<I>Haemulon Sciurus

BULLETIN OF MARINE SCIENCE, 80(3): 473–495, 2007

nearshore habitat use by gray snapper (lutjanus griseus) and bluestriped grunt (Haemulon sciurus): environmental gradients and ontogenetic shifts

Craig H. Faunce and Joseph E. Serafy

ABSTRACT Fringing mangrove forests and seagrass beds harbor high densities of juvenile snappers and grunts compared to bare substrates, but their occupancy of these habitats is not homogeneous at ecologically meaningful scales, thus limiting our ability to compare habitat value. Here, density and size information were used to determine how gray snapper, Lutjanus griseus (Linnaeus, 1758) and bluestriped grunt, Haemulon sciurus (Shaw, 1803), use vegetated habitats during their ontogeny, and how their use of mangrove forests varied with season across broad spatial scales and physicochemical conditions. Both species exhibited a three-stage ontogenetic strategy: (1) settlement and grow-out (8–10 mo) within seagrass beds, (2) expansion to mangrove habitats at 10–12 cm total length, and (3) increasing utilization of inland mangroves during the dry season and with increasing body size. For fishes inhabiting mangroves, multivariate tests revealed that the factors distance from oceanic inlet and water depth were stronger predictors of reef fish utilization than the factors latitude, temperature, or habitat width. These findings highlight that the nursery function of mangrove shorelines is likely limited to the area of immediately accessible habitat, and that more expansive forests may contain a substantial number of larger adult individuals.

The importance of coastal vegetated seascapes to the early life history stages of diadromous nekton has been well documented, and has led to the conclusion that many temperate and warm-water marine fishes are “estuarine dependent” G( unter, 1967; McHugh 1967; Day Jr. et al., 1989; Blaber, 2002). However, the spatial separation of juveniles from adults and their utilization of coastal vegetated habitats is not limited to estuarine waters. Day (1951) was among the first to postulate that the fauna observed within a coastal embayment were drawn to “quiet water” conditions rather than estuarine conditions per se. Expanding this idea further, Nagelkerken and van der Velde (2002) offered the idea of “bay habitat dependence”. Despite attempts to provide guidelines for assessing the comparative “importance value” among fish habitats (USDOC, 1996; Beck et al., 2001), the relative utilization of multiple habitats by fishes has been difficult to quantify. Although numerous studies of animal-environment relationships have been conducted, these studies largely have been limited by a focus on a single habitat type or life-history stage, and/or by sampling at spatial scales that are too small to adequately capture the full suite of habitat use by fishes during their ontogeny (Pittman and McAlpine, 2003). At latitudes that support coral reefs, natural seascapes are vegetated primarily by seagrass beds and coastal mangrove forests. Here, many exploited reef fishes occupy inshore regions as juveniles before migrating offshore to reproduce thereby under- going an ontogenetic pattern of habitat utilization similar to that exhibited by diadromous fishes in estuarine waters (Ogden, 1997). It has been postulated that the presence of mangroves and seagrass beds serve as extra “waiting room” habitats for juvenile coral reef fishes, and that adopting such a life-history strategy may buffer

against poor recruitment years (Bardach, 1959; Parrish, 1989). Recently, there have been renewed investigations into whether the presence of seagrass and mangrove habitats in the Caribbean enhance populations of fishes inhabiting adjacent coral reefs (e.g., Nagelkerken et al., 2002; Dorenbosch et al., 2004; Halpern, 2004; Mumby et al., 2004). In all of these studies, mangrove shorelines have been considered as ho- mogenous habitats. However, mangrove shoreline habitats with different geographic locations, physical structure, and associated physicochemical environments may not contribute equally to future adult populations of reef fishes. Despite these management implications, few field studies have investigated the factors responsible for the utilization of mangrove shoreline habitats by reef fishes during their ontogeny, and those few have limited analyses to simple correlations performed at the entire fish assemblage level (Faunce and Serafy, 2006). The preceding is important because there is growing evidence that the processes or factors underpinning observed patterns of habitat utilization are species- and scale-dependent. For example, Martino and Able (2003) suggested that patterns in fish assemblage structure across large spatial scales (i.e., > 10 km) are due to individual species responses to dominant environmental gradients, whereas smaller-scale (1 km) patterns are due to habitat selection, competition, and/or predator avoidance strategies. In addition, recent studies in the Netherlands Antilles have reported patterns in the ontogenetic utilization of seagrass and mangrove habitats by reef fishes as a function of geographic scale (Nagelkerken et al., 2000b; Cocheret de la Morinière et al., 2002). However, their results apply to a 3 km wide bay. Determining whether similar patterns of ontogenetic utilization occur across larger spatial scales in other regions of the western Atlantic is important, especially if reef fish can move seasonally over distances more than a few kilometers (Gillanders et al., 2003; Pittman and McAlpine, 2003). An expansive (ca. 1000 km2) series of lagoons and embayments are found in southeastern Florida bordered to the west by the Florida mainland and to the east by the Florida Keys. In general, the influence of fresh water increases to the west-southwest, whereas oceanic and tidal influences are greater to the east and northeast. Across this diverse hydrographic seascape, the bays of southeastern Florida are primarily vegetated by the same seagrass (i.e., Thalassia testudinumBanks ex. König, 1805) and mangrove (i.e., Rhizophora mangle Linnaeus) species. Therefore, the region provides the opportunity to examine how reef fish utilization of seagrass and mangrove habitats changes with ontogeny and how utilization of a single habitat type (i.e., mangroves) varies across broad spatial scales and physicochemical conditions. This study aims to address these issues for two reef fishes common to the region, gray snapper,L utjanus griseus (Linnaeus, 1758) and bluestriped grunt, Haemulon sciurus (Shaw, 1803).

Methods

Data Collection.—The southeastern Florida coast was defined as the area between Key Biscayne to the northeast (80°09Ń, 25°43.8´W), Russell Key to the southwest (25°4.2Ń, 80°38.4´W), and Tavernier Creek to the southeast (25°06Ń, 80°32.4´W). The mangrove-lined shorelines of southeastern Florida were divided into five spatially-distinct strata. In order of their general position from west to east these strata include: (1) the mainland (ML); (2) islands within bays (IS); (3) landbridges (LB) that connect portions of the ML and the Florida Keys; (4) the western (i.e., leeward) shorelines of the Keys FAUNCE AND SERAFY: NEARSHORE ONTOGENIC HABITAT SHIFTS BY TWO REEF FISHES 475

