Feeding ecology of two sympatric species of large-sized groupers ( Perciformes : Epinephelidae ) on Southwestern Atlantic coralline reefs

Red and black groupers are large-bodied opportunistic ambush predators commonly found in Southwestern Atlantic tropical reefs. We investigated the diet of both species in order to detail ontogenetic, spatial and temporal trends, and to assess the extent of overlap in resource use between these two sympatric predators on the Abrolhos Bank, Brazil. Decapods and fishes were the main food items of Epinephelus morio while fishes were the main prey of Mycteroperca bonaci. Both diets were significantly influenced by body size and habitat, but only smaller individuals of E. morio feed almost exclusively on crustaceans. While the two groupers rely on many of the same prey types, coexistence may be facilitated by E. morio feeding more heavily on crustaceans, particularly the blackpoint sculling crab Cronius ruber, while black grouper take comparatively few crustaceans but lots of fish prey. Predators like red and black groupers could trigger indirect effects in the community and influence a large range of ecological processes, such as linkages between top and intermediate predators, and intermediate predators and their resources.


Introduction
Epinephelids are demersal predatory fishes found in shallow to mesophotic waters of all tropical and subtropical oceans, especially in hard bottom habitats (Heemstra, Randall, 1993;Craig et al., 2011).The 160 or so species in the family encompass a wide range of body sizes and food habits, ranging from small zooplanktivores to some of the largest predators in tropical and subtropical reef ecosystems (Claro et al., 2001).
Black and red groupers are important predators in hard-bottom communities of the tropical Western Atlantic (Brulé, Canché, 1993;Brulé et al., 1994;Brulé et al., 2005).In the Abrolhos Bank, they are the two most abundant epinephelids, representing more than three times the biomass of all other groupers combined, and 20% of large carnivore's biomass, a category that also includes sphyraenids, lutjanids and carangids (Francini-Filho, Moura, 2008a).In the Northern Hemisphere, fishes, crustaceans, gastropods and cephalopods are important dietary components of both species (Moe, 1969;Bullock, Smith, 1991;Brulé, Canché, 1993;Brulé et al., 1994;Brulé et al., 2005), but there is no information on their food habits for the entire Southwestern Atlantic.
Groupers are predators thought to play important roles in ecosystem function through either direct or indirect interactions (Huntsman et al., 1999;Dulvy et al., 2004;Campbell, Perdede, 2006;Rizzari et al., 2014).Therefore investigations on the relationships among such large-bodied marine predators, their predator-prey interactions and the environment are important for an overall understanding of the mechanisms that structure populations and communities of reef fishes (Hixon, 1991;Hixon, Beets, 1993).Food partitioning among closely related co-occurring predators is crucial for their coexistence (Davies et al., 2007).
In the present study we investigated food habits of the black and red groupers obtained from eight years of artisanal fisheries monitoring program in the Abrolhos Bank region in Brazil.This comprehensive dataset included stomach contents from groupers collected from a variety of habitats and seasons, and therefore presumably reflects the general feeding habits of species examined.Information provided here represents one of the most extensive databases assembled over a relatively large period of time, particularly for a coralline reef environment, and provides a uniquely large amount of information to investigate ecological interactions among large-bodied marine predators, and to understand the role of predation in the regulation of reef fish communities.Specifically, our objectives were to determine important prey groups in the diets of these groupers and estimate their relative importance, identify any spatial, temporal or ontogenetic shifts in their diets, and assess the extent of overlap in the diet between these two sympatric predators.Data provided here is the first information on the feeding ecology of the red and black groupers in the Southwestern Atlantic.

