Reef Fish Community Structure in Bocas Del Toro (Caribbean, Panama): Gradients in Habitat Complexity and Exposure

Home , Aluterus scriptus, Chromis multilineata, Lactophrys, Serranus tortugarum, Stegastes adustus

Reef Fish Community Structure in Bocas del Toro (Caribbean, Panama): Gradients in Habitat Complexity and Exposure

ARTURO DOMINICI-AROSEMENA AND MATTHIAS WOLFF

Center for Tropical Marine Ecology (ZMT), Fahrenheitstrasse 6, 28359, Bremen, Germany. Corresponding author: [email protected]

ABSTRACT.—We compared the community-structure of reef-fish over different spatial scales, levels of exposure, and physical complexity in 12 study zones of Bocas del Toro, Panama. Two hundred and eighty- eight visual censuses were conducted on 48 benthic transects from April to September 2002. Substrate coverage and surface complexity was also recorded. We found 128 fish species in 38 families with increasing species richness from sheltered to expose and from low-complexity to intermediate and high-complexity zones. Only 7% of the species occurred in all zones. Gobies and pomacentrids were most abundant in sheltered areas and labrids at exposed zones. Eleven species showed significant size-segregations between zones, suggesting ontogenic movements, with smaller sizes in low-complexity zones, and larger-sizes in intermediate to high complexity areas. Species-richness and diversity are high in three of the four exposed zones and in the main areas of massive-coral reefs and significantly correlate with certain types of complex substrates. Highly mobile fish were more abundant in exposed rocky zones while sedentary fish were more abundant in sheltered massive and foliaceous corals zones. Towards the most exposed areas, the number of mobile invertebrate-feeding fish species greatly increased, while territorial herbivores increased in sheltered zones. Roving herbivores (scarids and acanthurids) showed lower frequency than territorial herbivores in all zones. Demersal zooplankton feeders were common in sheltered areas and oceanic planktivores in exposed areas. Omnivores were more abundant in zones of rubble and sand. Carnivores were less frequent, but contribute to the majority of species. We concluded that the species’ richness in Bocas del Toro relates to the structural complexity of the substrate rather than substrate type. While some species change their preferred habitat during ontogeny, general species diversity increased with habitat complexity. This increase was more pronounced in exposed zones. It seem that water current strength and waves, which select for swimming capacity, play an important but still little understood role in the organization of fish assemblages in rocky and coral reefs.

KEYWORDS.—Fish diversity, distribution, fish mobility, trophic groups, exposure level, lagoonal system, Mesoamerican Caribbean

INTRODUCTION area, and the Caribbean Mesoamerican region in general, only a few studies exist It has been widely stated that there are that address variation in the community- geographic gradients in the richness of fish structure of reef-fishes (Sierra and García species, corals and other coral reef biota, 1996; Clifton and Clifton 1998), and most of with a general decrease in diversity with these were conducted along Caribbean is- increasing latitude and distance from the lands away from the mainland (McGehee Indo-Philippine ‘centre’ (Goldman and Tal- 1994; McKenna 1997). Generally, species bot 1976; Veron 1995). In the Atlantic composition varies among different habi- Ocean, the Caribbean is thought the diver- tats (Talbot and Goldman 1972; McGehee sity centre for both fishes and corals (Briggs 1994) and it appears that a strong taxo- 1995; Veron 1995; Floeter and Gasparini nomic division between reef and non-reef 2000). The present study focuses on Bocas faunas is difficult. “Reef fishes” may thus del Toro, which forms part of the Southern be characteristic of—but not restricted to— Caribbean biogeographic province. For this coral reefs (Robertson 1998; Bellwood 613 614 A. DOMINICI-AROSEMENA AND M. WOLFF

1998). Support for this assertion can be geological changes it has been subjected to, found in different geographical regions, including the emergence of the Isthmus of such as the south-western Atlantic, where Panama. the most diverse fish faunas are along The study aimed to characterize fish as- rocky shores with low coral coverage semblages (including taxonomical compo- (Floeter et al. 2001). sition, abundance and species richness) Studies concerning reef fish have focused along spatial scales and gradients of sub- almost exclusively on coral reefs, not in- strate complexity to determine the relative cluding a wide spectrum of reef habitats importance of the physical structure pro- (Robertson 1998). They usually compare ei- vided by corals, rocks and benthic sessile ther diverse living coral-reef areas or rocky substrate-building organisms. We also de- shores. The structural complexity of a habi- termined characteristics such as mobility, tat is very important for the community or- home range, size, territoriality, and feeding ganization, since it provides physical struc- behaviour responsible for the selection of ture for juvenile and adult animals (Jones habitats by certain species and/or guilds 1988; Beck 1997 1998); it plays an important along gradients of wave exposure, reef type role in the regulation of foraging patterns (i.e., Coral reef vs. rocky reef) and general (Dolmer 1998; Erlandsson et al. 1999) and substrate complexity. predation (Hixon and Beets 1993); it also plays an important role in habitat selection MATERIALS AND METHODS (Levin and Grimes 2002), and in some cases relaxes competition (Diehl 1988). The role Study site of complexity seems clear for most re- searchers, but whether its difference among The study was conducted from April to areas is due to the degree of coverage of September 2002 in sheltered areas of Almi- living coral or other habitat characteristics rante Bay and the exposed areas around should be studied on a spatial scale that Isla Bastimentos at the Province of Bocas includes different habitats and levels of ex- del Toro (8°30Ј09°40ЈN, 82°56Ј081°08ЈW), posure to waves and currents. The presence (Fig. 1). of currents may affect larval dispersal and Intense rain and irregular seasonal pat- retention, thus influencing the populations’ terns characterise Bocas del Toro. The most connectivity (Cowen 2002). The relative important coastal current comes from an contribution of population growth, mortal- easterly direction of Nicaragua and Costa ity, food requirements, metabolic level to Rica (Greb et al. 1996). Waves and currents fitness, as well as the role of habitat struc- have a strong effect outside of the archi- ture and selection, requires further investi- pelago, but the major islands act as a bar- gation (Jones and McCormick 2002), In this rier, decreasing wind-strength, wave- context, edges of habitats have often been height, and tidal amplitude for the inner severely under sampled (Sagarin and water bodies (Bahía de Almirante and La- Gaines 2002). Swimming ability and tro- guna de Chiriquí). A sheltered semi- phic behaviour are also important proper- lagoonal system with tidal amplitudes < 0.5 ties of species that may influence com- m is created, which supports mangrove for- munity organization. More studies are ests. Changes in oceanic conditions of the necessary in this field. southern Caribbean during the late Mio- We compared the community-structure cene, due to the progressive constriction of of reef-fish in the Bocas del Toro region of the Atlantic-Pacific seaway weakened the the Tropical western Atlantic, which exhib- westward circulation and the southern Car- its one of the highest diversity and abun- ibbean flow (Collins et al. 1996). All these dance of corals in shallow Caribbean wa- changes resulted in a well-defined semi- ters (Guzmán and Guevara 1998a,b 1999). lagoonal system that is connected to the This Mesoamerican or “isthmian” Car- southern reef areas of the Caribbean from ibbean is a peculiar region due to the many Costa Rica (Cortés 1984). REEF FISH COMMUNITY STRUCTURE 615

FIG. 1. Sampling zones in Bocas del Toro, Caribbean. 1: North east of Cristóbal Island; 2: South of Solarte Island; 3: Crawl Key, west side of Bastimentos Island; 4: Northeast of Cristóbal Island; 5: South of Solarte Island; 6: Crawl Key; 7: Carenero Island; 8: Hospital point, northeast of Solarte Island; 9: Carenero Key; 10: Northwest of Solarte Island; 11: Wild Cane Key, north of Bastimentos Island; 12: Bastimentos Point, east of Isla Bastimentos.