(LK); and (5) the eastern (i.e., windward) shorelines of the Keys (WK; Fig. 1A). These strata were sampled using an iterative stratified random design during the wet (June–Au- gust) and dry (January–March) seasons of 2000 and 2001 following Cochran (1977) and detailed by Ault et al. (1999a; Fig. 1B). Sample locations, including five alternates, were selected each season by first dividing each stratum into sequentially numbered 30 m latitudinal sections, and using a random number generator to select the latitude of each sample. Samples consisted of (30 × 2 m) 60 m2 visual belt transects wherein a trained underwater observer identified, enumerated, and estimated the minimum, average, and maximum total length (TL) of each fish species to the nearest 2.5 cm. Only one individual conducted all fish surveys, thus eliminating inter-observer bias (St. John, 1990).P rior to the study, the observer’s ability to identify and rapidly enumerate fishes accurately was tested according to three criteria. First, the observer must have correctly identified all of the species common to the region (as defined from table 1 of Serafy et al., 2003) from digital photographs, including ontogenetic color variations where appropriate (e.g., parrotfishes). The observer must also have correctly estimated the identification of all species and number of fish within 10% from a 30 s digital video clip of a 30 m mangrove transect. Finally, following Bell et al. (1985), accurate and precise length assessment of fishes underwater was achieved by having the observer repeatedly estimate the sizes of plastic pipes of various lengths. To ensure fishes could be effectively observed within each transect, horizontal visibility was determined prior to surveys using a vertically mounted secchi disk and a mea- suring line. If visibility was < 2 m, the survey location was abandoned and an alternative location was visited. If no suitable alternative was available, locations were re-visited on a later date. Because fish observations in mangroves may be influenced by tidal stage and light levels, we attempted to control for this factor by conducting surveys between 0900 and 1700 on days that maximized the hours of daytime flood tide. A suite of environmental variables was also measured at each sample location. The factors water temperature and salinity were obtained using a Hydrolab® multi-probe instru- ment. Water depth and the width of mangrove canopy (WOC) were used as metrics of the vertical and horizontal components of the mangrove habitat, respectively. Depth was averaged from measurements at the mid- and end-points of each transect, while width of canopy was defined as the distance from the outer submerged mangrove drop- or prop-root to the upland edge. Latitude and distance to nearest oceanic inlet were used as landscape metrics. Distance to inlet (DI) was approximated with straight line measurements using Geographic Information System software (ArcView ver. 3.2). Because these distances sometimes crossed land barriers, they are considered here as indices and not potential routes of travel. Data Manipulations.—Length-frequency distributions from mangrove samples were constructed for L. griseus and H. sciurus following Meester et al. (1999). Briefly, this technique assumes the length-frequency distribution per sample has a mode at the mean length estimate with the two tails defined by the minimum and maximum lengths estimates. Length information was used to assign fish to one of three life-history components (LHC). Individuals smaller than the size at age 1 from published age-and-growth studies were designated as juveniles. Published size at sexual maturity (i.e., the size at which > 50% of individuals are sexually mature) was used to define the two remaining groups; individuals larger than their respective size at maturity were classified as adults, and remaining fish were classified as sub-adults. The literature sources, length ranges, and data manipulations used to define each LHC are given in Table 1. Data Analyses.—Average size and density information were generated for L. griseus and H. sciurus inhabiting seagrass beds and mangroves of southeastern Florida. Data 476 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007 95.1 119.0 253.3 247.1 (mm TL) (mm Final size

Source (L-L conversions) Source (L-L Billings and Munro, 1974 Billings and Munro, 1974 Haemulon plumeieri; Manooch III and Matheson III, 1981 95.1 TL 95.1 103.5 FL 220.0 FL 195.0 SL Input size none Edits Mean 2 and 3 Mean 6 and 7 Source (sizes) Manooch III and Matheson III, 1981 Billings and Munro, 1974 Males: Domeier et al., 1996 García-Arteaga, 1992 Billings and Munro, 1974 Females: Domeier et al., 1996 Mean 4 and 5 Both sexes: Stark II, 1970 Size (mm) 95.1 TL 95.1 105 FL 182 SL 102 FL 220 FL 198 SL 190 SL 200 SL 1 2 4 3 8 5 6 7 ID Source Table 1. Table Information used to determine length; cut-off fork sizes FL: for conversions; life-history length-length components L-L: (LHC) sub-adults. used defined in were analyses. in-between Individuals those less and than adults, the defined size were at maturity age-1 at were size definedthe as than larger those juveniles, TL: total length. SL: standard length; Species Age-1 Lutjanus griseus Haemulon sciurus Size at maturity Lutjanus griseus Haemulon sciurus FAUNCE AND SERAFY: NEARSHORE ONTOGENIC HABITAT SHIFTS BY TWO REEF FISHES 477

from seagrass beds were obtained from raw data used by Serafy et al. (1997), who sampled fishes at eight sites within Biscayne Bay over 14 consecutive months using paired “rollerframe” trawls (3-m wide × 1-m high with 10-mm mesh) and identified and measured all captured fishes for TL. Average density and size data were plotted for each month, compared to published spawning seasons for L. griseus (Domeier et al., 1996; García-Cagide et al., 2001) and H. sciurus (Munro et al., 1973) and used to elucidate the timing of first recruitment, defined as the month when a minimum average length was observed during the year. Mean densities of each L. griseus and H. sciurus LHC, were determined for each mangrove stratum / season combination using a delta-distribution mean estimator; a measure of fish abundance (hereafter density, or fish 60 m–2) that separately considers the proportion of occupied sites and the number of individuals within them (Pennington, 1983). Patterns in these density values were subsequently compared to determine the relative contribution of each LHC to the overall distribution of each species. In addition, the average size of L. griseus and H. sciurus was generated per stratum per season. These data were plotted against the average distance (to inlet) index. Multivariate analyses were performed on the LHC assemblage with the statistical package PC-ORD (ver. 4.3) following the general recommendations of McCune and Grace (2002). The average number of juveniles, sub-adults, and adults ofL . griseus and H. sciurus observed within each stratum per year, was calculated and ln(x+1) transformed prior to analyses. A graphical representation of the underlying structure of these data was performed using Non-metric Multidimensional Scaling (NMS). A six-dimensional solution to a Bray-Curtis ordination was used to seed the initial geometry for NMS. This ordination was calculated using Sorensen (Bray and Curtis) distances, minimum devia- tion endpoint selection, and Euclidean geometry. A six-dimensional NMS solution was iteratively reduced (500 maximum) using an instability criterion of 0.0001. One hundred runs were performed on real data and 50 runs were performed on data that had been randomized within columns (i.e., each LHC). These provided the basis for a Monte Carlo test between the stress of each dimension from real and randomized data (i.e., that expected by chance). The appropriate number of dimensions for the final NMS solution and the significance of Monte Carlo tests for each dimension were determined from visual inspection of the resulting scree-plot. One final run on real data was performed using the appropriate number of dimensions (axes) and the same number of iterations and instability criterion as for prior runs (McCune and Grace, 2002). Whereas NMS was used to elucidate relationships among species and life-history stages at broad spatial scales (i.e., among strata), Canonical Correspondence Analysis (CCA) was performed to graphically express each LHC in terms of the environmental variables measured at sampling (i.e., direct gradient analysis, ter Braak, 1986). This method con- strains the ordination of the LHC matrix by a linear multiple regression on the environmental variable matrix, so that each successive canonical axis accounts for a smaller proportion of the total variance. Fish data were ln(x+1) transformed and standardized to maximum (Minchin, 1987), whereas environmental data were screened for outliers and standardized to z-scores. To control for the influence of “double-zero” matches in the data, only samples with at least one individual belonging to an LHC were included in the analyses (Legendre and Legendre, 1983). The CCA was run with one hundred iterations with randomized site locations to facilitate Monte Carlo tests between (a) the eigenvalues and species-environment correlations for each axis that resulted from CCA and (b) those expected by chance. Canonical Correspondence Analysis produces a biplot where environmental variables are represented as arrows (vectors) radiating from the origin of the ordination. For analy- 478 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 1. Map of Southeastern Florida depicting (A) landscape features and strata used in sampling design and (B) locations of visual transect samples. FAUNCE AND SERAFY: NEARSHORE ONTOGENIC HABITAT SHIFTS BY TWO REEF FISHES 479

Figure 2. Monthly average (± 1 SE) densities (top panels), and sizes (lower panels) of Lutja- nus griseus and Haemulon sciurus within Biscayne Bay seagrass beds. Horizontal arrows depict spawning seasons (for sources see Materials and Methods). Vertical arrows depict timing of initial recruitment.

ses, row (i.e., site) and column (i.e., environmental factor or LHC) scores were centered by unit variance, and optimized so that LHC scores (i.e., biplot projection points) were weighted mean sample location scores. In this way distances between each LHC facilitated direct spatial interpretation with environmental variables in the biplot (McCune and Grace, 2002). The length of an environmental vector is related to strength of the relationship between (a) the environmental variable that vector represents, and (b) the species community (i.e., each LHC belonging to L. griseus and H. sciurus). The order of LHC projection points with respect to an environmental vector (measured by “drawing” a perpendicular line from the projection point to the vector) corresponds to the rank- ing of LHC weighted averages with respect to that environmental factor (Jongman et al., 1995).

