Mar Biol (2012) 159:1305–1315 DOI 10.1007/s00227-012-1911-z

ORIGINAL PAPER

Variability in nursery function of tropical seagrass beds during fish ontogeny: timing of ontogenetic habitat shift Yohei Nakamura • Keisuke Hirota • Takuro Shibuno • Yoshiro Watanabe

Received: 22 November 2011 / Accepted: 28 February 2012 / Published online: 21 March 2012 Ó Springer-Verlag 2012

Abstract Seagrass beds are often considered to be important nurseries for coral reef fish, yet the effectiveness of these nursery functions (refuge and food availability) at different juvenile stages is poorly understood. To understand how the demands of juvenile fish on seagrass nursery functions determines the timing of ontogenetic habitat shifts from seagrass beds to coral reefs, we conducted visual transect survey and field tethering and caging experiments on three different sizes of the coral reef fish Pacific yellowtail emperor (Lethrinus atkinsoni) during its juvenile tenure in seagrass beds at Ishigaki Island, southern Japan. The study showed that although the number of individual L. atkinsoni juveniles decreased by [90 % during their stay in the seagrass nursery, the shelter and/or food availability functions of the nursery, at least for a juvenile size of approximately 5 cm total length (TL), provided the best survival and growth option. The timing of ontogenetic migration to coral reefs of larger fish ([8 cm Communicated by D. Goulet. Y. Nakamura (&) Graduate School of Kuroshio Science, Kochi University, 200 Monobe, Nankoku, Kochi 783-8502, Japan e-mail: [email protected] K. Hirota Faculty of Agriculture, Kochi University, 200 Monobe, Nankoku, Kochi 783-8502, Japan T. Shibuno National Research Institute of Aquaculture, 422–1, Nakatsuhamaura, Minami-ise, Mie 516-0193, Japan Y. Watanabe Atmosphere and Ocean Research Institute, The University of Tokyo, 5-1-5, Kashiwanoha, Kashiwa-shi, Chiba 277-8564, Japan

TL) was attributed to foraging efficiency for larger food items in different habitats. Overall, the function of the seagrass bed nursery changed with juvenile body size, with marginally higher survival and significantly greater growth rates during early juvenile stages in seagrass beds compared to coral reefs. This would contribute to the enhancement in the number of individuals eventually recruited to adult populations.

Introduction Seagrass beds often cover extensive areas surrounding coral reefs, a conspicuous feature of the former being their high densities of juvenile fishes, the adults of many of which live on the adjacent reefs (Nagelkerken et al. 2000, 2002; Shibuno et al. 2008). Based on the spatial separation between juvenile and adult populations, seagrass beds have been assumed to function as important nursery habitats that produce relatively more adult recruits per unit area than other juvenile habitats (Beck et al. 2001). Such a function results from factors that singly or in combination influence the density, growth and survival of juveniles to a greater extent than in other habitats (Beck et al. 2001). Hypotheses proposed to explain the nursery function of tropical seagrass beds are based mainly on (1) shelter function—predation risk is less in seagrass beds than in coral reefs due to fewer predators and the presence of good refuges afforded by seagrass leaves, and (2) food availability function— food availability for some juvenile fishes is greater in seagrass beds than in coral reefs, which sometimes resulting in higher growth rates in the former (see review by Nagelkerken 2009). Although a combination of faster growth and lower predation risk has been presumed to underpin the importance of seagrass nurseries (Heck et al.

123

1306

2003), empirical data supporting this model for reef fishes has still not been presented. Reef fish juveniles grow in seagrass bed nurseries, and probably, move to adult coral reef habitats at such a time as the nursery functions are no longer necessary. The timing of ontogenetic habitat shift from nurseries to adult habitats often reflects behavioral decisions associated with the demands of predator avoidance, foraging or reproduction, because such demands change during ontogeny (e.g., due to increases in body size, e.g., Werner and Hall 1988; Dahlgren and Eggleston 2000). For example, Nassau grouper Epinephelus striatus settle and grow in algal-covered coral clumps until reaching ca. 3.5–4 cm total length (TL), at that time moving to a post-algal microhabitats (e.g., coral, rubble and sponges). Field experiments showed that growth rates for two size classes of juveniles (3.5–4 cm and 5–5.5 cm TL) were greater in the post-algal habitat, whereas the predation-induced mortality for smaller juveniles was lower in the algal habitat, indicating that primary mechanisms driving ontogenetic habitat use were predation risk (for smaller juveniles) and foraging efficiency (for larger juveniles) (Dahlgren and Eggleston 2000). Although a trade-off between predation risk and food availability may exist among reef fish juveniles that use seagrass bed nurseries, exactly how the demands of juveniles to these nursery functions determine the timing of ontogenetic habitat shift is poorly understood (Adams et al. 2006). Only a single study has provided possible clarification, ontogenetic migration of French grunt (Haemulon flavolineatum) from an embayment with mangroves and seagrass beds to reef habitats having been shown to be related to changes in predation risk with increased body size (Grol et al. 2011). The Pacific yellowtail emperor Lethrinus atkinsoni is a widely distributed coastal species in the subtropical/tropical Indo-Pacific region, being an important component of reef-based fisheries (Carpenter and Allen 1989). This species uses chemical cues to orient and settle in seagrass beds (Arvedlund and Takemura 2006; Nakamura et al. 2009). It grows in these beds until 7–12 cm TL is attained and then moves to coral reef habitats (Wilson 1998; Shibuno et al. 2008; Nakamura and Tsuchiya 2008). Because the onset of sexual maturity of L. atkinsoni occurs at ca. 20 cm folk length (Ebisawa 1999), the timing of movement from the seagrass beds to coral reefs is not attributed to reproduction, but may be related to foraging efficiency and/or predator avoidance. Accordingly, this study was designed to clarify the mechanisms underlying the timing of ontogenetic migration of L. atkinsoni from seagrass beds to coral reefs, utilizing underwater observations and field experimentation. The following questions were addressed specifically:

123

Mar Biol (2012) 159:1305–1315

(1) What is the decrease in juvenile number during growth in the seagrass nursery? (2) Does a seagrass bed provide juveniles with more survival and growth advantages than a coral reef? (3) How do the demands of juveniles on the shelter and food availability functions of the nursery determine the timing of ontogenetic migration from seagrass beds to coral reefs?