The selection of study sites was based on limp following the surface contour of the a preliminary-survey that identified areas substrate. In this way the structural com- with different but nevertheless characteris- plexity of the substrate was estimated as tic habitats of the region. From this survey, the ratio between the length of the chain 12 characteristic zones were classified in laid over the substrate and the direct linear terms of depth, substrate type, topography length from the beginning to the end of and cover. Within each of the 12 zones, four each transect. The number of links outlin- benthic 30 m transects were set up parallel ing the surface of the substrate was to the shore at approximately the same counted, noting the kind of substrate under depth. Substrate coverage and surface com- each segment of the chain; the length of the plexity was estimated using a link-chain contour for each portion was then esti- methodology (CARICOMP 2001; Rogers et mated at 1.6 cm per chain-link. al. 1994). The benthic surface measures A detailed description of the benthic re- were taken along the chain that would lay gions is in Guzmán and Guevara (1998a,b 616 A. DOMINICI-AROSEMENA AND M. WOLFF

1999). We classified the zones in ascending ture of living and dead massive coral of order of complexity (1 to 12) along with the different genera including Montastraea following general description (Fig. 2): cavernosa, Siderastrea siderea; also erect sponges and sand. Zones of sand and rubble SRU (1-2): Found Zones of massive volcanic rocks: Not com- in deeper areas adjacent to patches of the mon in Bocas del Toro and virtually ab- branching coral Porites furcata. sent in sheltered areas (pers. obs.). This Zones of turf-algae and dead branching kind of habitat was found on a single coral TA (3): A mixture of eroded skel- deeper profile (11) and in a single shal- etons of foliaceous and branching corals, low one (12). It has a component of ses- with a dense cover of different species of sile organisms (e.g., encrusting sponges turf-algae. These zones were found be- and algae) that tend to predominate in low the exposed reef-slope of fire-coral areas of high disturbance and strong cur- zones. rents and waves (Dethier 1994; Steneck Zones of madreporic branching coral BC and Dethier 1994). (4-5): Patchy zones dominated by the fin- ger-like coral Porites furcata. The species and abundance of fish was Zones of fire-coral FC (6): Continuous shal- assessed using standard methodology for low zones (reef flat) exposed to wave in- underwater visual-survey with SCUBA fluence and characterised by the pres- diving equipment (English et al. 1994; ence of the fire coral Millepora complanata, Khalaf and Kochzius 2002; McKenna 1997). macroalgae and other benthic compo- The fixed benthic transects (30*5m)used nents. for the study of the benthos were also used Zones of foliaceous coral FOLI (7-8): Shel- for the fish census. Monthly sampling was tered deeper reefs covered with living conducted at all transects (six census per and dead colonies of the scleractinean transect), amounting to 288 on 48 benthic coral Agaricia tenuifolia. transects. Two to three observers swam Zones of massive coral MC (9-10): Coral- along the transects, recording data for reef zones separated by sand, with a mix- fishes (including small cryptic individuals)

FIG. 2. Summary of habitat structure variables and results of the frequency analysis (X2) to test the differences in size distribution of species and habitat type, only the species that had significant differences of small and larger individuals are showed. REEF FISH COMMUNITY STRUCTURE 617 encountered within 2.5 m on both sides and ences in physical conditions (substrate di- 5 m above. The standard length (SL) of the versity), fish density, mobility pattern fishes was estimated to the nearest centime- groups and index of diversity (HЈ) of fish tre with a PVC ruler. From these data, assemblages (Zar 1996). An additional Stu- abundance (indiv./150 m2) and Shannon- dent-Newman-Keuls (SNK) test of mul- Wiener diversity (HЈ) were calculated tiple-comparisons of means was applied as based on relative abundance (Pielou 1975). a post-hoc test (Zar 1996). The relationships Each species was included in one of three between diversity indices, species density, mobility /home range size categories, as and physical parameters were examined applied by Floeter et al. 2004: Category 1 using Spearman rank-correlation (Siegel represents high mobility, including wide 1970; Peet 1974; Sokal and Rohlf 1980; Zar horizontal displacement species (e.g., rov- 1996). Association between the most abun- ing herbivores, mullids). Category 2 is rep- dant species, mobility pattern groups and resented by relatively sedentary and dem- their relationship with the habitat was ex- ersal species in close association with the amined using Canonical Correspondence reef substrate (e.g., serranids, haemulids, Analysis (CCA) (Ter Braak and Verdon- chaetodontids). Category 3 includes species schot 1995). For this CCA analysis a value with a small home range and a site attach- for the degree of exposure (score 0 to 4) was ment also expressed by territorial behav- determined by general field observations. iour (e.g., gobies, damselfishes). Differences in size distribution of all spe- Fish species were trophically classified cies and habitat types were examined following Ferreira et al. (2004) and Jones et based on length frequencies and applying al (1991) as: roving herbivores: fish that chi-square analysis (␹2) to the size groups. feed on detritus, turf algae and macroalgae Fish abundance for different zones were (scarids and acanthurids); territorial herbi- pooled and visually presented in a rank- vores: fish that feed on farmed turf-algae order of species according to their corre- within their territories; mobile-invertebrate sponding log numbers to easily visualize feeders: fish that feed primarily on crabs, the species richness and relative impor- molluscs and other benthic mobile inverte- tance of live coral coverage and low coral brates on hard and soft substrates; sessile- coverage substrates vs. complexity (zone invertebrate feeders: fish that eat sessile number) at the different study zones (Wolff benthic invertebrates; piscivores: fishes that and Alarcón 1993). prey on living fishes; carnivores: fish that feed on mobile benthic organisms and also fishes; planktivores: fish that consume pri- RESULTS marily macro and micro-zooplankton; omnivores: fishes that feed on a variety of or- The One-Way ANOVA (Table 1) showed ganisms, including both animal and plant significant differences of the Simpson’s di- material. versity index between some zones. It We used the Simpson’s diversity index to ranged close to zero (highest substrate di- estimate the diversity of organic and inor- versity) at many of the exposed zones, es- ganic habitat categories (Ferreira et al. pecially fire corals and rocky reef, and it 2001). The index ranges from zero (highest was close to one (monotony) in the shel- diversity) to one (monotony). An index tered zones that had similar index values value of one indicates coverage by a single (Fig. 2). kind of substrate. Since much of the data on One hundred twenty eight fish species in physical and biological parameters may not 38 families were found (Appendix). The to- meet the criteria for normality and homo- tal number of fish species increased from geneity of variances, parametric one-way sheltered to exposed zones. The number of ANOVA or non-parametric Kruskall- genera and species found at all locations Wallis tests (Kruskall-Wallis) were applied was not equal; 63% of the genera were re- to the data after testing for normality. stricted to certain zones, and only 7% of the Study sites were then compared for differ- species occurred in all zones. The number 618 A. DOMINICI-AROSEMENA AND M. WOLFF

TABLE 1. Results of parametric (One way ANOVA: F, MS) and non-parametric (Kruskal Wallis, H) ANOVA and multiple comparisons (Student Neumann Keuls, SNK) for diversity comparisons (fish and habitat), and mobility groups between study zones in Bocas del Toro. ND = Normal distribution data, NND = Non normal distribution data.

H Df F MS p Multiple comparison (SNK) Substratum diversity (Simpson’s diversity index) ND 11 7.90 0.04 <0.001 Other zones<8=7=5=4=2=1 Fish diversity (HЈ) ND 11 9.94 0.065 <0.001 12 = 11 = 6 > other zones Fishes/census 36.52 11 NND NND <0.001 5 = 2 = 1 > other zones Category 1 40.89 11 NND NND <0.001 12 > other zones Category 2 42.57 11 NND NND <0.001 12 = 11 = 2 = 1 > other zones Category 3 32.62 11 NND NND <0.001 10 = 9 = 8 = 2 = 1 > other zones of species per genus increased from low- larger sizes were found in zones of differ- complexity to intermediate and high- ent levels of complexity and according to complexity zones in the entire region. Re- no discernible pattern (Fig. 2). gardless of their densities, many of the Gobies and pomacentrids were the most genera and species in the families (e.g., abundant families. The goby Coryphopterus Blennidae, Labridae) and species in the personatus was the most common species in genera (e.g., Halichoeres spp.) were not sheltered zones along with some labrids, overlapping. They showed marked differ- such as Halichoeres bivittatus, the scarid Sca- ences in distribution between protected rus iseri, the pomacentrid Stegastes plani- and exposed zones (Appendix). frons, the serranid Serranus tortugarum and In the size frequency analysis 27 species the haemulid Haemulon plumierii (Fig. 3). were abundant enough to allow segrega- On the other hand, the species Thalassoma tion by size; all these species were habitat- bifasciatum, Chromis multilineata, Stegastes representative and 11 of them show signifi- adustus along with some acanthurids were cant size-segregation between different common at exposed zones (Fig. 4). zones (Fig. 2). Non-territorial species such As revealed by the results of Canonical as Scarus iseri, Halichoeres bivittatus, Correspondence Analysis (CCA), (Table 2, Acanthurus bahianus, Sparisoma viride, and 3 and Fig. 5), the association between fish Haemulon plumierii, had smaller sizes (juve- and environmental variables for the 12 niles and pre-adults) in low-complexity study zones show (at the right side of Axis zones such as sand rubble and turf algae; 2) a gradient of exposure. Most exposed while larger sizes increase proportionally zones were associated with rock-encrusting in zones of intermediate and high-com- red algae and fire coral (Zone 6, 1.22 m; plexity such as those of massive and fire Zone 11, 8.15 m; Zone 12, 3.88 m). Fishes coral, and also in some cases in complex associated with this axis were: Thalassoma exposed rocky zones. The genus Hypoplec- bifasciatum (TLUC), Halichoeres maculipinna trus showed a similar pattern but larger in- (HMAC), Chromis multilineata (CMUL), Ste- dividuals were scarce at any of the exposed gastes adustus (SADU), the acanthurids zones. Sedentary species such as Coryphop- Acanthurus bahianus (ABAH) and Acanthu- terus personatus and Stegastes partitus were rus chirurgus (ACHI) which had a closer as- also present at smaller sizes on sand-rubble sociation with this group at exposed zones. areas and at larger sizes in zones of inter- The positive side of Axis 1 showed shel- mediate complexity such as those of tered zones with branching corals (Zones 4 branching, foliaceous, and massive coral. and 5, Porites furcata, 2.5 m) and zones with Individuals of Stegastes adustus were foliaceous corals (Zones 7 and 8, Agaricia smaller at all rocky zones and larger at fire- tenuifolia 6.93 m). Here Stegastes planifrons coral locations. Individuals of Stegastes leu- (SPLA) and Scarus iseri (SISE) predomi- costictus had smaller sizes at zones of nated, along with Hypoplectrus spp. (HPUE, branching coral and turf-algae, but the HNIG), and some other species. Zone 3 REEF FISH COMMUNITY STRUCTURE 619