Results

Utilization of Seagrass Beds.—Recruitment of L. griseus and H. sciurus into Biscayne Bay seagrass beds was determined from temporal patterns in density and size (Fig. 2). First recruitment occurred during September–October (at the end of the spawning season) for L. griseus, when individuals averaging 7.8 mm TL were col- 480 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 3. Seasonal delta-distribution mean estimators (DME ± 1 SE) belonging to Lutjanus griseus and Haemulon sciurus observed within five mangrove shoreline strata of southeastern Flor- ida. Strata abbreviations: IS: Islands; LB: Landbridges; LK: Leeward Key; ML: Mainland; WK: Windward Key. Note: y-axes are not of equal scale. FAUNCE AND SERAFY: NEARSHORE ONTOGENIC HABITAT SHIFTS BY TWO REEF FISHES 481

Figure 4. Seasonal proportion of juvenile, sub-adult, and adult Lutjanus griseus and Haemulon sciurus observed within mangrove shoreline strata within southeastern Florida. Strata abbreviations follow Figure 3. lected in densities of 1.90 1000 m–2. For H. sciurus, first recruitment occurred during February (at the middle of the spawning season), when the smallest individuals of the year (averaging 5.5 mm TL) were collected in densities of 9.93 1000 m–2. For both species, monthly average density declined and average length increased during the period between first recruitment and approximately 7–8 mo afterward. The greatest average length collected from seagrass beds was 11.4 mm TL for L. griseus and 10.8 mm TL for H. sciurus. Utilization of Mangrove Shorelines.—Density patterns among mangrove strata differed by season and life-history stage for both species. Relatively high densities of all L. griseus were observed among the IS and LB during the dry season, and within the LK and WK during the wet season (Fig. 3). Although juvenile L. griseus were largely restricted to Keys strata (i.e., the LK and WK) during both seasons, sub- adults and adults exhibited seasonal patterns similar to those observed for all fish. For example, relatively high densities of sub-adult and adult L. griseus were observed during the dry season within the center-strata, where IS > LB for sub-adults, and IS >> LB for adults. During the wet season the greatest density of sub-adults was observed within the LK, and that of adults was observed within the LB. Compared to patterns in L. griseus, relatively high densities of all H. sciurus were observed within Keys-strata during both seasons (Fig. 3). This trend was also observed for juveniles and sub-adults of this species. However, whereas juveniles were rarely observed within the ML, IS, and LB, densities of sub-adults within these strata were more comparable to those within Keys-strata. Spatial differences in H. sciurus seasonal density patterns were most evident for adults. For example, relatively high 482 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 5. Size distributions of Lutjanus griseus and Haemulon sciurus observed within seagrass and mangrove shoreline habitats of southeastern Florida. Hashed areas depict range of monthly average size data for fishes collected from seagrass beds (depicted in Figure 3). Average (± 1 SE) sizes of fishes from each mangrove strata during the wet and dry seasons are plotted as a function of the average (± 1 SE) distance from the nearest oceanic inlet. Strata abbreviations follow Figure 3. For L. griseus, average size = 15.09 + 6.01[1 − e–0.36(distance from inlet)] (Adj. r2 = 0.33, P = 0.1002). For H. sciurus, average size = 10.26 + 11.05[1 − e–0.55(distance from inlet)] (Adj. r2 = 0.57, P = 0.0216). densities of adults were observed within the IS during the dry season and within the LK during the wet season (Fig. 3). These results demonstrate that utilization of western mangrove shorelines by L. griseus and H. sciurus was greater during the dry season compared to the wet, and was related to the ontogenetic stage of the individual: utilization of western mangrove shorelines was greatest for adults and least for juveniles. Because their patterns did not match, it is unlikely that juveniles were responsible for observed overall patterns of either L. griseus or H. sciurus among mangrove shorelines. Indeed, juveniles represented only a small portion of total fish observed during both seasons (Fig. 4). FAUNCE AND SERAFY: NEARSHORE ONTOGENIC HABITAT SHIFTS BY TWO REEF FISHES 483

Figure 6. Scree plot of non-metric multidimensional scaling eigenvalues and Monte Carlo test results. A lack of overlap between randomized data and real data allows rejection of the null hypothesis that the eigenvalues resulted from pure chance. From this plot a final solution with two axes was selected (point of “elbow”). Similar trends in the mean sizes of L. griseus and H. sciurus were observed. With the exception of the WK for H. sciurus, the average size of these species was always greater along mangroves than within seagrass beds (Fig. 5). Among mangrove shorelines, average size was always lowest within the WK and increased with average distance from oceanic inlet, where: L. griseus average size = 15.09 + 6.01[1 − e–0.36(DI)] (Adj. r2 = 0.33, P = 0.10); and H. sciurus average size = 10.26 + 11.05[1 − e–0.55(DI)] (Adj. r2 = 0.57, P = 0.02). Structure of LHC Assemblage.—The NMS scree plot revealed a reduction in the amount of final stress accounted for by each dimension after dimension 2, and that the amount of stress for the initial dimensions was significantly different than expected by chance (P = 0.02, Fig. 6). Therefore, a two-dimensional solution was selected for final NMS. This solution reached tolerance after 12 iterations, had a stress value of 6.76, and accounted for 96.5% of the variance in the original data set. The first dimension accounted for 80.1% of the original variance and separated the LHC assemblage among strata, whereas the second dimension accounted for 15.6% of the variance and separated the assemblage according to LHC (Fig. 7). Relationships Between LHC Structure and Environmental Factors.— Canonical correspondence analysis generated three canonical axes from continuous environmental variables that accounted for 21.2% of the variation in the species community; axis 1 accounted for 81.6% of the explained variance, axis 2 accounted for an additional 16.7%, and axis 3 accounted for 1.7% (Table 2). Monte Carlo tests revealed that the eigenvalues and species-environment correlations were significantly greater than expected by chance for the first two axes P( = 0.01). Because axis 3 accounted for a low proportion of the cumulative variance explained by environmental vari- 484 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Figure 7. Non-metric multidimensional scaling biplot depicting spatiotemporal relationships among species and life stages. Axis 1 separates strata (open circles), while the second axis separates life-history components (solid circles). Stratum labels are in the form season_strata_year, where D: dry season; W: wet season; ML: mainland, IS: island; LB: landbridge; LK: leeward key; WK: windward key; 00: 2000; 01: 2001. Life-history component labels are in the form life-history stage_species, where JUV: Juvenile; SUB: Sub-adult; MAT: Adult; G: Lutjanus griseus; S: Haemulon sciurus.