Materials and methods Study sites All experiments were conducted off northern Ishigaki Island, southern Ryukyu Islands (Fig. 1). Seagrass beds at Oganzaki and Kanashiro were selected for the determination of post-settlement decline of juvenile L. atkinsoni, since preliminary monthly visual transect surveys (1 m 9 30 m, n = 10, June–December 2006) at six seagrass bed sites around Ishigaki Island found that the former two sites harbored greater numbers of L. atkinsoni than the others (Fig. 1). Itona reef was selected as the field experimental site, a previous study having found clear ontogenetic habitat shift of L. atkinsoni from seagrass bed to coral reef on this reef (Shibuno et al. 2008). All of the study sites had very similar reef profiles with clearly defined reef flats and outer slopes. The fringing reef extended seaward for approximately 100 m (Oganzaki), 200 m (Kanashiro) and 250 m (Itona) from the shoreline to the reef edge, the inner reef flats of the three sites being dominated by Cymodocea serrulata and Thalassia hemprichii seagrasses (110–130 shoots m-2; leaf height, 8–15 cm; 2 ha in Oganzaki, 3.5 ha in Kanashiro, 11 ha in Itona), and the fore reefs by branching and tabular Acropora corals (hereafter, coral area). Water depth on each reef flat was 1.5–2 m, the seagrass beds remaining submerged during low tide. All sites were similar in certain key environmental characteristics, including water depth, water temperature (ca. 26 °C in May, 29 °C in August, 25 °C in November and 22 °C in February), salinity (ca. 35) and underwater visibility ([10 m). Post-settlement decline of juveniles in seagrass beds To clarify intersize class decline rate of seagrass nursery juveniles, the size structure of such was investigated by visual transect surveys. From March to December 2007, L. atkinsoni in the seagrass beds at Kanashiro and Oganzaki were visually censused each week, with the exception of September and December (1 week lost in each month due to bad weather). Even if no juveniles were found, the census was still conducted in January and February 2008. For each census, ten 30 m 9 1 m belt transects, separated

Mar Biol (2012) 159:1305–1315

1307

Fig. 1 Map of the study sites at Ishigaki Island, southern Ryukyu Islands. Dashed lines indicate reef margins. Numbers indicate total individual numbers of juvenile Lethrinus atkinsoni at the 6 seagrass bed sites during June–December 2006 (sum of the individual numbers encountered across 70 transects in each site)

from each other by at least 10 m, were established haphazardly over the seagrass beds using a scaled rope. The total length of all L. atkinsoni within each transect area were recorded on a white plastic slate using a plastic ruler. This nondestructive method allowed reasonable estimation of the density and body size of fish compared to seine net sampling in seagrass beds (Horinouchi et al. 2005). Because newly settled L. atkinsoni do not possess speciesspecific color patterns (semi-transparent body) until 72 h after settlement (McCormick et al. 2002), those individuals were not included for size structure analysis. Decline rate among size classes was estimated by comparing total individual numbers of each size class within an annual period. In this study, three key size classes were selected for analysis: small size class (2.0–3.5 cm TL) (corresponding to fish recently settled in the seagrass beds), large size class ([8.0 cm TL) (corresponding to the expected size of fish making an ontogenetic habitat shift from seagrass bed to coral reef) and medium size class (5.0–6.0 cm TL), being midway between the small and large size classes.

Seagrass bed fidelity To clarify whether the seagrass habitat fidelity of juveniles changed with growth, the home range of each size class (small, medium, and large) was investigated in the Oganzaki and Itona seagrass beds. In June and July 2011, after finding focal size juveniles in the beds, each juvenile was monitored for 30 min, and their positions were recorded every minute on a waterproof graph paper. The positions and substratum characteristics were then mapped using the grid squares for reference, and home range size was estimated by the convex polygon method. Five of each small and medium juvenile individuals and three large juvenile individuals were investigated at each site, and the mean ± standard deviation (SD) of the area of a home range was estimated for each size class at each site. For each individual, the percentage coverage of each substratum (e.g., seagrass, coral or Sargassum) within the home range represented the core area of activity, that is, the habitat within the home range of greatest use. During the census, a majority of juveniles showed foraging behavior,

123

1308

such as pecking the sea bottom or seagrass leaves, and moved within the same areas. No fish showed apparent avoidance or attraction against the diver. All observations were made during daytime (10:00–16:00 h) because it has been shown that the behavior of juveniles does not change during day and crepuscular hours and that they rest on the bottom within their home ranges at night (Nakamura and Tsuchiya 2008). Size-specific survival rate To test whether or not predation-induced mortality was lower in seagrass beds than in coral areas for smaller juveniles, relative survival rates between the habitats were compared for each size class of juveniles. Relative survival rate was estimated by tethering juvenile L. atkinsoni in the seagrass bed and coral area at Itona reef in July and August 2007. Fish were tethered though the muscle tissue above the abdominal vertebrae with 1.5 m of 0.128-mm diameter monofilament fishing line attached to an iron stake pushed into the substrate. Five tethering stations were established randomly in the seagrass bed and in the coral area, individual fish at each station being placed at least 10 m apart so as to ensure the independence of each predation event. Five replicates were conducted for the small and medium size classes (25 individuals for each fish size in each habitat), while three replicates were conducted for the large size class (15 individuals in each habitat), due to the difficulty in collecting large juveniles (very few individuals in the field). In each size class, similar-sized juveniles were tethered in each habitat in order to minimize the effects of prey size on predation (P = 0.86 between habitats for small juveniles with mean TL of 3.7 ± 0.2 SD cm, P = 0.68 for medium juveniles with 5.9 ± 0.5 cm TL, P = 0.27 for large juveniles with 10.0 ± 1.7 cm TL). Because none of the control fish (25 individuals) in predator exclusion mesh nets had broken free from their tethers, fish missing from the tethers were considered to have been predated. Each trial was conducted for a 60 min period, the presence or absence of tethered juveniles being recorded at the end of each. The percentage of juveniles present was used as the survival rate for each habitat in each trial. Independent t test was used to examine differences in survival rates between the two habitats in each size class. Prior to the analysis, data were transformed to arcsine Hx to produce homogeneous variances. All tethering experiments were made at high tide between 10:00 and 16:00 h. Although the tethering technique has the shortcoming of potentially producing experimental artifacts, such as changes in behavior or escape responses of tethered fish, resulting in increased vulnerability to predators (Peterson and Black 1994; Curran and Able 1998), the technique is