FIG. 3. Density (Ind/150 m2 ± SE) for the most abundant species found in sheltered zones (1, 2, 4, 5, 7, 8, 9, and 10) in Bocas del Toro, Panama. 620 A. DOMINICI-AROSEMENA AND M. WOLFF

FIG. 4. Density (Ind/150 m2 ± SE) for the most abundant species found at exposed zones (3, 6, 11, 12) and most of the zones in Bocas del Toro, Panama.

(deeper exposed zone 6.3 m) shows similar The results of One-Way ANOVA and habitat composition than 4 and 5, except for multiple comparisons (Student-Newman the lack of living coral coverage, and had a Keuls [SNK]) for fish parameters show sig- similar fish fauna as zones 4, 5, 7, and 8. nificant differences in the Shannon-Wiener The negative part of Axis 1 shows sand- index (H´ ) between some zones (Table 1). rubble areas (Zones 1 and 2, 6.36 m), which Fish diversity had higher and similar val- associate with Serranus tortugarum (STOR), ues in three of the four exposed zones, in- Serranus tigrinus (STIG), the goby Coryphop- cluding the rocky deeper zone (11), fol- terus dycrus (CDIC) and the labrid Halicho- lowed by shallow rocky zone (12), and eres bivittatus (HBIV). A fourth group closer shallow fire-coral zone (6). All other zones to the centre of the ordination diagram is had a lower diversity. comprised by species found in all study These differences are appreciated in zones: the pomacentrid Stegastes partitus more detail in the log-series model in Fig. 6. (SPAR), the scarids Sparisoma viride (SVIR), The flatness of the line (greatest evenness) Sparisoma aurofrenatum (SAUR) and the as well as its intersection with the x axis chaetodontid Chaetodon capistratus. (Species richness) was greatest in the more REEF FISH COMMUNITY STRUCTURE 621

TABLE 2. Dominant species. O = Omnivore; C = Carnivore; P = Piscivore; MI = Mobile Invertebrate feeder; SI = Sessile Invertebrate feeder; PL = Planktivore; RH = Roving Herbivore; TH = Territorial Herbivore.

Code Species Family Trophic group

ABAH Acanthurus bahianus ACANTHURIDAE RH ACHI Acanthurus chirurgus ACANTHURIDAE RH CCAP Chaetodon capistratus CHAETODONTIDAE SI CDIC Coryphopterus dicrus GOBIIDAE O CMUL Chromis multilineata POMACENTRIDAE PL CPER Coryphopterus personatus GOBIIDAE PL CCRU Cephalopholis cruentata SERRANIDAE C HAUR Haemulon aurolineatum HAEMULIDAE C HBIV Halichoeres bivittatus LABRIDAE MI HMAC Halichoeres maculipinna LABRIDAE MI HNIG Hypoplectrus nigricans SERRANIDAE C HPLU Haemulon plumierii HAEMULIDAE C HPUE Hypoplectrus puella SERRANIDAE C SADU Stegastes adustus POMACENTRIDAE TH SAUR Sparisoma aurofrenatum SCARIDAE RH SISE Scarus iseri SCARIDAE RH SPAR Stegastes partitus POMACENTRIDAE TH SPLA Stegastes planifrons POMACENTRIDAE TH STIG Serranus tigrinus SERRANIDAE C STOR Serranus tortugarum SERRANIDAE PL SVIR Sparisoma viride SCARIDAE RH TBIF Thalassoma bifasciatum LABRIDAE PL

TABLE 3. Canonical Correspondance Analysis. VCP Wallis) shows that fish abundance differs = Variance in cumulative percentage. SEV = Sums of significantly between zones. Zones of eigenvalues. SACCA = Species association (Fig. 5). MCCA = Mobility groups (Fig. 8). rubble (1 and 2) and branching corals (5) had the higher fish abundance (Table 1, see SACCA MCCA Fig. 3 and 4 for most abundant species). VCP 41.45 66.30 75.50 62.30 100.0 The Spearman rank-correlation analysis SEV 0.78 0.46 0.17 0.11 0.07 (Fig. 7) shows significant positive correlation between species-richness and index of diversity with habitat complexity and cer- complex zones 6, 9, 10, 11, and 12, indepen- tain types of substrates such as rocks with dent of their coral coverage. The lines of encrusting red algae, dead and living mas- zones 1 to 5, 7, and 8 that had predomi- sive coral, turf calcareous algae, rock- nance of a single type of substrate (e.g., encrusting sponges. An inverse correlation rubble, branching or foliaceous corals) had was significant for species richness and fish a higher slope and lower regression coeffi- diversity with the Simpson’s index. Mo- cient, indicative of a higher dominance notony (single type of substrate, value close within the assemblage and of a less diverse to 1) was related to a less diverse fish com- community respectively. All outlier points munity structure. Inverse correlations were above the lines (omitted in Fig. 6 for better also found with diversity parameters and graphic visualization) correspond to the dead and living branching coral, dead fo- low and medium complexity sheltered liaceous coral, and turf-algae, and (on a zones and represent the dominant species lower level) for living foliaceous coral, cal- of pomacentrids, gobiids and serranids careous algae, and rubble. (dwarf basses, Serranus sp.). More species Fish abundance (indiv./census) seem were in areas with complex massive coral positively correlated only with substrates and rocky reef. like dead massive coral, sand, and erect The non-parametric ANOVA (Kruskal sponges (Fig. 7), due to the high abundance 622 A. DOMINICI-AROSEMENA AND M. WOLFF

FIG. 5. Canonical Correspondence Analysis (CCA) for fish abundance vs. percentage of substrate coverage and environmental variables associated to respective zones (numbers) in Bocas del Toro, Panama. Species codes are shown on Table 2. of the gobiid Coryphopterus personatus and tal and vertical mobility (Category 1). Rela- the serranid Serranus tortugarum. Total fish- tively sedentary species (Category 2) had abundance was inversely correlated with similar abundance in rubble-sheltered hard substrates as living and dead branch- zones (1 and 2) and exposed rocky reef (11 ing coral, fire-coral with fleshy and turf and 12). Site attached species (Category 3) brown algae (Fig. 7). were more represented in rubble (1 and 2), As seen in Table 1, the rocky shallow foliaceous (7 and 8) and massive coral zone (12) had more fishes of high horizon- zones (9 and 10). This association of mobil- REEF FISH COMMUNITY STRUCTURE 623

FIG. 6. Species Rank vs. log-abundance (log-series model). Shallow living coral (1.22-3.05 m); Deep living coral (5.9-7.9 m); Low coral coverage (3.8-8.15 m). Outlier points omitted for better graphic visualization. 624 A. DOMINICI-AROSEMENA AND M. WOLFF