Table 2. Results of canonical correspondence analysis (CCA). * = significant. ns: non-significant

Axis 1 2 3 Total inertia 1.4408 Eigenvalue Real data 0.250 0.051 0.005 Monte Carlo randomized data (100 iterations) Minimum 0.008 0.004 0.001 Mean 0.021 0.009 0.004 Maximum 0.052 0.020 0.011 P (0.01) .* .* .ns Cumulative percentage variance of species data 17.4 20.9 21.2 of species-environment relation 81.6 98.3 100 Species-environment correlations Pearson 0.660 0.416 0.158 Monte Carlo randomized data (100 iterations) Minimum 0.131 0.105 0.072 Mean 0.226 0.177 0.132 Maximum 0.352 0.267 0.205 P (0.01) .* .* .ns FAUNCE AND SERAFY: NEARSHORE ONTOGENIC HABITAT SHIFTS BY TWO REEF FISHES 485

Figure 8. Biplot of canonical correspondence analysis ordination depicting relationships between environmental factors and three life-history stages of Lutjanus griseus and Haemulon sciurus. WOC: Width of canopy. Life-history stages are separated by factors strongly correlated with axis 1 whereas the species are separated by factors strongly correlated with axis 2 (see Table 2 for de- tails). Correlations between individual environmental vectors and each axis is related to the length and direction of the vector; the length of vector is related to the strength of correlation, whereas the direction of the vector depicts which axis that vector is most related to.

ables and failed the Monte Carlo test, further consideration of results will be limited to the first two axes. The relationships between each LHC and the environmental variables are depicted in a species-environment biplot (Fig. 8). Whereas the LHC of L. griseus and H. sciurus were separated along axis 1, the origin effectively separated these species from one another along axis 2. The factors distance index and water depth were important correlates of the first two axes (Table 3), and LHC of both species were ordered juveniles < sub-adults < adults along corresponding vectors. Differences in the mean position of each LHC were less influenced by salinity, whereas separation of species was less influenced by latitude and width of canopy (Table 3, Fig. 8).

Discussion

The ontogenetic utilization of seagrass beds and mangrove shorelines byL . griseus and H. sciurus consist of three phases: (1) settlement into seagrass beds, (2) movement or expansion to mangrove habitats, and (3) increasing inland utilization with size. 486 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

Table 3. Weighted inter-set correlations between environmental variables and CCA axes. Correla- tions > 0.50 are underlined.

Canonical axis Environmental variable 1 2 3 Temperature 0.185 0.281 −0.291 Salinity 0.321 −0.159 −0.873 Depth −0.630 −0.555 −0.367 Width of canopy −0.134 −0.373 0.030 Distance index −0.711 0.579 0.336 Latitude 0.104 0.416 −0.090

Utilization of Seagrass Beds.—Peak recruitment of L. griseus and H. sciurus into seagrass beds was observed during the fall (September–October) and winter (February), respectively, and residence by these fishes was indicated by a steady increase in average size during the subsequent 8 mo. The timing and size that these species recruit to seagrass beds are corroborated by other available data from the region. Working on the Atlantic Ocean side of Lower Matecumbe Key (24˚51´N, 88˚44´W), Springer and McErlean (1962) observed the smallest L. griseus and H. sciurus in monthly samples during August–September and December, respectively. In addition to an earlier timing of recruitment, these authors observed much smaller individuals than in the present study. Gear differences between studies are likely at play here; Springer and McErlean (1962) utilized push nets (~1 mm mesh) in addition to seines (10 mm mesh), but in the present study trawls (10 mm mesh) were utilized. The size of individuals (~15 mm SL) captured in push nets by Springer and McErlean (1962) closely matched the proposed size of ingress and settlement for these species (Lindeman et al., 2001; Tzeng et al., 2003). From their data, offshore spawning and seagrass recruitment occurred 1–2 mo earlier than indicated in the present study, i.e., during August–September for L. griseus, and during December–January for H. sciurus. Therefore, although the residence time of new recruits within Biscayne Bay seagrass beds was estimated at approximately 8 mo (as suggested from monthly average size plots), the age of these fishes was likely closer to 10 mo; thus exclusive use of seagrass beds is probably restricted to the first year of life for the fishes studied. Utilization of Mangrove Shorelines.—Lutjanus griseus and H. sciurus above a certain size also utilize mangrove shoreline habitats. The maximum size of these fishes within seagrass beds matched the smallest average size of individuals within mangrove shorelines, which strongly suggests that mangroves are utilized as a secondary or sequential habitat by these species. Data from prior studies in south Florida substantiate that larger snappers and grunts inhabit mangroves than within seagrass beds (Serafy et al., 1997, 2003; Ley et al., 1999). Data from the present study indicate that a shift from seagrass to mangroves occurs after approximately 8–10 mo and at a size of 10.5–12 cm TL. Because the “fingerbars” used on rollerframe trawls exclude fishes above a certain size, it is alternatively possible that gear selectivity was responsible for observed size differences between habitats. However, previous studies using identical gear types in both mangroves and seagrass habitats have reported size differences of L. griseus and H. sciurus similar to those reported here (Cocheret de la Morinière et al., 2002). FAUNCE AND SERAFY: NEARSHORE ONTOGENIC HABITAT SHIFTS BY TWO REEF FISHES 487

That L. griseus and H. sciurus within mangrove shorelines are larger than those within seagrass beds does not necessarily indicate that mangroves are exclusively utilized by fish above a certain size. Rather, these data likely represent an expansion of habitat breadth with ontogeny (Beck et al., 2001; Cocheret de la Morinière et al., 2002; Serafy et al., 2003). In particular, mangroves appear to supplant the role of predator refugia for reef fishes first provided by seagrasses. The shade and struc- turally heterogeneous root habitat provided by mangroves reduces the visual effectiveness of larger-bodied predators as well as their ability to capture prey (Helfman, 1981; Laegdsgaard and Johnson, 2001; Cocheret de la Morinière et al., 2004). The importance of mangroves as predator refugia is further evidenced by behavioral and stable isotope studies that demonstrated mangroves are primarily utilized as “daytime resting sites” while seagrass beds are utilized as nocturnal foraging grounds (Rooker and Dennis, 1991; Vega-Cendejas et al., 1994; Nagelkerken et al., 2000a; Val- dés-Muñoz and Mochek, 2001). Expansion of habitat utilization with ontogeny not only occurs between seagrass and mangrove habitats, but among spatially separated mangrove shorelines as well. These patterns of habitat utilization can be remarkably similar between species. In the present study, the NMS analysis of the LHC community separated the ML from other strata along the first axis, and this axis accounted for 80% of the variance. Analysis of stratum-specific densities revealed that L. griseus and H. sciurus were much lower in the ML stratum compared to the other mangrove shorelines studied. Therefore, the first NMS axis represented a “filter” for the presence of these fishes, whereas the second axis separated the data points based on their densities. Separa- tion along the second axis resulted in groupings by LHC and not by species, indicat- ing that these two fishes exhibited similar ontogenetic habitat utilization patterns. That L. griseus and H. sciurus utilize more westward and inland mangrove shorelines with ontogeny was supported by two sources of data. First, juvenile density and contribution to the total number of both species declined from the WK (the easternmost strata) to the LK (the adjacent strata to the west), and additional western and inland strata were almost exclusively utilized by sub-adults and adults. Second, average size increased among mangrove shoreline strata in a westward and inland direction during both seasons. Only one other study has examined the relationship between the size of reef fish within a coastal embayment and geographic distance to the bay/ocean interface in the western Atlantic. Nagelkerken et al. (2000b) found a positive relationship between these metrics for Acanthurus chirurgus (Bloch, 1787), Scarus iserti (Bloch, 1789), and Haemulon flavolineatum (Desmarest, 1823) inhabiting seagrass beds at scales up to 3 km in Curaçao, and concluded that the stability of such patterns must be tested in further studies. The present study demonstrates that positive relationships between body size and distance to inlet are also evident for L. griseus and H. sciurus inhabiting mangrove shorelines at scales > 10 km. Relationships Among LHC and Environmental Factors.—Martino and Able (2003) suggested that patterns in estuarine fish assemblage structure across large spatial scales (i.e., > 10 km) are due to individual species responses to dominant environmental gradients, whereas smaller-scale (≤ 1 km) patterns are due to individual behaviors. Previous analyses of coastal fish assemblages using CCA have found that distance from inlet and either salinity (Kupschus and Tremain, 2001; Martino and Able, 2003), temperature (Hajisamae and Chou, 2003), or depth (Araujo et al., 2002) controlled species composition and structure across broad spatial scales. In the 488 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

