OPEN 3 ACCESS Freely available online © PLOSI o -

Overview on the Diversity of Sounds Produced by Clownfishes (Pomacentridae): Importance of Acoustic Signals in Their Peculiar Way of LifeOrphal Co 11 eye", Eric ParmentierLaboratory of Functional and Evolutionary Morphology, University of Liège, Liège, Belgium

Abstract

Background: Clownfishes (Pomacentridae) are brightly colored coral reef fishes well known for their mutualistic symbiosis with tropical sea anemones. These fishes live in social groups in which there Is a size-based dominance hierarchy. In this structure where sex is socially controlled, agonistic interactions are numerous and serve to maintain size differences between Individuals adjacent In rank. Clownfishes are also prolific callers whose sounds seem to play an important role in the social hierarchy. Here, we aim to review and to synthesize the diversity of sounds produced by clownfishes In order to emphasize the Importance of acoustic signals In their way of life.

Methodology/Principal Findings: Recording the different acoustic behaviors Indicated that sounds are divided Into two main categories: aggressive sounds produced In conjunction with threat postures (charge and chase), and submissive sounds always emitted when fish exhibited head shaking movements (I.e. a submissive posture). Both types of sounds showed size-related Intraspeclflc variation In dominant frequency and pulse duration: smaller Individuals produce higher frequency and shorter duration pulses than larger ones, and inversely. Consequently, these sonic features might be useful cues for individual recognition within the group. This observation Is of significant Importance due to the size-based hierarchy In clownflsh group. On the o ther hand, no acoustic signal was associated with the different reproductive activities.

Conclusions/Significance: Unlike o ther pom acentrus , sounds are not produced for mate attraction In clownfishes but to reach and to defend the competition for breeding status, which explains why constraints are not Important enough for promoting call diversification In this group.

Citation: Colleye O, Parmentier E (2012) Overview on the Diversity of Sounds Produced by Clownfishes (Pomacentridae): Importance of Acoustic Signals in Their Peculiar Way of Life. PLoS ONE 7(11): e49179. doi:10.1371/journal.pone.0049179

Editor: Stephan C. F. Neuhauss, University Zürich, Switzerland

Received August 31, 2012; Accepted October 9, 2012; Published November 7, 2012

Copyright: © 2012 Colleye, Parmentier. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: This research was supported by the Takeda Science Foundation and a 21st Century Center o f Excellence project entitled "The Comprehensive Analyses on Biodiversity in Coral Reef and Island Ecosystems in Asian and Pacific Regions" from the University of the Ryukyus via a Visiting Fellowship, by a grant entitled "Concours des bourses de voyage 2009" from "Ministère d e la Communauté française de Belgique" attributed to OC, and by a grant from Fonds de la Recherche Fondamentale Collective (FRFC) (no. 2.4.535.10.F) delivered by the Belgian National Fund for Scientific Research. OC was supported by a grant from the F.R.S.- FNRS (Bourse de Doctorat F.R.S.- FNRS). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Com peting Interests: The authors have declared that no competing interests exist.

* E-mail: 0.Colleye@ulg.ac.be

Introduction

In teleost fishes, the ability to produce sounds was developed independently in d istant phylogenetic taxa [1], T o date, m ore than 100 fish families include species with the ability to em it sounds [2,3]. T h e m ajority o f acoustic signals are used in different behavioral contexts such as aggressive behavior (territorial defense, p re d a to r/p re y interactions, com petitive feeding) o r reproductive activities (mate identification an d choice, courtship, synchroniza­tion o f gam ete release) [2,4,5]. T h e diversity o f sounds produced by fishes is not as rem arkable as in o ther taxa; m ost fishes show poor am plitude an d frequency m odulation in their sounds [6,7,8] and have relatively lim ited acoustic repertoires. O nly few fish species em it m ore th an one or two distinct sound types. However, the calling characteristics provide sufficient inform ation for species identification and com m unication. For exam ple, the rainbow cichlid Herotilapia multispinosa emits four distinct sound types (thumps, growls, whoofs and volley sounds) that w ould result

from two different sound-producing m echanism s [9]. T he Lusitanian toadfish Halobatrachus didactylus produces the boatwhistle advertisem ent call that is related to the b reeding season and at least three o ther sounds during agonistic encounters: grunts, croaks and double croaks [10,11],

Damselfishes (Pom acentridae) a re one of the best-studied families for the use o f acoustic com m unication during courtship and agonistic interactions, w ith some species such as Dascyllus albisella and D. flavicaudus showing a great diversity an d complexity in their acoustic repertoire. T h ey are know n to produce pulsed sounds during num erous behaviors including signal ju m p , m atin g / visiting, chasing conspecifics and heterospecifics, fighting conspe- cifics an d heterospecifics, an d nest cleaning [12,13]. All these sounds seem to be constructed on the basis o f the same m echanism since they display the same type of sound spectrum and show few differences in term s of pulse duration. O n the o ther hand, differences in the num ber o f pulses and pulse period could be due

PLOS ONE I www.plosone.org 1 Novem ber 2012 | Volume 7 | Issue 11 | e49179

Sound Diversity in Clownfishes

to the fish physiology reflecting the behavior and its m otivational state [13].

Clownfishes (Pom acentridae) live in social groups in w hich there is a size-based dom inance hierarchy [14,15]. W ithin each group, num erous agonistic interactions occur and they ap p ear to play an im portan t role by m aintain ing size differences betw een individuals adjacent in rank [14,15]. Intraspecific encounters are com m on and sometimes ra th e r severe in their intensity. L arger fishes chase smaller ones, w hich m eans that the smallest one is the recipient o f num erous charges [15]. All clownfish species have evolved ritualized th rea t and submissive postures that presum ably serve to circum vent physical injury during intraspecific quarreling [16], For exam ple, the “head shaking” is considered as a submissive state exhibited by fish in reaction to aggressive interactions [16,17,18], This behavior consists in a lateral quivering o f the body that begins a t the head and continues posteriorly.

Clownfishes are know n to p roduce aggressive sounds while displaying charge an d chase tow ards ano ther specim en during agonistic interactions [16,18,19], M ore recently, Colleye et al. [20] conducted further studies on aggressive sounds in the skunk clownfish Amphiprion akallopisos; they highlighted a size-related intraspecific variation in dom inant frequency and pulse duration: smaller individuals p roduce h igher frequency and shorter duration pulses th an larger ones. Surprisingly, the relationship betw een fish size and b o th dom inant frequency and pulse duration is not only species-specific. T hese relationships are also spread out over the entire tribe o f clownfishes by being found am ong 14 different species that are situated on exactly the same slope, w hich m eans the size o f any Amphiprion can be predicted by b o th acoustic features [21].

Besides these aggressive sounds, Schneider [18] docum ented a second type o f sound that was associated with “ head shaking” and was em itted by fishes in conjunction w ith submissive posture. Later, Allen [16] reported the presence o f head shaking m ovem ents associated w ith sound emission during agonistic interactions betw een group m em bers. Lhifortunately, the lack of detailed acoustic da ta an d the small sample sizes o f the behavioral observations require further investigations to differentiate these sounds from aggressive ones and to bette r understand the scope of these acoustic signals w ithin a social group o f clownfishes.

A dditionally, it was reported that clownfishes m ight produce sounds during courtship. C ourtship in clownfishes is generally stereotyped and ritualized, an d is typically accom panied by different activities such as nest cleaning, courtship, spawning and nest care [16], Basically, studies th a t describe the courtship sounds in clownfishes are lim ited in num ber. T o date, sound production during reproductive period has been reported in three clownfish species (A. ocellaris, A. frenatus, A . sandaracinos) by T akem ura [22]. How ever, these observations need to be carefully considered since, according to the au thor, the sounds w ere hardly heard and sometimes they do no t seem to be directly related to spawning behavior [22]. T herefore, deeper a ttention m ust be paid to confirm the im plication o f acoustic signals during reproduction in this group.