FIG. 7. Spearman rank correlations between fish community structure and habitat variables;n=288,#=p< 0.0001. ity groups or categories with certain habitat represented by Serranus tortugarum and this attributes and zones was revealed by the category is also related to exposed zones CCA-Analysis (12 zones; all species in- (11 and 12) where relatively sedentary cluded Table 3; Appendix and Fig. 8), families were more diverse (e.g., Serrani- which shows the environmental variables dae, Balistidae) (Fig. 3, 4 and 8, Appendix). by arrows and the Categories in the differ- As compared to sheltered areas, in most ent zones in a single diagram. At the right exposed and complex areas, such as Punta (positive) side of Axis 2 a gradient of expo- Bastimentos (12), Wild Cane (11), and sure is found (clockwise) with most ex- Crawl Cay (6), fish of specific trophic posed zones (11 and 12) associated with a groups (e.g., mobile-invertebrate feeders) habitat of rock-encrusting red algae. Fishes increase in abundance and species numbers in Category 1 (high mobility species) were Herbivores accounted for 77% of the rela- associated with this axis mainly by the tive fish abundance in Bocas del Toro with presence of the planktivorous Chromis mul- a predominance of territorial herbivores in tilineata, Thalassoma bifasciatum and almost all zones. Their frequency increased acanthurids. The area between the negative toward sheltered zones; however roving side of Axis 2 and negative side of Axis 1 herbivores (scarids and acanthurids) were shows, in an anti-clockwise direction, the less frequent in all zones. sheltered zones with foliaceous corals Along with herbivores, demersal zoo- (Agaricia tenuifolia) and the presence of plankton feeders (in sheltered areas) and massive and shallow branching corals reef planktivores that use a larger part of the (Porites furcata). Here, the Category 3 (site water column (at exposed zones) were attached species) is mostly represented by common trophic groups in Bocas del Toro, Coryphopterus personatus and Stegastes plani- but of low species numbers. Omnivores frons. Category 2 (relatively sedentary spe- were more abundant in zones of rubble and cies) in rubble-sheltered zones is mostly sand; mobile-invertebrate feeders increased REEF FISH COMMUNITY STRUCTURE 625

FIG. 8. Canonical Correspondence Analysis (CCA) mobility groups (categories) vs. percentage of substrate coverage and environmental variables associated to respective zone (numbers) in Bocas del Toro, Panama. in relative abundance and diversity from California. The smaller individuals on low- sheltered to exposed zones. complexity substrates may be able to hide Carnivores are less frequent compared to while predators and other large fish may the other mentioned trophic groups, but in- lack retreats. We consider these patterns cluded the majority of species. Their spe- also to play an important role for the orga- cies number did not vary significantly nization of reef fish assemblage in Bocas among zones. Sessile-invertebrate feeders del Toro. and piscivores were represented by few Our results indicate that in many fish species and were found in low numbers families species numbers increased toward only (Fig. 7). more complex zones and densities of specialized feeders (e.g., herbivores) increase DISCUSSION in sheltered zones. This pattern is marked in Labridae (particularly the genus Halicho- Along with other factors, the distribution eres), Scaridae (Scarus sp.) and Pomacentri- pattern of a species varies with ontogenetic dae (Stegastes sp.). Similar to what was morphological changes and shifts in habitat found in some southern Caribbean reefs requirements (García-Charton and Pérez- such as in Venezuela (Rodríguez and Villa- Ruzafa 2001; Nagelkerken et al. 2000a,b). mizar 2000), the scarid Scarus iseri and the Some species ontogenetic movement be- pomacentrid Stegastes planifrons are two of tween habitats may be directed towards the most common fishes particularly in more complex habitats and related to sheltered zones in Bocas del Toro, feeding swimming ability of post recruits (see re- on green turf algae that grows between spective section). Few studies have been branches or the coral Porites spp. and Agari- conducted in this field, but Aburto- cia spp. However, with the exception of S. Oropeza, and Balart (2001) found spatial iseri, none of the dominant species in Bocas variation and size-discrimination for eight del Toro are also dominant further North in of the most abundant species in the Gulf of Honduras (Clifton and Clifton 1998). 626 A. DOMINICI-AROSEMENA AND M. WOLFF

FIG. 9. Frequency, in percent, of the principal trophic groups in the different study zones; numbers in parenthesis indicate the number of species in each group. E = Exposed zone; P = Protected zone.

The species associations at exposed com- but different species of hamlets along with plex rocky shores and fire-coral reefs in Bo- schools of roving herbivores that can over- cas del Toro (right side the CCA diagram) come the aggressiveness of damselfishes were similar to those of exposed coral reefs (Robertson et al. 1976). Haemulids were in Puerto Rico, but here there are no re- also present, and they may migrate from ported records of invertebrate feeders such adjacent areas of mangroves. And the ge- as Halichoeres poeyi (HPOE) and Halichoeres nus Serranus spp and Coryphopterus dicrus maculipinna (HMAC) at exposed zones and (CDIC) are typical inhabitants of sand- in the same study an association between rubble areas. According to Floeter et al Hypoplectrus chlorurus (hamlets), Stegastes (2004), some of the species that were not planifrons (SPLA), and Holocentrus rufus in associated to any particular habitat in our areas of low water movements were found study are considered highly mobile and (McGeehee 1994). In Bocas del Toro, this relatively sedentary (Sparisoma spp. and association is replaced with similar genera Chaetodon capistratus respectively). Even REEF FISH COMMUNITY STRUCTURE 627

Stegastes partitus is considered as territorial energy, while coral community structure, herbivore was in most zones. We believe and to a greater extent fish community that this is due to the fact that S. partitus is structure, seems a complex, indirect and an omnivore that can either feed on non-linear consequence of reef structure zooplankton, small invertebrates and/or and environmental conditions (Bradbury benthic algae at different exposed or shel- and Young 1981). In some studies, no cor- tered zones (Booth and Hixon 1999). Differ- relation was found between fish diversity ences in feeding behaviour among species and habitat complexity (Ault and Johnson of the same genus are an important factor 1998; Clarke 1988). Studies in the Red Sea in defining distributional patterns in Bocas also suggest that coral cover has little influ- del Toro as well. Randall (1967) found that ence on the species richness and abundance the two species of blennies Ophioblennius of fish (Roberts and Ormond 1987; Luck- atlanticus and Parablennius marmoreus have hurst and Luckhurst 1978). These findings different feeding behaviour; the former are confirmed in Bocas de Toro due to the feeding on algae growing in rocky shores high abundance of small serranids and go- and the latter, besides feeding on algae, bids in little complex rubble sand areas. We feeds on many soft substrate animals that did, however, find positive correlations of can often be found on patchy sandy habi- species richness with hard substrates (mas- tats. The same author also found that Hali- sive corals and rocky complex reefs) along choeres bivittatus shows a higher feeding with an inverse correlation with monotony plasticity compared to other species on the (less diverse substrate). It is suggested that genus and this is indicated by the many the diverse strata in Bocas del Toro provide prey items found in their stomachs. This more niches and shelter, and thus increase may be the reason why it is widely distrib- species richness. Coral cover is also impor- uted on the different exposed and pro- tant in enhancing the physiographic struc- tected zones in Bocas del Toro. The more ture in many of the more complex sheltered specialized foragers of the genus such as coral reefs in Bocas del Toro (e.g., zones 9 Halichoeres radiatus and H. maculipinna are and 10). Recent studies on the Red Sea, by present in the more complex exposed zones Khalaf and Kochzius (2002) also found total were they feed on particular food re- fish abundance and species richness to sources. positively correlate with hard substrate and The species richness in Bocas del Toro is habitat diversity. intermediate when compared with other Waves and tides may play an important studies from the Caribbean (Florida Keys role in modifying the habitat structure (and 74 spp. McKenna 1997; Puerto Rico 71 spp. by this increasing complexity) due to shap- McGhee 1994; South-eastern Brazil 91 spp. ing of rocks and morphology of corals Ferreira et al. 2001; Gulf of México 153 spp. (Bradbury and Young 1981). In Bocas del Pattengill et al. 1997; Honduras 214 spp. Toro, exposed zones are physically more Clifton and Clifton 1998; Colombia 273 spp. complex and the fish community is more Mejía et al. 1998) and similar to that along diverse. But besides the influence of expo- the Pacific coast of the Isthmus (Gulf of sure in shaping the habitat, the presence of Chiriquí, Dominici-Arosemena and Wolff diverse and abundant food resources (e.g., unpublished data). Many of the regions of oceanic plankton) is notorious in these similar or lower species richness seem to lie exposed zones. Under these conditions, close to the mainland and being affected by also sessile invertebrates (e.g., sponges) the runoff of rivers and environmental sea- and algae increase the substrate diversity sonality (Venezuela 68 spp. Rodríguez and and allow, along with the rich food supply, Villamizar 2000; Costa Rica 48 spp. Sierra for the occurrence of additional fish spe- and García 1996). This situation also holds cies. for Bocas del Toro. The variability in reef fish larval trans- Besides the consequences of geological port and settlement is determined by the changes (Coates et al. 1992), reef structure interaction of water masses and the effects is regarded as a direct consequence of wave of external forcing such as winds and tides. 628 A. DOMINICI-AROSEMENA AND M. WOLFF

