey seal (Halichoerus grypus) of Sable Island, Nova Scotia, Canada. Journal of Zoology, 251, 449e455. http://dx.doi.org/10.1111/j.1469-7998.2000.tb00800.x. McCulloch, S., Pomeroy, P. P., & Slater, P. J. B. (1999). Individually distinctive pup vocalizations fail to prevent allo-suckling in grey seals. Canadian Journal of Zoology, 77, 716e723. http://dx.doi.org/10.1139/cjz-77-5-716. McGregor, P. K., Catchpole, C. K., Dabelsteen, T., Falls, J. B., Fusani, L., Gerhardt, H. C., et al. (1992). Design of playback experiments: The Thornbridge Hall NATO ARW consensus. Playback and studies of animals communication, Series A: Life Sciences (Vol. 228). New York, NY: Plenum Press. Miller, E. H. (1991). Communication in pinnipeds, with special reference to non acoustic signalling. In D. Renouf (Ed.), The behaviour of pinnipeds (pp. 128e235). London, U.K.: Chapman & Hall. Nowak, R. M. (2003). Walker's marine mammals of the world. Baltimore, MD: Johns Hopkins University Press. Numan, M. (1994). Maternal behaviour. In E. Knobil, & J. D. Neill (Eds.), The physiology of reproduction (pp. 221e302). New York, NY: Raven. Nyakatura, K., & Bininda-Emonds, O. R. P. (2012). Updating the evolutionary history of Carnivora (Mammalia): a new species-level supertree complete with divergence time estimates. BMC Biology, 10, 12. http://dx.doi.org/10.1186/1741-700710-12. Perry, E. A., & Renouf, D. (1988). Further studies of the role of harbor seal (Phoca vitulina) pup vocalizations in preventing separation of mother pup pairs. Canadian Journal of Zoology, 66, 934e938. Petrinovich, L. (1974). Individual recognition of pup vocalizations by northern elephant seal mothers. Zeitschrift für Tierpsychologie, 34, 308e312. Pike, T. W. (2011). Egg recognition in Japanese quail. Avian Biology Research, 4, 231e236. http://dx.doi.org/10.3184/175815511x13228268534598. Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D., & R Development Core Team. (2012). nlme: Linear and nonlinear mixed effects models. Vienna, Austria: R Foundation for Statistical Computing. Pitcher, B. J., Harcourt, R. G., & Charrier, I. (2012). Individual identity encoding and environmental constraints in vocal recognition of pups by Australian sea lion mothers. Animal Behaviour, 83, 681e690. http://dx.doi.org/10.1016/ j.anbehav.2011.12.012. Pitcher, B. J., Harcourt, R. G., Schaal, B., & Charrier, I. (2011). Social olfaction in marine mammals: wild female Australian sea lions can identify their pup's scent. Biology Letters, 7, 60e62. http://dx.doi.org/10.1098/rsbl.2010.0569. Plush, K. J., Hebart, M. L., Brien, F. D., & Hynd, P. I. (2011). The genetics of temperament in Merino sheep and relationships with lamb survival. Applied Animal Behaviour Science, 134, 130e135. R Development Core Team. (2011). R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. Reale, D., Gallant, B. Y., Leblanc, M., & Festa-Bianchet, M. (2000). Consistency of temperament in bighorn ewes and correlates with behaviour and life history. Animal Behaviour, 60, 589e597. http://dx.doi.org/10.1006/anbe.2000.1530. Reiter, J., Stinson, N. L., & Leboeuf, B. J. (1978). Northern elephant seal development: transition from weaning to nutritional independence. Behavioral Ecology and Sociobiology, 3, 337e367. http://dx.doi.org/10.1007/bf00303199. Renouf, D. (1984). The vocalization of the harbor seal pup Phoca vitulina and its role in the maintenance of contact with the mother. Journal of Zoology, 202, 583e590. Renouf, D. (1985). A demonstration of the ability of the harbor seal Phoca vitulina to discriminate among pup vocalizations. Journal of Experimental Marine Biology and Ecology, 87, 41e46. http://dx.doi.org/10.1016/0022-0981(85)90190-x. Riedman, M. (1990). The pinnipeds: Seals, sea lions, and walruses. Berkeley, CA: University of California Press. Robillard, A., Lesage, V., & Hammill, M. O. (2005). Distribution and abundance of harbour seals (Phoca vitulina concolor) and grey seals (Halichoerus grypus) in the estuary and Gulf of St Lawrence. 1994e2001. Canadian Technical Report of Fisheries and Aquatic Sciences, 2613, 1e152.
, C. C., Beauplet, G., Hammill, M. O., & Charrier, I. (in press). Acoustic analysis Sauve of airborne, underwater, and amphibious mother attraction calls by wild harbour seal pups (Phoca vitulina). Journal of Mammalogy. http://dx.doi.org/10. 1093/jmammal/gyv064. Sayigh, L. S., Tyack, P. L., Wells, R. S., Solow, A. R., Scott, M. D., & Irvine, A. B. (1999). Individual recognition in wild bottlenose dolphins: a field test using playback experiments. Animal Behaviour, 57, 41e50. http://dx.doi.org/10.1006/ anbe.1998.0961. Schulz, T. M., & Bowen, W. D. (2004). Pinniped lactation strategies: evaluation of data on maternal and offspring life history traits. Marine Mammal Science, 20, 86e114. http://dx.doi.org/10.1111/j.1748-7692.2004.tb01142.x. Stewart, R. E. A., & Fay, F. H. (2001). Walrus. In D. Macdonald (Ed.), The new encyclopedia of mammals (pp. 174e179). Oxford, U.K.: Oxford University Press. Stirling, I. (1970). Observations on the behavior of the New Zealand fur seal (Arctocephalus forsteri). Journal of Mammalogy, 51, 766e778.

