Environmental Biology of Fishes (2005) 72: 33–43 Ó Springer 2005

Distribution and abundance of labrids in northeastern New Zealand: the relationship between depth, exposure and pectoral fin aspect ratio

Christopher M. Denny Leigh Marine Laboratory, P.O. Box 349, Warkworth, New Zealand (e-mail: [email protected])

Received 28 August 2003 Accepted 3 March 2004

Key words: labridae, temperate, wrasse, fetch, spatial, reef fish

Synopsis

Physical factors influencing the distribution and abundance of seven common labrid fishes were examined over four rocky reef locations in northeastern New Zealand. Depth and exposure for each species (both within and among sexes) were related to pectoral fin aspect ratio. Each of the four locations (two mainland and two island) displayed distinct labrid assemblages, which were consistent over time, likely due to the influence of the East Auckland Current. There was a consistent depth-related trend for most species, regardless of location. Several species also showed a sex related depth difference. There was also a trend for some species to be associated with certain levels of wave exposure. For most species, the relationship between pectoral fin aspect ratio and the above physical variables was not as strongly evident in this temperate assemblage as has been previously found in tropical reef fish systems. Although some species did follow the predicted shifts in fin aspect ratio with depth and/or exposure, the observed trends were unre- lated to fin aspect ratio for many other species. These findings suggest that wave exposure may not be as important for labrids on northeastern New Zealand reefs as it may be in tropical coral reefs systems. The lower fin aspect ratios for New Zealand labrids, compared to tropical labrids, suggest that New Zealand labrids represent a subset of the total pectoral fin diversity in the Labridae. Consequently, the potential for distinct trends in fin aspect ratio and physical variables to be evident may be reduced.

Introduction studies examining the spatial distribution and abundance of fishes on temperate reefs are rela- Understanding the processes which are important tively rare (Larson & DeMartini 1984, Choat in structuring reef fish assemblages is a central goal & Ayling 1987, Francis 1996, Brook 2002) with of reef fish ecologists. To determine these processes many studies investigating only a single species it is important to identify the distribution and (e.g. Lowry & Suthers 1998). Temperate fish abundance patterns of the relevant species over the assemblages have been shown to be influenced by spatial scales of interest. Previous studies have wave exposure (Thorman 1986), macroalgae cover shown that the abundance and distribution of (Ebeling & Laur 1985, Anderson 1994) or water coral reef fishes can be related to a variety of temperature (Stephens & Zerba 1981). In contrast habitat characteristics (see review in Jones & Syms to coral reef studies, workers have found no con- 1998). For instance, on coral reefs, reef fish sistent depth related trends (Bell et al. 1992, Cole abundance has been correlated with depth (Green et al. 1992). Furthermore, studies have failed to 1996, Richardson 1999), exposure (Russ 1984, find a strong correlation between reef fish density Williams 1991, Gust 2002) or live coral cover and habitat zones on coral (e.g. Green 1996, Gust (Bouchon-Navaro & Bouchon 1989). In contrast, 2002) and temperate reefs (Holbrook et al. 1993). 34

