RESEARCH ARTICLE

C. J. Fulton Æ D. R. Bellwood

Wave exposure, swimming performance, and the structure of tropicaland temperate reef fish assemblages

Received: 26 May 2003 / Accepted: 2 September 2003 / Published online: 14 October 2003� Springer-Verlag 2003

Abstract We examined the relationship between swim-ming performance, wave exposure, and the distributionpatterns of labrids on temperate rocky reefs, in com-parison with previous functional analyses of a tropicalassemblage. Visual censuses of the distribution andabundance of labrids across two major gradients ofwave exposure (depth and aspect to prevailing winds)were made at two offshore islands near Port Stephens,New South Wales, Australia. Distinct shifts in speciescomposition and abundance were evident between highand low wave exposure habitats on temperate rockyreefs, particularly between deep and shallow habitats onexposed reef fronts. The swimming performances oftemperate labrids were assessed through examination ofpectoral fin shape (aspect ratio) and in situ swimmingspeeds. A diversity of pectoral fin morphologies wasexhibited within this temperate assemblage, rangingfrom rounded to tapered fins (aspect ratios of 0.52 and1.43, respectively). Fin shape was strongly correlated(Pearson�s correlation 0.884, P<0.001) with swimmingspeed (ranging from 1.05 and 3.06 body lengths s)1), in arelationship comparable to that observed in tropical la-brids. Inter-specific differences in swimming ability pro-vided some explanation for differences in the distributionand abundance of temperate labrids in relation to waveexposure. However, our findings suggest that althoughcoral reef labrids appear to predominantly use high as-pect-ratio fins to successfully occupy wave-exposedhabitats, temperate labrids appear to be using an en-hanced swimming ability through increased body size.

Introduction

Wave exposure has been suggested as one of the keyphysical factors influencing community structure inshallow aquatic habitats. Evidence from a wide range ofintertidal and subtidal marine habitats has indicated thatconcurrent changes in wave exposure and the distribu-tion and abundance of species are evident for bothmobile and sessile taxa (e.g. Lewis 1968; Denny 1988). Insessile organisms, this relationship has often beenattributed to differences in the ability of each species towithstand the water movements produced by the crashand surge of breaking waves. Such water movementshave been found to inflict physical damage throughimpact and abrasion (Shanks and Wright 1986; Bodkinet al. 1987), affect rates of growth (Dennison and Barnes1988; Trussell 2002), and increase rates of mortalitythrough the detachment and removal of individualsfrom habitats (Ebeling et al. 1985; McQuaid and Lind-say 2000). Wave-swept sessile organisms often displayfunctional characteristics to counteract these effects,including morphologies that reduce drag and lift forcesin echinoderm tests and mollusc shells (Denny 1994;Denny and Gaylord 1996), increased tensile tissuestrengths and simpler branching in algal thalli (Gaylordet al. 1994; Friedland and Denny 1995), and robustcolony architectures in calcareous corals (Done 1983).Consequently, the functional morphology of residentsessile organisms has often been correlated with levels ofwave exposure, particularly in high wave energy habitats(Denny 1994; Vogel 1994).

Similar functional explanations for the influence ofwave exposure on the distribution patterns of moremobile organisms have not been so readily identified.Previous studies have generally focused on the indirecteffects of wave exposure, without considering the po-tential for wave action to directly affect the mobility ofthese organisms. However, recent studies on a highlymobile group of organisms, reef fishes, have indicatedthat their distribution patterns across a range of wave

Marine Biology (2004) 144: 429–437DOI 10.1007/s00227-003-1216-3

Communicated by G.F. Humphrey, Sydney

C. J. Fulton (&) Æ D. R. BellwoodCentre for Coral Reef Biodiversity,Department of Marine Biology,James Cook University, 4811 Townsville,QLD, AustraliaE-mail: christopher.fulton@jcu.edu.auFax: +61-7-47251570

exposures may be directly related to their swimmingabilities. Strong correlations were found to exist betweenthe swimming performance of labrid fishes (family La-bridae) and their distribution and abundance amongstcoral reef habitats of differing wave exposure, bothwithin (Fulton et al. 2001) and among reefs (Bellwoodand Wainwright 2001). These patterns applied to bothadults and juveniles (Fulton and Bellwood 2002), andwere consistent for several tropical localities over globalbiogeographical scales (Bellwood et al. 2002). On coralreefs, swimming performance appears to be a particu-larly useful predictor for understanding the assemblagestructure of labrid fishes in relation to wave exposure.But what of rocky reef systems in temperate latitudes?Whilst wave energy has been noted as an importantphysical factor for the community composition of bothcoral and rocky reefs (Ebeling and Hixon 1991), they arecharacterised by very different benthic communities andtaxonomically distinct ichthyofaunas (Ebeling and Hi-xon 1991; Meekan and Choat 1997). Very little is knownof the swimming abilities of temperate reef fishes. Couldthe same functional systems be operating in these twodistinct reef ecosystems? Using comparisons with pre-vious functional analyses of a tropical system, we ex-plored the extent to which the same functional systemsapplied in two taxonomically distinct reef fish assem-blages occurring over a latitudinal scale.