123

Mar Biol (2012) 159:1305–1315

recognized as an effective tool for quantifying relative predation pressures in habitats and has been used in numerous earlier studies (e.g., Laegdsgaard and Johnson 2001; Nakamura and Sano 2004; Chittaro et al. 2005; Grol et al. 2011; Nanjo et al. 2011). In field observations prior to the experiment, we confirmed that tethering treatment did not severely influence the behavior of fish. For example, fish rarely became entangled in the seagrasses and were able to exhibit antipredator maneuvers such as hiding behind seagrass leaves, remaining stationary on the sediment or fleeing for 1 h. Although nighttime predation may be a partial factor in juvenile mortality and should be addressed in future research (Danilowicz and Sale 1999), the density of piscivores has been shown to be lower at nighttime compared to daytime, with the number of nocturnal piscivores in the seagrass bed being lower than that in the coral areas at Itona reef (Nakamura personal observation). Accordingly, we consider that seagrass beds provide a safer sleeping area for juveniles compared to coral areas. During the period of the tethering experiments, the densities of potential piscivores in the seagrass bed and coral area at Itona reef were assessed by underwater visual transect surveys. In each habitat, fifteen 20 m 9 1 m belt transects, separated from each other by at least 10 m, were established at random, and density and total length of all piscivorous fishes within the transect area were recorded. Because opportunistic piscivory can impact on juvenile abundance (Sheaves 2001), potential piscivores included not only strict piscivores but also opportunistic piscivores (e.g., species for which fish form a minor component of the diet, Sano et al. 1984; Nakamura et al. 2003). The minimum body size of each piscivorous fish relative to the size class of prey was determined on the basis of gape size. Each census was made at high tide between 10:00 and 16:00 h. Independent t test was used to test for significant differences in the density of piscivorous fishes between the habitats. Prior to the analysis, data were transformed to log (x ? 1) to produce homogeneous variances. Size-specific foraging rate To test whether or not food availability for juveniles was greater in seagrass beds than in the coral areas, relative foraging rate (expressed as a percentage of individuals with food items in the stomach) in each habitat was investigated for each size class by a caging experiment. In July and August 2007, enclosure cages (1.5 m height, 1.2 m diameter) with 1-mm-mesh sides and open at the top were secured firmly to the substratum in the seagrass bed and coral area at Itona reef. Juvenile L. atkinsoni were collected from seagrass beds (because no small juveniles were available in other habitats, their use in this experiment was

Mar Biol (2012) 159:1305–1315

impracticable), starved for 2 days in laboratory water tanks continuously supplied with filtered sea water, released individually into each cage for 70 min and then removed using a dip net. Test duration (70 min) was fixed by our preliminary experiment in accordance with the method described by Laegdsgaard and Johnson (2001). Twelve replicates were conducted for small and medium size classes in each habitat and six replicates for the large size class in each habitat because of the difficulty in collecting large juveniles. After the foraging rate experiment, we investigated the stomach contents of these juveniles to clarify whether difference in the foraging rate between the seagrass bed and coral area in each size class could be explained by the ontogenetic food preference and the density of food items in each habitat. In the laboratory, stomach content items from each fish were identified to the lowest possible taxon. The percentage volume of each food item in the diet was visually estimated under a stereomicroscope using a 1 9 1 mm grid slide. Food resource use was expressed as mean percentage composition of each item by volume, being calculated by dividing the sum total of the individual volumetric percentages for the item by the number of specimens examined. Specimens with empty stomachs were excluded from the analysis. Independent t test was used to compare the percent volume of each food item in the diet in the seagrass bed and coral area for each size class after arcsine Hx data transformation. Mann–Whitney U test was performed instead of the t test for heterogeneous variances. The densities of important food items for L. atkinsoni juveniles were compared between the seagrass bed and coral area at Itona reef. Epifauna (invertebrates living on seagrasses and corals) and infauna (invertebrates living in the top layer of the substratum) were collected from five points located randomly (at least 10 m apart) in each habitat, using a hand-closing net (net mesh 0.4 mm, net length 74 cm, mouth opening 46 cm 9 46 cm) for epifauna and a cylindrical core sampler (13 cm in diameter and 2 cm in depth) for infauna according to the method described by Nakamura and Sano (2005). All samples were then preserved in 10 % formalin with Bengal Rose to color invertebrates and facilitate quantification. In the laboratory, individual numbers of each food item were counted using a stereomicroscope. Mean densities of each food item were expressed per m2. Independent t test was used to test for significant differences in the density of each food item between the habitats after log (x ? 1) data transformation. Mann–Whitney U test was performed instead of the t test for heterogeneous variances.