T h e present study aims to review and to synthesize the diversity o f sounds p roduced by clownfishes in o rder to em phasize the im portance o f acoustic signals in their way o f life. T h e purpose of this study is 1) to record an d to analyze sounds associated with head shaking in o rder to determ ine their role in the social structure o f clownfishes; 2) to determ ine w hether clownfishes use acoustic signals to synchronize one or several o f their reproductive activities and 3) to m ake further analyses o f some results previously obtained for the aggressive sounds (see [20]) with the aim o f determ ining w hether some acoustic features m ay contribute to individuality.

Materials and Methods

Different species and different types o f da ta were collected in fish tanks an d in the field w ith the aim of covering all the behaviors that could be associated with sound production. E xperim ental and anim al care protocols followed all relevant international guidelines and were approved by the ethics comm ission (no. 728) o f the Lhiiversity o f Liège.

Agonistic soundsT h ree groups being each com posed of four individuals o f

Amphiprion frenatus (S tandard Length, SL: 44 -112 nini) were collected by scuba diving on the fringing reef a round Nakijin village (26"40 'N - 127"59'E ; O kinaw a, Japan) during M ay and Ju n e 2009. All fish were then brough t back w ith their host (Entacmaea quadricolor) to Sesoko Station, T ropical Biosphere R esearch C enter, Lhiiversity o f the Ryukyus w here they were transferred to a com m unity tank (3.5 x 2 .0 x 1.2 m) filled with runn ing seawater at am bient tem perature (28 to 30.5"C). All fish were kept under natural pho toperiod and fed once daily with food pellets ad libitum. T h e social rank o f each individual was attested using size differences. Basically, groups were com posed of a b reeding pa ir and two non-breeders (Table 1).

Recordings were m ade in a smaller glass tank (1 .2 x 0 .5 x 0 .6 m) filled w ith runn ing seawater m aintained a t 28"C by m eans o f a G E X cooler system (type G X C -2 0 1 x , Osaka, Japan) for having standardized conditions. For the sound recordings, all individuals o f a group and their host were first p laced in the tank for an acclim ation tim e o f 2 days. T w enty sessions (each lasting around 45 minutes) were recorded during w hich interactions betw een group m em bers were observed and no ted in o rder to identify the sound em itter. O nly sounds associated with head shaking m ovem ents were taken into account in the analyses because the aim o f this pa rt was to give a concise physical description o f this type o f sounds in o rder to com pare it w ith clownfish aggressive sounds (see [20,21,23]).

Sound recordings were m ade using a Brüel & K jaer 8106 hydrophone (sensitivity: —173 dB re. 1 V /p P a ) connected via a N ex u s™ conditioning amplifier (type 2690) to a T ascam H D -P2 stereo audio recorder (recording bandw ith: 20 H z to20 k H z ± 1 .0 dB). T h i system has a flat frequency response over wide range betw een 7 H z and 80 kH z. T h e hydrophone was placed ju st above the sea anem one (± 5 cm).

Sounds w ere digitized at 44.1 kH z (16-bit resolution) and analysed w ith AviSoft-SAS Lab Pro 4.33 software (1024 point H ann ing w indow ed fast Fourrier transform (FFT)). R ecord ing in small tanks induces potential hazards because o f reflections and tank resonance [24], A relevant equation [24] was thus used to calculate the resonant frequency of the tank, and a low pass filter o f 2.05 kH z was applied to all sound recordings. T em poral

Table 1. Standard length (SL) and size order groups of Amphiprion frenatus.

in the different

Size order SL (mm)

Group 1 Group 2 Group 3

a(female) 105 110 112

ß(male) 76 81 83

y(non-breeder) 63 65 75

S(non-breeder) 44 50 53

doi:10.1371 /journal.pone.0049179.t001

PLOS ONE I www.plosone.org 2 Novem ber 2012 | Volume 7 | Issue 11 | e49179

Sound Diversity in Clownfishes

features were m easured from the oscillograms w hereas frequency param eters were obtained from pow er spectra (filter bandw idth 300 H z, FF T size po in t 256, tim e overlap 96.87% and a flat top window). G enerally speaking, the recorded sounds were com posed o f a series o f sounds being m ultiple-pulsed. T h e following sonic features were m easured: pulse duration in ms, pulse period in ms (the average peak to peak interval betw een consecutive pulse units in a series), num ber o f pulses per sound, sound duration in ms, sound period in ms (the average peak to peak interval betw een consecutive sounds in a train), num ber o f sounds pe r tra in and dom inant frequency in H z (frequency com ponent with the m ost energy). Note th a t sounds p roduced sim ultaneously by several individuals were excluded from acoustic analyses.

Reproductive soundsR ecordings were m ade bo th in aquarium and in the field.In captivity, three species (A. akyndinos, A . melanopus an d A.

ocellaris) reared for several years a t O ceanopolis A quarium in Brest (France) were recorded during Ju ly 2008. O n e m ating pa ir per species was studied. E ach m ating p a ir was m ain tained in separate glass tanks (0 .7 0 x 0 .4 5 x 0 .5 0 m) filled w ith runn ing seawater at am bien t tem perature (26°C). T h e different reproductive activities (nest p reparation , courtship, spawning an d eggs care) were observed and recorded. E ach recording session lasted approxi­m ately 2 h, and four recording sessions with a 1-hour interval were carried ou t by day in o rder to cover the whole daytim e from daw n to dusk. Behaviors o f m ale an d female were observed and noted. In addition, recordings in captivity w ere also m ade in Sesoko Station. O ne m ating p a ir o f A. clarkii was kept in a tank (1.2 x0.5 x0 .6 m) filled with runn ing seawater m ain tained at 28°C. R ecordings were carried ou t during sum m er season 2009 (between M ay and July) because reproduction is lim ited to this period w hen seawater tem peratures are w arm er.

In bo th cases, sound recordings were m ade using the same Brüel & K jaer 8106 hydrophone (see above for details on m aterial characteristics), an d sounds w ere analyzed according to the p rocedure previously described.

Field recordings w ere m ade on the fringing reef in front o f H izuchi beach (26°11 'N - 127°16'E; Akajim a, K eram a Islands, Japan) in August 2009. T hey were m ade using a SO N Y H D D video cam era p laced in a housing (HC3 series) coupled with an external hydrophone (High Tech. Inc.) w ith a flat response of 20 H z to 20 kH z an d a nom inal calibration o f —164 dB re. 1 V / pP a (Loggerhead Instrum ents Inc.). R ecordings were m ade by placing the housing in front o f the inhab ited sea anem one (distance o f betw een 50 cm and 1 m) th a t lived a t a dep th o f betw een 5 m and 10 m. E ach recording session lasted from 1 to 4 h.

Behaviors associated w ith sound production were described and sounds were extracted in .wav files using the AoA audio extractor setup freeware (version 1.2.5). Sounds w ere digitised a t 44.1 kH z (16-bit resolution), low-pass filtered a t 1 kH z an d analysed using AviSoft-SAS Lab Pro 4.33 software (1024 po in t H ann ing w indow ed fast Fourrier T ransform (FFT)). O nly sounds with a good signal to noise ratio w ere included in the analyses.

