Feeding ecology of Lutjanus analis ( Teleostei : Lutjanidae ) from Abrolhos Bank , Eastern Brazil

Diet and feeding ecology of the mutton snapper Lutjanus analis were investigated in the Abrolhos Bank, Eastern Brazil, the largest and richest coral reefs in the South Atlantic, where about 270 species of reef and shore fishes occur. To evaluate seasonal and ontogenetic shifts in the diet, specimens of L. analis were obtained through a fish monitoring program in four cities in southern Bahia State, from June 2005 to March 2007. Stomachs from 85 mutton snappers that ranged in size from 18.1 to 74.0 cm TL were examined. Prey were identified to the lowest possible taxon and assessed by the frequency of occurrence and volumetric methods. Variations in volume prey consumption were evaluated using non-metric multi-dimensional scaling ordination, analysis of similarity, and similarity percentage methods. Significant differences in diet composition among size classes were registered, whereas non significant differences between seasons were observed. Considering size-classes, food items consumption showed important variations: juveniles (<34.0 cm TL) fed mostly on crustaceans, sub-adults (34.1-50.0 cm TL) showed a diversified diet and adults (>50.1 cm TL) consumed basically fish, mostly Anguiliformes. Lutjanus analis is an important generalist reef predator, with a broad array of food resources and ontogenetic changes in the diet. This snapper species plays an important role on the trophic ecology of the Abrolhos Bank coral reefs.


Introduction
The diets and trophic relationships of top predators and mesopredators are important for understanding their ecological roles.Addressing this lack of data has become more important with the increasing recognition that fish mesopredators may be important in transmitting indirect effects of top predators (e.g.Myers et al., 2007;Heithaus et al., 2008;Baum & Worm, 2009;Ferretti et al., 2010).In coral reefs this role is represented by sharks, jacks (Carangidae), groupers (Serranidae), and snappers (Lutjanidae) (Polovina, 1984;Meyer et al., 2007).
Fishes from family Lutjanidae are among the most important fishery resources in tropical and subtropical regions (Grimes, 1987;Duarte & Garcia, 1999).They are heavily exploited by commercial fisheries in Northeastern and Central Brazil since the introduction of "pargueiras" (hook and line) by Portugueses during the 1950's and 1960's as a consequence of declining stocks of tuna fishery (Resende et al., 2003).
The mutton snapper Lutjanus analis (Cuvier, 1828) is found in the tropical Western Atlantic, from Massachusetts (USA) to Southeastern Brazil, with relatively high abundances recorded in Florida, Bahamas, and the Antilles (Figueiredo, 1980;Allen, 1985;Menezes & Anderson, 2003).Sub-adults and juveniles are found in a variety of habitats, such as sand bottoms, seaweed dominated reefs, bays, mangroves, and estuaries (Allen, 1985;Claro & Lindeman, 2004).Adults are found on hard substrates, generally on offshore deep reefs and less commonly on coastal environments (Cocheret de la Morinière et al., 2003;Frédou & Ferreira, 2005).
In order to understand the role that L. analis plays in the trophic food web of the Abrolhos Bank, the present study examined the diet and feeding ecology of specimens obtained through a fish monitoring program.Emphasis was placed on the assessment of seasonal and ontogenetic shifts on feeding habits.This dietary information is critical for modeling trophic pathways and assessing the implications of predator removal from a system (Farmer & Wilson, 2010), and combined with fisheries data, information on predator/prey abundance and consumption rates of predators can be used to assess the magnitude of trophic fluxes.