In areas of low tidal variation, such as Bo- and fire coral fringing reef (zone 6) areas, cas del Toro, winds may play an important most of the sheltered zones can be consid- intermittent role in larvae input to inshore ered as relatively isolated discontinuous sheltered zones while currents may play an patches, separated mostly by sand, sea important role for settlement and recruit- grass and mangroves (Guzmán and Gue- ment in exposed zones (Cowen 2002). vara 1998b). Even if we pool all these zones Many horizontal patterns of community and count the total number of species, it structure were described with respect to will be lower than that of massive corals shore proximity, including occasional, and exposed rocky and fire coral reef ones. completely isolated embayment (or la- This suggests that these types of habitat are goonal) assemblages. While water physics vitally important for the fish species rich- is important, larval distribution is not ness in Bocas de Toro due to their complex- merely the result of passive dispersion by ity, level of exposure, and interconnectivity currents and tides. Particularly in areas of with other substrates. It is also important to low current flow and small tidal range, mention that in the province of Bocas del many studies have found a high concentra- Toro we can find a particular “nestedness” tion of pomacentrid/gobiid larvae (see re- of select families of obligated reef fishes in view by Cowen 2002). There is no informa- sheltered zones (e.g., aggregation of gobi- tion of larvae assemblages in Bocas del ids and pomacentrids). These nesting pat- Toro, but our results indicate that these terns seem to be associated with recruit- aforementioned families are the most abun- ment limitation at the reef scale (McLain dant post-recruits in sheltered zones. From and Pratt 1999). a comparative Mesoamerican perspective, Particular mobility guilds were found in patterns of spatial variation could mark- Bocas del Toro, which were specific to cer- edly differ between Tropical Eastern Pa- tain zones: better swimmers are adapted to cific (TEP) and Tropical Western Atlantic exposed rocky zones of strong currents. (TWA), since different oceanographic con- Small site attached species and territorial ditions could lead to differences in spatial herbivores are present in high numbers in patterns diversity. The large tidal fluctua- sheltered zones of massive and foliaceous tion observed in the TEP (which can be sev- coral reefs with little current influence. eral meters) in comparison with TWA (0.5 Most studies regarding these mobility m) may facilitate larval transportation, guilds focused only on labrids and tried to settlement, and recruitment across a wider link their swimming performance to fin range of habitats than in the TEP. morphology (Bellwood and Wainwright Another reason for the spatial difference 2001; Bellwood et al. 2002). There is still no in the number of genera and species of reef knowledge about the swimming ability of fish in the study area may be the “habitat many of the other reef fish, however, and patchiness effect”. Recent studies have questions regarding their ability for vertical found that the degree and the scale of habi- and horizontal movements after settlement tat patchiness may be associated with fish and recruitment on different types of reef diversity. Acosta and Robertson (2002) em- can as yet not be answered. We found a phasize that species assemblages of small marked difference in the presence/absence isolated habitat patches will constitute only of species (within families) between ex- a subset of the assemblage of a large patch posed and sheltered zones in Bocas de reef, and the patterns of reef fish diversity Toro. Besides labrids, spatial differences may thus be highly scale dependent. In our were found in species of territorial herbi- study area the community differences were vores and scarids. While improved swim- clearly marked between exposed and shel- ming performance is definitely an advan- tered zones, although the total area per tage in exposed zones with higher zone included in our sampling units were structural complexity and shelter availabil- of the same size (600 m2 in each habitat). ity, there must be a trade of between the We suggest that, with the exception of the energy loss for this high swimming activity continuous rocky reef (zones 11 and 12) and the energy gain through the high food REEF FISH COMMUNITY STRUCTURE 629 abundance. Further research needs to de- be used to organize sustainable fishery termine if the distribution of certain species practices for the region. This study has pro- is more related to their swimming perfor- vided the first description of the reef fish mance or/and type of food. community structure, which can be used to It seems that in the northern part of the help define areas of protection (Sale 2002). Caribbean such as in the Gulf of México Even though more relevant studies are re- (including the Florida Keys), planktivores quired to evaluate these ideas, we suggest and invertebrate feeders dominate, while that Bocas del Toro is an isolated site, herbivores are low in numbers compared to where the population are likely to be main- Bocas del Toro (Bohnsack and Bannerot tained by self-recruitment and this types of 1986; Pattengil et al. 1997). Here territorial regions needs particular guidelines to es- herbivorous fishes are abundant and the ar- tablish management plans (Sponaugle et al. eas they defend can cover over 80% of the 2002). Most of the smallest post recruit and surface of some reef habitats (Robertson juveniles that we found in our transects in and Lassig 1980; Ferreira et al. 1998; Cec- sheltered zones belong to species that may carelli et al. 2001). Ferreira et. al. (1998) have shorter pelagic larval duration (e.g., mentioned that this trophic group may Gobiidae, Pomacentridae, Serranidae) in greatly depend on the physical structure of comparison to muraenids, balistids and te- the coral reef and the distribution of the traodontids that are families with the long- associated benthic organisms. An impor- est larval duration of reef fish families and tant part of the diet of some herbivores were present occasionally only in exposed (e.g., Scarus iseri) is detritus and/or calci- zones or in very lower numbers if not vir- fied material (Randall 1967). Browsing scar- tually absent in sheltered zones (Floeter ids (e.g., Sparisoma) that feed on some types and Gasparini 2000; Leis and McCormick of macro algae increase in abundance along 2002; Robertson 2001). In addition, in with this food source at exposed and het- closed, shallow regions such as the internal erogeneous complex zones. Roving herbi- sheltered zones of Bocas del Toro with low vores such as acanthurids have preference water circulation and changes in surface sa- for certain types of algae common in ex- linity may negatively affect the species posed reef flats (Sluka and Miller 2001). richness of the Ichthyoplankton (Dominici- Planktivores (e.g., Serranus tortugarum)at Arosemena et al. 2000; Romero 1992). De- sheltered areas account for 20 to 70% of spite a few interconnected coral reefs in fishes in certain zones of rubble-sand. They Southern Costa Rica, much of the remain- swim above the bottom to mainly feed ing coastline stretching to the Southern on demersal plankton (e.g., amphipods boarder of Nicaragua is unsuitable for reef and harpadicoids copepods), found in low development due to sandy substrate, current areas (Randall 1967; Morales, Al- coastal marshes in the North, and the input varo pers. comm.). This type of habitat is of freshwater by San Juan River (Cortés similar to that proposed by Parrish and Bo- pers. comm.). The abundance of larvae be- land (2004) who called it “banks with infre- ing transported from the reefs in Southern quent relief features”, where planktivorous Nicaragua southwards are likely to decline are most numerous and benthic carnivores resulting in low recruitment in Bocas del increased with an increased substrate com- Toro. Thus, the distance from Northern reef plexity. develop areas and the presence of suitable Bocas del Toro has been subjected to an- interconnected substrate for larval settle- thropogenic disturbances in recent years, ment in neighbouring regions, may consti- due to its increasing importance as a popu- tute an important function of settlement lar tourist destination in the Republic of and recruitment in Bocas del Toro. We sug- Panama. This will increase the human ac- gest that as a result of its particular oceano- tivities such as artisanal extractions, sport graphic regime, larval recruitment may be fishing and the use of the environmental low in Bocas del Toro and even what landscape due to diving activities. This re- would normally be considered an average gion lacks previous baseline data that could fishing pressure may be unsustainable. The 630 A. DOMINICI-AROSEMENA AND M. WOLFF highest-populated centres of the Bocas re- Smithsonian Tropical Research Institute gion are located at Isla Bastimentos, Isla (STRI) at the Bocas del Toro station. We Colón, and Carenero island, where some of thank Ross Robertson for his advice at the massive coral and rocky reefs are lo- STRI; and Carlos L. Ferreira, Sergio Floeter, cated. In our study we confirm that rocky Marc Kochzius, Matthias Birkicht, Suzanne zones are rich in species and should be in- Lao, Tom Nicolai, Fernando Zapata, Carlos cluded within the limits of the Bastimentos Jiménez, Sabine Dittmann, Jorge Cortés, National Park (e.g., rocky shores of Basti- Hector Guzmán, Jorge Ventocilla, Nélida mentos Island and areas of spear or other Gómez, David Kline, Gustavo Alcides Con- artisanal fisheries). In other studies, the re- cheiro Pérez, Pedro Alcolado, Carlos Gam- duction of three-dimensional structure has boa, Ghislain Rompré, Marc Taylor, Scott been identified as one cause for the shift in Smith, Alexander Schröder, Lukas species composition (McKenna 1997), a Schäerer, Werner Ekau, Carlos Guevara, process that is currently underway due to Uta Berger, Christian Jakob and Steve Rob- physical destruction of habitat (e.g., reef ert Niedzielski who also contributed with fishing, anchoring); which will directly or their advice. Juan Gabriel Domínguez and indirectly affect the fish community struc- Irving Bethancourt helped on field and ture. Our results indicate a negative corre- data management, Arcadio Castillo and lation of fish diversity with either the per- Willie Pomare, also helped on the field. We centage of dead coral substrates and/or also thank Sabine Kadler, Christa Müller, low substrate diversity zones (rubble, dead Silke Meyerholz, Andreas Hanning, Kai corals live branching and foliaceous corals) Bergmann, Stefanie Bröhl, Dieter Peterke that normally result from habitat destruc- (ZMT), Harry Barnes, Luis Mou, Adriana tion or the growing of opportunistic species Bilgray, Marissa Batista, Ernesto Peña, of corals due to the high sediment deposition (Cortés 1984). Our study supports the Edgardo Ochoa, Reynaldo Tapia and Mer- need for: 1. the protection of many reef ar- cedes Denis (STRI), for logistic and admin- eas and their benthic fauna (especially istrative support. Map and profile figures those zones of massive and fire coral cover) drawn by Marco Luque Parigi. (for delineation of these areas, see Guzmán and Guevara 1998a,b, 1999); 2. the educa- tion of tourists to not destroy or extract the LITERATURE CITED benthic flora and fauna; and 3. restrict the Acosta, C. A., and D. N. Robertson. 2002. Diversity fishing activities in coral reef areas. There is in coral reef fish communities: the effects of habi- also a need to delimit all reef zones to avoid tat patchiness revisited. Mar. Ecol. Prog. Ser. 227:87- anchorage of boats in reef areas. In addi- 96. tion, other anthropogenic impacts (e.g., Arburto-Oropeza, O., and E. Balart. 2001. Community sediment deposition from extensive banana structure of reef fish in several habitats of a rocky plantations) are also affecting this area. For reef in the Gulf of California. Mar. Ecol. 22(4):283- these regions, habitat degradation of the 305. benthic fauna and flora could be faster than Ault, T. R., and C. R. Johnson. 1998. Spatially and expected. The present situation of Bocas del temporary predictable fish communities on coral Toro may already represent a fish commu- reefs. Ecol.l Monogr. 68(1):25-46. nity structure subject to human induced Beck, M. W. 1997. A test of the generality of the effects changes. of shelter bottlenecks in four stone crabs populations. Ecology 78:2487-2503. Acknowledgments.—This research was Beck, M. W. 1998. Comparison of the measurement and effects of habitat structure on gastropods in sponsored by the World Wildlife Fund rocky intertidal and mangrove habitats. Mar. Ecol. (BMZ, Mesoamerican Biological corridor Prog. Ser. 169:165-178. fellowship), with cooperation of the Centre Bellwood, D. R. 1998. What are reef fishes? Comment for Marine Tropical Ecology in the Univer- on the report by D. R. Robertson: Do coral reef fish sity of Bremen (ZMT), The German Aca- faunas have a distinctive taxonomic structure? demic Exchange Service (DAAD), and the (Coral Reefs. 17:179-186) Coral Reefs 17:187-189. REEF FISH COMMUNITY STRUCTURE 631