Statistical analysesC orrelations analyses were used to exam ine changes in all

acoustic features o f submissive sounds across SL. T h e da ta used in these analyses were m ean values o f all recorded sounds for each individual. Tw o statistical analyses w ere th en perform ed to test the influence of social rank on sonic features. First, a full A N C O V A was ru n to test differences betw een social ranks (ß, y, 8; see T able 1) for the sonic variables correlated w ith SL. In this test, sonic variables are considered as variâtes, SL as a covariate and

social rank is the grouping factor. Secondly, sonic variables not correlated w ith SL, w hich failed the test for norm al distribution (Shapiro-W ilk W Test), were analyzed using a non-param etric Kruskal-W allis one-way A N O V A by ranks with subsequent D u n n ’s test for pair-wise com parisons to test differences betw een social ranks. All statistical analyses were carried ou t w ith Statistica 7.1. Results are presented as m eans ± standard deviation (S.D.). Significance level was determ ined a t ƒ><().05.

M ean ± S.D. values were calculated for each acoustic feature o f agonistic sounds for all individuals. O verall m eans, S.D. and range values were subsequently calculated using each individual m ean value for each variable. In o rder to com pare between-individuals with within-individuals variability for each acoustic feature, the within-individuals coefficient o f variance (C.V.W = S .D ./m ean) was calculated and com pared with the betw een-individuals coefficient o f variation (C.V.b). T h e C.V.b was obtained by dividing the overall S.D. by the respective overall m ean. T h e ratio C .V .b / C.V.W was then calculated to obtain a m easure o f relative betw een- individuals variability for each acoustic feature. W hen this ratio assumes values larger th an one, it suggests th a t an acoustic feature could be used as a cue for individual recognition [25,26,27]. Differences betw een individuals for each acoustic variable were tested using a Kruskal-W allis analysis due to the lack of hom ogeneity o f variance. Note th a t this statistical test was run betw een individuals from a same group in the case o f submissive sounds. In addition, this test was also run betw een the 14 individuals o f the skunk clownfish Amphiprion akallopisos for which aggressive sounds were previously recorded (see [20]), in o rder to determ ine if some acoustic features m ay contribute to individu­ality.

Results

Agonistic soundsSubmissive sounds w ere always associated w ith head shaking

m ovem ents (Fig. 1), b u t fish could sometimes carry ou t these m ovem ents w ithout vocalizing. Submissive sounds (jV= 285 sounds analyzed for all individuals o f the different groups; see T able 2) were p roduced w hen subordinates displayed submissive posture as a reaction to charge an d chase by dom inants, w hich m eans that these sounds were never recorded for the dom inan t females (rank 1) during this study. G enerally speaking, submissive sounds are com pletely different from aggressive ones. T hey are always com posed o f several pulses w hereas aggressive sounds are com posed o f a single pulse unit th a t can be em itted alone o r in series (Fig. 2). T hey also exhibit shorter pulse periods and shorter pulse durations than aggressive sounds. In A. frenatus, submissive sounds can be p roduced alone or in series (2-9 sounds, 3 .0±0.56), and are m ultiple-pulsed (2-6 pulses, 3 .2±0 .26). Pulse period averaged 1 1 .8± 2 .4 ms an d pulse duration ranged from 4.7 to 10.3 ms (7 .9± 2 .15 ms). Sound period averaged 197.0 =±26.6 ms and sound duration ranged from 23.5 to 50.6 ms (35 .9± 9 .59 ms). Pulses had peak frequency of 591 ± 115 H z and m ost sound energy ranged from 454 to 778 Hz.

D om inan t frequency an d pulse duration were highly correlated with SL. D om inan t frequency significantly decreased (R = —0.98, j&< 0.0001 ; Fig. 3A) w hereas pulse duration significantly increased (R = 0.98, j6<0.0001; Fig. 3B) with increasing SL. Pulse period was correlated across SL (R = 0.96, j6<0.0001; Fig. 3C) and this sonic variable was also significantly correlated with pulse duration (R = 0.95, p = 0.0001). Additionally, sound duration was correlated with increasing SL (R = 0.97, /)<0.0001 ; Fig. 3E); this acoustic feature being also significantly correlated w ith bo th pulse duration (R = 0.93, p = 0.0003) an d pulse period (r= 0 .9 8 , j&CO.OOOl).

PLOS ONE I www.plosone.org 3 Novem ber 2012 | Volume 7 | Issue 11 | e49179

Sound Diversity in Clownfishes

B

F igu re 1. B ehavioral p o stu r e s a sso c ia te d w ith v o c a liz a t io n s an d ex h ib ite d b y Amphiprion frenatus d u rin g a g o n is t ic in terac­t io n s . A) D o m in an t indiv idual ch asin g su b o rd in a te w hile p ro d u c in g a g g re s s iv e s o u n d s . B) H ead s h a k in g m o v e m e n ts d is p la y e d by s u b o rd in a te w h ile p ro d u c in g su b m iss iv e s o u n d s in re a c tio n to ag g re ss iv e a c t by d o m in a n t. N ote th a t w iggly lines in d ica te th e s o u n d -p ro d u c in g indiv idual, an d arrow s p o in t o u t th e receiver o f th e ag g re ss iv e act.

Likewise, sound period was correlated with increasing SL (R = 0.86, p = 0.0031; Fig. 3F), being significantly correlated with sound duration (i? = 0.82, p = 0.0063). T he num ber o f pulses per sound did no t change significantly (R = 0.10, p = 0.7917; Fig. 3D) across SL, as well as the num ber o f sounds pe r train (i?= 0 .06 , p = 0.8686; Fig. 3G).

A com parison o f social rank values using SL as a covariate showed that the dom inant frequency (ANC OV A, test for com m on slopes: F,2 3) = 3.677, p = 0.156; test for intercepts: F,25)= 1.204, p = 0.374) an d pulse duration (AN COVA , test for com m on slopes: F,2 3) = 3.644, p = 0.175; test for intercepts: F,2 5) = 4.204, p = 0.137) did not differ betw een individuals o f different social ranks. T hereby, differences betw een social ranks in these acoustic features exclusively resulted from size differences. In addition, the influence o f fish size on acoustic features was enhanced by com paring them betw een individuals o f the same social rank bu t from different groups. All the acoustic features were significantly different betw een individuals o f the same rank (Table 3), except w hen these ones had similar SL. In this case, acoustic features did not differ (D unn’s test, />>0.05).

K ruskal-W allis one-way A N O V A revealed th a t m eans were significantly different betw een social ranks for pulse period (H = 6.489, d f = 2 , p = 0.0390) and sound duration (H = 7.200, df= 2, p = 0.0273), bu t not for sound period (H = 3.822, df= 2, p = 0.1479), num ber o f pulses pe r sound (H = 1.898, df= 2, p = 0.3871) and num ber o f sounds pe r train (H = 2.508, df= 2, p = 0.2853).

Pairwise com parisons showed that pulse period an d sound duration were h igher in rank 2 (D unn’s test, /><0.05; T able 2) than in rank 4, w hereas no significant differences were observed betw een ranks 2 an d 3 (D unn’s test, />>0.05; T able 2), and betw een ranks 3 an d 4 (D unn’s test, />>0.05; T able 2) due to considerable overlap.

Aggressive and submissive sounds presented some acoustic features that displayed C .V .W 0 .1 0 (Tables 4, 5), suggesting a

strong hom ogeneity o f these variables. All the acoustic features analyzed had C.V.b/C.V .w r a tio s > l , showing a h igher variability am ong th an w ithin individuals. Consistently, the Kruskal-W allis analyses revealed significant differences am ong individuals for alm ost all features (Tables 4, 5), indicating that these acoustic variables (except the num ber o f pulses pe r sound and the num ber o f sounds pe r train in groups 1 and 2 o f A . frenatus', T ab le 5) can potentially provide recognition cues to identify the sound em itter. T h e larger relative betw een-individuals variability (larger C.V.b/ C.V.W ratios) corresponded to the dom inant frequency, pulse d uration and pulse period (Tables 4, 5).