Pectoral fin shape has been linked to swimming per-formance in a wide range of labriform (i.e. using solelythe pectoral fins to produce thrust, Webb 1994) swim-ming fishes (Blake 1979; Drucker and Jensen 1997;Walker and Westneat 2000). Recent examination of thisfunctional relationship in an entire assemblage of labridsfrom the Great Barrier Reef identified a considerablediversity of fin shapes, which was reflected in a similarlydiverse range of swimming performances. Species werefound to be spread on a continuum between the twoextremes of rounded (low aspect ratio) and tapered (highaspect ratio) fins, with field measures of swimming per-formance indicating that these extremes were stronglycorrelated to sustained swimming speeds (Wainwrightet al. 2002). Application of this functional informationto aspects of their ecology revealed consistent links be-tween fin morphology, swimming performance andhabitat-use patterns. Species that displayed a high as-pect-ratio pectoral fin and faster sustained swimmingspeeds were the most abundant in wave-swept habitats,whilst the slower swimming (low fin aspect ratio) specieswere either rare or absent from such habitats (Bellwoodand Wainwright 2001; Fulton et al. 2001; Wainwrightet al. 2002). Although allometric effects also had aninfluence on the overall swimming speeds produced byindividuals (Wainwright et al. 2002; Walker and West-neat 2002), size appeared to be relatively unimportantfor differences in the among-habitat distributions ofthese tropical species. Instead, differences in fin mor-phology and the relative performance advantages asso-ciated with the different labriform modes of thrust wereconsidered to be the most important factors (Bellwood

and Wainwright 2001; Fulton et al. 2001; Wainwrightet al. 2002).

Given this detailed information on the swimmingabilities and habitat-use patterns of tropical labridfishes, we can predict that the distribution of temperatelabrids across a range of wave exposures will be corre-lated to their swimming performances, particularly, thatspecies exhibiting higher sustained swimming speeds willoccupy the most wave-exposed habitats in greaterabundance than their slower swimming counterparts. Totest this prediction, the present study aimed to: (1)determine if temperate labrids display distinct patternsof distribution and abundance amongst rocky reefhabitats of different wave exposures, (2) establish ifdifferences in pectoral fin morphology and swimmingperformance exist among temperate labrids and (3)evaluate the extent to which the distribution of speciesacross habitats of different wave exposure conform withdifferences in their sustained swimming performance.Overall, these aspects will be compared to patternspreviously reported for a tropical coral reef assemblageto evaluate the wider utility of swimming abilities forexplaining the distribution patterns of reef fishes inrelation to wave exposure.

Materials and methods

Study sites

The study was conducted between February and November 2001 atPort Stephens (32�43¢S; 152�11¢E), New South Wales, Australia.Study sites were located on the reefs surrounding two offshore is-lands (Broughton Island and Cabbage Tree Island), and were eitherexposed or sheltered in relation to incident SE wave energy (Fig. 1).This is the prevailing wave direction in the region for much of theyear (Short and Trenaman 1992), and there are indications thataverage levels of wave energy in this region are of comparablemagnitude to those incident on outer-shelf reefs of the GreatBarrier Reef (Young 1989). These sites were chosen for their similarbathymetry, substratum complexity, and orientation to prevailingwinds. Both exposed sites consisted of a steep rock face off theisland, very large boulders (>3 m diameter) forming a patchy reefflat in shallow water (1–5 m depth), interspersed with medium (1–3 m diameter) to small (<1 m diameter) boulders that continueddown on a moderately steep slope to sand at 14–18 m depth.Sheltered sites consisted of a conglomeration of large, medium andsmall boulders on a rock shelf at 1–6 m in depth, which then dis-sipated to sand at 6–9 m depth. Benthic flora largely consisted ofencrusting and turfing algae in the shallow habitats (both exposedand sheltered), with patchy areas of large laminarian algae at eachof the censused depths.