Bellwood, D. R., and P. C. Wainwright. 2001. Locomo- Diehl, S. 1988. Foragins efficiency of three freshwater tion in labrid fishes: implications for habitat use fishes: effects of structural complexity and light. and cross-shelf biogeography on the Great Barrier Oikos 3:207-214. Reef. Coral Reefs 20:139-150 Dolmer, P. 1998. The interactions between bed struc- Bellwood, D. R., P. C. Wainwright, C. J. Fulton, and A. ture of Mytilus edilus L. and the predator Asterias Hoey. 2002. Assembly rules and functional groups rubens L. J. Exp. Mar. Biol. Ecol. 228:137-150. at global biogeographical scales. Functl. Ecol. 16: Dominici-Arosemena A., E. Brugnoli-Olivera, S. So- 557-562. lano-Ulate, and A. Ramírez-Coghi. 2000. Ictio- Bohnsack, J. A., and S. P. Bannerot. 1986. A stationary plancton en la zona portuaria de Limón, Costa visual census technique for quantitatively assess- Rica. Rev. Biol. Trop. 48(2):439-442. ing community structure of coral reef fishes. Survey NOAA Tech. Rep. NMFS Cir. 41:1-15. English, S., C. Wilkinson, and V. Baker. 1994. manual for tropical marine resources. Townsville: Booth, D. J., and M. A. Hixon. 1999. Food ration and Australian Institute of Marine Science. condition affect early survival of the coral reef damselfish, Stegastes partitus. Oecologia 121(3):364- Erlandsson, J., V. Kostylev, and G. A. Williams. 1999. 368. A field technique for estimating the influence of Bradbury, R. H., and P. C. Young. 1981. The effects of surface complexity on movement tortuosity in the Cellana grata Ophelia a major forcing function, wave energy, on a coral tropical limpet Gould. 50:215- reef ecosystem. Mar. Biol. Prog. Ser. 5:229-241. 224. Briggs, J. C. 1995. Global Biogeography. Developments Ferreira, C. E. L., A. C. Peret, and R. Coutinho. 1998. in Paleontology and Stratigraphy, 14, Amsterdam: Seasonal grazing rates and food processing by Elsevier. tropical herbivorous fishes. J. Fish Biol. 53:222-235. Caribbean Coastal Marine Productivity (CARICOMP). Ferreira, C. E. L., J. E. A. Gonçalves, and R. Coutinho. 2001. Methods Manual Levels 1 and 2. Manual of 2001. Community structure of fishes and habitat methods for mapping and monitoring of physical complexity in a tropical rocky shore. Env. Biol. Fish and biological parameters in the coastal zone of the 61:353-369. Caribbean. University of the West Indies Mona, King- Ferreira, C. E. L., S. R. Floeter, J. L. Gasparini, B. P. ston Jamaica and Florida Institute of Oceanography. Ferreira, and J. C. Joyeux. 2004. Trophic structure 90 p. patterns of brazilan reef fishes: a latitudinal com- Ceccarelli, D. M., G. P. Jones, and L. J. McCook. 2001. parison. J. Biogeogr. 31:1-13. Territorial damselfishes as determinants of the Floeter, S. R., and J. L. Gasparini. 2000. The southwest- structure of benthic communities on coral reefs. ern Atlantic reef fish fauna: composition and zoo- Oceanogr. Mar. Biol. Annu. Rev. 39:355-389. geographic patterns. J. Fish Biol. 56:1099-1114. Clarke, R. D. 1988. Chance and order in determining Floeter, S. R., et al. 2001. Geographic variation in reef- fish species composition on small coral patches. J. fish assemblages along the Brazilian coast. Global Exp. Mar. Biol. Ecol. 115:197-212. Ecol. Biogeogr. 10(4):423-433. Clifton, K E., and L. M. Clifton.1998. A survey of fishes from various coral reef habitats within the Cayos Floeter, S. R., C. E. L. Ferreira, A. Dominici-Arosemena, Cochinos Marine Reserve, Honduras. Rev. Biol. and I. Zalmon. 2004. Latitudinal gradients in At- Trop. 46:109-124. lantic reef fish communities: trophic structure and spatial use patterns. J. Fish Biol. 64:1-20. Coates, A. G., et al. 1992. Closure of the Isthmus of Panamá: the near-shore marine record of Costa García-Chartron, J. A., and A. Pérez-Ruzafa. 2001. Spa- Rica and Western Panamá. Geol. Soc. Amer. Bull. tial pattern and the habitat structure of a Mediter- 104:814-828. ranean rocky reef fish local assemblage. Mar. Biol. Collins, L. S., A. F. Budd, and A. G. Coates. 1996. Ear- 138:917-934. liest evolution associated with closure of the Tropi- Goldman, B., and F. H. Talbot. 1976. Aspects of the cal American Seaway. P. Natl. Acad. Sci. 93:6069- ecology of coral reef fishes. In Biology and geology of 6072. coral reef, ed. O. A. Endean, 125-154. New York: Cortés, J., and M. J. Risk. 1984. El arrecife coralino del Academic Press. Parque Nacional Cahuita, Costa Rica. Rev. Biol. Greb, L., et al. 1996. Ökologie ind Sedimentologie Trop. 32(1):109-121. eines rezenten Rampensystem an der Karibikküste Cowen, R. K. 2002. Larval dispersal and retention and von Panama. Profil Band 10. Universität Stuttgart, consequences for population conectivity. In Coral 168 p. reef fishes: dynamics and diversity in a complex ecosys- Guzmán, H. M., and C. A. Guevara. 1998a. Arrecifes tem, ed. P. F. Sale, 149-170. San Diego, CA.: Aca- coralinos de Bocas del Toro, Panamá: I. Distribu- demic Press. ción,estructura, diversidad y estado de conserva- Dethier, M. N. 1994. The ecology of intertidal algal cion de los arrecifes continentales de la Laguna de crusts: variation within a functional group. J. Exp. Chiriquí y Bahia Almirante. Rev. Biol. Trop. 46(3): Mar. Biol. Ecol. 177:37-71. 601-622. 632 A. DOMINICI-AROSEMENA AND M. WOLFF