1309

the growth rates of small and medium juveniles were compared between those habitats at Itona reef using experimental cages in June and July 2011. Three cages (60 9 40 9 20 cm) with 3-mm nylon mesh sides and separated from each other by at least 3 m were set in the seagrass bed and coral area, respectively. The bottom was uncovered and exposed to the natural substratum. A closable window (10 cm 9 10 cm) was constructed on the upper surface of the cage so as to introduce and remove fish from the cages. L. atkinsoni juveniles were caught in the seagrass beds by seine net and total length (0.1 mm) measured in the field before introduction into the cages. A single juvenile was introduced into each cage because intraspecific competition for food could affect individual growth rates. Moreover, an experimental period of 1 week was established in order to prevent natural mortality of caged juveniles, based on a preliminary caging experiment in 2007–2008. However, one trial was conducted for only 4 days (1 set of medium juveniles), owing to the onset of bad weather. For each trial, similar-sized juveniles (no more than 1.5 mm in the difference in total length) were used (P = 0.69, t test between habitats for small juveniles of mean TL of 3.3 ± 0.2 SD cm, P = 0.78 for medium juveniles of 4.8 ± 0.3 cm TL). During the experiment, the cages were checked every day and cleaned of silt and algae. At the end of the experiment, fish were recovered from the cages by dip net and their length remeasured before release to the sea. Cages were removed after each trial in each habitat. Growth experiments were not conducted for large juveniles because the experimental cage was too small and the foraging rate of large juveniles was similar between the habitats. Finally, six replicates of small and medium size juveniles were successfully conducted in each habitat at the same time. For each juvenile, the growth rate (GR) was calculated as: GR ¼ ðLe  LsÞt1 where Ls and Le are the lengths of the fish at the beginning and end of the experiment, respectively, and t is the duration of the experiment in days (basically 7 days) for each individual fish. Mean growth was calculated by pooling all of the individuals per habitat for each size class and testing for statistical differences between the habitats by t test.

Results Post-settlement decline of juveniles in seagrass beds

Growth rate To confirm whether or not young juvenile growth rates were higher in the seagrass beds than in the coral areas,

Newly settled juveniles (\3 cm TL) were observed from late April to early September with the highest numbers occurring in May at Kanashiro (Fig. 2a) and June at

123

1310

Mar Biol (2012) 159:1305–1315

(a)

60

Small juveniles Medium juveniles Large juveniles

40

0

Total individual numbers

Mar April May June July Aug Sep Oct Nov Dec Jan Feb

50

(b)

40

Small juveniles Medium juveniles Large juveniles

30

9

Rock Coral Sargassum

6

Seagrass 3

0

Itona

10

Mar April May June July Aug Sep Oct Nov Dec Jan Feb

120

Ogan

Small

20

0

Total individual numbers

Tethering experiments for each juvenile size class demonstrated marginally higher survival rates in the seagrass bed than in the coral area for small juveniles even though the difference was not significant (t test, P = 0.10). No significant differences were obtained between habitats for medium (t test, P = 0.68) and large juveniles (all juveniles

2

20

Survival, foraging and growth rates between seagrass bed and coral habitats

Mean home range (m ) + SD

80

size between sites) had mean home ranges of only 0.6 m2. The home range size increased abruptly to a mean of 3.7 m2 for medium juveniles (4.9 ± 0.2 cm TL, n = 10, P = 0.87) and 6.5 m2 for large juveniles (7.8 ± 0.2 cm TL, n = 6, P = 0.88). Home range size did not differ significantly between the two sites for any of the juvenile size classes (t test, small juveniles, P = 0.87; medium juveniles, P = 0.80; large juveniles, P = 0.88). Small and medium juveniles were widely distributed in the seagrass beds and showed strong seagrass habitat fidelity, whereas large juveniles were often found around the edge of the seagrass beds and sometimes moved into the adjacent coral–algal complex habitat (Fig. 3).

(c) Kanashiro

80

Oganzaki

60 40 20 0 2

3

4

5

6

7

8

9

10

11

12

13

Total length (cm)

Itona

Ogan

Large

80 60 40 20 0 Sg

Fig. 2 Total individual numbers of three size classes of juvenile Lethrinus atkinsoni in the seagrass beds at a Kanashiro and b Oganzaki in each month from March 2007 to February 2008. c Length-frequency distributions of juvenile Lethrinus atkinsoni in the two seagrass bed sites during March 2007–February 2008

123

Ogan

Medium

100

100

1

Itona

Fig. 3 Mean home range size and substratum percentages within each home range of three size classes of juvenile Lethrinus atkinsoni in the seagrass beds at Itona and Oganzaki (Ogan)

Mean survival rate + SD

Total individual numbers

Oganzaki (Fig. 2b). The occurrence of medium size juveniles was confirmed from June to September at both sites, and large juveniles were observed from July to November, suggesting that most juveniles spent ca. 3–4 months in the beds. Decrease in number was very high between the small and medium size classes, but low between the medium and large size classes (Fig. 2c). Total individual numbers of each size class showed mean decreases of 69.2 % (82.3 % for Kanashiro and 56.1 % for Oganzaki) between the 2–3 cm and 5 cm TL classes, 88.4 % (90.0 % for Kanashiro and 86.8 % for Oganzaki) between the 2–3 cm and 8 cm TL classes, and 95.8 % (97.7 % for Kanashiro and 93.9 % for Oganzaki) between the 2–3 cm and 10 cm TL classes. All size classes lived in the seagrass beds and had a fairly stable home range (Fig. 3). Small juveniles (3.4 ± 0.1 SD cm TL, n = 10; P = 0.54, t test for body

Cr

Small

Sg

Cr

Medium

Sg

Cr

Large

Fig. 4 Mean survival rates in the seagrass bed (Sg) and coral area (Cr) of small, medium and large size classes of juvenile Lethrinus atkinsoni (tethering experiments)

Mar Biol (2012) 159:1305–1315

1311

Table 1 Mean (± SD) density and total length (cm) of piscivorous fishes per one transect (20 m2, n = 15) in the seagrass bed and coral area at Itona reef Family