Among-habitat distributions

Visual censuses of among-habitat distributions were conducted inthree different habitats at each of the two islands: exposed deep(9 m depth), exposed shallow (4 m) and sheltered (4 m). Thesehabitats were chosen to approximate the depth, aspect to prevailingwinds, and gross geomorphology of the slope, crest and back reefzones found on a coral reef (Bellwood and Wainwright 2001). Eachcensus consisted of a 10-min-timed swim, recording all labridindividuals >5 cm total length (TL) within a 5-m-wide lineartransect that was run parallel to the low-tide mark in each habitat.

430

Fish total lengths were estimated and placed into 5 cm TL sizeclasses. This procedure was repeated three times within each hab-itat, with a minimum of 20 m separating adjacent censuses.

Fin morphology and swimming performance

Pectoral fin morphology was examined in all of the temperatespecies recorded in the distribution censuses. Fishes were collectedusing hand spears or barrier nets, placed in an ice-water slurrywithin 1 h of capture; TL and body mass were then taken prior todissection. Pectoral fins were removed at the base as close to thebody as possible, spread but not stretched on a sheet of foam,pinned, then fixed in position with concentrated formalin solution(40% formaldehyde). A digital image was then taken and lateranalysed using ScionImage (v4.0.2, Scion) to measure the length ofthe leading edge and the total fin area following Fulton and Bell-wood (2002). Pectoral fin shape was then expressed as an aspectratio (AR), which was calculated as the square of the leading edgedivided by the total fin area of a single pectoral fin. This procedurewas repeated for at least three adult individuals of each species.

Swimming speeds attained by each species were estimated usingan in situ technique following Wainwright et al. (2002). Individualfish were observed and timed with a stopwatch as they swamundisturbed on the reef in an approximately linear path. Thebeginning and end of each trial were marked as the fish travelledpast noted landmarks on the reef. Immediately after each trial thedistance between these landmarks was measured to the nearest5 cm and recorded, along with the species and estimated TL of theindividual. This procedure was repeated for a minimum of six adultindividuals from each species. Whenever possible, additionalobservations on an expanded size range of individuals were alsomade to examine size effects.

Statistical analyses

After initial data exploration, all departures from normality andhomoscedasticity were corrected using log10(x+1) transformations.Among-habitat distributions were examined using principal com-ponent analysis on a covariance matrix of the transformed species

abundances within each of the three habitats at each of the twosites. Differences in the mean abundance of species among habitatsand sites were tested using a three-way analysis of variance(ANOVA), with site, habitat and species as fixed factors. Sizecomposition of these distribution patterns were also examinedthrough size-frequency plots, whereby individuals from all specieswere pooled into six size classes for each habitat, at each of the twosites.

Differences in pectoral fin shape among temperate species wereexamined using a one-way ANOVA comparing mean pectoral finAR, with species as a fixed factor. Sources of variation in pectoralfin shape were examined further using Pearson�s correlation coef-ficient to compare mean values of fin AR with fin leading edge, andfin AR with fin area. Average absolute swimming speeds (cm s)1)were calculated for each species, and found to display a strongcorrelation with mean size (Pearson�s correlation 0.867, n=10,P<0.01). Consequently, speeds were converted to body lengths persecond (BL s)1) to minimise these body size effects (Pearson�scorrelation 0.596, n=10, P>0.05). Differences in mean swimmingspeed (BL s)1) among species were tested using a one-way ANOVA,with species as a fixed factor. Correlations between fin morphologyand swimming performance were then assessed using Pearson�scorrelation coefficient, comparing the mean values of pectoral finAR and swimming speed (BL s)1) from all temperate species.Relationships between size and swimming performance wereexamined further through least-squares linear regression of abso-lute speed against total length for each species.

Results

Among-habitat distributions

Labrid assemblages censused on the two offshore rockyreefs were dominated by temperate taxa (ten species,83.5% of all individuals), with six predominantly trop-ical species also occurring on these reefs in low numbers.Distinct trends were evident in the distribution of speciesamong the three habitats censused (Fig. 2). In particu-lar, a marked reduction in the number of species (78.6%)was evident in the exposed shallow habitats when com-pared to the other habitats censused (Fig. 2). Principalcomponent analysis highlighted this general trend, withthe first principal component (explaining 59.8% of thevariation in species distributions) indicating a majorseparation between the exposed shallow habitats and allother habitats censused (Fig. 3a), most likely as a con-sequence of the absence of species in exposed shallowhabitats (Figs. 2, 3b). A division was also evident be-tween the exposed deep and sheltered habitats along thesecond principal component (Fig. 3a). This appeared tobe largely due to the restriction of a few species to eitherexposed or sheltered locations (Fig. 2), particularly Eu-petrichthys angustipes and Pictilabrus laticlavius in thesheltered habitats (Fig. 3b). Three-way analysis of vari-ance indicated a significant difference among habitatsand among species, a significant interaction betweenspecies and habitats, but no significant interactions be-tween any other combination of factors (Table 1). Giventhat the mean abundances of the widespread speciesremained relatively constant among habitats, this sig-nificant interaction between species and habitats wasmost likely due to the restriction of a few species toeither exposed or sheltered habitats, combined with the

Fig. 1 Map of study sites located around the two offshore islandsoff Port Stephens, New South Wales, Australia. Each of theexposed (E) and sheltered (S) study sites are indicated

431

relative lack of species in the exposed shallow habitats(Fig. 2).