120

C. C. Sauv
e et al. / Animal Behaviour 105 (2015) 109e120

Sueur, J., Aubin, T., & Simonis, C. (2008). Seewave, a free modular tool for sound analysis and synthesis. Bioacoustics, 18, 213e226. Suryan, R. M., & Harvey, J. T. (1999). Variability in reactions of Pacific harbor seals, Phoca vitulina richardsi, to disturbance. Fishery Bulletin, 97, 332e339. Thompson, P. M., Tollit, D. J., Wood, D., Corpe, H. M., Hammond, P. S., & Mackay, A. (1997). Estimating harbour seal abundance and status in an estuarine habitat in north-east Scotland. Journal of Applied Ecology, 34, 43e52. http://dx.doi.org/ 10.2307/2404846. Trillmich, F. (1981). Mutual motherepup recognition in Galapagos fur seals and sea lions: cues used and functional significance. Behaviour, 78, 21e42. http:// dx.doi.org/10.1163/156853981x00248. Trimble, M., & Insley, S. J. (2010). Mothereoffspring reunion in the South American sea lion Otaria flavescens at Isla de Lobos (Uruguay): use of spatial, acoustic and olfactory cues. Ethology, Ecology and Evolution, 22, 233e246. http://dx.doi.org/ 10.1080/03949370.2010.502318. Van Opzeeland, I. C., Van Parijs, S. M., Frickenhaus, S., Kreiss, C. M., & Boebel, O. (2012). Individual variation in pup vocalizations and absence of behavioral signs of maternal vocal recognition in Weddell seals (Leptonychotes weddellii). Marine Mammal Science, 28, E158eE172. http://dx.doi.org/10.1111/j.17487692.2011.00505.x.

Van de Walle, J. (2013). De la naissance au sevrage: influence des conditions environnementales et des caract
eristiques individuelles chez le phoque commun (Phoca
Laval. vitulina) du Saint-Laurent (M.Sc. Dissertation). Universite Weinstein, T. A. R., Capitanio, J. P., & Gosling, S. D. (2008). Personality in animals. In O. P. John, R. W. Robins, & L. A. Pervin (Eds.), Handbook of personality: Theory and research (3rd ed.). (pp. 328e350). New York, NY: Guilford Press. Wheatley, K. E., Bradshaw, C. J. A., Harcourt, R. G., & Hindell, M. A. (2008). Feast or famine: evidence for mixed capital-income breeding strategies in Weddell seals. Oecologia, 155, 11e20. http://dx.doi.org/10.1007/s00442-007-0888-7. Wiley, R. H., & Richard, D. G. (1982). Adaptations for acoustic communication in birds: sound transmission and signal detection. In D. E. Kroodsma, & E. H. Miller (Eds.), Acoustic communication in birds (pp. 131e181). New York, NY: Academic Press. Wilson, D. S., Clark, A. B., Coleman, K., & Dearstyne, T. (1994). Shyness and boldness in humans and other animals. Trends in Ecology & Evolution, 9, 442e446. http:// dx.doi.org/10.1016/0169-5347(94)90134-1. Zeilis, A., & Hothorn, T. (2002). Diagnostic checking in regression relationships. R News, 2, 7e10.