Material and Methods
Study area.The study was carried out in the Abrolhos Bank (16º40'-19º40' S; 39º10'-37º 20' W), a region with 42,000 km 2 and depths rarely exceeding 30 m, with a shelf edge at about 70 m depth.Abrolhos harbors an extensive mosaic of benthic megahabitats (Moura et al., 2013).Rhodolith beds comprise the largest megahabitat (~21,000 km²), followed by unconsolidated sediments (~20,000 km 2 ) and reefs (~9,000 km 2 ).The two groupers studied are predominantly found on reefs and less often on rhodolith beds, being absent from unconsolidated bottoms.Most reefs in this region assume the characteristic and highly peculiar form of mushroom-shaped pinnacles (chapeirões), which rise 5 to 25 m height above the bottom and extend from 20 to 300 m across their tops (Francini-Filho, Moura, 2008a,b).
There are four marine protected areas (MPAs) in the region, all of them poorly enforced.The Extractive Reserve of Cassurubá (ERC; 1,006 km 2 ) includes large mangroves in which fishing is small-scale artisanal or for subsistence.The other three MPAs encompass reefs: the no-take National Marine Park of Abrolhos (NMPA; 913 km 2 ), the multiple-use Marine Extractive Reserve of Corumbau (MERC; 895 km 2 ), and the Environmental Protected Area Ponta da Baleia/Abrolhos (EPA; 3,460 km 2 ).
Samples were obtained in two cross-shelf strata, Inner Shelf (IS) and Outer Shelf (OS) (Fig. 1).The IS includes a no-take zone of the NMPA (Timbebas Reefs) and other unprotected reefs such as the Parcel das Paredes (the largest continuous reef in the South Atlantic), Sebastião Gomes, Coroa Vermelha and Viçosa.Reefs in this IS arc consist of banks with flat and exposed tops, and mushroom-shaped pinnacles, with depths reaching up to 20 m.This area is subjected to the highest fishing pressure in the region, with ~200 boats operating regularly with hand lines, 3 e160047 [3] spears and various types of nets (Francini-Filho, Moura, 2008a).Siliciclastic (= terrigenous) contents in peri-reefal sediments are higher in the IS, while the carbonate fraction is higher in the OS.The OS includes the best-enforced area of the NMPA, including the Abrolhos islands and the Parcel dos Abrolhos.Besides this protected area with emerging pinnacles, the OS as also encompasses a large realm of mesophotic reefs, rhodoliths beds and seasonal fleshy algal pavements, from 25 to 90 m depths (Moura et al., 2013).The OS also harbors non-emergent pinnacles and coalesced reef structures (drowned reefs), paleo valleys and channels, as well as depressions (similar to sinkholes but formed differently) locally known as "buracas" (Bastos et al., 2013;Moura et al., 2013).
Sampling and analytical procedures.Specimens were obtained monthly through a fish landing monitoring program that target fleets on hook and line, longline and spear fishing in the Cities of Nova Viçosa, Caravelas, Alcobaça, and Prado (Fig. 1), between June 2005 and September 2012.After capture, fishes were immediately stored on ice on board, and transferred to the laboratory where they were kept until processing.Fishes were measured (Total Length -TL cm; Standard Deviation -SD), weighed, and their stomachs were immediately fixed in 10% formalin for 24 h, and subsequently transferred and stored in 70% alcohol.Voucher specimens were deposited in the ichthyological collection of Museu de História Natural Capão da Imbuia (MHNCI 12657 for E. morio and MHNCI 12658 for M. bonaci).Stomach contents were examined in the laboratory using a stereomicroscope.The identification of food items was performed as refined as possible according to literature data (Melo, 1996;Amaral et al., 2006) and consultation with experts.All identifiable prey items were enumerated, weighed and identified to the lowest possible taxon (LPT).If gut contents were too digested for identification, the material was weighed and classified as "remains".Frequency of occurrence (%FO) (i.e., percentage of stomachs in which a food item occurred), proportion by weight (%W) (i.e., percentage participation of each item in the total food weight), as well as the proportion in number (%N) (i.e., percent of each item within total food items) were determined (Hyslop, 1980;Bowen, 1996).These variables were used to calculate the Index of Relative Importance (IRI) (cf.Pinkas et al., 1971), which establishes the order of importance of food items in the diet.IRI values were standardized to percentages (Cortés, 1999) and were calculated according to the equation IRI = FO (%W + %N).Feeding strategy diagrams (Amundsen et al., 1996) were used to examine the dietary importance of a particular prey item.These diagrams allow for a visual assessment of niche width and the importance of different prey items, as well as the predators' strategies (specialization vs. generalization), using occurrence and prey-specific weight.