Species

Seagrass bed

Coral area

Density

Total length

Density

Total length

Synodontidae

Synodus variegatus

–

–

0.1 ± 0.3

14.0

Serranidae

Cephalopholis urodeta

–

–

0.5 ± 0.6

12.3 ± 2.4

Epinephelus fasciatus Epinephelus merra

– –

– –

0.1 ± 0.3 0.2 ± 0.4

9.0 15.0 ± 5.0

Epinephelus polyphekadion

–

–

0.1 ± 0.3

22.5 ± 10.6

Pseudochromidae

Labracinus cyclophthalmus

–

–

0.3 ± 0.6

12.2 ± 4.4

Malacanthidae

Malacanthus latovittatus

–

–

0.1 ± 0.3

22.0

Carangidae

Caranx papuensis

0.1 ± 0.3

45.0

–

–

Lutjanidae

Lutjanus fulvus

–

–

0.1 ± 0.3

20.0

Lutjanus monostigma

–

–

0.1 ± 0.3

20.0

Labridae

Cheilio inermis

0.3 ± 0.5

15.8 ± 2.8

–

–

Pinguipedidae

Parapercis hexophtalma

–

–

0.1 ± 0.4

13.0

Mean density of species

0.4 ± 0.5

1.4 ± 1.0

Mean density of individuals

0.4 ± 0.5

1.5 ± 1.1

survived in each habitat) (Fig. 4). The survival rate increased with increasing body size in the seagrass bed (Tukey HSD test, P \ 0.05 between small and large juveniles, P \ 0.05 between medium and large juveniles, P = 0.98 between small and medium juveniles). Species and individual numbers of piscivores were significantly higher in the coral area than in the seagrass bed (t test, P \ 0.01 for both), large size piscivores being particularly abundant in the coral area (Table 1). Foraging experiments for each size class identified higher foraging rates in the seagrass bed than in the coral area for all size classes, the differences in rates between the habitats decreasing with increasing fish size (Fig. 5). For small juveniles, seagrass bed individuals fed mainly on small crustaceans, such as harpacticoid copepods, amphipods and isopods. However, the percentage volume of these food items in the diets did not differ significantly between the habitats (Mann–Whitney U test, P = 0.51, 0.31, and 0.41). Coral area individuals took only small gastropods, the percentage of the latter in the diets being significantly higher than in the seagrass bed individuals (Mann–Whitney U test, P \ 0.05). For medium juveniles, seagrass bed individuals fed mainly on amphipods and polychaetes, the percentage of these food items in the diets being higher than in the coral area individuals (Mann– Whitney U test, P = 0.10 for each), which took mainly isopods and gastropods (isopod percentage significantly higher than in seagrass bed individuals; Mann–Whitney U test, P \ 0.05). Large juveniles in the seagrass bed fed mainly on gastropods, polychaetes, fish and shrimps, the percentage volume of the latter 3 items in the diet not

Percentage volume of food items

– not observed

100

83

17

92

50

83

67

Sg

Cr

Sg

Cr

Sg

Cr

80 60 40 20 0

Small

Medium

Bivalve

Polychaete

Shrimp

Gastropod

Fish

Isopod

Ostracoda

Crab

Harpacticoida

Large Amphipod

Fig. 5 Foraging rates and percentage volume of food items in the diets of three size classes of juvenile Lethrinus atkinsoni in the seagrass bed (Sg) and coral area (Cr). Foraging rate, expressed as a percentage of individuals that were experimentally allowed to forage for a set time (70 min), is listed above each column

differing significantly between the habitats (Mann–Whitney U test, P = 0.37, 0.37 and 0.66). Coral area individuals took mainly gastropods (percentage in the diet marginally significantly higher than in seagrass bed individuals; Mann–Whitney U test, P = 0.08). For epifauna, the seagrass bed harbored significantly higher densities of amphipods, harpacticoid copepods and polychaetes than the coral area, all being important food

123

1312

Mar Biol (2012) 159:1305–1315

Table 2 Density (mean ± SD, n = 5) per square meter of the commonly consumed invertebrates by Lethrinus atkinsoni juveniles in the seagrass bed and coral area at Itona reef Density in top layer of bottom m-2

Food items

Harpacticoid copepods

Seagrass bed

Coral area

Density in seagrass and coral surfaces m-2 P value

Seagrass bed

Coral area

P value \0.01

21,197 ± 20,956

19,780 ± 10,960

0.73

270 ± 92

21.8 ± 14.0

Ostracods

166 ± 202

60.3 ± 98.3

0.40

15.1 ± 21.0

2.8 ± 4.2

Gammaridean amphipods Tanaids

769 ± 592 136 ± 188

678 ± 315 75.4 ± 131

0.87 0.57

105 ± 65.0 23.7 ± 21.7

18.9 ± 11.6 0

Isopods

45.2 ± 67.4

60.3 ± 135

0.93

1.9 ± 2.6

136 ± 112 15.1 ± 33.7

30.2 ± 67.4

0.76

Crabs

166 ± 98.2

136 ± 112

0.56

8,081 ± 5,069

8,804 ± 2,401

30.1 ± 67.4

151 ± 53.3

\0.01

90.5 ± 82.6

\0.05

Errant polychaetes Gastropods

Mean growth (body length increase) (mm d-1) + SD

Bivalves

0

\0.05

Shrimps Hermit crabs

0

0.5 0.4 0.3 0.2 0.1 0 Sg

Cr

Small

Sg

Cr

Medium

Fig. 6 Mean daily growth rates of small and medium size Lethrinus atkinsoni juveniles in the seagrass bed (Sg) and coral area (Cr) (from caging experiments)

items for small and medium juveniles (Table 2). For infauna, on the other hand, densities of gastropods and bivalves, both dietary items of large juveniles, were higher in the coral area, but no difference was observed between the seagrass bed and coral area for harpacticoid copepods, amphipods and polychaetes. Mean growth (body length) of juveniles was significantly greater in the seagrass bed than in the coral area (Fig. 6) for both small (t test, P = 0.05) and medium size classes (t test, P \ 0.01). However, although small juveniles grew faster than medium juveniles in the seagrass bed, no significant difference was found between the two size classes (t test, P = 0.22).