Apart from these differences in species composition, adistinct shift in the size-frequency distribution of indi-viduals between habitats was also evident (Fig. 4).Individuals occurring in the exposed shallow habitats atboth sites were almost exclusively large (>20 cm totallength) (Fig. 4). Conversely, a bias towards smaller sizes(<15 cm total length) was apparent in the exposed deepand sheltered habitats, most notably in the latter(Fig. 4).

Fin morphology and swimming performance

Pectoral fin shape, when expressed as an AR, was foundto be significantly different among the temperate speciesexamined (Table 2). Mean pectoral fin AR for eachspecies varied between a low of 0.52 in Eupetrichthysangustipes to a high of 1.43 in Coris picta (Fig. 5). Asignificant correlation between fin AR and fin leading

edge (Pearson�s correlation 0.695, n=10, P<0.02), butnot between fin AR and fin area (Pearson�s correlation0.695, n=10, P=0.19) suggested that differences inpectoral fin AR among species were primarily due tovariations in the length of the leading edge for a given finarea.

Swimming speeds differed significantly among species(Table 3), with average speeds ranging from a low of1.05 BL s)1 in E. angustipes to a high of 3.06 BL s)1 inC. picta. These differences in swimming performanceappeared to be related to pectoral fin shape in the speciesexamined (Fig. 5), which was supported by a strongsignificant correlation between the two variables (Pear-son�s correlation 0.884, n=10, P<0.001). Size also ap-peared to be related to swimming performance, withabsolute swimming speed displaying a positive

Fig. 2 Among-habitat distribution and abundance of labrid fishesat two offshore rocky reef sites off Port Stephens. Abundances aremean number of individuals per 10-min-timed swim (n=3).Asterisks indicate predominantly tropical species

Fig. 3 Principal component analysis of mean abundance of labridfishes across three habitats at the two rocky reef sites of BroughtonIsland (B) and Cabbage Tree Island (CT): a ordination plot ofhabitat zones on principal components 1 (PC1) and 2 (PC2) and bspecies vector plot on PC1 and PC2. Full species names are given inFig. 2

432

relationship with size in all of the species examined(Table 4; Fig. 6). This trend of increasing speed withincreasing size, combined with differences in the maxi-mum observed size of each species, resulted in a broadrange of maximum speeds observed (Table 4).

Discussion

Among-habitat distributions and wave exposure

Labrid fishes displayed distinct patterns in their distri-bution and abundance among rocky reef habitats ofdifferent wave exposure. Several species were restrictedto sheltered habitats (e.g. Pictilabrus laticlavius, Eupe-trichthys angustipes), and very few species were found to

be abundant in the exposed shallow habitats (primarilyOphthalmolepis lineolata and Notolabrus gymnogenis),the latter habitat being characterised by a 79% reduc-tion in species richness when compared to each of theother habitats censused. These trends were similar, albeiton a reduced scale of diversity, to those reported forlabrids in tropical localities, where most species weregenerally found in either exposed deep or shelteredhabitats, and shallow, high wave energy habitats werecharacterised by relatively few abundant species (Green1996; Bellwood and Wainwright 2001; Fulton et al.2001, Bellwood et al. 2002). Differences in fin mor-phology and swimming performance were suggested tobe important determinants of these patterns on coralreefs (Bellwood and Wainwright 2001; Fulton et al.2001; Bellwood et al. 2002), and this study demonstratesa comparable link between swimming performance andlabrid distributions in a temperate reef ecosystem.