Variations on diet were visualized with respect to cross-shelf strata (IS and OS), size classes (E.morio: juveniles 15-50.0cm TL and adults 50.5-90 cm TL; M. bonaci: juveniles 26.1-62 cm TL and adults 68-147 cm TL; according to sex and reproductive development presented in Freitas, 2014) and seasons (cool and warm).Seasons were characterized from sea surface temperature (SST) historical data (November 1981-January 2008; http://nomad3.ncep.noaa.gov).The cool season has median SSTs < 26 o C and spans from June to November (min.SST 24.5 o C, August).The warm season has median SSTs >26 o C, spanning from December to May (max.SST 28 o C, February).The warm season is characterized by prevailing NE winds and higher water visibility, while the cold season is characterized by polar front intrusions causing strong sediment re-suspension (Segal et al., 2008).
A two-way crossed permutational multivariate analysis of variance (PERMANOVA) using a populationwide dissimilarity metric was employed to examine the effects of habitat, ontogeny and season (factors) on the standardized and transformed (Logx+1) weight contribution of each prey item (LPT).Significant factors were further analyzed using PERMANOVA pair-wise comparisons.The Bray-Curtis dissimilarity was used in all tests, with 999 permutations under a reduced model.Similarity of percent contribution (SIMPER) analysis was used to examine the prey categories that are most responsible for between-factors separation (Clarke, Gorley, 2001).All analyses were performed using the PRIMER/PERMANOVA 6.0 software (Plymouth Marine Laboratory, Plymouth, England).
For E. morio several less abundant taxa clustered in the lower left quadrant of the feeding strategy diagram, representing items with low contribution to the diet in terms of prey-specific weight.For juveniles (Fig. 2a), the main prey items were C. ruber, fish remains, brachyuran remains, O. vulgaris, and several decapods (Leptopisa setirostris, Macrocoeloma camptocerum, Mithraculus forceps, Notolopas brasiliensis and Microphrys antillensis).Among adults (Fig. 2b) the dominant identifiable items included Gymnothorax spp., Acanthurus spp.and Scarinae.For M. bonaci juvenile samples the feeding strategy diagram showed a dominance of fish remains in the upper right portion (Fig. 2c), which suggested a more specialized feeding strategy than that of E. morio.This is also supported by the presence of other fish species (S. trispinosus and H. brasiliensis), and also of P. argus in the lower left quadrant.Adult samples of M. bonaci were similar, with presence of A. monocerus and H. aurolineatum (Fig. 2d).
The diet of the red and black groupers was significantly influenced by body size and habitat (PERMANOVA P=0.001 and P=0.015, respectively; Tab. 2).We also found significant dietary differences in pairwise comparisons between cross shelf strata (PERMANOVA, P= 0.026) and species/stages of juvenile and adult E. morio (PERMANOVA, P= 0.011), juvenile E. morio and M. bonaci (PERMANOVA, P= 0.001), adult E. morio and juvenile M. bonaci (PERMANOVA, P= 0.001) and adult E. morio and M. bonaci (PERMANOVA, P= 0.016).Fish as prey were the main contributors to the dietary dissimilarities between species, with a greater importance for M. bonaci diet (Tab.3).SIMPER analyses showed that fish remains (23.86%), C. ruber (12.18%) and brachyura remains (11.5%) (Tab.4), where the main contributions to the dietary dissimilarities between IS and OS cross-shelf strata.