Discussion The present study demonstrated that the greatest decrease in juvenile number occurred long before they moved out of the nursery, with individual numbers declining by 70 % during growth from 3 to 5 cm in body length, and

123

0.91

40.7 ± 16.9 1.9 ± 2.6

0 12.3 ± 8.6 0

0.16 \0.01 \0.01 0.31 \0.05 0.31

34.1 ± 25.8

\0.05

155 ± 45.3

12.3 ± 7.9

\0.01

5.7 ± 5.2

11.4 ± 8.6

0.33

5.7 ± 10.3

9.5 ± 3.3

0.15

0

probably, more than 90 % of individuals were lost during the juvenile stage in seagrass beds. High densities of conspecifics in a nursery can influence the timing of ontogenetic habitat shifts because of the potential for density to limit the amount of available food (Adams et al. 2006; Snover 2008). However, our study implies that the densitydependent effect on ontogenetic habitat shift was negligible because the density of large juveniles was very low in seagrass beds (8-cm size class; mean of 0.1–0.5 individuals per 30 m-2 for Oganzaki and 0.1–0.4 individuals per 30 m-2 for Kanashiro in each weekly census during the study period). As movement of juvenile L. atkinsoni was found to be negligible in the seagrass bed, the steep decline in individual numbers was considered to be directly related to mortality, particularly predation. Indeed, the tethering experiment showed that the survival rate in the seagrass bed was lowest for small juveniles. Similarly, Shulman and Ogden (1987) and Watson et al. (2002) estimated that 79–90 % of Haemulon flavolineatum and Ocyurus chrysurus individuals disappeared within 1 month following settlement in a seagrass nursery, believing the primary cause to be predation. On the other hand, declines in individual numbers were low in the[5 cm TL size classes. Tethering experiments in our study also showed survival rates to be higher for larger juveniles. Because most piscivores in seagrass beds are small and larger juveniles have better swimming ability, the latter may be less susceptible to predation. Although post-settlement mortality was high for early stage juveniles in the seagrass beds, the tethering experiments showed that the survival rate of small juveniles was marginally higher in the seagrass bed than in the coral area. A similar phenomenon has been reported in other juvenile fish in this region (Nakamura and Sano 2004), suggesting that predation pressure is potentially lower in the seagrass

Mar Biol (2012) 159:1305–1315

bed than in the coral habitat for small juveniles. Indeed, density of piscivores was significantly lower in the seagrass bed than in the coral area. Moreover, the foraging rate of small juveniles was considerably higher in the seagrass bed than in the coral area. This phenomenon may be attributable to the high density of important food items on the seagrass leaves and on the sea bottom. Clearly, small juveniles in the seagrass beds have the benefit of both increased acquisition rates of food and increased protection. Although their survival rate did not differ between the two habitats, medium juveniles had a higher foraging rate in the seagrass bed than in the coral area. These higher foraging rates in the seagrass bed may be largely due to the greater density of epiphytic small crustaceans and polychaetes in the beds compared to the coral area, leading to significantly higher growth rates in the seagrass beds, as shown in our growth experiment. It is possible that juveniles collected from seagrass beds in caging experiments had to adapt to a novel environment, that is, going from seagrass bed to coral habitat, which may have diminished foraging intensity and growth rates in the latter habitat. However, all the juveniles showed a consistent manner in both habitats immediately after release making this possibility unlikely. It is also known that foraging gains of smaller fish are often balanced against the risk of predation, that is, protection from predators allows more time for feeding (Mittelback 1984). In this study, all juveniles were released in fine mesh enclosure cages; therefore, the foraging activity of each juvenile was not affected by the density of predators. Foraging intensity in coral habitats may be less under natural conditions than in experimental cages given that smaller juveniles fear predators during foraging. Although caging experiments by Grol et al. (2008) showed that growth rates of early juvenile Haemulon flavolineatum were lower in mangrove and seagrass nurseries than in coral reefs, being directly opposite to the present findings, this was due to juvenile dietary differences in the two species, L. atkinsoni feeding on seagrass-bed-rich invertebrates and H. flavolineatum taking planktonic copepods, which are more abundant on coral reefs. Thus, differences in the food habits of resident species would lead to species-specific differences in seagrass nursery function and mechanisms underlying the timing of ontogenetic habitat shift. All large tethered juveniles survived in the seagrass bed and coral area (total 30 individuals), indicating that predation was not a major limiting factor for the distribution patterns observed in the two habitats. Indeed, we did not find large piscivores that could prey on large size juveniles in the habitats (Table 1). Moreover, large juveniles moved around, not only in the seagrass beds but also in adjacent coral–algal complex habitats. This indicated that juveniles need not fear predators during foraging because larger fish

1313

are less vulnerable to predation (Werner et al. 1983). The difference in foraging rate was smallest between the seagrass bed and the coral area, with large juveniles beginning to feed on larger items such as bivalves, gastropods and crabs, all of which were more abundant in the coral area. The shift in food preference to large crustaceans or hardshelled animals is partly the consequence of morphological changes including the development of jaw crushing strength and an increase in mouth gape size with body size (Wainwright 1988). In addition, effective energy gain minimizes the energy cost–benefit ratio, with larger predators selecting less large preys in preference to numerous small preys (Stephens and Krebs 1986). The positive relationship between fish size and home range size in the seagrass beds in addition to the expanded home range of large juveniles in the coral–algal complex habitat indicated that large juveniles sought large food items in the beds and began to adapt to coral habitats to capture this prey. Indeed, we observed large juveniles sometimes pecking the sea bottom in the coral–algal complex habitat. Wild larger juveniles ([9 cm TL), immediately before moving from the seagrass nursery as well as living in coral areas, fed exclusively on large crustaceans and hard-shelled animals (Nakamura et al. 2003) despite small prey animals being abundant in both habitats (Nakamura and Sano 2005). Therefore, the timing of the ontogenetic habitat shift from seagrass beds to coral reef occurs when the density of larger food items, such as crabs and gastropods, in the seagrass beds is not sufficient for the needs of large juveniles. These energy demands are not satisfied by the numerous small preys in the beds. Similar shifts in diet that are related to coral reef migration have also been observed in Caribbean reef fish living in a seagrass/mangrove nursery (Cocheret de la Moriniere et al. 2003). The body size of the timing of ontogenetic habitat shift is 7–12 cm TL for L. atkinsoni (Wilson 1998; Shibuno et al. 2008; Nakamura and Tsuchiya 2008). Therefore, the optimal size for the ontogenetic habitat shift out of the seagrass nursery within this size range may be affected by the size of coral habitat predators and the degree of large food requirements at the time. The ontogenetic habitat shift out of the seagrass nursery is size-dependent and driven by foraging and refuge needs. Although some of our experiments had relatively small sample size, we assume this would not have significantly undermined the major finding of our study because the results of each experiment were fairly consistent and were representative of the surrounding environments (e.g., densities of food and predators). Moreover, although the present study was conducted at a single location in the southern Ryukyu Islands, we predict that the mechanisms underlying ontogenetic habitat shifts that we observed are representative of the general pattern in other locations in