Guzmán, H. M., and C. A. Guevara. 1998b. Arrecifes Serrana y Roncador, Caribe occidental, Colombia. coralinos de Bocas del Toro, Panamá: II. Distribu- Bol. Ecotrópica 32:25-41. ción, estructura, diversidad y estado de conser- Nagelkerken, I., M. Dorenbosch, W. C. E. P. Verberk, vación de los arrecifes de las Islas Bastimentos, E. Cocheret de la Morinière, and G. van der Velde. Solarte, Carenero y Colón. Rev. Biol. Trop. 46(4): 2000a. Importance of shallow-water biotopes of a 889-912. Caribbean bay for juvenile coral reef fishes: pat- Guzmán, H. M., and C. A. Guevara. 1999. Arrecifes terns in biotope association, community structure coralinos de Bocas del Toro, Panamá: III. Distri- and spatial distribution. Mar. Ecol. Progr. Ser. 202: bución, estructura, diversidad y estado de con- 175-192. servación de los arrecifes de las islas Pastores, Cristobal, Popa y Cayo Agua. Rev. Biol. Trop. 47(4): Nagelkerken, I., et al. 2000b Importance of mangroves, 659-676. seagrass beds and the shallow coral reef as a nurs- Hixon, M. A., and J. P. Beets. 1993. Predation, prey ery for important coral reef fishes, using a visual refuges, and the structure of coral-reef assem- census technique. Estuar. Coast. Shelf Sci. 51:31- blages. Ecol. Monogr. 63:77-101. 44. Jones, G. P. 1988. Experimental evaluation of the ef- Parrish, F. A., and R. C. Boland. 2004. Habitat and fects of habitat structure and competitive interac- reef-fish assemblages of banks in the North West- tions on the juveniles of two coral reefs fishes. J. ern Hawaiian Islands. Marine Biology www Exp. Mar. Biol. Ecol. 123:115-126. .springerlink.com/index/WY3L2F2VX6U83AQP Jones, G. P., D .J. Ferrel, and P. F. Sale.1991. Fish pre- .pdf dation and its impact on the invertebrates of coral Pattengill, C. V., B. X. Semmens, and S. R. Gittings. reefs and adjacent sediments. In The ecology of fishes 1997 Reef fish trophic structure at the Flower Gar- on coral reefs, ed. P. F. Sale, 156-179. San Diego, CA.: dens and Stetson Bank, NW Gulf of Mexico. Proc. Academic Press. 8th Int. Coral Reef Sym. 1:1023-1028. Jones, G. P., and M. I. McCormick. 2002. Numerical Peet, R. K. 1974. The measurement of species diversity. and energetic processes in the Ecology of coral reef Ann. Rev. Ecol. Syst. 5:285-307. fishes. In Coral reef fishes: dynamics and diversity in a complex ecosystem, ed. P. F. Sale, 221-238. San Diego, Pielou, E. C. 1975. Ecological diversity. New York: John CA.: Academic Press. Wiley and Sons. Khalaf, M. A., and M. Kochzius. 2002. Changes in tro- Randall, J. E. 1967. Food habits of reef fishes of the phic community structure of shore fishes at an in- West Indies. Stud. Trop. Oceanogr. 5:665-847. dustrial site in the Gulf of Aqaba. Mar. Ecol. Progr. Roberts, M. R., and R. F. G. Ormond. 1987. Habitat Ser. 239:287-299. complexity and coral reef diversity and abundance Leis, J. M., and M. I. McCormick. 2002. The Biology, on Red Sea fringing reefs. Mar. Ecol. Prog. Ser. 41: behaviour, and Ecology of the pelagic, larval stage 1-8. of coral reef fishes. In Coral reef fishes: dynamics and Robertson, D. R., H. P. A. Sweatman, E. A. Fletcher, diversity in a complex ecosystem, ed. P. F. Sale, 171- and M. G. Cleland. 1976. Schooling as a mecha- 199. San Diego, CA.: Academic Press. nism for circumventing the territoriality of com- Levin, P. S., and C. B. Grimes. 2002. Reefs fish Ecology petitors. Ecology 57:1208-1220. and grouper conservation and management. In Coral reef fishes: dynamics and diversity in a complex Robertson, D. R., and B. Lassig. 1980. Spatial distribu- ecosystem, ed. P. F. Sale, 377-389. San Diego, CA.: tion patterns and coexistence of a group of territo- Academic Press. rial damselfishes from the Great Barrier Reef. Bull. Mar. Sci. 30:187-203. Luckhurst, B. E., and K. Luckhurst. 1978. Analysis of the influence of substrate variables on coral reef Robertson, D. R. 1998. Do coral-reef fish faunas have a fish communities. Mar. Biol. 49:317-323. distinctive taxonomic structure? Coral Reefs McGehee, A. 1994. Correspondence between assem- 17:1-8. blages of coral reef fishes and gradients of water Robertson, D. R. 2001. Population maintenance among motion depth, and substrate size off Puerto Rico. tropical reef fishes: Inferences from small-island Mar. Ecol. Prog. Ser. 105:243-255. endemics. PNAS. 98:5667-5670. McKenna, J. 1997. Influence of Physical Disturbance Rodríguez, J., and E. Villamizar 2000. Estructura de la on the Structure of Coral Reef Fish Assemblages in comunidad de peces arrecifales de playa Mero, the Dry Tortugas. Caribb. J. Sci.. 33(1-2):82-97. Parque Nacional Morrocoy, Venezuela. Rev. Biol. McLain, D. K., and A. E. Pratt. 1999. Nestedness of Trop. 48(1):107-113. coral reef fishes accross a set of fringing reefs. Oikos Rogers, C. S., G. Garrison, R. Grober, Z. M. Hillis, and 85:53-67. M. A. Franke. 1994. Coral reef monitoring manual Mejía, L. S., J. Garzón-Ferreira, and A. Acero. 1998. for the Caribbean and Western Atlantic. Southern Peces registrados en los complejos arrecifales de Regional Office National Park Service. The Nature Con- los cayos Courtown, Albuquerque y los bancos servancy WWF. 100 pp. REEF FISH COMMUNITY STRUCTURE 633