Discussion
Several stomachs of both species caught by hook and line, longline and spear fishing were empty.Despite the fact that some fish species regurgitated their prey(s) after capture, as already stated for some large-fish species (Randall, 1967), including groupers (Randall, Brock, 1960;Nakai et al., 2001;Dierking, Meyer, 2009), we suspect that epinephelids regularly experience long periods of empty stomachs (e.g.Condini et al., 2011;López, Orvay, 2005;Reñones et al., 2002), as a combined result of resource availability, digestion rates and energy balance (Arrington et al., 2002).Both species are opportunistic (Smith, 1971;Randall, 1967;Sierra et al., 2001), "sit and wait" ambush predators (Bullock, Smith, 1991;Brulé et al., 2005), and the regular occurrence of empty stomachs provides support to the idea that they are intermittent rather than continuous feeders.
Decapods and fishes were the main food items of E. morio, while fishes were the main prey of M. bonaci in the Abrolhos Bank.This pattern corroborates previous studies in the Northern Hemisphere (Gulf of Mexico and Caribbean), where red groupers also feed on brachyurans, stomatopods, molluscs, and small fishes (Randall, 1967;Moe, 1969;Brulé, Canché 1993;Brulé et al., 1994), while black groupers feed mainly on fishes, and secondarily on crustaceans (Brulé et al., 2005).Results were consistent with the notion that both species are important mesocarnivores on West Atlantic reefs (Randall, 1967;Parrish, 1987;Sierra et al., 2001).
The diet of red grouper changed significantly with size.Smaller individuals fed almost preferentially on crustaceans (especially Cronius ruber and brachyurans generally), and also on a wide range of rare items including teleosts, decapods, octopuses, shrimps, and stomatopods.Reef fragments and filamentous algae were recorded in small quantities and most probably represent accidental ingestion during predation (Linde et al., 2004;Brulé et al., 1994;Machado et al., 2008).Cronius ruber and brachyuran remains were also important components of the adult diet.Studies in the Northern Hemisphere (Brulé, Canché, 1993 -12.1 to 40 cm TL; Brulé et al., 1994 -13 to 36 cm FL) also recorded such a preference for crustaceans and mollusks in addition to small fishes by juveniles, even considering sizerelated differences among previous studies.
The increased consumption of fishes by larger individuals may represent a general trend in groupers (Smith, 1971;Brulé, Canché, 1993), and also for other large carnivorous reef fishes (Sierra et al., 2001;Linde et al., 2004), observations already noted for mutton snapper and goliath grouper on Abrolhos Bank (Freitas et al., 2011b(Freitas et al., , 2015)).Such ontogenetic trend is clear for the more generalist red grouper, and is aligned with optimal foraging theory, i.e. maximum energetic return obtained when predator selects high quality prey (Gerking, 1994).Besides reducing the competition for food, or even meeting the higher energetic demand from migration and reproduction (Gerking, 1994;Sierra et al., 2001), ontogenetic shifts in diet are often accompanied by shifts in habitat (Sierra et al., 2001), possibly contributing to reduced niche overlap (Machado et al., 2008).