123

1314

the Ryukyu Islands and elsewhere in the Western Pacific. This prediction is based on the knowledge that size distribution patterns in seagrass and coral habitats of L. atkinsoni are similar in this huge area (Wilson 1998; Nakamura and Tsuchiya 2008) and that the ontogenetic food preference of L. atkinsoni (Nakamura et al. 2003) and distribution patterns of predators and commonly consumed invertebrates are also similar in other locations in the Ryukyu Islands (Nakamura and Sano 2004, 2005). Overall, the present study showed that although individual numbers of L. atkinsoni juveniles decreased by [90 % during their stay in the seagrass nursery, the shelter and/or food availability functions of the latter, at least to a juvenile size of ca. 5 cm TL, provided the best survival and growth option, the timing of ontogenetic migration to coral reefs at larger size being attributed to foraging efficiencies on larger food items in the different habitats. For prereproductive animals, fitness is maximized by the minimization of time to reproductive maturity (i.e., maximization of growth rates) and maximization of probability of survival to reproductive maturity (i.e., minimization of mortality rates; Snover 2008). Because mortality rates during benthic life stages of marine fishes are generally highest in the youngest/smallest classes, survival and growth during the early juvenile stage will drive adult population size. Accordingly, any loss of specific nursery habitats will dampen the speed and sustainability of reproductive cycles of nursery-reliant species (Nagelkerken et al. 2002; Mumby et al. 2004; Nakamura 2010). The identification, conservation and restoration of nursery habitats is urgently required to meet increasing demands of fisheries resources and other marine ecosystem services impacted upon by human population growth. Acknowledgments We are grateful to O. Abe, T. Takada, T. Hayashibara, A. Nishihara, M. Horinouchi, S. Nakamura and T. Miura for assistance in fieldwork. Constructive comments on the manuscript from G. Hardy, J. Fodrie and three anonymous reviewers were much appreciated. This study was supported by grunts from the Japan Society for the Promotion of Science (No. 1810371) and the Ministry of Education, Culture, Sports, Science and Technology of Japan (No. 21780178) to Y.N.

References Adams AJ, Dahlgren CP, Kellison GT, Kendall MS, Layman CA, Ley JA, Nagelkerken I, Serafy JE (2006) Nursery function of tropical back-reef systems. Mar Ecol Prog Ser 318:287–301. doi:10.3354/ meps318287 Arvedlund M, Takemura A (2006) The importance of chemical environmental cues for juvenile Lethrinus nebulosus Forsska˚l (Lethrinidae, Teleostei) when settling into their first benthic habitat. J Exp Mar Biol Ecol 338:112–122. doi:10.1016/j.jembe. 2006.07.001 Beck MW, Heck KL Jr, Able KW, Childers DL et al (2001) The identification, conservation, and management of estuarine and

123

Mar Biol (2012) 159:1305–1315 marine nurseries for fish and invertebrates. Bioscience 51:633– 641. doi:10.1016/S0006-3207(00)00212-3 Carpenter KE, Allen GR (1989) FAO species catalogue, Vol 9 Emperor fishes and large-eye breams of the world (family Lethrinidae). An annotated and illustrated catalogue of lethrinid species known to date. FAO Fish. Synop. No. 125, vol. 9. FAO, Rome Chittaro PM, Usseglio P, Sale P (2005) Variation in fish density, assemblage composition and relative rates of predation among mangrove, seagrass and coral reef habitats. Environ Biol Fish 72:175–187. doi:10.1007/s10641-004-9077-2 Cocheret de la Moriniere E, Pollux BJA, Nagelkerken I, Hemminga MA, Huiskes AHL, van der Velde G (2003) Ontogenetic dietary changes of coral reef fishes in the mangrove-seagrass-reef continuum: stable isotopes and gut-content analysis. Mar Ecol Prog Ser 246:279–289. doi:10.3354/meps246279 Curran MC, Able KW (1998) The value of tethering fishes (winter flounder and tautog) as a tool for assessing predation rates. Mar Ecol Prog Ser 163:45–51. doi:10.3354/meps163045 Dahlgren CP, Eggleston DB (2000) Ecological processes underlying ontogenetic habitat shifts in a coral reef fish. Ecology 81:2227–2240. doi:10.1890/0012-9658(2000)081[2227:EPUOHS] 2.0.CO;2 Danilowicz BS, Sale PF (1999) Relative intensity of predation on the French grunt, Haemulon flavolineatum, during diurnal, dusk, and nocturnal periods on a coral reef. Mar Biol 133:337–343. doi: 10.1007/s002270050472 Ebisawa A (1999) Reproductive and sexual characteristics in the Pacific yellowtail emperor, Lethrinus atkinsoni, in waters off the Ryukyu Islands. Ichtyol Res 46:341–358. doi:10.1007/s00442010-1833-8 Grol MGG, Dorenbosch M, Kokkelmans EMG, Nagelkerken I (2008) Mangroves and seagrass beds do not enhance growth of early juveniles of a coral reef fish. Mar Ecol Prog Ser 366:137–146. doi:10.3354/meps07509 Grol MGG, Nagelkerken I, Rypel AL, Layman CA (2011) Simple ecological trade-offs give rise to emergent cross-ecosystem distributions of a coral reef fish. Oecologia 165:79–88. doi: 10.1007/s00442-010-1833-8 Heck KL Jr, Hays G, Orth RJ (2003) Critical evaluation of the nursery role hypothesis for seagrass meadows. Mar Ecol Prog Ser 253:123–136. doi:10.3354/meps253123 Horinouchi M, Nakamura Y, Sano M (2005) Comparative analysis of visual censuses using different width strip-transects for a fish assemblage in a seagrass bed. Estuar Coast Shelf Sci 65:53–60. doi:10.1016/j.ecss.2005.05.003 Laegdsgaard P, Johnson C (2001) Why do juvenile fish utilize mangrove habitats? J Exp Mar Biol Ecol 257:229–253. doi: 10.1016/S0022-0981(00)00331-2 McCormick MI, Makey L, Dufour V (2002) Comparative study of metamorphosis in tropical reef fishes. Mar Biol 141:841–853. doi:10.1007/s00227-002-0883-9 Mittelback G (1984) Predation and resource partitioning in two sunfishes (Centrarchidae). Ecology 65:499–513. doi:10.2307/ 1941412 Mumby PJ, Edwards AJ, Arias-Gonzalez JE, Lindeman KC et al (2004) Mangrove enhance the biomass of coral reef fish communities in the Caribbean. Nature 427:533–536. doi:10.1038/ nature02286 Nagelkerken I (2009) Evaluation of nursery function of mangroves and seagrass beds for tropical decapods and reef fishes: patterns and underlying mechanisms. In: Nagelkerken I (ed) Ecological connectivity among tropical coastal ecosystems. Springer, New York Nagelkerken I, Dorenbosch M, Verberk WCEP, Cocheret de la Morinie`re E, van der Velde G (2000) Importance of shallow-