Reproductive soundsA total o f eight spawning events were observed. All reproductive

patterns including nest p reparation , courtship, spawning and paren tal care were once observed an d recorded in -4. akindynos, -4. melanopus and -4. percula during Ju ly 2008 a t O ceanopolis A quarium . Amphiprion clarkii spawned four times betw een M ay and Ju ly 2009 in Sesoko Station, an d all the reproductive activities were observed an d recorded. Spaw ning always occurred in the afternoon from 2:00 to 5:00 p .m ., w hatever the species. In addition, one com plete spawning sequence in -4. perideraion was observed and recorded in the field for approxim ately 80 m inutes in August 2009 (11:20 to 12:40 a.m.).

Overall, the m ost striking observation was the com plete absence o f sound production th roughout all activities o f the reproductive period in the different clownfish species recorded, and w hatever the recording environm ent (aquarium or field). How ever, some other typical behaviors seem to be responsible for the synchroni­zation o f the reproductive activities.

1. T h e arrival o f spawning period was indicated by an increase o f cleaning activity by the m ale, and by the belly o f the female that was noticeably distended (especially in -4. percula and -4. perideraion). T hese features becam e usually distinct three o r four days before spawning. Occasionally, fish chased each ano ther o r engaged in fast side-by-side swim m ing an d belly touching; the female being the initiator in m ost o f these encounters. Sometimes, the female entered the nest and pressed her belly against the rock (spawning ground). Pecking m ovem ent o f the m ale at the surface o f nest becam e m ore vigorous from about two hours before spawning; this m ovem ent was continued until ju st before spawning.

A bout 15 m inutes before spawning, nest-cleaning activities was m ore rigorously carried ou t by the female, w hich pecked the surface o f the spawning ground. Also, she pressed her belly against the substrate. These activities seem ed to aim at m aking sure o f the com pletion o f the spawning ground. At the onset o f spawning, the whitish cone-shaped ovipositor o f the female was clearly apparen t.

2. T h e spawning was carried out in the following way: the female entered the nest, pressed her belly against the spawning ground and swani slowly in a circular path; the m ale followed closely beh ind an d fertilized the spawn. L ocom otion during the spawning passes was achieved by rap id fluttering of the pectoral fins. T h e m ale frequently m outhed the eggs during the spawning period. Both fish also nibbled on the tips o f the anem one tentacles for preventing the spawn from entering into contact w ith them .

3. After the spawning, the incubation period took place an d was characterized by paren tal care, w hich lasted usually six to seven days. T h e m ale assum ed nearly the full responsibility o f tending the nest. Except an initial m oderate level o f activity a t spawning, there were no cleaning activities the first two days. T hen , an ab rup t increase occurred the next few days until hatching. Two basic nest-caring behavior patterns were observed. Fanning was the m ost com m on an d was m ainly perform ed by fluttering the pectoral fins. M outh ing the eggs an d substrate biting a t the

PLOS ONE I www.plosone.org 4 Novem ber 2012 | Volume 7 | Issue 11 | e49179

Sound Diversity in Clownfishes

1—1

-12

NI

> ,o3<D3CT"<L>I-

[J-.

----- >

¿ i ' í í l ü in i'Ü 'tP lii t i l lin in 1 1 lm U nJW

iL m i ni i i'nM !m 'lii' b i l l i

i l « i i

ii Wl 'I W lí1 if iUii1 Ur Ei I I è i i l i20 60 100 140 180 220 260 300 340 380 420 460 500 540

Time (ms)

B>■S 80

^ 40

&Enj<L>>A-tnj"ÖJ

-40

-80

— i 1 - -------------------------------------------i

HN

M 3

o 2 C3 1e r2

tuV W

20 60 100 140 180 220 260 300 340 380 420

Time (ms)

F igu re 2 . E xam p le o f a g o n is t ic so u n d s p ro d u ced b y A m p h ip rio n fre na tu s d u rin g in te ra ctio n s . A) O scillogram (top) an d sp ec tro g ram (b o tto m ) o f subm iss ive s o u n d s p ro d u c e d by s u b o rd in a te d u rin g h e a d shak ing m o v e m e n ts . B) O scillogram (top) an d sp ec tro g ra m (b o tto m ) o f ag g re ss iv e s o u n d s p ro d u c e d by d o m in a n t w hile d isp lay ing c h arg e an d chase . N ote th e d ifferen ces in (1) p u lse d u ra tio n an d (2) p u lse p e rio d . The aco u s tic v ariab le m e a su re d in (3) re p re s e n ts th e s o u n d d u ra tio n in th e case o f subm iss ive so u n d s , an d th e tra in d u ra tio n in th e case o f ag g ress iv e so u n d s . The co lo u r sca le c o rre sp o n d s to th e in ten sity a sso c ia ted w ith th e d iffe ren t freq u e n c ie s. do i:10 .1371/jo u rn a l.p o n e .0 0 4 9 1 7 9 .g 0 0 2

periphery o f the nest were also exhibited. D ead eggs were regularly rem oved as indicated by bare patches a t the nesting surface.

Discussion

In clownfishes, different types o f sounds such as “th reaten ing” and “ shaking” [18], “ click” and “g run t” [16] o r “p o p ” and “ ch irp” [19,23] have already been described during interactions

PLOS ONE I www.plosone.org 5 Novem ber 2012 | Volume 7 | Issue 11 | e49179

Tabl

e 2.

Sum

mar

y (m

ean

± S.D

.) of

the

seve

n ac

oust

ic

feat

ures

an

alyz

ed

from

subm

issi

ve

soun

ds

prod

uced

by

9 A

mph

iprio

n fre

natu

s.

Sound Diversity in Clownfishes

ql/i

<un £ .£ 3 re Z £

q1/1

O' à O'

betw een conspecifics. A lthough a dichotom y in sounds was reported in each case, these term s have been inconsistently applied and w ere no t always supported by appropria te da ta which create confusion [10], C onsequently, it rem ained difficult to m atch a type o f sound w ith a given behavior. H ow ever, our multiple observations highlight that submissive sounds (i.e. chirps, see [23]) are clearly different from aggressive sounds (i.e. pops, see [23]). Aggressive sounds are m ainly p roduced by dom inants during chases an d th rea t displays betw een conspecifics [20], whereas submissive sounds are always em itted w hen subordinates exhibit h ead shaking m ovem ents in reaction to aggressive displays by h igher-ranking individuals. Therefore, bo th types o f sounds seem to be an integral p a rt o f the agonistic behavior in clownfishes. G iven th a t they present some differences in sound spectra and shape o f the tem poral envelope, it is im portan t to em phasize that these two types o f sounds w ould result from two different m echanism s. Aggressive sounds result from jaw teeth snapping [21,28] bu t the sound-producing m echanism of submissive sounds is still unknown.

■o■o

qui

Qui

K.0 T3

I I

qi/i+i

q

£ vT0 p n S

qi/i+1

■Ö>

<u 2fC I"-. 3 t—

c 8CUro c 3 o 13 Q- "rö ~o C

.£ 3<4- Oo

Importance o f size-related acoustic signals for the group hierarchy

Interestingly, dom inant frequency and pulse duration o f submissive sounds display size-related variation in some acoustic features. T h e m ore fish size increases, the m ore dom inant frequency decreases, an d the m ore pulse duration increases. T he same relationships have already been found for aggressive sounds am ong 14 different species [21]. Differences in bo th sound characteristics w ere related to fish size and no t to sexual status or social rank. How ever, size, sex and social rank are extrem ely related to each o ther due to the size-based h ierarchy within each group [14,15]. In -4. percula, Buston and C ant [29] dem onstrated that individuals adjacent in rank are separated by body size ratios whose distribution is significantly different from the distribution expected under a null m odel: the grow th of individuals is regulated such that each dom inant ends up being about 1.26 times the size o f its im m ediate subordinate. T h e same kind o f ratio (—1.30) is observed in the different groups o f A . frenatus o f this study (Fig. 4). T h e respect o f this ratio w ithin groups highlights that dom inant frequency and pulse duration can be signals conveying inform a­tion on the social rank of the em itter w ithin the group.