Relationship between fin shape, performanceand habitat use

Temperate labrids displayed a range of pectoral finmorphologies that were strongly correlated to theirswimming speeds on the reef. Specifically, species with a

Table 1 Summary of three-way ANOVA on mean abundance perhabitat for 16 labrid species across three habitats from two reefsites, with site, habitat and species as fixed factors (NS non-sig-nificant)

Source of variation df SS MS F P

Site 1 5.12·10)4 5.12·10)4 0.02 NSHabitat 2 6.53 3.27 114.40 <0.01Species 15 27.64 1.84 64.57 <0.01Site·Habitat 2 0.09 0.05 1.60 NSSite·Species 15 0.52 0.03 1.22 NSHabitat·Species 30 7.06 0.24 8.24 <0.01Site·Habitat·Species 30 1.13 0.04 1.32 NSError 192 5.48 0.03Total 287 48.45

Table 2 Summary of one-way ANOVA comparing the mean pec-toral fin aspect ratios from adult individuals of ten temperate labridspecies, with species as a fixed factor

Source of variation df SS MS F P

Species 9 0.36 0.04 143.59 <0.01Error 74 0.02 2.79·10)4

Total 83 0.38

Fig. 4 Size-frequency distributions of labrid fishes across threehabitats at each of the two rocky reef sites of Broughton Island(closed bars) and Cabbage Tree Island (open bars). Individualswithin each size class are pooled from all species observed in eachhabitat

433

low pectoral fin AR were found to display a loweraverage in situ swimming speed for their size than spe-cies with a higher fin AR. These trends in performanceclosely resembled the functional relationship seen intropical labrids, whereby species with a high fin ARtended to use a lift-based flapping of the fins to produceand maintain higher swimming speeds than could beattained by their drag-based, low fin AR counterparts(see Wainwright et al. 2002; Walker and Westneat 2002).

When compared to the locomotor diversity of labrids oncoral reefs, however, temperate labrids represent a rel-atively small subset of the total diversity in the family, interms of both morphology and performance. Spanningmuch of the same range of body sizes seen in tropicalassemblages, temperate labrids displayed a reducedrange of pectoral fin morphologies, particularly at theupper limit (Fig. 7). Although the lowest AR displayedin the temperate assemblage (0.52 in Eupetrichthys an-gustipes) was comparable to that reported for a tropicallabrid (0.56 in Pseudocheilinus evanidus, Wainwrightet al. 2002), the highest AR reported for a temperatelabrid (1.43 in Coris picta) was markedly lower than thehighest found on the Great Barrier Reef (2.1 in Stetho-julis bandanensis, Wainwright et al. 2002). Similarly, thehighest average swimming speed displayed by a tem-perate labrid (3.0 BL s)1) was less than half the highestseen in labrids from the Great Barrier Reef (6.8 BL s)1,Wainwright et al. 2002). Collectively, this evidenceindicates that although temperate labrid assemblages doencompass a range of locomotor morphologies andperformances, when compared to tropical systems, thereappears to be an under-representation of high fin ARspecies that predominantly use lift-based locomotion.

Despite this reduced locomotor diversity in the tem-perate assemblage, differences in fin morphology andswimming performance provided some explanation fortheir observed distribution patterns in relation to waveexposure. In particular, species which were restricted tothe most sheltered habitats (Eupetrichthys angustipes,Austrolabrus maculatus and Pictilabrus laticlavius) dis-played much lower fin ARs and relative swimmingspeeds than species that were more widespread. Underconditions of high water movement in exposed reefhabitats, the ability to undertake daily tasks may bestrongly dependent on locomotor ability, specificallyrequiring an ability to efficiently maintain high sustainedswimming speeds on a daily basis. Given that the abovespecies displayed fin morphologies and swimming speedsthat indicate that they are almost exclusively drag-basedlocomotors (Wainwright et al. 2002; Walker and West-neat 2002), these species might be unable to maintain therequired speeds, or at the very least, be at a disadvantageto their faster swimming counterparts, and must there-fore avoid habitats of high water movement. A similarmechanism has been proposed for the habitat-use pat-

Fig. 5 Mean in situ swimming speed against mean pectoral finaspect ratio for ten species of temperate labrid fishes. Means arebased on measurements taken from adult individuals only. Fullspecies names are given in Fig. 2 (n>3 for aspect ratio means, n>5for swimming speed means)

Table 3 Summary of one-way ANOVA comparing mean in situswimming speeds (body lengths s)1) from adult individuals of tentemperate labrid species, with species as a fixed factor

Source of variation df SS MS F P

Species 9 1.33 0.15 35.60 <0.01Error 106 0.44 4.16·10)3

Total 115 1.77

Table 4 Summary of leastsquares linear regressions ofswimming speed against size forten species of temperate labrid,ranked by maximum speed.Max. size and max. speed referto the maximum total lengthand in situ swimming speedobserved for that species,respectively