Red grouper juveniles (20-40 cm) are commonly found in crevices and under ledges in 5-25 m depths, in the inner and mid shelf.At 40-50 cm adults migrate to deeper water (50-300m), where they also occur over sandy or mud bottoms (Craig et al., 2011).The juvenile habitat for black grouper appears to vary geographically.Juvenile habitat in the Yucatán Peninsula mainly consists of sandyrocky bottoms with some ridges and crevices (Renán et al., 2003) while in the north Atlantic and Gulf of Mexico it is largely seagrass beds in estuarine areas and on coral reefs (Bullock, Smith, 1991;Sluka et al., 1994;Ross, Moser, 1995).In our study, juveniles of E. morio and M. bonaci came from exclusively the IS while the adults mostly came from the OS.This pattern suggests that E. morio and M. bonaci exhibit size segregation according to depth, showed to dog snapper Lutjanus jocu (Bloch, Schneider, 1801), mutton snapper Lutjanus analis (Cuvier, 1828), goliath grouper Epinephelus itajara (Lichtenstein, 1822) and Cephalopholis fulva (Linnaeus, 1758), another carnivorous fish in Abrolhos Bank (Freitas et al., 2011a(Freitas et al., , 2015;;Moura et al., 2011;Gathaz et al., 2013), and as noted by Moe (1969) for the Gulf of Mexico, and by González et al. (1974) and Valdés, Padrón (1980) at the Campeche Bank, Mexico.On the IS, juveniles of both species inhabited chapeirões, reef structures with a characteristic mushroom-shaped form, clustered at depths of 5 to 25m.This could explain the high occurrence of crustaceans in red grouper diet and reef fishes in black grouper diet in this particular reef habitat, because such structures increase the heterogeneity and complexity of habitat types (Dutra et al., 2005).In contrast, middle and outer shelf habitats comprise a much larger realm of mesophotic reefs, rhodolith beds, and fleshy algal pavement (Moura et al., 2013) in depths from 25-90 m.While the two groupers rely on many of the same prey types, coexistence may be facilitated by E. morio feeding more heavily on crustaceans, particularly C. ruber, while black grouper take comparatively few crustaceans but lots of fish prey.Furthermore, prey captured by young red grouper are generally slow-moving benthic species (principally decapods) while those consumed by black grouper are less bottom-dependent, and faster-moving organisms (reeffishes) (Brulé et al., 2005), featuring a pattern similar to E. itajara, on Abrolhos (Freitas et al., 2015).The annual rhythm of fish feeding intensity is strongly associated with environmental conditions and their effect on the food supply (John, 2001;Sierra et al., 2001).Feeding habits and diet composition of reef fishes are usually highly variable, in part because prey availability change seasonally (e.g.Roos, Moser, 1995;Monteiro et al., 2009;Pimentel, Joyeux, 2010).The lack of dietary temporal variation in this study (i.e.no significant seasonal variation in diet e160047 [10] composition to E. morio and M. bonaci) is consistent with the results of Brulé et al. (1994Brulé et al. ( , 2005) ) and others groupers (John, 2001) and we believe that the absence of major seasonal differences (e.g.mean and seasonal changes in SST) may explain our results.In fact, SST's along the Abrolhos Bank are very homogeneous over the year, with fluctuations not exceeding 4ºC (see Freitas et al., 2011b).
Red and black groupers from Abrolhos Bank did not feed heavily on other species that are commercial harvested.Their diet is likely most influenced by variation in prey availability and diversity in different areas.In marine ecosystems, trophic cascades have been observed in hard bottom environments such as coral reefs (Pinnegar et al., 2000).Predators like E. morio and M. bonaci could trigger indirect effects in the community (e.g.Heithaus et al., 2008Heithaus et al., , 2010;;Ferretti et al., 2010) and influence a large range of ecological processes (Babcock et al., 1999;Pinnegar et al., 2000;Willis, Anderson, 2003;Silveira et al., 2015), such as linkages between top and intermediate predators, and intermediate predators and their resources (Pace et al., 1999).For example, the high consumption of scarines, by both M. bonaci and E. morio, illustrate how these groupers could have a controlling role in the abundance of parrotfishes, which in turn play key roles as grazers and sand producers on coral reefs.

Fig. 1 .
Fig. 1.Map of the Abrolhos Bank showing the municipalities where samples were collected and emerging coralline reefs are shown in dark grey.MER= Marine Extractive Reserve; ER= Extractive Reserve.
e160047[5] Tab. 1. Diet composition of Epinephelus morio and Mycteroperca bonaci showing number percentage (N), frequency of occurrence (FO), weight percentage (W) and percentage of Index of Relative Importance (IRI) of food items consumed in two different size classes (Juvenile=J and Adult=A) from the Abrolhos Bank, Brazil.T= Total.