Recent studies have found correlations between waters in northern New Zealand (Francis &Evans pectoral fin morphology, swimming mode and 1993). The aim of this study was to examine if patterns of habitat use in labrid fishes (Bellwood depth and exposure influence the distribution and & Wainwright 2001, Fulton et al. 2001, Fulton & abundance of labrid fishes in northeastern New Bellwood 2002, 2004). Labrids rely on their pec- Zealand and the relationship between swimming toral fins during sustained swimming and generally ability (as examined through pectoral fin mor- use two thrust-producing mechanisms: drag-based phology) and these environmental variables. Po- rowing and lift-based flapping. Whilst the drag- tential differences between sexes in response to based mechanism provides good acceleration and these variables were also explored. This paper thrust at low speeds, lift-based thrust is mechani- makes the assumption that, in labrids, the fin as- cally much more efficient at higher speeds (Vogel pect ratio correlates with performance (Drucker 1994, Walker & Westneat 2000). A tendency to & Jenson 1996, Wainwright et al. 2002, Walker & predominantly use either mechanism, and the Westneat 2002). However, this assumption may associated implications for their swimming abili- not always hold true. ties, is reflected in the pectoral fin morphology of each species. Rounded fins (low aspect-ratio) are generally indicative of drag-based swimming at Methods slow speeds, whereas tapered fins (high aspect-ra- tio) indicate a tendency towards lift-based swim- Study locations ming and generally higher sustained swimming speeds (Drucker & Jenson 1996, Walker & West- Three locations were surveyed biannually in spring neat 1997, Wainwright et al. 2002, Walker & and autumn between 1999 and 2002; two island Westneat 2002). Recent studies have utilised this locations, the Poor Knights and Mokohinau Is- functional relationship to link swimming abilities lands, and a coastal location, Cape Brett. An in the labrid fishes to their patterns of distribution additional survey was conducted at Mimiwhang- across several spatial scales. For example, Fulton ata in autumn 2002 to give a further coastal et al. (2001) and Bellwood & Wainwright (2001) location for spatial comparisons (see Denny & revealed that labrids on the Great Barrier Reef Babcock (2004) for location details). General (GBR) with high pectoral fin aspect ratios domi- descriptions of coastal and island habitats in nated exposed high-energy locations, whereas northern New Zealand are given in Choat & Schiel species with low pectoral fin aspect ratios were (1982). All locations are influenced, to various relatively rare or absent from such locations. Such extents, by the East Auckland Current (EAC). The studies have suggested that the swimming abilities EAC, derived from the subtropical East Austra- of species may be limiting their access to high-en- lian Current, flows offshore parallel to the ergy locations, and may be a significant factor Northland coast and brings water that is clearer, influencing their habitat use and biogeography on and is often a few degrees warmer (2.5°C) than coral reefs, from regional (Bellwood & Wainwright nearby coastal areas (Stanton et al. 1997). 2001) to global scales (Bellwood et al. 2002). This work is an extension of a study that Study species investigated temporal changes in fish populations in response to the establishment of a no-take Labrids were selected as the study species in this marine reserve in northeastern New Zealand study as they are one of the most abundant and (Denny et al. 2003). During that study, fish widespread families of reef fish in New Zealand, assemblages were found to vary between locations, with 25 species recorded. Labrids also show little a result consistent with previous workers in this response to the presence of SCUBA divers which region (Choat & Ayling 1987, Brook 2002). These makes visual census techniques a reliable measure differences have been attributed mainly to broad- of their distribution, abundance and population scale hydrological patterns, particularly the East density (Samoilys & Carlos 2000). The seven spe- Auckland Current, that facilitates the dispersal of cies selected for study are the most abundant larvae from subtropical regions to the temperate labrid species in northern New Zealand; these were 35

Bodianus unimaculatus Gunther 1862 (pigfish), does not take into account differences in wave Coris sandageri Hector 1884 (sandagers wrasse), spectra among areas (Denny 1988) but does permit Notolabrus celidotus Bloch & Schneider 1801 tentative conclusions to be made (Thomas 1986). (spotty), Notolabrus fucicola Richardson 1840 The use of fetch as a proxy for exposure in (banded wrasse), Pseudolabrus luculentus Rich- northeastern New Zealand appears suitable as the ardson 1848 (orange wrasse), Pseudolabrus miles longest fetch and largest swells are from the Schneider & Forster 1801 (scarlet wrasse) and northeast/east. Suezichthys aylingi Russell 1985 (crimson cleaner- The following categories of wave exposure were fish). Labrids are usually sexually dichromatic, defined: Very Protected (VP, fetch < 300 km), each sex having a distinctive colour phase. The Protected (P, 301 and 1000 km), Semi-protected relationship between colour phase and sex was (SP, 1001–1500 km), Semi-exposed (SE, 1501– confirmed for the above species by histological 2500 km) and Exposed (E, fetch > 2501 km). examinations, except for N. fucicola where no The average wave height and direction at all relationship exists between colour phase and sex locations was estimated from data obtained from (Denny & Schiel 2002). the DATAWELL wave-rider buoy moored on the northern side of the Mokohinau Islands. The data Census technique for each record were obtained by taking the average of a 20-min record of the sea surface The distribution and abundance of labrids were height, which is transmitted to software running visually censused within 25 m long 5 m wide on a computer connected to the radio receiver at (125 m2) transects following Denny & Babcock the Leigh Marine Laboratory. (2004). Between 19 and 23 sites were surveyed at each location (except at Mimiwhangata where 16 Fin aspect ratio sites were surveyed), with nine transects being completed at each site. Each transect was con- The study species were captured at various sites ducted by first attaching a tape measure to the around northeastern New Zealand by handspear; substratum, then an observer swam for 25 m whilst B. unimaculatus (n ¼ 40), C. sandageri (n ¼ 23), N. counting all of the labrid individuals encountered celidotus (n ¼ 6), N. fucicola,(n ¼ 6), P. luculentus within 2.5 m either side of the transect line, and (n ¼ 9), P. miles (n ¼ 10) and S. aylingi (n ¼ 1). noting their colour phase (including juvenile C. Only one specimen of S. aylingi was taken, as this sandageri). A 5 m lead in was swum before com- species is relatively rare outside the Poor Knights, mencing counts to avoid including fish attracted to Three Kings and Kermadec Islands (Francis et al. the diver while the tape was being attached. Dur- 1987). The standard length (SL, mm), wet weight ing each transect observers recorded the depth (g) and colour phase of each fish was recorded. within each of six categories (0–5, 6–10, 11–14, 15– The right pectoral fin was removed, spread and 19, 20–24 and 25+ m). pinned on white paper, and traced around in pencil. The length of the leading edge and the total Exposure fin area were measured using digital image analysis in order to calculate the fin aspect ratio; aspect Wave exposure at each site was estimated using an ratio ¼ length of leading edge2 divided by total fin index of potential fetch. Fetch, defined as the area area. To allow comparisons with the aspect ratios of the sea surface over which waves are generated of other organisms, the fin aspect ratio was mul- by a wind having a constant direction and speed, tiplied by two (see Wainwright et al. 2002). was calculated using the Fetch Effect Analysisâ computer program (Version 1.1) (E. Villouta, Statistical analysis personal communication). Fetch was calculated by summing the potential fetch for each 10° sector of A non-parametric MANOVA was conducted to the compass rose. For open sectors of water, the examine if the labrid assemblage differed signifi- potential fetch was arbitrarily set to 300 km. The cantly between surveys and between locations use of fetch as an estimate of relative exposure (Anderson 2001). Pairwise a posteriori 36 comparisons were completed to isolate where any (A) significant differences occurred between locations Poor Knights 0.2 Cape Brett or surveys. To visualise differences between loca- Mokohinau Is. tions an unconstrained and constrained multi- Mimiwhangata dimensional scaling (MDS) was performed in the 0.1