Species Slope r2 Max. size (cm) Max. speed (cm s)1) n

Coris picta 2.35 0.31 27 81.7 22Notolabrus gymnogenis 0.85 0.20 32 80.5 31Ophthalmolepis lineolata 1.27 0.37 32 72.0 21Achoerodus viridis 0.03 0.02 90 54.9 7Pseudolabrus luculentus 1.51 0.03 15 38.5 6Pseudolabrus guentheri 1.13 0.18 17 31.5 17Austrolabrus maculatus 1.38 0.59 17 28.2 16Eupetrichthys angustipes 1.51 0.68 18 24.8 21Pictilabrus laticlavius 1.12 0.26 16 23.3 9Suezichthys arquatus 2.63 0.56 12 20.5 13

434

terns of some stream fishes, where differences in mor-phometrics and swimming capacity between species havebeen suggested for their differential use of habitats offast and slow water flow (Sagnes et al. 2000).

For more widespread species, however, among-habi-tat differences in distribution and abundance did notappear to be obviously linked to differences in pectoralfin morphology. One of the most striking trends in thedistribution patterns of labrids on these offshore rockyreefs was the presence of very few species in the mostwave-affected, shallow habitats on the exposed reeffronts. These same species (Ophthalmolepis lineolata,Notolabrus gymnogenis and Achoerodus viridis) wereoften the dominant taxa in each of the other habitats aswell. However, they displayed fin morphologies (1.23–1.34 mean fin ARs) similar to several other species in theassemblage that were lacking in the exposed shallowhabitats. This suggests that swimming performancesbased on differences in fin morphology are relativelyuninformative for examining the distribution patterns ofthese widespread temperate species, particularly forthose occurring in the most wave-affected habitats.

Body-size effects

Size effects on swimming performance, however, mayprovide an explanation. Whilst various advantages areassociated with fin shape and the use of different modesof thrust, these relate to relative differences in theswimming performance of individuals of a given size(Vogel 1994). Absolute swimming speeds, however, areoften influenced to a greater extent by the overall size ofan individual. Larger individuals of a given species,whilst producing relatively slower speeds in terms oftheir size, are capable of higher absolute swimmingspeeds than smaller individuals, both within and amongtaxa (Wardle 1977; Beamish 1978)—a trend which wasstrongly evident in the majority of temperate taxaexamined in this study. Essentially this means that in theabsence of mechanical advantages associated withchanges in fin morphology and stroke kinematics, sub-stantial increases in performance may be gained throughattaining a greater size and muscle mass to producefaster overall swimming speeds. On temperate reefs, itappears that this strategy is being utilised by labrid fishesto access a greater range of habitats, particularly themost wave-exposed habitats. Temperate labrids dis-played distinct differences in their size distributionamongst habitats, with fishes in the exposed shallowhabitats being strongly biased towards larger sizes(>25 cm TL). Notably, with only one exception (C.picta), the only three species that were abundant in the

Fig. 6 Relationship between size and swimming speed in two(Ophthalmolepis lineolata and Notolabrus gymnogenis) of the tenlabrid species examined. Details of the least squares linearregressions for these species are given in Table 4

Fig. 7 Plot of pectoral fin shape against size for labrid speciesfound in tropical assemblages on coral reefs (Tropical) andtemperate assemblages on rocky reefs (Temperate), with speciesoverlapping between the two assemblages indicated (Tropical–Temperate). Morphospace for the temperate labrid species isbounded by the dashed line. Morphometric data for the tropicaland tropical–temperate species from the Great Barrier Reef weremodified after Wainwright et al. (2002)

435

exposed shallow habitats were the only taxa that at-tained a maximum size of >20 cm TL, and displayed amaximum absolute swimming speed >50 cm s)1. Theremay be some threshold of size and swimming speed re-quired to access these most wave-affected rocky reefhabitats, and the aforementioned three taxa are able todo so by attaining sizes and maximum sustained swim-ming speeds that exceeded most of the other observedspecies.

Size versus fin shape: alternate strategiesfor wave-swept habitats

In contrast to this temperate rocky reef assemblage, sizeappears to be relatively unimportant in tropical assem-blages. Whilst tropical labrids were found to have arather uniform average size amongst habitats, temperateindividuals displayed marked differences in size distri-bution across habitats equivalent (in terms of depth andaspect) to coral reef zones (Fig. 8). Rather than a shift insize between habitats, tropical labrids displayed distinctdifferences in their fin morphologies. In particular, coralreef labrids that were abundant in high wave energyhabitats displayed an extremely tapered fin shape, indi-cating a predominant use of lift-based thrust (Bellwoodand Wainwright 2001; Fulton et al. 2001; Bellwood et al.2002). These fishes were often so numerically dominantin these habitats that population sizes were often anorder of magnitude larger than their drag-based coun-terparts (Bellwood et al. 2002).