Aggressive an d submissive sounds are involved in interactions betw een group m em bers (Video S I, S2). Being associated with a specific display, they m ight have a different function w ithin the group. Indeed, aggressive sounds could possess a deterrent function by giving a rem inder signal o f dom inance during interactions w hereas submissive sounds could possess an appease­m ent function by expressing the lower rank status during interactions. Likewise, two different types o f sounds are em itted by the grey gu rn ard Eutrigla gurnardus depending on the in terac­tions betw een individuals: knocks are p roduced during low levels o f aggression and grunts m ainly while perform ing frontal displays to opponents [30],

C lear differences were found am ong aggressive and submissive sounds a ttribu ted to different individuals. All acoustic variables were significantly m ore variable betw een than w ithin individuals and thus could all potentially provide cues to identify individuals. Furtherm ore, the m ost im portan t variables to allow individual identification w ere dom inan t frequency and pulse duration for b o th types o f sounds (Tables 4, 5). Pulse period, in a lesser extent, was also consistently im portan t for discrim inating am ong individ­uals in the case o f submissive sounds (Table 5). In o rder to be good candidates for individual recognition, these acoustic features should propagate th rough the environm ent, an d should be

PLOS ONE I www.plosone.org 6 Novem ber 2012 | Volume 7 | Issue 11 | e49179

Sound Diversity in Clownfishes

A

¡3 so°" > .g 700- <l>IÖ 600-

eö

on 400-

B

-öoua

3d .

16-

15-

14-

13-

1 2 -

11-

1 0 -

9-

SL (nun)

i -------------r60 70

SL (mm)

5 0 -

45_c.2 4 ° - 'S

I 35‘'S 30-

SL (mm)

t» 2 -

2 4 0 -

SL (nun)

a 220-T3 ■§ 200- wP-

’S 180- G3

(S) 160-

SL (mm) SL (mm)

3.5-

c

Bsl

T3 C

§« 3.0H q-Hoo

X)EZ

2 .5 -

” T“50

□ A

” I“60

~~r~70

” T“80

SL (mm)

"T”90

F igu re 3 . In flu en ce o f fish s iz e (SL) o n a c o u stic fe a tu r e s o f su b m iss iv e so u n d s in A m p h ip rio n frenatus. C orre la tions o f (A) d o m in a n t freq u en cy , (B) pu lse d u ra tio n , (C) p u lse p e rio d , (D) n u m b e r o f p u lses p e r so u n d , (E) s o u n d d u ra tio n , (F) s o u n d p e rio d an d (G) n u m b e r o f s o u n d s p e r tra in a g a in s t SL. Fish ra n g e d from 44 to 112 m m in SL (N = 9). R esults a re ex p re ssed as m ean va lu es o f all re c o rd e d pu lses fo r e ach individual (Ö = rank 4, □ = rank 3, A = rank 2). do i:10 .1371/jo u rn a l.p o n e .0 0 4 9 1 7 9 .g 0 0 3

PLOS ONE I www.plosone.org 7 Novem ber 2012 | Volume 7 | Issue 11 | e49179

Sound Diversity in Clownfishes

Table 3. Compa sounds between different groups

ris o n o f t h e a c o u s t ic f e a tu r e s in d iv id u a ls o f t h e s a m e s o c ia o f A m p h ip r io n frena tus .

o f s u b m is s iv e ra n k b u t f ro m

Acoustic variables Social rank H p-value

Dominant frequency (Hz) 2 108.4 <0.0001

3 42.11 <0.0001

4 49.54 <0.0001

Pulse duration (ms) 2 79.16 <0.0001

3 37.72 <0.0001

4 139.0 <0.0001

Pulse period (ms) 2 121.5 <0.0001

3 35.13 <0.0001

4 64.19 <0.0001

Sound duration (ms) 2 6.418 0.0404

3 15.18 0.0005

4 10.27 0.0059

Sound period (ms) 2 9.995 0.0068

3 22.80 <0.0001

4 12.31 0.0021

H values are the result of the Kruskal-Wallis test (df= 2, n = pulses analyzed.doi:10.1371 /journal.pone.0049179.t003

300). n, number of

detected by the receiver. Sound propagation in shallow w ater can result in signal degradation over short distances, including sound pressure level and frequency attenuation , and sound duration loss [31]. How ever, the effect o f environm ental a ttenuation and signal degradation should no t impose a m ajor restriction w ithin a group o f clownfishes since all individuals inhab it a restricted territory (the sea anem one) and spend m ost o f the tim e in close vicinity o f their host.

Since dom inant frequency an d pulse duration o f bo th aggressive and submissive sounds are size-dependent, tem poral an d spectral inter-individual differences m ight be detected by m em bers within the group. Teleost fishes such as Gobius niger (Gobiidae) and Sparus annularis (Sparidae) are able to discrim inate tonal sounds differing in frequency o f approxim ately 10%; the frequency discrim ination ability at 400 H z is approxim ately 40 H z [32]. In the clownfish el. akallopisos, aggressive sounds em itted by non-breeders, m ales and females differ in dom inant frequency by > 1 0 % [20], In A. frenatus, sounds em itted by individuals o f different social ranks also differ in dom inant frequencies by > 1 0 % (Table 2). Such an ability to discrim inate frequency differences has a lready been observed in

some pom acentrids. In the damselfish Abudefduf saxatilis, fish size has a significant effect on auditory sensitivity [33] : all fish are most sensitive to the lower frequencies (10CM-00 Hz) bu t the larger ones are m ore likely to respond to h igher frequencies (1000-1600 Hz). T h e effect o f fish size on hearing abilities was also supposed in three different clownfish species [34], A lthough the best hearing sensitivity is a round 100 H z, small individuals were m ore sensitive to a larger frequency interval (100-450 Hz), an d thus they are m ore sensitive to the frequencies em itted by larger conspecifics.

No inform ation related to fish size can be extracted from num ber o f pulses pe r sound or num ber o f sounds pe r train. Differences in these acoustic features appear to be related to a difference in m otivation. M otivation is know n for playing a role in damselfishes, regarding their sounds p roduced during aggression. In Dascyllus albisella and D . flavicaudus, aggressive sounds are different according to w hether they are em itted towards conspe­cifics or heterospecifics, being m ultiple-pulsed or single-pulsed, respectively [12,13]. In Pomacentrus partitus, the frequency o f sounds by a territorial resident is relatively low at the territorial border, bu t it rapidly increases as in truder approaches the residence [35]. In the clownfish A. akallopisos, the m ost aggressive males were characterized by a h igher num ber o f pulses per sound an d a shorter pulse period (pers. obs.). T h e smallest individuals (rank 4) in A . frenatus groups em itted the highest num ber o f sounds per train (Table 2). All these variations in acoustic features m ight be related to the willingness to express the position w ithin the group hierarchy. For exam ple, low er-ranking individuals m ight produce m ore submissive sounds in o rder to lim it aggressive acts from dom inants.