Canonical Analysis of Principal coordinates = 0.327) 2 (CAP) statistical package on a Bray–Curtis dis- δ similarity matrix of log ðx þ 1Þ transformed data 0.0 summed at the site level (Anderson & Willis 2003). A bi-plot showing the correlation between location and each species was also generated. Similarity -0.1 percentages (SIMPER) analysis was carried out to Canonical Axis 2 ( provide a measure of the average contribution each species made to the average similarity/dis- -0.2 similarity between locations and the contribution of each species to different exposure levels. A -0.2 -0.1 0.0 0.1 0.2 general linear model (GLM) was used to identify if Canonical Axis 1 (δ2 = 0.822) significant differences existed between labrid spe- cies (and between sexes) for the factors Location 1 (with Site nested), Depth and Exposure. When (B) significant interactions occurred a Tukeys test was conducted to determine where the significant P. miles interactions were. When several tests of signifi- B. unimaculatus cance are carried out simultaneously, the proba- bility of committing a type 1 error becomes larger. Accordingly the nominal value was set at a ¼ 0:01. -1 1 P. luculentus ANOVA was conducted to examine differences in N. celidotus S. aylingi the fin aspect ratio between species and a Spear- man correlation test was used to examine the N. fucicola relationship between standard length and the fin C. sandageri aspect ratio.