Mar Biol (2012) 159:1305–1315 water biotopes of a Caribbean bay for juvenile coral reef fishes: patterns in biotope association, community structure and spatial distribution. Mar Ecol Prog Ser 202:175–192. doi:10.3354/ meps202175 Nagelkerken I, Roberts CM, van der Velde G, Dorenbosch M, van Riel MC, Cocheret de la Morinie`re E, Nienhuis PH (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. doi:10.3354/meps244299 Nakamura Y (2010) Patterns in fish response to seagrass bed loss at the southern Ryukyu Islands, Japan. Mar Biol 157:2397–2406. doi:10.1007/s00227-010-1504-7 Nakamura Y, Sano M (2004) Is there really lower predation risk for juvenile fishes in a seagrass bed compared with an adjacent coral area? Bull Mar Sci 74:477–482 Nakamura Y, Sano M (2005) Comparison of invertebrate abundance in a seagrass bed and adjacent coral and sand areas at Amitori Bay, Iriomote Island, Japan. Fish Sci 71:543–550. doi:10.1111/ j.1444-2906.2005.00998.x Nakamura Y, Tsuchiya M (2008) Spatial and temporal patterns of seagrass habitat use by fishes at the Ryukyu Islands, Japan. Estuar Coast Shelf Sci 76:345–356. doi:10.1016/j.ecss.2007.07.014 Nakamura Y, Horinouchi M, Nakai T, Sano M (2003) Food habits of fishes in a seagrass bed on a fringing coral reef at Iriomote Island, southern Japan. Ichthyol Res 50:15–22. doi:10.1007/ s102280300002 Nakamura Y, Shibuno T, Lecchini D, Watanabe Y (2009) Habitat selection by emperor fish larvae. Aquat Biol 6:61–65. doi: 10.3354/ab00169 Nanjo K, Nakamura Y, Horinouchi M, Kohno H, Sano M (2011) Predation risk for juvenile fishes in a mangrove estuary: a comparison of vegetated and unvegetated microhabitats by tethering experiments. J Exp Mar Biol Ecol 405:53–58. doi: 10.1016/j.jembe.2011.05.016 Peterson CH, Black R (1994) An experimentalist’s challenge: when artifacts of intervention interact with treatments. Mar Ecol Prog Ser 111:289–297

1315 Sano M, Shimizu M, Nose Y (1984) Food habits of teleostean reef fishes in Okinawa Island, southern Japan. Univ Mus Univ Tokyo Bull 25:1–128 Sheaves M (2001) Are there really few piscivorous fishes in shallow estuarine habitats? Mar Ecol Prog Ser 222:279–290. doi: 10.3354/meps222279 Shibuno T, Nakamra Y, Horinouchi M, Sano M (2008) Habitat use patterns of fishes across the mangrove-seagrass-coral reef seascape at Ishigaki Island, southern Japan. Ichthyol Res 55:218–237. doi:10.1007/s10228-007-0022-1 Shulman MJ, Ogden JC (1987) What controls tropical reef fish populations: recruitment or benthic mortality? An example in the Caribbean reef fish Haemulon flavolineatum. Mar Ecol Prog Ser 39:233–242 Snover ML (2008) Ontogenetic habitat shifts in marine organisms: influencing factors and the impact of climate variability. Bull Mar Sci 83:53–67 Stephens DW, Krebs JR (1986) Foraging theory. Princeton University Press, Princeton Wainwright PC (1988) Morphology and ecology: functional basis of feeding constraints in Caribbean labrid fishes. Ecology 69:635– 645. doi:10.2307/194102 Watson M, Munro JL, Gell FR (2002) Settlement, movement and early juvenile mortality of the yellowtail snapper Ocyurus chrysurus. Mar Ecol Prog Ser 237:247–256. doi:10.3354/meps 237247 Werner EE, Hall DJ (1988) Ontogenetic habitat shifts in bluegill: the foraging rate-predation risk trade-off. Ecology 69:1352–1356. doi:10.2307/1941633 Werner EE, Gilliam JF, Hall DJ, Mittelbach GG (1983) An experimental test of the effects of predation risk on habitat use in fish. Ecology 64:1540–1548. doi:10.2307/1937508 Wilson GG (1998) A description of the early juvenile colour patterns of eleven Lethrinus species (Pisces: Lethrinidae) from the Great Barrier Reef, Australia. Records of the Australian Museum 50:55–83. doi:10.3853/j.0067-1975.50.1998.1274

123