No reproduction-related soundUnlike observations m ade by T akem ura [22], no acoustical

behavior was observed during reproductive activities in clown­fishes. M oreover, T ak em u ra’s da ta are som ewhat doubtful since-4. ocellaris, A . frenatus and -4. sandaracinos w ould em it sounds w ith high frequency com ponent o f m ore th an 2 kH z during reproduction [22]. A ccording to hearing sensitivity in clownfishes (A. frenatus, A. ocellaris and -4. clarkii), the frequency range over w hich they can detect sounds is betw een 75 an d 1800 Hz, and they are the most sensitive to frequencies below 200 H z [34], T herefore, this finding raises the question over the interest o f clownfishes to produce sounds they could no t detect during reproductive activities. It rem ains these sounds could ju st be a by-product o f the nest cleaning activities.

In the field, clownfishes spawn on average from 1 .0± 0 .5 to 0 .6 ± 0 .1 times per m onth depending on w hether they live in tropical waters [16,36] or in m ore tem perate regions [37,38,39]. In captivity, the spawning frequency is h igher an d on average 2 5 ± 5 .3 times pe r year [40,41], This frequency was observed at

Table 4. Means, ± S.D., range, within-individuals variability (C.V.W) and between-individuals variability (C.V.b) for the four acoustic features analyzed from aggressive sounds produced by 14 Amphiprion akallopisos.

Acoustic variables Overall Mean ± S.D. (range) C.V.w (Mean) C .V .w (range) c.v.b C .V .b/C .V .w H* />value

Dominant frequency (Hz) 663 ±199 (346-1207) 0.10 0.07-0.21 0.30 2.73 1577 <0.001

Pulse duration (ms) 13.4±3.9 (3.8-22.9) 0.10 0.06-0.18 0.29 2.90 1614 <0.001

Pulse period (ms) 75.9±13.4 (32.4-121.9) 0.15 0.09-0.21 0.23 1.53 203.9 <0.001

Number of pulses per sound 3.9±2.2 (2-14) 0.50 0.30-0.68 0.56 1.12 25.55 0.0195

*Results of Kruskall-Wallis test (df= 13, n = 1818) comparing differences between 14 individuals of A. akallopisos for each acoustic feature. Note that these values were calculated based on acoustic data obtained from Colleye e t al. (2009). n, number of pulses analyzed. doi:10.1371 /journal.pone.0049179.t004

PLOS ONE I w ww.plosone.org Novem ber 2012 | Volume 7 | Issue 11 | e49179

Sound Diversity in Clownfishes

T a b le 5. Means, ± S.D., range, within-individuals variability (CV.W) and between-individuals variability (C.V.b) for t h e seven acoustic features analyzed from submissive sounds p ro d u ced by 9 Amphiprion frenatus.

Acoustic variablesGroupnum ber Overall Mean ± S.D. (range)

C.V.W(Mean) C.V.W (range) C.V.b

C.V.b/C.V.W p-value

593 ±139 (431-862) 0.10 0.08-0.12 0.22 2.20 251.0 < 0.001

Pulse duration (ms) 7.4 ±2.1 (3.8-11.2) 0.10 0.09-0.11 0.28 2.80 232.4 < 0.001

9.1 ±2.8 (4.5-13.9) 0.08 0.06-0.12 0.31 3.87 233.2 < 0.001

11.9±2.5 (7.4-1 7.< 0.10 0.08-0.11 0.21 2.10 167.4 < 0.001

Number of pulses per sound 3.1 ±0.6 (2-5) 0.20 0.19-0.23 0.21 1.05 5.36

3.1 ±0.8 (2-5) 0.25 0.22-0.29 0.27 1.08 12.86 < 0.01

36.9 ±14.1 (16.6-72.4) 0.28 0.22-0.33 0.38 1.36 25.94 < 0.001

Sound period (ms) 176.4±36.9 (92.2-295.3) 0.16-0.21 0.21 16.53 < 0.01

201.7 ±31.5 (118.3-266.3) 0.13 0.08-0.16 0.16 1.23 < 0.01

3.2 ±1.3 (2-7) 0.35 0.30-0.41 0.41 4.93

552 ±105 (345-776) 0.09 0.08-0.11 0.19 239.9 < 0.001

Number of sounds per train 3.3 ±1.5 (2-9) 0.39 0.28-0.53 0.44 4.36

206.3 ±40.9 (135.6-280.9) 0.15 0 .12- 0.1 0.20 1.33 21.07 < 0.001

Sound duration (ms) 29.2 ±8.4 (9.8-48.2) 0.22 0.18-0.24 0.29 1.32 59.24 < 0.001

3.4±0.9 (2-6) 0.26 0.21-0.31 0.27 1.04 2.59

12.7±2.6 (8.2-18.4) 0.08 0.05-0.09 0.21 2.62 176.2 < 0.001

36.6 ±14.1 (15.0-73.8) 0.24 0.13-0.35 0.38 1.58 49.90 < 0.001

Pulse period (ms) 10.7±2.0 (6.7-15.4) 0.10 0.08-0.12 0.19 1.90 143.Í < 0.001

>.5 ±2.8 (3.6-15.1) 0.10 0.09-0.11 0.34 3.40 250.Í < 0.001

Dominant frequency (Hz) 631 ±135 (431-1036) 0.10 0.09-0.15 0.21 1.90 213.7 < 0.001

2.6±0.8 (2-5) 0.23 0.18-0.27 0.30 1.30 14.39 < 0.01

*Results of Kruskal-Wallis test (df= analyzed.doi:10.1371 /journal.pone.0049179.

-2, n = 300) comparing differences between individuals in a group of A. frenatus for each acoustic feature, n, num ber of pulses

t005

Oceanopolis Aquarium with captive clownfishes for which spawning occurred approximately every two weeks. In this context, it could be argued that the captivity modulates some aspects of the behavior [40] such as sound production during reproduction. However, 1) the pair o f A. clarkii reared a t Sesoko Station showed the same spawning frequency (~every 2 weeks), although its reproduction was restricted to summer season and it

was reared under semi-natural conditions (i.e. outdoor tank filled with running seawater and maintained under natural photoperi­od); 2) spawning was witnessed in the field for A. perideraion, and no sound was produced by the mating pair during the reproductive event. Yet, recording of aggressive sounds during the same session supports the fact that the recording material worked well.

A120-

110 -

100-90-80-

6 0 -

50-4 0 -

2 43

Social rank

B

ao

o

sUO

1.4-

c53

1.2 -

2 3

Group

F igu re 4 . Fish s iz e (SL) an d s iz e ra tio s o f in d iv id u a ls a d ja c en t in rank w ith in ea ch g r o u p o f A m p h ip rio n frenatus. A) T he o b se rv ed d is trib u tio n o f fish size (SL) w ith in e ac h g ro u p . B) D istribution o f b o d y size ra tios b e tw e e n indiv iduals a d ja c e n t in rank w ith in each g ro u p . R esults a re e x p re ssed as m ean ± S.D. va lu es ( □ = g r o u p 1, A = g ro u p 2, 0 = g ro u p 3). do i:10 .1371/jo u rn a l.p o n e .0 0 4 9 1 7 9 .g 0 0 4

PLOS ONE I www.plosone.org 9 Novem ber 2012 | Volume 7 | Issue 11 | e49179

Sound Diversity in Clownfishes

O verall, sound production does no t seem to be involved in the reproductive behavior o f clownfishes, w hich m ight be explained by some particular aspects related to their way of life.