Results -1

The seven labrid species examined in this study Figure 1. (A) CAP ordination for the four locations, Poor Knights, Cape Brett, Mokohinau Islands and Mimiwhangata, accounted for 15 953 individuals. Other labrids in autumn 2002 comparing the labrid assemblages and (B) bi- (e.g. Anampses elegans, Coris picta, Notolabrus plot showing the correlation of seven labrids with the two CAP inscriptis, Suezichthys arquatus and Thalassoma axes. spp.) were censused, however, these species were seldom observed. The labrid assemblage structure varied signifi- each species contributes towards the separation of cantly between locations for each survey locations (Figure 1B). For example, N. celidotus (NP-MANOVA, all p < 0:05). This is supported correlates with Mimiwhangata and S. aylingi and by the CAP ordination from autumn 2002 where P. luculentus are correlated with the Poor Knights. the labrid assemblages showed a degree of sepa- In addition, the GLM found that all species ration, albeit with some overlap (Figure 1A). All showed a significant location effect (all, p < 0:01). other surveys showed a similar pattern in both the The SIMPER analysis found that N. celidotus constrained and unconstrained ordination. The contributed 86% of the total abundance at Mim- bi-plot shows the magnitude and direction that iwhangata, distinguishing this location from the 37 other three. These three locations had a high island locations and at the sheltered sites at the contribution by C. sandageri (35–38%) but were coastal locations (Figure 3). characterised by particular species. The Poor Pectoral fin shape, expressed by fin aspect ratio, Knights had a high percentage contribution by P. varied from a mean of 1.92 for N. fucicola to 4.0 luculentus (21%) and B. unimaculatus (22%) with S. for S. aylingi (Figure 4). There was no significant aylingi also mainly found at this location. The difference between the fin aspect ratio for most coastal Cape Brett location had a high contribu- species, with the exception of N. fucicola and C. tion by N. fucicola (28%) and N. celidotus (19%) sandageri, both of which were significantly differ- and the Mokohinau Islands were characterised by ent from all other species (P < 0:05). There was B. unimaculatus (28%) and P. miles (12%). also a significant difference in the fin aspect ratio There was a clear depth distribution pattern for between P. luculentus and B. unimaculatus most species, with a similar trend found at all (p ¼ 0:014). locations. All species were found at each depth There was a weak correlation between fin aspect strata, albeit in different numbers (Figure 2). There ratio and standard length for species with enough was, however, a significant Loc(site) Depth data to draw a trend (r2 < 0:14), except for C. interaction term for five of the seven species (ex- sandageri (r2 ¼ 0:4). There was also a trend for the cept for N. celidotus and N. fucicola), and reflects males of C. sandageri and P. miles to have higher the variation in species abundance between loca- fin aspect ratios than females, however these dif- tions. In spite of this interaction, most spe- ferences were not statistically significant (Figure 5). cies showed consistent trends at most locations. Three species, B. unimaculatus (F53; 831 ¼ 12:38, p < 0:001), P. miles (F53; 831 ¼ 5:7, p < 0:001) and Discussion S. aylingi (F53; 831 ¼ 3:39, p ¼ 0:0046) were signifi- cantly more common in deeper waters whereas N. This study was interested in the processes impor- fucicola (F53; 831 ¼ 16:26, p < 0:001) was mainly tant in structuring labrid assemblages in northern found in shallow waters (10 m). The remaining New Zealand. To do this the relationship between species also showed trends either towards deeper, variables like exposure and depth, and swimming P. luculentus, or shallower waters, C. sandageri ability (as indicated by pectoral morphology) were and N. celidotus, but these were not significant. examined. Differences in the labrid assemblages Except for three species, both sexes followed a were found between the four locations, likely due similar pattern with depth. Female B. unimaculatus to the varying influence of the East Auckland occurred in significantly higher numbers in deeper Current (EAC) (Ward & Roberts 1986, Choat & waters compared to males who were found equally Ayling 1987). For example, the EAC has the least distributed across all depths (F58; 504 ¼ 3:91, influence at Mimiwhangata, which has the lowest p ¼ 0:0015). Conversely, N. celidotus number of ‘subtropical’ species (e.g. S. aylingi and (F58; 504 ¼ 15:65, p < 0:001) and C. sandageri P. luculentus ). Conversely, the EAC flows directly (F58; 504 ¼ 3:06, p ¼ 0:009) females were common past the Poor Knights bringing with it warmer in shallow waters, with males found equally dis- waters and a higher number of these ‘subtropical’ tributed across all depths (Figure 2). species. It is interesting to note that the coastal Exposure (fetch) ranged from 16 to 6609 km, Cape Brett location had a similar distribution of with 83% of sites having a fetch less than 3000 km. labrids as the offshore Mokohinau Islands, sug- There was no significant fetch effect for any spe- gesting that the EAC has a similar influence at cies, except B. unimaculatus (F43; 857 ¼ 4:6, both these locations. The spatial variation between p ¼ 0:001), but trends were apparent for several locations was consistent over time, a finding con- species (Figure 3). At each location, the abundance sistent with other workers in northern New Zea- of N. celidotus and C. sandageri was related to a land (Choat et al. 1988), the Great Barrier Reef low exposure whereas the abundance of P. miles, (Green 1996) and the Indian Ocean (Letourneur S. aylingi and B. unimaculatus were related to a 1996). higher exposure. Interestingly, N. fucicola was When finer scale habitats were examined most found more often at the more exposed sites at the species were associated with certain depth strata. 38

Bodianus unimaculatus Coris sandageri 1.0 1.6 Male Male 1.4 Female 0.8 Female 1.2 Juvenile 0.6 1.0 0.8 0.4 0.6 0.4 0.2 0.2 0.0 0.0

0-5m 0-5m 6-10m11-14m15-19m20-24m 25+m 6-10m11-14m15-19m20-24m 25+m

Notolabrus celidotus Notolabrus fucicola 1.0 0.4 Male TP 0.8 Female 0.3 IP 0.6 0.2 0.4

0.1 0.2

(± SE) 0.0 0.0 2 0-5m 0-5m 6-10m11-14m15-19m20-24m 25+m 6-10m11-14m15-19m20-24m 25+m

Pseudolabrus luculentus Pseudolabrus miles 0.5 0.25 Male Male 0.4 Female 0.20 Female Mean density per 125m 0.3 0.15

0.2 0.10

0.1 0.05

0.0 0.00

0-5m 0-5m 6-10m11-14m15-19m20-24m 25+m 6-10m11-14m15-19m20-24m 25+m

Suezicthys aylingi 0.25 Male 0.20 Female

0.15

0.10

0.05

0.00

0-5m 6-10m11-14m15-19m20-24m 25+m Depth (m)

Figure 2. Mean density per 125 m2 (+ SE) of seven species of male and female labrids at six depth strata. Note the presence of juveniles for C. sandageri and the initial phase (IP) and terminal phase (TP) labels for N. fucicola. 39

Poor Knights Mokohinau Is. Cape Brett Mimiwhangata High E P. miles P. miles B. unimaculatus P. miles P. miles

N. fucicola SE S. aylingi P. luculentus N. fucicola B. unimaculatus P. luculentus B. unimaculatus S. aylingi

SP C. sandageri S. aylingi

B. unimaculatus Exposure (fetch) gradient C. sandageri N. fucicola P N. celidotus P. luculentus N. celidotus P. luculentus N. fucicola

N. celidotus N. celidotus Low VP C. sandageri C. sandageri

Figure 3. Average densities versus exposure (fetch) for seven species at four locations in northeastern New Zealand.