If high fin AR, lift-based taxa are so successful inoccupying wave-swept habitats on tropical reefs, whyare these fishes, or at least a functional equivalent,lacking in the temperate assemblage? Phylogeny mayplay a major role in the lack of lift-based species ontemperate reefs. Fishes that are the characteristic,numerically dominant taxa in high wave energy habitatson coral reefs (e.g. Thalassoma and Gomphosus, Bell-wood et al. 2002) were either rare or entirely lacking inthis temperate assemblage, and no equivalent lift-basedspecies were evident in the corresponding high waveenergy rocky reef habitats. It appears that a lineage oflift-based taxa, which probably evolved within the tro-pics, has not expanded to temperate locations, and afunctionally analogous lineage has not evolved withinthis temperate assemblage. Furthermore, temperate la-brids appeared to display a generally larger average sizethan their tropical counterparts (Fig. 8). The use of theirinherently larger size may represent the most readilyavailable mechanism for these temperate fishes to exploitwave-exposed locations, and thus preclude the need for(or selective advantages of) other mechanisms. Alter-natively, the advantages of lift-based swimming may notapply within a temperate reef ecosystem. For example,differences in temperature may ultimately affect theenergetic advantage that is thought to be associated withthe long-term use of lift-based thrust (Walker andWestneat 2000; Wainwright et al. 2002). Lower ambient

sea temperatures in temperate latitudes would essentiallymean lower average metabolic rates in these poikilo-thermic animals (Schmidt-Nielsen 1997). If efficiency inlocomotion and conservation of energy are the primaryadvantages of using lift-based thrust for sustainedlocomotion in high wave energy habitats, a lower met-abolic rate may proportionally reduce this advantage.An examination of the functional characteristics of amid-latitude reef fish assemblage, with a considerableoverlap in the presence of tropical and temperate taxa inthe one locality, may help to clarify the relative impor-tance of these issues.

Regardless of these differences in strategy, this studydemonstrates an underlying trend—wave energy appearsto be an important physical factor for the distributionand abundance of fishes on reefs. Furthermore, theevidence presented herein indicates that swimming per-formance may play a major role in this relationship inboth tropical and temperate reef ecosystems. Whilst itwas found that fishes that occupy the most wave-swepthabitats in both localities exhibited greater swimmingperformance than taxa that reside in more shelteredlocations, the means by which this increased swimmingperformance was achieved differed in each locality. Oncoral reefs, labrids appeared to use morphologicallyassociated mechanical efficiencies and different modes ofthrust, whereas increased speed through increasedoverall size appears to be the dominant strategy ontemperate reefs. Whilst the relative importance of phy-logeny and abiotic conditions for these differences in

Fig. 8 Average total length of labrid individuals in three tropicaland three temperate reef habitats, based on three censuses (10-minswim, 5-m-wide transect) in each habitat at each of two reef sites.Tropical census data is from two outer-shelf reefs of the GreatBarrier Reef, modified after Bellwood and Wainwright (2001)

436

strategy are yet to be determined, these results highlightthe utility of functional characteristics in examining theecological patterns of two taxonomically distinctassemblages.

Acknowledgements We thank J. Ackerman, T. Fulton and A.Thomas for assistance in the field, and J. Simm for his generouscontribution of time and logistical support in Port Stephens. Wealso thank J. Walker, P. Wainwright and M. Westneat for helpfuldiscussions. Financial support was provided by the AustralianResearch Council and James Cook University. JCU AnimalExperimentation Ethics approval no. A656-01. This is contributionno. 083 of the Centre for Coral Reef Biodiversity.

References

Beamish FWH (1978) Swimming capacity. In: Hoar WS, RandallDJ (eds) Fish physiology, vol VII. Academic, New York,pp 101–187

Bellwood DR, Wainwright PC (2001) Locomotion in labrid fishes:implications for habitat use and cross-shelf biogeography onthe Great Barrier Reef. Coral Reefs 20:139–150

Bellwood DR, Wainwright PC, Fulton CJ, Hoey A (2002)Assembly rules and functional groups at global biogeographicalscales. Funct Ecol 16:557–562

Blake RW (1979) The mechanics of labriform locomotion. I.Labriform locomotion in the angelfish (Pterophyllum eimekei):an analysis of the power stroke. J Exp Biol 82:255–271

Bodkin JL, Van Blaricom GR, Jameson RJ (1987) Mortalities ofkelp forest fishes associated with large oceanic waves off centralCalifornia, 1982–1983. Environ Biol Fishes 18:73–76