T h e reproductive behavior o f pom acentrids is subdivided into the following m ajor categories [42,43]: 1) establishm ent o f territory, 2) selection o f nest site w ithin the territory, 3) p reparation o f the nest site, 4) courtship and pa ir form ation, 5) spawning and fertilization, an d 6) paren tal care. Clownfishes conform to this general p a tte rn b u t are distinctive w ith regards to form ation of perm an en t p a ir bonds th a t usually last for several years in m ost species [16,44], In o ther damselfishes, one m ale m ay m ate with several females during a single spawning episode [42,44], In clownfishes, m ale does no t need to exhibit typical courtship behavior for a ttracting female. Pair-bonding is very strong and is correlated by the small size o f their territories (centered on actinians) th a t is, in turn , correlated with the unusual social h ierarchy existing in each social group. O n the o ther hand , it seems th a t o ther cues such as visual signals m ight be useful for synchronizing reproductive activities. Ju s t before spawning occurs, the female joins the m ale an d becom es m ore insistent in the nest- cleaning activities, p robably in o rder to convey visual cues about its readiness to spawn. Likewise, it is possible th a t the m ale regulates its level o f nest-caring activity in response to visual stimuli received w hen inspecting eggs [16], A visual stimulus o f this sort w ould signal the stage o f egg developm ent an d the need for increased fanning and m outhing activities. Allen [16] experim en­tally dem onstra ted th a t strong agitation o f the eggs is a requisite for hatching. H e also noted th a t there was a p ronounced increase in the am oun t o f m ale nest care on day six o f incubation. O n that day, the em bryos are well developed with one o f the m ost noticeable features being the large eyes w ith their silvery pupils. Such a feature m ight serve as an appropria te visual cue. Therefore, o ther cues such as visual and perhaps chem ical signals m ight be involved in reproductive activities. H ow ever, new behavioral tests w ould need to be run to determ ine the p roper role o f such signals during the reproduction o f clownfishes.

Conclusion

Unlike o ther pom acentrids, sounds are no t p roduced for m ate a ttraction in clownfishes. It is likely an evolutionary outcom e related to their peculiar way o f life: these fishes form small social groups including only one m ating pair, inhabit a restricted territory (the sea anem one), spend m ost o f the tim e in close vicinity o f their host and rarely in teract w ith o ther species on the reef. O n the o ther hand , sounds seem to be im portan t to reach an d to defend the com petition for b reeding status. A lthough they are

References1. S ch n e id e r H (1967) M o rp h o lo g y a n d physio logy o f so u n d -p ro d u c in g m ec h a ­

nism s in teleost fishes. In: T av o lg a W N , ed ito r. M a rin e B io-acoustics, vol. 2. O x fo rd : P e rg am o n Press, p p . 135—158.

2. H aw k in s A D (1993) U n d e rw a te r so u n d a n d fish b eh av io u r. In : P itc h e r T J , ed ito r. B eh av io u r o f teleost fishes. L o n d o n : C h a p m a n a n d H aii. p p . 129—169.

3. S lab b ek o o rn H , B o u to n N , O p z e e la n d IV , C o ers A, C a te C T , e t al. (2010) A noisy spring: th e im p a c t o f g lobally rising u n d e rw a te r so u n d levels o n fish. T re n d s in E co l a n d E vol 25: 4 1 9 -4 2 7 .

4. T a v o lg a W N (1971) S o u n d p ro d u c tio n a n d d e tec tio n . In: H o a r W S , R a n d a ll D J, ed ito rs. F ish physio logy, vol. 5. N ew Y ork: A cad em ic Press, p p . 135—205.

5. W in n H E (1964) T h e b io log ical significance o f fish sounds. In T av o lg a W N , ed ito r. M a rin e B io-A coustics, vol. 2. N e w Y ork: P e rg am o n Press, pp . 213—231.

6. C ra w fo rd J D , C o o k A P , H e b e rle in AS (1997) A coustic co m m u n ica tio n in a n e lectric fish, Pollimyrus isidori (M orm yridae). J C o m p Physiol A 159: 297—310.

7. L ad ich F (1997) A gonistic b eh a v io u r a n d significance o f sounds in vocaliz ing fish. M a r F resh w B eh av Physio l 29: 8 7 -1 0 8 .

8. L ugli M , T o rrice lli P , P a v a n G , M a in a rd i D (1997) S o u n d p ro d u c tio n d u rin g co u rtsh ip a n d sp aw n in g a m o n g fre sh w a te r gobiids (Pisces, G obiidae). M a r F resh w B eh av Physio l 29: 1 0 9 -1 2 6 .

restricted to agonistic interactions only, acoustic signals seem to be an integral p a r t o f their daily behaviors. T h e im plication of acoustic signals in agonistic interactions m ay be an interesting strategy with an econom ic way for preventing conflicts which otherwise m ight escalate to a severe outcom e.

Clownfish sounds can be divided into two m ain categories: aggressive sounds produce in conjunction with th rea t postures (charge an d chase), and submissive sounds always em it when subordinates exhibit head shaking m ovem ents in reaction to aggressive displays by dom inants. Both types o f sounds show intraspecific differences related to fish size, highlighting th a t some acoustic features (i.e. dom inant frequency and pulse duration) m ight be useful cues for individual recognition within the group. T hese observations are o f significant im portance because the social structure o f clownfishes strictly relies on a size-based dom inance hierarchy.

Supporting Information

Video SI Implication o f aggressive sounds during agonistic interactions between group members in the field. N ote th a t a fish is chasing ano ther one (smaller) while producing a series o f aggressive sounds.(AVI)

Video S2 Behavioral posture (head shaking move­ments) exhibited by subordinates while producing submissive sounds. Note th a t fish m ake sounds while doing lateral quivering of the body th a t begins a t the head.(AVI)

AcknowledgmentsThe authors would like to thank Prof. M asaru N akam ura (Sesoko Station, Tropical Biosphere Research Center, University of the Ryukyus, Japan) and Dr. Kenji Iwao (Akajima Marine Science Laboratory, AMSL, Japan) for helping to collect fishes and for providing hospitality and laboratory facilities. We are also indebted R. Suwa, M. Alam, S. Nakam ura, Y. Kojima, Y. Nakajima, R. M urata, R. Nozu and N. Pandit (Sesoko Station, TBRC) for technical support. M any thanks to Dominique Barthélémy (Oceanopolis, France) and Dr. Christian Michel (Aquarium Dubuisson, Belgium) for providing free access to fishes during recordings.

Author ContributionsConceived and designed the experiments: EP O C. Perform ed the experiments: O C. Analyzed the data: O C . C ontributed reagents/ materials/analysis tools: EP. W rote the paper: O C EP.

9. B ro w n D H , M a rs h a llJ A (1978) R e p ro d u c tiv e b eh a v io u r o f th e ra in b o w cichlid , Herotilapia multispinosa (Pisces, C ichlidae). B eh av io u r 67: 299—322.

10. A m o rim M C P (2006) D iversity o f so u n d p ro d u c tio n in fish. In: L ad ich F, Collin SP, M ö lle r P , K a p o o r B G , ed ito rs. C o m m u n ic a tio n in fishes, vol. 1. Enfield: Science Publishers, p p . 71—105.

11. S an tos M d , M o d esto T , M a to s B J , G ro b e r M S , O liv e ira R F , e t al. (2000) S o u n d p ro d u c tio n b y th e lu sitan ian toadfish , Halobatrachus didactylus. B ioacoustics 10: 3 0 9 -3 2 1 .

12. M a n n D A , L o b el PS (1998) A coustic b eh a v io r o f the dam selfish Dascyllus albisella: b e h a v io ra l a n d g eo g rap h ic v a ria tio n . E n v iro n Biol F ish 51: 4 2 1 —4-28.

13. P a rm e n tie r E , K év er L , C asad ev all M , L ecch in i D (2010) D iversity a n d c o m plex ity in th e acoustic b e h a v io u r o f Dascyllus flavicaudus (P om acen tridae). M a r Biol 157: 2 3 1 7 -2 3 2 7 .