This depth related pattern was consistent at all not investigated, however, depth may be related to locations, even though the overall assemblage diet as urchins, an important prey item for N. ce- make-up differed. A reason for this depth stratifi- lidotus, are mainly found in 4–10 m depth stratum cation may be competition for food (Holbrook & (Choat & Ayling 1987). Alternatively, the depth Schmitt 1986). Several species in this study have a related differences might be due to species degree of dietary separation (Denny, unpublished displacement (Daan 1980). However, this data). The extent to which the availability of prey hypothesis was not tested as inter-specific aggres- determines the vertical distribution of labrids was sion was seldom observed. Three species also showed a depth related pat- tern with sex. A sex related change in depth pref- erence was found for another common reef fish, 4 Cheilodactylus spectablis (McCormick 1989). The reasons for the sex differences observed in this study are unclear but may be due to an ontoge- netic change in preferred prey items or reduced 3 predation risk with changing depth. Ontogenetic migrations between (Green 1996) and within (Fulton & Bellwood 2002) habitat zones have been recorded in some tropical labrids. 2 The use of fetch as a proxy for exposure does not take into account the complexity of wave Mean pectoral fin aspect ratio (± SE) 0 theory (Denny 1988). In spite of this limitation, a P. miles N. fucicola S. aylingi conservative estimate of how fish assemblages are P. luculentusN. celidotus C. sandageri B. unimaculatus structured in relation to wave energy can be pro- Species vided. Patterns were observed for several species where their abundance decreased with increased Figure 4. Mean pectoral fin aspect ratio (SE) for seven labrid levels of exposure and vice versa. This finding is species. 40

Bodianus unimaculatus Coris sandageri 400 400

300 300

200 200

100 Female 100 Female 2 Male 2 Male r = 0.1 r = 0.4 0 0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0

Pseudolabrus luculentus Pseudolabrus miles 400 400

Standard length (mm) Female Female 300 Male 300 Male

200 200

100 100 2 2 r = 0.096 r = 0.14 0 0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0