present study, CCA revealed that distance from inlet and water depth were primarily responsible for the separation of LHC of L. griseus and H. sciurus, and that these species were separated to a lesser extent by latitude, salinity, temperature, and width of canopy. Following Martino and Able (2003), the present CCA results demonstrate a hierarchy of factors associated with the distribution of fishes. Whereas prior studies have shown that dominant environmental gradients influence the composition of entire fish assemblages over broad spatial scales, this study is the first to demonstrate that such gradients also result in the geographic sorting of the ontogenetic stages of two reef fishes. Water depth, rather than temperature or salinity, represented the dominant environmental spatial gradient in the present study. Temperatures may not have reached extremes to exclude fishes from mangroves, and the development of the South Dade Conveyance System has resulted in a reduced salinity gradient from west to east in the region studied (Light and Dineen, 1994). With the exception of the WK during the dry season, water depth tended to decrease among strata from east to west during both seasons (i.e., exhibited a gradient at large spatial scales), but was not significantly correlated with distance from inlet at the level of individual samples (weighted Pearson Correlation Coefficient −0.02). Water depth has an influence on the amount of coastal vegetated habitat available to fishes at numerous scales. At large scales, depth may be correlated with more meaningful metrics related to wetland area. For example, Lee (2004) demonstrated that the extent of intertidal areas and organic matter availability (as represented by tidal amplitude) have a stronger influence on prawn catch from mangroves than mangrove area per se. At smaller scales, water depth dictates the vertical area available to fishes. For example, Faunce et al. (2004) noted a positive relationship between fish size and water depth, and it has been suggested that shallow depths regulate predation on small fishes by excluding their potential predators (Vance et al., 1996; Rönnbäck et al., 1999; but see Sheaves, 2001). Corroborating evidence that water depth sorts fishes by body size comes from the present CCA that showed adults resided in the deepest waters, sub-adults in moder- ately deep waters, and juveniles in the shallowest waters. Latitudinal gradients also played a part in separating the species in the CCA biplot. This analysis suggested thatL . griseus had greater associated latitudes than H. sciurus. The waters of Biscayne Bay comprise the higher latitudes of the southeastern Florida study area. Expressing the number of L. griseus and H. sciurus observed within Bis- cayne Bay as a proportion of the total number of these species (within southeastern Florida) revealed that proportions within Biscayne Bay were substantially greater for L. griseus than for H. sciurus (Table 4). Juvenile L. griseus and H. sciurus within the LK of Biscayne Bay accounted for a respective 0.83–1.00 and 0.70–0.86 of the total number of juveniles within southeastern Florida. Furthermore, the proportion of total fish observed within the LK of Biscayne Bay was greater for juveniles than for sub-adults and least for adults, and this pattern was observed for both species during both seasons. These data lend additional support to the earlier finding that more eastern located mangrove strata are utilized earlier in fish ontogeny. Because the ML and LK strata within Biscayne Bay represent 34.7% and 42.1% of the total 2-m wide shoreline area of their respective strata within the study area, the proportion of reef fishes, especially juveniles, observed within Biscayne Bay is substantial. It is likely that the comparatively large opening south of Key Biscayne (i.e., the “Safety Valve”) led to increased water exchange, westward net flow, and ulti- FAUNCE AND SERAFY: NEARSHORE ONTOGENIC HABITAT SHIFTS BY TWO REEF FISHES 489

Table 4. Proportion of the total Lutjanus griseus and Haemulon sciurus life-history components of southeastern Florida observed within the mainland (ML) and leeward key (LK) portions of Biscayne Bay.

Strata ML LK Species Lutjanus griseus Haemulon sciurus Lutjanus griseus Haemulon sciurus Dry season Juvenile 0.00 No fish 0.83 0.70 Sub-adult 0.04 0.00 0.59 0.28 Mature 0.00 0.00 0.56 0.06

Wet season Juvenile 1.00 No fish 1.00 0.86 Sub-adult 0.92 0.90 0.59 0.60 Mature 0.77 1.00 0.58 0.25

mately increased utilization of Biscayne Bay mangrove shoreline habitats by juvenile reef fishes. This topic warrants further research in the region, such as the biophysical simulation modeling approach used by Ault et al. (1999b) and Wang et al. (2003) for pink shrimp (Farfantepenaeus duorarum Burkenroad, 1939) and spotted seatrout (Cynoscion nebulosus Cuvier, 1830), and highlights the influence of bay-ocean connectivity on local recruitment patterns. In addition to latitude, width of canopy influenced separation the two reef species in the CCA. Width of canopy was expected to be an important habitat metric because it reflects the structural complexity and extent of shelter provided by mangrove root systems. This complexity in turn has been shown to provide shading and visual advantages for forage fishes H( elfman, 1981), and to reduce predator efficacy (Primavera, 1997; Laegdsgaard and Johnson, 2001). Ley and McIvor (2001) used CCA to examine habitat relationships between several species of reef fishes, including L. griseus and H. sciurus within the LB stratum of the present study. They found that for H. sciurus, mangrove tree height and water depth were significantly correlated with the first axis, whereas for L. griseus, mangrove fringe width, height, prop-root density, and water depth were significantly correlated with the second axis. Results of the present study closely resemble those of Ley and McIvor (2001) with the exception that we examined different life-history components. Both studies demonstrated a separation of species according to gradients in habitat complexity including mangrove fringe width. In addition, results from these field surveys agree with the experimental manipulations of Cocheret de la Morinière et al. (2004) and Verweij et al. (2006) that show abundances of fishes including H. sciurus increase as a function of both habitat complexity and shading.