Material and Methods
This study was carried out in the Abrolhos Bank, Bahia State, which includes a wide portion (42,000 km 2 ) of the continental shelf, with depth rarely exceeding 30 m and a shelf edge at about 70 m depth (16º40'-19º40'S; 39º10'-37º 20'W) (Fig. 1).The region comprises the largest and richest coral reefs in the South Atlantic, as well as an extensive mosaic of algal bottoms, mangrove forests, beaches, and vegetated sandbanks (Leão, 1999;Leão & Kikuchi, 2005).About 270 species of reef and shore fishes occur at the Abrolhos Bank (Moura & Francini-Filho, 2006).Despite its importance, with nearly 10.000 artisanal fishermen operating in the area, the region's reef fisheries are still poorly known and were not included in Frédou et al. (2006) revision of Northeastern Brazil reef fisheries.
Specimens of L. analis were obtained through a fish landing monitoring program that target fleets on hand line, longline and spear fishing in the Cities of Nova Viçosa, Caravelas, Alcobaça, and Prado (Fig. 1) between June 2005 and March 2007, together with the communities where they landed the craft caught.Morphometric data and weight were recorded to the nearest centimeter and decigrams respectively.Stomach were removed and immediately fixed in 10% formalin solution for 24 h and subsequently transferred and stored in 70% alcohol.
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., 2005) and consultation with experts.The frequency of occurrence (i.e., percentage of stomachs in which a food item occurs considering all stomachs examined) and the proportion by volume (i.e., percentage participation of each item in the total food) of a given food item were determined (Hyslop, 1980;Bowen, 1996).
For temporal analysis, the seasons were defined as follow: winter (June -August), spring (September -November), summer (December -February), and autumn (March -May).
Differences in prey consumption using quantitative data (volume prey) among seasons and sizes classes were tested with cluster based on Bray-Curtis similarities and hierarchical agglomerative group-average clustering was constructed with the PRIMER program (ver.6.0, Plymouth Marine Laboratory, Plymouth, England).Data was investigated using analysis of similarity (ANOSIM) and similarity percentage (SIMPER) methods.
A Non-metric Multi-Dimensional Scaling (NMDS) ordination was performed in order to better visualize the similarities between seasons and size classes.Samples in the NMDS plot were grouped according to the results of the cluster analysis using the single linkage method (Clarke & Warnick, 2001).An analysis of similarity percentages (SIMPER) considering the different food items was conducted to examine potential differences in the diet.This latter method helps in determining which food item is responsible for differences recorded.