Fin aspect ratio

Figure 5. Plots of fin aspect ratio against standard length (mm) for both sexes of B. unimaculatus, C. sandageri, P. luculentus and P. miles. consistent with studies that found varying wave species, with the exception of P. luculentus. This exposure is a principal cause in the change in the species had a low fin aspect ratio and was found composition and abundance of fish in the surf zone mainly in deeper areas with a moderate exposure, (Romer 1990) and on tropical reefs (Williams & consistent with studies in tropical systems (Bell- Hatcher 1983, Gust 2002). However, as surveys wood & Wainwright 2001, Fulton et al. 2001). were undertaken when the weather was relatively Interestingly, for at least one of the physical vari- calm (<1.5 m swell), the observed patterns are ables, some species did follow the hypothesis that likely to provide a conservative measure of the fish with a high fin aspect ratio should be mainly impact of wave action. For example, during large found in shallow areas with a high exposure and storm events which can produce 12 m swells at the vice versa. Variation in these results show that, offshore islands (based on wave-rider buoy data), although there are consistent patterns for depth fish remaining in shallow waters are likely to suffer and exposure for some species, these trends are mortality from physical abrasion, gill damage, etc. often unrelated to the fin aspect ratio, and in some Therefore, during severe weather conditions fish cases show the reverse pattern. communities are likely to respond by shifting from The average fin aspect ratios found in New shallow more exposed areas into deeper waters Zealand labrids were, on average, lower compared that provide refuges from heavy seas (Friedlander to labrids from the GBR (Fulton et al. 2001, & Parrish 1998). Wainwright et al. 2002). The lower fin aspect ratios Trends were found in the abundance of several observed for labrids in New Zealand, may largely species with different depth and exposure gradients be due to the reduced diversity of this assemblage, but no obvious relationship between pectoral fin combined with the possibility that labrids on New aspect ratio and the physical variables for most Zealand reefs have appeared to have not evolved 41 the high fin aspect ratios observed for their tropi- greater impact on differences in swimming ability cal counterparts (Bellwood et al. 2002). The lower among species than differences in fin morphology. number of species in New Zealand (compared to The small sample size for some species in the the GBR) coupled with a generalist diet, a degree present study may have hidden some of these of dietary separation, and an ability to forage trends, with fin shape and swimming performance widely means that different species can co-exist data from an expanded size range of individuals with little antagonism. Alternatively, wave expo- being needed to resolve this issue of size effects. sure might not be as important an influence on The relationship between pectoral fin aspect ratio New Zealand’s rocky reefs as on coral reefs of the and physical variables was not as strongly evident GBR. This is in spite of the similar average wave for most labrids in this temperate assemblage as has amplitude in both regions, 0.8 m in New Zealand been previously found in tropical reef fish systems (Pickrill & Mitchell 1979) vs. 0.75 m on the GBR (Bellwood & Wainwright 2001, Fulton et al. 2001, (P. Kench pers. com.). There are, however, higher Bellwood et al. 2002). Although some species did maximum waves in extreme weather conditions in follow the predicted shifts in fin aspect ratio with New Zealand, 12 m (wave-rider buoy data) vs. depth and/or exposure, the observed trends were 7.5 m on the GBR (Allen & Callaghan 2000). La- unrelated to fin aspect ratio for many other species. brids in New Zealand may be better able to find Whilst these findings suggest that wave exposure suitable shelter during these storms events whereas may not be as important for labrids assemblages on shelter may be limited on the GBR. However, northeastern New Zealand reefs as it may be in given the high habitat heterogeneity of coral reefs, tropical coral reefs systems, the overall magnitude it is likely there is suitable shelter available for of fin aspect ratios in these New Zealand labrids was many species. Alternatively, because labrids in low compared to tropical labrids. Consequently, the New Zealand tend to prey on benthic invertebrates labrids in this study represent a subset of the total (Russell 1983, Denny & Schiel 2001) slow-speed pectoral fin diversity (and associated range in manoeuvring around the reef is probably more swimming abilities) in the Labridae, possibly important than high-speed cruising. This idea is reducing the potential for distinct trends in the fin consistent with Fulton et al. (2001) who indicated aspect ratio and depth/exposure to be evident. that species with low fin aspect ratios tended to occur in closer association with the structure of the reef. Two species that can both act as cleanerfish, S. aylingi and C. sandageri, had the highest fin Acknowledgements aspect ratios. This is consistent with other studies where cleanerfish may spend considerable amounts Thanks to Phil Bendle, Brady Doak, Murray of time swimming higher in the water column, an Birch, and Geordie Murman for skippering their environment where a higher fin aspect ratio may various vessels in sometimes suspect weather. be more suitable (Fulton et al. 2001). Thanks to the divers that contributed their time There was little change in pectoral fin mor- collected data: these are Kiley Bloxham, Daniel phology with increasing size for most species, ex- Egli, Dave Feary, Neil Hart, Timmy Langlois, cept for C. sandageri. This suggests that most Greg Nesbitt, Darren Parsons, Angela Rapson, species displayed a similar relative swimming Laura Richards, Phil Ross, Justine Saunders, ability for a given size. This is in contrast with Evan Skipworth, Tracey Smith, Schannel Van Fulton & Bellwood (2002), who demonstrated that Dijken, Jarrod Walker, Caroline Williams and five labrid species (out of eight) from the GBR James Williams. Thanks to Varenka Lorenzi for displayed marked shifts in pectoral fin aspect ratio her laboratory skills, Chris Fulton, Nick Tolimieri, with size, these species typically being those which Richard Ford and Russ Babcock for their com- displayed a high fin aspect ratio as adults. Nota- ments on this manuscript, Eduardo Villouta for bly, (Fulton & Bellwood in press) revealed that allowing the use of the Fetch Analysis Index swimming speed varies strongly with overall size in Program and Trevor Willis for initiating these temperate labrids from New South Wales, with the surveys. Finally thanks to the Department of maximum size of the species ultimately having a Conservation for providing funding. 42