Summary and Implications

The structure and composition of coastal fish assemblages are thought to be a function of individual species’ responses to environmental gradients that operate at multiple spatial scales. Although several studies have examined the relationship between the entire fish assemblage and environmental gradients, such an approach is less likely to capture individual species’ responses as they change with ontogeny. In the present study, two ontogenetic patterns of habitat utilization were exhibited by L. 490 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

griseus and H. sciurus within southeastern Florida. First, size-distributions averaging > 13 cm TL were exclusively observed during daytime within mangrove habitats, whereas smaller fish were commonly collected within seagrass beds. Such size dis- crepancies between habitats are likely due to the inability of seagrasses to effectively shelter comparatively large fish from predation (Jackson et al., 2001). Second, an increased utilization of inland mangrove shoreline strata was observed with ontogenic development. Such patterns are likely due to an expansion of home range and mobil- ity with body size (e.g., Kramer and Chapman, 1999). As such, the increase in mean size observed here as a function of distance from inlet likely represents the exclusion of smaller bodied fishes, rather than habitat resource selection by larger individuals. One application of these data is that they may explain observed fish distributions within Florida estuarine “no-take” reserves. Two such reserves exist on the east coast of Florida, the Crocodile Sanctuary of Everglades National Park (CS-ENP), and the No Take Zone of Merritt Island National Wildlife Refuge (NTZ-MINWR). Stud- ies conducted within these reserves have noted a high proportion of large-bodied individuals relative to surrounding waters (Johnson et al., 1999; Faunce et al., 2002). Recently, Tremain et al. (2004) evaluated tag-return data from the Indian River La- goon System and the NTZ-MINWR and concluded that a greater percentage of fish immigrated to the no-take zone than emigrated from it. Distances between these reserves and permanently accessible oceanic inlets range from 27–29 km for the CS- ENP to 75–100 km for the NTZ-MINWR, and are substantially larger than the 10 km scale commonly cited for fishes that have spatial separation of juvenile and adult populations (Gillanders et al., 2003; Pittman and McAlpine, 2003). For these rea- sons, the relatively large size of fishes observed within these coastal no-take reserves relative to surrounding areas may partially be the result of differences in motility, i.e., only larger-bodied fishes with greater home ranges and migratory capabilities can access the remotely located estuarine reserves, or similarly, the time it takes to slowly reach these locations begets a larger body size upon arrival. While the mechanism by which larger fish occur within remote inland waters remains unclear, it seems neces- sary to assess the contribution of these fish towards the overall spawning population. Whether larger fishes present in remote bays or estuaries seasonally emigrate to join the spawning population elsewhere, spawn within remote areas without emigration, or merely subsist without spawning or emigration should be considered by decision makers as to whether such areas remain no-take and/or no-access. A second application of the present study is that it provides the basis for prioritization of areas con- taining the same habitat type (seagrass or mangrove) for preservation/protection, as in the United States, there are inadequate resources to fully manage and protect all habitats all of the time. From our results, habitat value (as defined by daytime density values) for L. griseus and H. sciurus during ontogeny is a function of distance from oceanic inlet, with seagrass beds being the primary habitat for fish during their first year, and mangroves thereafter. The expansion of inland habitat utilization with ontogeny observed in the present study may be either accompanied by an equivalent range expansion eastward and offshore, or be followed by a rapid long-distance offshore migration facilitated by a large body size. A steady increase in average size with distance from inlet eastward towards the coral reef would support the first scenario. Such data are lacking for L. griseus. However, larger mean sizes of H. sciurus have been collected from inshore reefs compared to mangrove shorelines (Acosta, 1997), suggesting that this fish likely FAUNCE AND SERAFY: NEARSHORE ONTOGENIC HABITAT SHIFTS BY TWO REEF FISHES 491 undergoes regular seasonal movements into and out of coastal embayments. In con- trast, L. griseus appear to remain within mangrove habitats for an extended period, as a greater proportion of mature-sized individuals have been recorded within mangrove habitats than within adjacent coral reefs (Nagelkerken et al., 2000b; Cocheret de la Morinière et al., 2002; Serafy et al., 2003). Indeed, new tagging efforts using acoustic telemetry show that circular movements of gray snapper between offshore reefs and inshore mangroves can occur in a matter of days (Luo et al., 2006). Evi- dence from these studies demonstrates that although H. sciurus and L. griseus utilize inshore coastal vegetated habitats, their occupancy is not homogeneous, but is location-, season-, and life-history stage-dependent. In addition, seasonal forays between inshore and offshore locations may be of widely different durations between species, lasting longer for H. sciurus than for L. griseus, which may be considered more of a coastal species than a reef species.

Acknowledgments

An earlier version of this manuscript was originally submitted in partial fulfillment of a doctoral dissertation by C. Faunce (RSMAS, University of Miami). We thank O. Becerio, J. Luo, G. Schmalzle, and D. Jones for help with analyses, and J. Kool for his GIS assistance and valuable discussions regarding landscape ecology. Primary funding was provided by the Perry Institute for Marine Science, Environmental Defense, the United States Geological Survey (Biological Resources Division, Restoration Ecology Branch) and the National Park Service (Biscayne National Park). Additional funding was provided by the American Institute for Ma- rine Studies, the Florida Keys Audubon Society, and the Captain Bob Lewis and Paul Kopp Memorial scholarships awarded through the University of Miami.

Literature Cited

Acosta, A. 1997. Use of multi-mesh gillnets and trammel nets to estimate fish species composition in coral reef and mangroves in the southwest coast of Puerto Rico. Caribb. J. Sci. 33: 45–57. Araujo, F. G., M. C. C. De Azevedo, M. A. Silva, A. L. M. Pessanha, I. D. Gomes, and A. G. Da Cruz-Filho. 2002. Environmental influences on the demersal fish assemblages in the Sepe- tiba Bay, Brazil. Estuaries 25: 441–450. Ault, J. S., G. A. Diaz, S. G. Smith, J. Luo, and J.E. Serafy. 1999a. An efficient sampling survey design to estimate pink shrimp population abundance in Biscayne Bay, Florida. North Am. J. Fish. Manage. 19: 696–712. ______, J. G. Luo, S. G. Smith, J. E. Serafy, J. D. Wang, R. Humston, and G. A. Diaz. 1999b. A spatial dynamic multistock production model. Can. J. Fish. Aquat. Sci. 56: 4–25. Bardach, J. E. 1959. The summer standing crop of fish on a shallow Bermuda reef. Limnol. Oceanogr. 4: 77–85. Beck, M. W., K. L. Heck, K. W. Able, D. L. Childers, D. B. Eggleston, B. M. Gillanders, B. Halp- ern, C. G. Hays, K. Hoshino, T. J. Minello, R. J. Orth, P. F. Sheridan, and M. R. Weinstein. 2001. The identification, conservation, and management of estuarine and marine nurseries for fish and invertebrates. BioScience 51: 633–641. Bell, J. D., G. C. S. Craik, D. A. Pollard, and B. C. Russell. 1985. Estimating length frequency- distributions of large reef fish underwater. Coral Reefs 4: 41–44. Billings, V. C. and J. L. Munro. 1974. The biology, ecology, exploitation and management of Caribbean reef fishes. Scientific report of the ODA/UWI fisheries ecology research project, Port Royal Marine Laboratory, Jamaica, 1969–1973. Part 5e. The biology, ecology and bio- 492 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