Romero, M. 1992. Ictioplancton de la región nororien- Sokal, R. R., and F. G. Rohlf. 1980. Biometry. San Fran- tal de la plataforma de Cuba (Zona D). Cienc. Biol. cisco, CA.: Freeman. 8:82-94. Sponaugle, S., et al. 2002. Predicting self-recruitment Sagarin, R. D., and S. D. Gaines 2002. The abundance in marine populations: Biophysical correlates and ‘Centre’ distribution: To what extend is it a biogeo- mechanisms. Bull. Mar. Sci. 70(1):341-375. graphical rule? Ecol. Lett. 5:137-147. Steneck, R. S., and M. N. Dethier. 1994. A functional Sale, P. F. 2002. The science we need to develop for group approach to the structure of algal dominated more effective management. In Coral reef fishes: dy- communities. Oikos 69:476-498. namics and diversity in a complex ecosystem, ed. P. F. Talbot, F. H., and B. Goldman. 1972. A preliminary Sale, 361-376. San Diego, CA.: Academic Press. report on the diversity and feeding relationships of Siegel, S. 1970. Estadística no-paramétrica aplicada a las the reef fishes of One Tree Island, Great Barrier ciencias de la conducta., México, D.F.: Trillas. Reef system. Proc. First Intl. Coral Reef Symp. 1:425- Sierra, L. M., R. Claro, J. P. García-Arteaga, and E. 440. Valdés-Muñoz. 1990. Estructura trófica de las co- Ter Braak, C. J. F., and P. F. M. Verdonschot. 1995. munidades de peces que habitan en diferentes bió- Canonical correspondance analisys and related topos del Golfo de Batabanó.InAsociaciones de multivariate methods in aquatic ecology. Aquat. peces en el Golfo de Batabanó, ed. R. Claro, 83-95, La Sci. 57:255-289. Habana, Cuba: Editorial Academia. Sierra, L. M., and J. P. García. 1996. Estructura de las Veron, J. E. N. 1995. Corals in space and time. Ithaca, comunidades de peces y la pesca artesanal en los NY.: Cornell University Press. arrecifes coralinos del Caribe costarricense. Wolff, M., and E. Alarcón 1993. Structure of a scallop Resúmenes del 8vo simposio de arrecifes de coral, Pan- Argopecten purpuratus (Lamarck, 1819) dominated amá, p. 182. subtidal macro-invertebrate assemblage in North- Sluka, R. D., and M. W. Miller. 2001. Herbivorous fish ern Chile. J. Shellfish Res. 12(2):295-304. assemblages and herbivory pressure on Laamu Zar, J. H. 1996. Biostatistical analysis,3rd ed. Upper Atoll, Republic of Maldives. Coral Reefs 20:255-262. Saddle River, New Jersey: Prentice Hall. 634 A. DOMINICI-AROSEMENA AND M. WOLFF

APPENDIX. List of species found and their distribution in the study zones (1-12). TL = Trophic Level; C = Carnivorous; I = mobile Invertebrates; O = Omnivore; P = Piscivore; PL = Planctivore; RH = Roving Herbivore; SI = Sessile Invertebrate feeder; TH = Territorial Herbivore.

Family/Name TL 123456789101112

ACANTHURIDAE Acanthurus bahianus RHXXXXXXXXXX X X Acanthurus chirurgus RHXXXXXX XX X X Acanthurus coeruleus RHX XXXXX XX X X APOGONIDAE Apogon townsendi MI X Apogon pseudomaculatus MI X AULOSTOMIDAE Aulostomus maculatus C X XXXXX BALISTIDAE Aluterus scriptus OX Balistes vetula SI X Cantherhines pullus OXX BLENNIDAE Ophioblennius atlanticus TH X X X Parablennius marmoreus OXX X XX BOTHIDAE Bothus lunnatus C X CARANGIDAE Carangoides bartholomaei PX XXX Caranx crysos P XX XXXX X X Caranx ruber P XXXX XXXXX X X Chloroscombrus chrysurus Pl X Decapterus punctatus CXXXX Decapterus sp. C X X CHAETODOTIDAE Chaetodon capistratus SIXXXXXXXXXX X Chaetodon ocellatus O XX X XXXXX X Chaetodon striatus SI X XXXXXXX X CIRRHITIDAE Amblycirrhitus pinos CX DASYATIDAE Dasyatis americana OXX XXX DIODONTIDAE Diodon holocanthus MI X Diodon hystrix MI X X ELOPIDAE Megalops atlanticus CX GERREIDAE Gerres cinereus MI X X X GINGLYMOSTOMATIDAE MI Ginglymostoma cirratum CXX GOBIIDAE Coryphopterus dicrus OXX X XX Coryphopterus personatus PLXX XX XXXX Gnatholepis thompsoni OX Ctenogobius saepepallens OXX Elacatinus dilepsis OX Elacatinus xanthiprora OXX X XXX Elacatinus sp. O X GRAMMATIDAE Gramma loreto MI X REEF FISH COMMUNITY STRUCTURE 635

APPENDIX. Continued.

Family/Name TL 123456789101112

HAEMULIDAE Anisotremus surinamensis CX Anisotremus virginicus MIX X XX XXXX Haemulon aurolineatum MI X X XXX X X Haemulon carbonarium MI X X X X Haemulon chrysargyreum MI X X Haemulon flavolineatum MI XXXXXXXX X X Haemulon macrostomum MI X XX XXX Haemulon parra CX Haemulon plumierii C XXXXXXXXXX X X Haemulon sciurus CX X XX HOLOCENTRIDAE Holocentrus adscensionis C XXXXXXXXX X X Holocentrus rufus C X XXX Myripristis jacobus CXXXX Neoniphon marianus MI X Sargocentron vexillarium C XXX KYPHOSIDAE Kyphosus sectator RH X X X LABRIDAE Bodianus rufus MI X X Halichoeres bivittatus MIXXXXXX XXX X X Halichoeres garnoti MI XXXX X X Halichoeres maculipinna MI X X X X X Halichoeres pictus MI X X Halichoeres poeyi MI X X X X Halichoeres radiatus MI X X Lachnolaimus maximus MI X Thalassoma bifasciatum PL XXXXXXXX X X LABRISOMIDAE Labrisomus nuchipinnis CX Malacoctenus triangulatus MI X X XXX LUTJANIDAE Lutjanus analis C XX XX XXXX Lutjanus apodus C XX X XXXXX Lutjanus cyanopterus C X Lutjanus griseus CXXXX Lutjanus jocu CXX X Lutjanus mahogoni CX X X X X Lutjanus synagris CX X Ocyurus chrysurus C XXXXXXXXXX X X MALACANTHIDAE Malacanthus plumieri CX MULLIDAE Mulloidichthys martinicus MIXX XXX XX X X Pseudupeneus maculates C XXXXX XXXX X X MURAENIDAE Echidna catenata C X Gymnothorax miliaris C X OPISTOGNATHIDAE Opistognathus aurifrons PLXXX X X OSTRACIIDAE Lactophrys triqueter O X 636 A. DOMINICI-AROSEMENA AND M. WOLFF

APPENDIX. Continued.

Family/Name TL 123456789101112

PEMPHERIDAE Pempheris schomburgkii PL X X POMACANTHIDAE Holacanthus ciliaris OXXXXX Holacanthus tricolor OX XXX Pomacanthus arcuatus O XXXX XXXX POMACENTRIDAE Abudefduf saxatilis O XXX X X X Chromis cyanea PL X Chromis multilineata PL X X Microspathodon chrysurus TH XXXX X X Stegastes adustus TH XXXX X X Stegastes leucostictus THXXXXXXXXXX X X Stegastes partitus THXXXXXXXXXX X X Stegastes planifrons THXXXXXXXXXX X X SCARIDAE Scarus coelestinus RH X XXXX X Scarus iseri RHXXXXXXXXXX X X Scarus taeniopterus RH X Scarus vetula RH X Sparisoma aurofrenatum RH XXXXXXXXX X X Sparisoma chrysopterum RHXX XXXXX X X Sparisoma radians RH X X X X Sparisoma rubripinne RH XX XXXXX X X Sparisoma viride RHXXXXXXXXXX X X SCIAENIDAE Equetus punctatus MIX XXX Odontoscion dentex C X XXXX X SCOMBRIDAE Scomberomorus maculatus PX X Scomberomorus regalis PX XX X XX SERRANIDAE Epinephelus adscensionis CXX Cephalopholis cruentata C XXXXXXXXXX X X Cephalopholis fulva CXXX Epinephelus guttatus CXXX Epinephelus striatus C X Hypoplectrus aberrans CX Hypoplectrus guttavarius CX Hypoplectrus indigo CX X Hypoplectrus nigricans C XXXXX XXXX Hypoplectrus puella C XXXXXXXXXX Hypoplectrus sp. C XXXXXXX X Hypoplectrus unicolor C XXXXX XXXX Mycteroperca venenosa CX Rypticus maculatus CX Rypticus saponaceus C X Serranus baldwini MI X Serranus tigrinus MIXXXXX XX X X Serranus tortugarum PL X X X SPARIDAE Archosargus rhomboidalis OX X X Calamus penna MI X X X X SPHYRAENIDAE Sphyraena barracuda PX XXX X REEF FISH COMMUNITY STRUCTURE 637

APPENDIX. Continued.

Family/Name TL 123456789101112

SYNODONTIDAE Synodus intermedius PX X XXX TETRAODONTIDAE Canthigaster rostrata O XXXXX XXXX X X Sphoeroides spengleri OX X UROLOPHIDAE Urobatis jamaicensis CX