References cola (Labridae) around Kaikoura, New Zealand. New Zea- land J. Mar. Freshwater Res. 36: 555–563. Denny, C.M., T.J. Willis & R.C. Babcock. 2003. Effects of the Allen, M.A. & J. Callaghan. 2000. Extreme wave conditions for Poor Knights Islands Marine Reserve on demersal fish the South Queensland Coastal Region. p. 17, Environment populations. Department of Conservation Science Internal Technical Report 32, Environmental Protection Agency, Series 142: 57 pp. Brisbane. Denny, M.W. 1988. Biology and the mechanics of the wave- Anderson, M.J. 2001. A new method for non-parametric mul- swept environment. Princeton University Press, Princeton, tivariate analysis of variance. Austral Ecol. 26: 32–46. New Jersey. 329 pp. Anderson, M.J. & T.J. Willis. 2003. Canonical analysis of Drucker, E.G. & J.S. Jenson. 1996. Pectoral fin locomotion in principal coordinates: a useful method of constrained ordi- the striped surfperch I. Kinematic effects of swimming speed nation for ecology. Ecology 84: 511–525. and body size. J. Exp. Biol. 199: 2235–2242. Anderson, T.W. 1994. Role of macroalgal structure in the Ebeling, A.W. & D.R. Laur. 1985. The influence of plant cover distribution and abundance of a temperate reef fish. Mar. on surfperch abundance at an offshore temperate reef. Ecol. Prog. Ser. 113: 279–290. Environ. Biol. Fishes 12: 169–179. Bell, J.D., D.J. Ferrell, S.E. McNeill & D.G. Worthington. Francis, M.P. 1996. Geographic distribution of marine reef 1992. Variation in assemblages of fish associated with deep fishes in the New Zealand region. New Zealand J. Mar. and shallow margins of the seagrass Posidonia australis. Mar. Freshwater Res. 30: 35–55. Biol. 114: 667–676. Francis, M.P. & J. Evans. 1993. Immigration of sub-tropical Bellwood, D.R. & P.C. Wainwright. 2001. Locomotion in and tropical animals into north-eastern New Zealand. pp. labrid fishes: implications for habitat use and cross-shelf 131–136. In: C.N. Battershill, D.R. Schiel, G.P. Jones, R.G. biogeography on the Great Barrier Reef. Coral Reefs 20: Creese & A.B. MacDiarmid (eds.) Proceedings of the 2nd 139–150. International Temperate Reef Symposium, NIWA Marine, Bellwood, D.R., P.C. Wainwright, C.J. Fulton & A. Hoey. Wellington. 2002. Assembly rules and functional groups at global bi- Francis, M.P., R.V. Grace & C.D. Paulin. 1987. Coastal fishes ogeographic scales. Functional Ecol. 16: 557–562. of the Kermadec Islands. New Zealand J. Mar. Freshwater Bouchon-Navaro, Y. & C. Bouchon. 1989. Correlations be- Res. 21: 1–13. tween chaetodontid fishes and coral communities of the Gulf Friedlander, A.M. & J.D. Parrish. 1998. Temporal dynamics of of Aquba (Red Sea). Environ. Biol. Fishes 25: 47–60. fish communities on an exposed shoreline in Hawaii. Envi- Brook, F.J. 2002. Biogeography of near-shore reef fishes in ron. Biol. Fishes 53: 1–18. northern New Zealand. J. Roy. Soc. New Zealand 32: 243– Fulton, C.J. & D.R. Bellwood. 2002. Ontogenetic habitat use in 274. labrid fishes: an ecomorphological perspective. Marine Choat, J.H. & A.M. Ayling. 1987. The relationship between Ecology Progress Series 236: 255–262. habitat structure and fish faunas on New Zealand reefs. Fulton, C.J. & D.R. Bellwood. 2004. Wave exposure, J. Exp. Mar. Biol. Ecol. 110: 257–284. swimming performance, and the structure of tropical and Choat, J.H., A.M. Ayling & D.R. Schiel. 1988. Temporal and temperate reef fish assemblages. Marine Biology 144: 429– spatial variation in an island fish fauna. J. Exp. Mar. Biol. 437. Ecol. 121: 91–111. Fulton, C.J., D.R. Bellwood & P.C. Wainwright. 2001. The Choat, J.H. & D.R. Schiel. 1982. Patterns of distribution and relationship between swimming ability and habitat use in abundance of large brown algae and invertebrate herbivores wrasses (Labridae). Marine Biology 139: 25–33. in subtidal regions of northern New Zealand. J. Exp. Mar. Green, A.L. 1996. Spatial, temporal and ontogenetic patterns of Biol. Ecol. 60: 129–162. habitat use by coral reef fishes (Family Labridae). Mar. Ecol. Cole, R.G., R.G. Creese, R.V. Grace, P. Irving & B.R. Jackson. Prog. Ser. 133: 1–11. 1992. Abundance patterns of subtidal benthic invertebrates Gust, N. 2002. Scarid biomass on the northern Great Barrier and fishes at the Kermadec Islands. Mar. Ecol. Prog. Ser. 82: Reef: the influence of exposure, depth and substrata. Envi- 207–218. ron. Biol. Fishes 64: 353–366. Daan, N. 1980. A review of replacement of depleted stocks by Holbrook, S.J. & R.J. Schmitt. 1986. Food acquisition by other species and the mechanisms underlying such replace- competing surfperch on a patchy environmental gradient. ment. Rapport et proces-verbaux des reunions. Conseil Environ. Biol. Fishes 16: 135–146. permanent international pour l’exploration de la mer 177: Holbrook, S.J., S.L. Swarbrick, R.J. Schmitt & R.F. Ambrose. 405–421. 1993. Reef architecture and reef fish: correlates of popu- Denny, C.M. & R.C. Babcock. 2004. Do partial marine reserves lation densities with attributes of subtidal rocky envi- protect reef fish assemblages? Biol. Conser. 116: 119–129. ronments. pp. 99–106. In: C.N. Battershill, D.R. Schiel, Denny, C.M. & D.R. Schiel. 2001. Feeding ecology of the G.P. Jones, R.G. Creese & A.B. MacDiarmid (eds.) banded wrasse Notolabrus fucicola (Labridae) in southern Proceedings of the 2nd International Temperate Reef New Zealand: prey items, seasonal differences, and ontoge- Symposium, NIWA Marine, Wellington, Auckland, New netic variation. New Zealand J. Mar. Freshwater Res. 35: Zealand. 925–933. Jones, G.P. & C. Syms. 1998. Disturbance, habitat structure Denny, C.M. & D.R. Schiel. 2002. Reproductive biology and and the ecology of fishes on coral reefs. Australian J. Ecol 23: population structure of the banded wrasse, Notolabrus fuci- 287–297. 43