nomics of Caribbean reef fishes: Pomadasyidae (grunts). Res. Rep. Zool. Dept., Univ. West Indies. 128 p. Blaber, S. J. M. 2002. ‘Fish in hot water’: the challenges facing fish and fisheries research in tropical estuaries. J. Fish. Biol. 61: 1–20. Cocheret de la Morinière, E., I. Nagelkerken, H. van der Meij, and G. van der Velde. 2004. What attracts juvenile coral reef fish to mangroves: habitat complexity or shade? Mar. Biol. 144: 139–145. ______, B. J. A. Pollux, I. Nagelkerken, and G. van der Velde. 2002. Post- settlement life cycle migration patterns and habitat preference of coral reef fish that use seagrass and mangrove habitats as nurseries. Estuar. Coast. Shelf Sci. 55: 309–321. Cochran, W. G. 1977. Sampling Techniques. Wiley, New York. 448 p. Day, J. H. 1951. The ecology of South African estuaries Part I. General considerations. Trans. Royal Soc. S. Afr. 33: 53–91. Day Jr., J. W., C. A. S. Hall, W. M. Kemp, and A. Yáñez-Arancibia. 1989. Estuarine ecology. Wiley Interscience, New York. 558 p. Domeier, M. L., C. Koenig, and F. Coleman. 1996. Reproductive biology of the gray snapper (Lutjanus griseus), with notes on spawning for other western Atlantic snappers (Lutjani- dae). Pages 189–201 in F. Arreguin-Sanchez, J. L. Munro, M. C. Blagos, and D. Pauly, eds. Biology, fisheries and culture of tropical snappers and groupers. ICLARM Conf. Proc. Intl. Ctr. Liv. Aquat. Res. Mgt., Manila. 48. Dorenbosch, M., M. C. van Riel, I. Nagelkerken, and G. van der Velde. 2004. The relationship of reef fish densities to the proximity of mangrove and seagrass nurseries. Estuar. Coast. Shelf Sci. 60: 37–48. Faunce, C. H. and J. E. Serafy. 2006. Mangroves as fish habitat: 50 years of field studies. Mar. Ecol. Prog. Ser. 318: 1–18. ______, J. E. Serafy, and J. J. Lorenz. 2004. Density-habitat relationships of mangrove creek fishes within the southeastern saline Everglades (USA), with reference to managed freshwater releases. Wetl. Ecol. Mgt. 12: 377–394. ______, J. J. Lorenz, J. A. Ley, and J. E. Serafy. 2002. Size structure of gray snapper (Lut- janus griseus) within a mangrove ‘no-take’ sanctuary. Bull. Mar. Sci. 70: 211–216. García-Arteaga, J. P. 1992. Edad y crecimiento del ronco amarillo, Haemulon sciurus (Shaw) (Pisces: Haemulidae) en la plataforma suroccidental de Cuba. Cienc. Biol. 25: 104–116. García-Cagide, A., R. Claro, and B. V. Koshelev. 2001. Reproductive patterns of fishes of the Cuban shelf. Pages 73–114 in R. Claro, K. C. Lindeman, and L. R. Parenti, eds. Ecology of the marine fishes of Cuba. Smithsonian Institution Press, Washington. Gillanders, B. M., K. W. Able, J. A. Brown, D. B. Eggleston, and P. F. Sheridan. 2003. Evidence of connectivity between juvenile and adult habitats for mobile marine fauna: an important component of nurseries. Mar. Ecol. Prog. Ser. 247: 281–295. Gunter, G. 1967. Some relationships of estuaries to the fisheries of the Gulf of Mexico. Pages 621–638 in G. H. Lauff, ed. Estuaries. Am. Assoc. Adv. Sci., Washington, D.C. Hajisamae, S. and L. M. Chou. 2003. Do shallow water habitats of an impacted coastal strait serve as nursery grounds for fish? Estuar. Coast. Shelf Sci. 56: 281–290. Halpern, B. 2004. Are mangroves a limiting resource for two coral reef fishes? Mar. Ecol. Prog. Ser. 272: 93–98. Helfman, G. S. 1981. The advantage to fishes of hovering in shade. Copeia 198: 392–400. Jackson, E. L., A. A. Rowden, M. J. Attrill, S. J. Bossey, and M. B. Jones. 2001. The importance of seagrass beds as a habitat for fishery species. Oceanogr. Mar. Biol. 39: 269–303. Johnson, D. R., N. A. Funicelli, and J. A. Bohnsack. 1999. Effectiveness of an existing estuarine no-take fish sanctuary within the Kennedy Space Center, Florida. North Am. J. Fish. Man- age. 19: 436–453. Jongman, R. H. G., C. J. F. ter Braak, and O. F. R. Van Tongeren. 1995. Data Analysis in Com- munity and Landscape Ecology. Cambridge University Press, Cambridge. 299 p. FAUNCE AND SERAFY: NEARSHORE ONTOGENIC HABITAT SHIFTS BY TWO REEF FISHES 493

Kramer, D. L. and M. R. Chapman. 1999. Implications of fish home range size and relocation for marine reserve function. Environ. Biol. Fish. 55: 65–79. Kupschus, S. and D. Tremain. 2001. Associations between fish assemblages and environmental factors in nearshore habitats of a subtropical estuary. J. Fish. Biol. 58: 1383–1403. Laegdsgaard, P. and C. Johnson. 2001. Why do juvenile fish utilize mangrove habitats? J. Exp. Mar. Biol. Ecol. 257: 229–253. Lee, S. Y. 2004. Relationship between mangrove abundance and tropical prawn production: a re-evaluation. Mar. Biol. 145: 943–949. Legendre, L. and P. Legendre 1983. Numerical Ecology. Elsevier, New York. 853 p. Ley, J. A. and C. C. McIvor. 2001. Linkages between estuarine and reef fish assemblages: en- hancement by the presence of well-developed mangrove shorelines. Pages 539–562 in J. W. Porter and K. G. Porter, eds. The Everglades, Florida Bay, and coral reefs of the Florida Keys: an ecosystem sourcebook. CRC Press, Boca Raton. ______, ______, and C. L. Montague. 1999. Fishes in mangrove prop-root habitats of northeastern Florida Bay: distinct assemblages across an estuarine gradient. Estuar. Coast. Shelf Sci. 48: 701–723. Light, S. S. and J. W. Dineen. 1994. Water control in the Everglades: a historical perspective. Pages 47–84 in S. M. Davis and J. C. Ogden, eds. Everglades: the ecosystem and its restoration. St. Lucie Press, Boca Raton. Lindeman, K. C., T. N. Lee, W. D. Wilson, R. Claro, and J. S. Ault. 2001. Transport of lar- vae originating in southwest Cuba and the Dry Tortugas: evidence for partial retention in grunts and snappers. Proc. Gulf Caribb. Fish. Inst. 52: 732–747. Luo, J., S. Sponaugle, J. E. Serafy, and J. J. Lorenz. 2006. Multiple habitat utilization by a coastal fish: diel, seasonal, and ontogenetic movement of gray snapper (Lutjanus griseus), Florida Sea Grant Final Report, Project Number R/C-RE-48. 96 p. Manooch III, C. S. and R. H. Matheson III. 1981. Age, growth, and mortality of gray snapper collected from Florida waters. Pages 331–344 in Proc. of the 35th Ann. Conf. Southeastern Assoc. Fish and Wildlife Agencies, Tulsa. Martino, E. J. and K. W. Able. 2003. Fish assemblages across the marine to low salinity transi- tion zone of a temperate estuary. Estuar. Coast. Shelf Sci. 56: 969–987. McCune, B and J. B. Grace. 2002. Analysis of Ecological Communities. MjM Software Design, Gleneden Beach. 300 p. McHugh, J. L. 1967. Estuarine Nekton. Pages 581–620 in G. H. Lauff, ed. Estuaries. American Association for the Advancement of Science, Washington, D. C. Meester, G. A., J. S. Ault, and J. A. Bohnsack. 1999. Visual censusing and the extraction of average length as an indicator of stock health. Naturalista Siciliana 23: 205–222. Minchin, P. R. 1987. An evaluation of the relative robustness of techniques for ecological ordination. Vegetatio 69: 89–107. Mumby, P. J., A. J. Edwards, J. E. Arias-Gonzales, K. C. Lindeman, P. G. Blackwell, A. Gall, M. I. Gorczynska, A. R. Harborne, C. L. Pescod, H. Renken, C. C. C. Wabnitz, and G. Llewellyn. 2004. Mangroves enhance the biomass of coral reef fish communities in the Caribbean. Nature 427: 533–536. Munro, J. L., V. C. Gaut, R. Thompson, andP . H. Reeson. 1973. The spawning seasons of Carib- bean reef fishes. J. Fish Biol. 5: 69–84. Nagelkerken, I. and G. van der Velde. 2002. Do non-estuarine mangroves harbour higher densities of juvenile fish than adjacent shallow-water and coral reef habitats in Curacao (Nether- lands Antilles)? Mar. Ecol. Prog. Ser. 245: 191–204. ______, M. Dorenbosch, W. C. E. P. Verberk, E. Cocheret de la Morinière, and G. van der Velde. 2000a. Day-night shifts of fishes between shallow-water biotopes of a Caribbean bay, with emphasis on the nocturnal feeding of Haemulidae and Lutjanidae. Mar. Ecol. Prog. Ser. 194: 55–64. ______, ______, ______, ______, and ______. 2000b. Importance of shallow-water biotopes of a Caribbean bay for juvenile coral 494 BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