Dennison WC, Barnes DJ (1988) Effect of water motion on coralphotosynthesis and calcification. J Exp Mar Biol Ecol 115:67–77

Denny MW (1988) Biology and the mechanics of the wave-sweptenvironment. Princeton University Press, Princeton

Denny MW (1994) Roles of hydrodynamics in the study of life onwave-swept shores. In: Wainwright PC, Reilly SM (eds) Eco-logical morphology: integrative organismal biology. Universityof Chicago Press, Chicago, pp 169–204

Denny M, Gaylord B (1996) Why the urchin lost itsspines—hydrodynamic forces and survivorship in three echi-noids. J Exp Biol 199:717–729

Done TJ (1983) Coral zonation: its nature and significance. In:Barnes DJ (ed) Perspectives on coral reefs. Clouston, Manuka,pp 107–147

Drucker EG, Jensen JS (1997) Kinematic and electromyographicanalysis of steady pectoral fin swimming in the surfperches. JExp Biol 200:1709–1723

Ebeling AW, Hixon MA (1991) Tropical and temperate reef fishes:comparison of community structures. In: Sale PF (ed) Theecology of fishes on coral reefs. Academic, Sydney, pp 509–563

Ebeling AW, Laur DR, Rowley RJ (1985) Severe storm distur-bances and reversal of community structure in a southernCalifornia kelp forest. Mar Biol 84:287–294

Friedland MT, Denny MW (1995) Surviving hydrodynamic forcesin a wave-swept environment: consequences of morphology inthe feather boa kelp, Egregia menziesii (Turner). J Exp Mar BiolEcol 190:109–133

Fulton CJ, Bellwood DR (2002) Ontogenetic habitat use in labridfishes: an ecomorphological perspective. Mar Ecol Prog Ser236:255–262

Fulton CJ, Bellwood DR, Wainwright PC (2001) The relationshipbetween swimming ability and habitat use in wrasses (Labri-dae). Mar Biol 139:25–33

Gaylord B, Blanchette CA, Denny MW (1994) Mechanical con-sequences of size in wave-swept algae. Ecol Monogr 64:287–313

Green AL (1996) Spatial, temporal and ontogenetic patterns ofhabitat use by coral reef fishes (family Labridae). Mar EcolProg Ser 133:1–11

Lewis JR (1968) Water movements and their role in rocky shoreecology. Sarsia 34:13–36

McQuaid CD, Lindsay TL (2000) Effect of wave exposure ongrowth and mortality rates of the mussel Perna perna: bottom-up regulation of intertidal populations. Mar Ecol Prog Ser206:147–154

Meekan MG, Choat JH (1997) Latitudinal variation in abundanceof herbivorous fishes: a comparison of temperate and tropicalreefs. Mar Biol 128:373–383

Sagnes P, Champagne JY, Morel R (2000) Shifts in drag andswimming potential during grayling ontogenesis: relations withhabitat use. J Fish Biol 57:52–68

Schmidt-Nielsen K (1997) Animal physiology: adaptation andenvironment, 5th edn. Cambridge University Press, Cambridge

Shanks AL, Wright WG (1986) Adding teeth to wave action: thedestructive effects of wave-borne rocks on intertidal organisms.Oecologia 69:420–428

Short AD, Trenaman NL (1992) Wave climate of the Sydney re-gion, an energetic and highly variable ocean wave regime. AustJ Mar Freshw Res 43:765–791

Trussell GC (2002) Evidence of countergradient variation in thegrowth of an intertidal snail in response to water velocity. MarEcol Prog Ser 243:123–131

Vogel S (1994) Life in moving fluids: the physical biology of flow,2nd edn. Princeton University Press, Princeton

Wainwright PC, Bellwood DR, Westneat MW (2002) Ecomor-phology of locomotion in labrid fishes. Environ Biol Fishes65:47–62

Walker JA, Westneat MW (2000) Mechanical performance ofaquatic rowing and flying. Proc R Soc Lond Ser B Biol Sci267:1875–1881

Walker JA, Westneat MW (2002) Performance limits of labriformpropulsion and correlates with fin shape and motion. J Exp Biol205:177–187

Wardle CS (1977) Effects of size on the swimming speeds of fish. In:Pedley TJ (ed) Scale effects in animal locomotion. Academic,London, pp 299–313

Webb PW (1994) The biology of fish swimming. In: Maddock L,Bone Q, Rayner JMV (eds) Mechanics and physiology of ani-mal swimming. Cambridge University Press, Cambridge,pp 45–62

Young IR (1989) Wave transformation over coral reefs. J GeophysRes 94:9779–9789

437