Larson, R.L. & E.E. DeMartini. 1984. Abundance and verti- Thomas, M.L.H. 1986. A physically derived exposure index for cal distribution of fishes in a cobble-bottom kelp forest marine shorelines. Ophelia 25: 1–13. off San Onofre, California. Fishery Bull. 82: 37–53. Thorman, S. 1986. Physical factors affecting the abundance and Letourneur, Y. 1996. Dynamics of fish communities on species richness of fishes in the shallow waters of the southern Reunion fringing reefs, Indian Ocean. I. Patterns of spatial Bothnian Sea, Sweden. Estuarine, Coastal Shelf Sci. 22: 357– distribution. Journal of Exp. Mar. Biol. Ecol. 195: 1–30. 370. Lowry, M.B. & I.M. Suthers. 1998. Home range, activity and Vogel, S. 1994. Life in moving fluids: the physical biology distribution patterns of a temperate rocky-reef fish, Cheilo- of flow. Princeton University Press, Princeton, NJ. dactylus fuscus. Mar. Biol. 132: 569–578. 467 pp. McCormick, M.I. 1989. Spatio-temporal patterns in the abun- Wainwright, P.C., D.R. Bellwood & M.W. Westneat. 2002. dance and population structure of a large temperate reef fish. Ecomorphology of locomotion in labrid fishes. Environ. Biol. Mar. Ecol. Prog. Ser. 53: 215–225. Fishes 65: 47–62. Pickrill, R.A. & J.S. Mitchell. 1979. Ocean wave characteristics Walker, J.A. & M.W. Westneat. 1997. Labriform propulsion around New Zealand. New Zealand J. Mar. Freshwater Res. in fishes: kinematics of flapping aquatic flight in the bird 13: 501–520. wrasse Gomphosus varius (Labridae). J. Exp. Biol. 200: 1549– Richardson, D.L. 1999. Correlates of environmental variables 1569. with patterns in the distribution and abundance of two Walker, J.A. & M.W. Westneat. 2000. Mechanical performance anemone fishes (Pomacentridae: Amphiprion) on an eastern of aquatic rowing and flying. Trans. Roy. Soc. Lond. Ser. B Australian sub-tropical reef system. Environ. Biol. Fishes 55: Biol. Sci. 267: 1875–1881. 255–263. Walker, J.A. & M.W. Westneat. 2002. Performance limits of Romer, G.S. 1990. Surf zone fish community and species re- labriform propulsion and correlates with fin shape and mo- sponse to a wave energy gradient. J. Fish Biol. 36: 279–287. tion. J. Exp. Biol. 205: 177–187. Russ, G. 1984. Distribution and abundance of herbivorous Ward, C. & L. Roberts. 1986. The East Auckland Current: one grazing fishes in the central Great Barrier Reef Australia 1. explanation for the distribution patterns of the coastal and Levels of variability across the entire continental shelf. Mar. offshore island fish faunas of north-eastern New Zealand. pp. Ecol. Prog. Ser. 20: 23–34. 211–219. In: A.E. Wright & R.E. Beever (eds.), Proceedings Russell, B.C. 1983. The food and feeding habits of rocky reef of a symposium convened by the Offshore Islands Research fish of north-eastern New Zealand. New Zealand J. Mar. Group, Wellington, Department of Lands and Survey Freshwater Res. 17: 121–145. Information Series, Auckland, 10–13 May 1983. Samoilys, M.A. & G. Carlos. 2000. Determining methods of Williams, D.M. 1991. Patterns and processes in the distribution underwater visual census for estimating the abundance of of coral reef fishes. pp. 437–474. In: P.F. Sale (eds.), The coral reef fishes. Environ. Biol. Fishes 57: 289–304. Ecology of Fishes on Coral Reefs, Academic Press, San Stanton, B.R., P.J.H. Sutton & S.M. Chiswell. 1997. The East Diego. Auckland Current, 1994–95. New Zealand J. Mar. Fresh- Williams, D.M. & A.I. Hatcher. 1983. Structure of fish com- water Res. 31: 537–549. munities on outer slopes of inshore, mid-shelf and outer shelf Stephens, J.S. & K.E. Zerba. 1981. Factors affecting fish diver- reefs of the Great Barrier Reef. Mar. Ecol. Prog. Ser. 10: sity on a temperate reef. Environ. Biol. Fishes 6: 111–121. 239–250.