14. B usto n P M (2003) Size a n d g ro w th m o d ifica tio n in clow nfish. N a tu re 424: 145— 146.

15. F ricke H W (1979) M a tin g system , reso u rce defense a n d sex ch a n g e in the a n em o n efish Amphiprion akallopisos. Z T ie rp sy ch o l 50: 313—326.

16. A llen G R (1972) T h e A nem onefishes : th e ir C lassification a n d Biology. N ew Jersey : T F H P ub lica tions In c., N e p tu n e C ity . 288 p.

PLOS ONE I www.plosone.org 10 Novem ber 2012 | Volume 7 | Issue 11 | e49179

Sound Diversity in Clownfishes

30. A m o rim M C P , S tra to u d ak is Y, H aw kins A D (2004) S o u n d p ro d u c d o n d u r in g co m p etitiv e feed in g in th e grey g u rn a rd , Eutrigla gurnardus (Triglidae). J F ish Biol 65: 18 2 -1 9 4 .

31. M a n n D A (2006) P ro p a g atio n o f f is h sounds. In : L ad ich F , C o llin SP, M ö lle r P, K a p o o r B G , ed ito rs. C o m m u n ic a tio n in F ishes, vol. 1. Enfie ld : Science P ublishers, p p . 10 7 -1 2 0 .

32. F ay R R (1988) H e a r in g in V er te b ra te s , A P sychophysics D a tab o o k . Illinois: H ill- F ay A ssociates.

33. E g n e r SA, M a n n D A (2005) A u d ito ry sensitivity o f se rg e a n t m ajo r dam selfish Abudefduf saxatilis f ro m p o s t-se td em en t ju v en ile to adu lt. M a r E co l P ro g Ser 285: 2 1 3 -2 2 2 .

34. P a rm e n tie r E , C olleye O , M a n n D A (2009) H e a r in g ab ility in th re e clow nfish species. J E x p Biol 212: 2 0 2 3 -2 0 2 6 .

35. R iggio R (1981) A coustical co rre la tes o f agression in th e b ico lo r dam selfish, Pomacentrus partitus. M .S c. T hesis, U n iversity o f M iam i, M iam i, FL.

36. R oss R M (1978) R e p ro d u c tiv e b eh a v io r o f th e a n em o n efish Amphiprion melanopus o n G u am . C o p e ia 1978: 1 03-107 .

37. M o y er J T , Bell LJ (1976) R e p ro d u c tiv e b e h a v io r o f th e an em o n efish Amphiprion clarkii a t M iyake-Jim a, J a p a n . J p n J Ich th y o l 23: 23—32.

38. O c h i H (1989) M a tin g b e h a v io r a n d sex c h an g e o f th e a n em o n efish , Amphiprion clarkii, in the te m p e ra tu re w aters o f J a p a n . E n v Biol F ish 26: 257—275.

39. R ic h a rd so n D L , H a rr iso n P L , H a r r io t t V J (1997) T im in g o f sp aw n in g a n d fecu n d ity o f a tro p ica l a n d su b tro p ica l a n em o n efish (P o m acen trid ae : Amphiprion) o n a h ig h la titu d e re e f o n the east co ast o f A ustra lia . M a r E co l P ro g Ser 156: 1 7 5 -1 8 1 .

40. G o rd o n A K , B ok A W (2001) F req u en cy a n d p e rio d icity o f sp aw n in g in the c low nfish Amphiprion akallopisos u n d e r a q u a riu m co nd itions. A q u a r iu m Science a n d C o n se rv a tio n 3: 307—313.

41. H o ff F H (1996) C o n d itio n in g , sp aw n in g a n d rea r in g o f fish w ith em phasis o n m arin e clow nfish. D a d e C ity: A q u a c u ltu re C o n su ltan ts, 212 p p .

42. A llen G R (1991) D am selfishes o f th e W o rld . M elle: M erg u s P ublishers. 271p.43. R eese E S (1964) E th o lo g y a n d m arin e zoology. O c e a n o g r M a r B iol A n n u R es 2:

4 5 5 —488.44. F a u tin D G , A llen G R (1992) F ield gu ide to anem onefishes a n d th e ir ho st sea

a n em o n es. P erth : W e ste rn A u s tra lia n M useum . 160 p.

PLOS ONE I w ww.plosone.org 11 Novem ber 2012 | Volume 7 | Issue 11 | e49179

17. F ricke H W (1974) Ö k o -e th o lo g ie des m o n o g am e n A nem onefisches Amphiprion bicinctus (F re iw asseru n tersu ch u n g aus d e m R o te n M eer). Z T ie rp sy ch o l 36: 429— 512.

18. S ch n e id e r H (1964) B ioakustische U n te rsu c h u n g e n a n A n em o n en fisch en d e r G a ttu n g Amphiprion (Pisces). Z M o rp h O k o l T ie re 53: 45 3 —174.

19. C h e n K -C , M o k H - K (1988) S o u n d P ro d u c tio n in the A nem onfishes, Amphiprion clarkii a n d A. frenatus (P om cen tridae), in C ap tiv ity . J p n J Ich th y o l 35: 90—97.

20. C olleye O , F red eric ii B, V an d ew alle P , C asad ev a ll M , P a rm e n tie r E (2009) A gonistic sounds in th e skunk clow nfish Amphiprion akallopisos: s ize-related v a ria tio n in acoustic fea tu res. J F ish Biol 75: 9 0 8 —916.

21. C olleye O , V an d ew alle P, L a n te rb e c q D , E e c k h a u t I, L ecch in i D , e t al. (2011) In te rspec ific v a ria tio n o f calls in clow nfishes: d eg ree o f sim ilarity in closely re la te d species. B M C E vol B iol 11: 365.

22. T a k e m u ra A (1983) S tudies o n th e U n d e rw a te r S o u n d - V III. A coustical b eh a v io r o f clow nfishes (.Amphiprion spp.). Bull F ac Fish N ag asak i U n iv 54: 21—27.

23. P a rm e n tie r E , L a g a rd è re J P , V an d ew alle P, F in e M L (2005) G eo g rap h ica l v a ria tio n in s o u n d p ro d u c tio n in th e a n e m o n e fish Amphiprion akallopisos. P ro c R Soc L o n d B 272: 1 6 9 7 -1 7 0 3 .

24. A k am atsu T , O k u m u ra T , N o v a rin i N , Y a n H Y (2002) E m p irica l refinem ents a p p licab le to th e rec o rd in g o f fish sounds in sm all tanks. J A co u st Soc A m 112: 3 0 7 3 -3 0 8 2 .

25. A m o rim M C P , V asconce los R O (2008) V ariab ility in th e m a tin g calls o f the L u s ita n ia n to ad f ish Halobatrachus didactylus: cu es fo r p o te n tia l in d iv id u a l reco g n itio n . J F ish Biol 73: 1 2 6 7 -1 2 8 3 .

26. B ee M A , K o v ich C E , Blackw ell K J, G e rh a rd t H C (2001) In d iv id u a l v a ria tio n in ad v e rtissem en t calls o f te rr ito ria l m ale g ree n frogs, Rana clamitans: im plications fo r in d iv id u a l reco g n itio n . E tho logy 107: 64—84.

27. C h ristie PJ, M en n ill D J, R atcliffe L M (2004) C h ick ad ee song s tru c tu re is ind iv idua lly d istinctive over lo n g b ro a d c a s t distances. B eh av io u r 141: 101—124.

28. P a rm e n tie r E , C olleye O , F ine M L , F red eric ii B, V an d ew alle P , e t al. (2007) S o u n d P ro d u c tio n in th e C low nfish Amphiprion clarkii. Science 316: 1006.

29. B usto n P M , C a n t M A (2006) A n ew persp ec tiv e o n size h ie ra rch ie s in n a tu re : p a tte rn s , causes a n d consequences . O eco lo g ia 149: 362—372.