Results
A total of 85 stomachs of Lutjanus analis were analyzed.The food items were grouped into four main categories: Gastropoda, Stomatopoda, Decapoda, and Fish.The category that showed the greatest diversity (i.e.Crustacea) was further refined to Stomatopoda, Decapoda Dendrobranchiata (shrimps), and Decapoda Brachyura (crabs and soft crab) (Table 1).
Crustaceans, particularly Decapods, were by far the most common and representative food item recorded, followed by Fish.Decapoda Brachyura (crabs and soft-crabs), had a great abundance in the analysis and was represented by five families.Gastropods were represented only by Fissurellacea.Stomatopods could not be identified to more refined taxonomic categories due to the high degree of digestion.Only one family (Sergestoidea) of Dendrobranchiata Decapoda (shrimps) was recorded.Fishes most commonly recorded were, in decreasing order: Anguilliformes, Perciformes, and Gobiesociformes.
The NMDS ordination and similarity analysis (ANOSIM) showed no significant differences between the seasons of the year in the diet of the species (Global R = 0.056, P =10.6%).On the other hand, the composition of the diet between the classes of length (ontogenic) showed significant differences (Global R = 0.308, P = 2.7%) (Fig. 2).Considering the main life stages (juvenile, sub-adult, and adult), some food items were determinant for the resulting ordination, such as Decapoda (Fig. 3) and Fish (Fig. 4).
The diet based mostly on pelagic crustaceans (specially Decapod) and Fish may reflect the prey availability and abundance in the environment (Wootton, 1990;Moyle & Cech, 1996;Claro & Lindeman, 2004), or are related with the Lutjanidae ability to exploit it (Claro, 1981;Heck & Weinstein, 1989;Duarte & Garcia, 1999;Claro & Lindeman, 2004;Rojas et al., 2004;Molina et al., 2005).For L. analis we believe that this ability can be related to morphological changes during the ontogenetic development (e.g.mouth gap), affecting prey size selectivity, a result already recorded for L. apodus (Roocker, 1995).In the other hand, it is important to consider that juveniles and adults of L. analis display different foraging styles, where small individuals displayed proportionally higher picking and midwater strikes during morning and evening, whereas individuals large winnowed proportionally more often than small or medium fish during evening (Mueller et al., 1994).
Despite the fact that Monteiro et al. (2009) and Pimentel & Joyeux (2010) suggested temporal changes in the diet composition of snappers in the Brazilian coast, we did not observe seasonal variation for L. analis.We believe that this variation is related with the local changes in prey availability (Randal, 1967).In fact, Monteiro et al. (2009) and Pimentel & Joyeux (2010) worked in mangrove environments, which are often subjected to more acute environmental changes than are reef faunas (Sierra et al., 2001).
Considering size-classes analyses, food items consumption showed important variations: juveniles consumed preferentially crustaceans, sub-adults showed a diversified diet, and adults feed basically on fish.Even though few stomachs of larger individuals were analyzed, similar feeding pattern was also observed for L. analis in the West Indies (Randall, 1967), in the Caribbean coast of Panama (Heck & Weinstein, 1989) in the Cuban platform (Claro, 1981;Sierra et al., 2001) and Southeastern Brazil (Pimentel & Joyeux, 2010).This adaptation probably aims to reduce the competition for food or meet physiological needs that the fish may have during its ontogenetic development in terms of migration, sexual maturation and/or reproduction (Braga & Braga, 1987;Gerking, 1994;Sierra et al., 2001).
On the other hand, ontogenetic shifts in the diet can also occur when juveniles and adults occupy different regions (Zavala-Camin, 1996;Bertoncini et al., 2003;Machado et al., 2008).Allen (1985), Lindeman et al. (1998), and Claro & Lindeman (2004) have noted two basic types of habitat selection for snappers: the juveniles of some species are usually found on shallow estuaries, while adults inhabit bays, estuaries, and reef environments of shelf waters.Starck (1970), for example, observed changes in diet of juveniles and adults of Lutjanus griseus (Linnaeus, 1758) related with the migration from estuarine areas to deeper waters.Our data generally agree with this statement, because juveniles and sub-adults of L. analis occur in estuarine and marine coastal environments (Randall, 1967;Allen, 1985;Claro & Lindeman, 2004;Pimentel & Joyeux, 2010), while adults inhabit hard substrates in offshore waters and reef environments (Cocheret de la Morinière et al., 2003;Frédou & Ferreira, 2005).
According to our data, this mesopredator plays an important role on the trophic ecology of the Abrolhos Bank coral reefs, because it is an important reef predator with a wide range of food resources.Despite classified as vulnerable in the IUCN red list of endangered species (Huntsman, 1996), L. analis is a commercially important species in the Brazilian Northeastern coast, and previous stock assessments in the study region (Klippel et al., 2005) indicated that it is moderately overfished.Impacts of chronic overfishing are evident in population depletions worldwide, and the removal of predatory fishes is also likely to have significant indirect effects on marine ecosystems (Farmer & Wilson, 2010), influencing a range of ecological processes (Babcock et al., 1999;Pinnegar et al., 2000;Willis & Anderson, 2003).

Fig. 1 .
Fig. 1.Map of the study region showing the Abrolhos Bank coral reefs, Eastern Brazil, as well as the marine protected areas and shallow (<50 m) coral reefs.

Fig. 2 .
Fig. 2. Ordination resulting from Non-metric Multi-dimensional Scaling analyses (NMDS) with cluster overlay (volume prey) among sizes classes of Lutjanus analis from Abrolhos Bank coral reefs, Eastern Brazil, between June 2005 and March 2007.

Fig. 3 .
Fig. 3. Contribution of the food item Decapoda to Lutjanus analis diet, from Abrolhos Bank coral reefs, Eastern Brazil, between June 2005 and March 2007.The larger the bubble, the greater the percent participation of volume.J = juveniles; SA = sub-adults; A = adults.

Fig. 4 .
Fig. 4. Contribution of the food item Fish to the Lutjanus analis diet, from Abrolhos Bank coral reefs, Eastern Brazil, between June 2005 and March 2007.The larger the bubble, the greater the percent participation of volume.J = juveniles; SA = subadults; A = adults.

Table 1 .
Diet composition showing frequency of occurrence (% FO) and relative volume (V%) of the food items consumed by Lutjanus analis from Abrolhos Bank coral reefs, Eastern Brazil, between June 2005 and March 2007.

Table 2 .
SIMPER analysis of similarity between classes of length of the Lutjanus analis, captured in the Abrolhos Bank coral reefs, Eastern Brazil, between June 2005 and March 2007.(Crust, Den, Por, Xan, corresponds respectively to Crustacea remains, Dendrobranchiata, Portunidae, and Xanthidae).