reef fishes: patterns in biotope association, community structure and spatial distribution. Mar. Ecol. Prog. Ser. 202: 175–192 ______, C. M. Roberts, G. van der Velde, M. Dorenbosch, M. C. van Riel, E. Cocheret de la Morinière, and P. H. Nienhuis. 2002. How important are mangroves and seagrass beds for coral-reef fish? The nursery hypothesis tested on an island scale. Mar. Ecol. Prog. Ser. 244: 299–305. Ogden, J. C. 1997. Ecosystem interactions in the tropical coastal seascape. Pages 288–297 in C. Birkeland, ed. Life and Death of coral reefs. Chapman & Hall, New York. Parrish, J. D. 1989. Fish communities of interacting shallow-water habitats in tropical oceanic regions. Mar. Ecol. Prog. Ser. 58: 143–160. Pennington, M. 1983. Efficient estimators of abundance, for fish and plankton surveys. Biomet- rics 39: 281–286. Pittman, S. J. and C. A. McAlpine. 2003. Movements of marine fish and decapod crustaceans: process, theory and application. Adv. Mar. Biol. 44: 205–294. Primavera, J. H. 1997. Fish predation on mangrove-associated penaeids: the role of structures and substrate. J. Exp. Mar. Biol. Ecol. 215: 205–216. Rönnbäck, P., M. Troell, N. Kautsky, and J. H. Primavera. 1999. Distribution pattern of shrimps and fish among Avicennia and Rhizophora microhabitats in the Pagbilao Mangroves, Phil- ippines. Estuar. Coast. Shelf Sci. 48: 223–234. Rooker, J. R. and G. D. Dennis. 1991. Diel, lunar and seasonal changes in a mangrove fish assemblage off southwestern Puerto Rico. Bull. Mar. Sci. 49: 684–698. Serafy, J. E., C. H. Faunce, and J. J. Lorenz. 2003. Mangrove shoreline fishes of Biscayne Bay, Florida. Bull. Mar. Sci. 72: 161–180. ______, K. C. Lindeman, T. E. Hopkins, and J. S. Ault. 1997. Effects of freshwater canal dis- charge on fish assemblages in a subtropical bay: field and laboratory observations. Mar. Ecol. Prog. Ser. 160: 161–172. Sheaves, M. J. 2001. Are there really few piscivorous fishes in shallow estuarine habitats? Mar. Ecol. Prog. Ser. 222: 279–290. Springer, V. G. and A. J. McErlean. 1962. Seasonality of fishes on a south Florida shore. Bull. Mar. Sci. 12: 39–60. St. John, J., G. R. Russ, and W. Gladstone. 1990. Accuracy and bias of visual estimates of num- bers, size structure and biomass of a coral reef fish. Mar. Ecol. Prog. Ser. 64: 253–262. Stark II, W. E. 1970. Biology of the gray snapper, Lutjanus griseus (Linnaeus), in the Florida Keys. Pages 11–150 in W. E. Stark II and R. E. Schroeder, eds. Investigations on the gray snapper, Lutjanus griseus. University of Miami Press, Coral Gables. ter Braak, C. J. F. 1986. Canonical correspondence analysis: a new eigenvector technique for multivariate direct gradient analysis. Ecology 67: 1167–1179. Tremain, D. M., C. W. Harnden, and D. H. Adams. 2004. Multidirectional movements of sport- fish species between an estuarine no-take zone and surrounding waters of the Indian River Lagoon, Florida. Fish. Bull. 102: 533–544. Tzeng, M. W., J. A. Hare, and D. G. Lindquist. 2003. Ingress of transformation stage gray snapper, Lutjanus griseus (Pisces: Lutjanidae) through Beaufort Inlet, North Carolina. Bull. Mar. Sci. 72: 891–908. United States Department of Commerce (USDOC). 1996. Magnusen-Stevens Fishery Conser- vation and Management Act as amended through October 11, 1996., National Oceanic and Atmospheric Administration Technical Memorandum NMFS-F/SPO-23. 121 p. Valdés-Muñoz, E, and A. D. Mochek. 2001. Behavior of marine fishes of the Cuban shelf.P ages 73–114 in R. Claro, K. C. Lindeman, L. R. Parenti, eds. Ecology of the marine fishes of Cuba. Smithsonian Institution Press, Washington. Vance, D. J., M. D. E. Haywood, D. S. Heales, R. A. Kenyon, N. R. Loneragan, and R. C. Pendrey. 1996. How far do prawns and fish move into mangroves? Distribution of juvenile banana prawns Penaeus merguiensis and fish in a tropical mangrove forest in northern Australia. Mar. Ecol. Prog. Ser. 131: 115–124. FAUNCE AND SERAFY: NEARSHORE ONTOGENIC HABITAT SHIFTS BY TWO REEF FISHES 495

Vega-Cendejas, M. E., U. Ordóñez, and M. Hernández. 1994. Day–night variation of fish population in the mangrove of Celestun lagoon, Mexico. Intl. J. Ecol. Env. Sci. 20: 99–108. Verweij, M. C., I. Nagelkerken, D. de Graaff, M.P eeters, E. J. Bakker, and G. van der Velde. 2006. Structure, food and shade attract juvenile coral reef fish to mangrove and seagrass habitats: a field experiment. Mar. Ecol. Prog. Ser. 306: 257–268. Wang, J. D., J. G. Luo, and J. S. Ault. 2003. Flows, salinity, and some implications for larval transport in South Biscayne Bay, Florida. Bull. Mar. Sci. 72: 695–723.

Addresses: (C.H.F.) Florida Fish and Wildlife Research Institute, Tequesta Field Labora- tory, 19100 SE Federal Highway, Tequesta, Florida 33469. (J.E.S.) Southeast Fisheries Science Center, National Marine Fisheries Service,75 Virginia Beach Drive, Miami, Florida 33149. Corresponding Author: (C.H.F.) E-mail: .

&lt;I&gt;Haemulon Sciurus

<I>Haemulon Sciurus