Journal of Fish Biology (2014) doi:10.1111/jfb.12499, available online at wileyonlinelibrary.com
Trophic niche and habitat shifts of sympatric Gerreidae J. A. A. Ramos* † ‡, M. Barletta† ‡ §, D. V. Dantas†‡, A. R. A. Lima†‡ and M. F. Costa†‡ *Instituto Federal de Educação, Ciência e Tecnologia da Paraíba (IFPB)-Campus Cabedelo, Rua Santa Rita de Cássia, s/n, Jardim Jericó, Cabedelo, Paraíba, CEP 58310-000, Brazil, †Laboratório de Ecologia e Gerenciamento de Ecossistemas Costeiros e Estuarinos, Departamento de Oceanografia, Universidade Federal de Pernambuco, Cidade Universitária, Recife, Pernambuco, CEP 50740-550, Brazil and ‡Instituto de Ecologia e Gerenciamento de Ecossistemas Costeiros e Estuarinos (IEGEA), P. O. Box 8132, Recife, Pernambuco, CEP 51020-970, Brazil (Received 15 January 2014, Accepted 9 July 2014) The diet and mouth growth rates of three Gerreidae species (Eugerres brasilianus, Eucinostomus melanopterus and Diapterus rhombeus) were assessed at different ontogenetic phases (juveniles, sub-adults and adults) in order to detect allometric growth, and whether they are related to habitat and seasonal changes in the Goiana Estuary, north-east Brazil. The importance of each prey for each ontogenetic phase was described using the index of relative importance. The three species showed seasonal ontogenetic shifts in diet and allometric growth of mouth morphology. They also had an exclusively zoobenthic diet, comprising mainly Polychaeta, Copepoda, Ostracoda, Gastropoda and Bivalvia. Mouth development showed a possible influence on diet changes for E. melanopterus. Significant interactions (P < 0⋅01) were detected among seasons, areas and ontogenetic phases for the most important prey for E. brasilianus and E. melanopterus. Diet overlaps are evidence of intra and interspecific competition among gerreids for specific prey. A conceptual model of the competition and seasonal diet shifts among ontogenetic phases of gerreids is given. The sediment ingested due to the feeding mechanisms of Gerreidae species could also partially explain the ingestion of synthetic items observed for all ontogenetic phases, which indicates one of a myriad effects of human activities (e.g. artisanal fishery) in this estuary. © 2014 The Fisheries Society of the British Isles
Key words: allometric growth; Brazil; conceptual model of feeding; diet shift; Goiana Estuary; South America.
INTRODUCTION Estuaries are ecosystems with high productivity and biological diversity that provide protection, reproduction and feeding grounds for several species of fishes, which may be resident or spend only part of their life cycles in this ecosystem (Barletta-Bergan et al., 2002a; Barletta et al., 2005, 2008; Dantas et al., 2012a). Within estuaries, mangrove forests have an important ecological role, providing the fauna with a high diversity of habitats (e.g. intertidal mangrove creeks, tidal pools, mangrove roots and crab holes) and other resources (Barletta et al., 2000, 2003; Barletta-Bergan et al., 2002b; §Author to whom correspondence should be addressed. Tel.: +55 81 2126 7223; email: [email protected]
1 © 2014 The Fisheries Society of the British Isles
2
J . A . A . R A M O S E T A L.
Krumme et al., 2005; Ramos et al., 2011). In addition, many species that occur in these areas are economically important, which confers an important argument for the conservation of these areas (Lenanton & Potter, 1987; Barletta & Costa, 2009; Barletta et al., 2010). The fishes of the family Gerreidae are an important marine resource for commercial and artisanal fisheries (Austin, 1971; Barletta & Costa, 2009). Their main characteristics are their highly protrusible mouth and their occurrence in coastal areas, seagrass beds, intertidal mangrove creeks and estuaries (Cyrus & Blaber, 1982; Carpenter, 2002). Many studies of fish ecology have reported their occurrence in coastal areas around the world, such as the Embley Estuary, north Australia (Barletta & Blaber, 2007), their wide distribution and abundance in the Kosi Estuary system, South Africa (Cyrus & Blaber, 1982), in the Sine-Saloum Estuary in Senegal (Gning et al., 2008) and along the western coast of the tropical Atlantic Ocean, as e.g. in the Caeté Estuary, eastern Amazon (Barletta & Blaber, 2007), the Mandaú/Manguaba Lagoon system (Teixeira & Helmer, 1997), the Goiana Estuary, north-east Brazil (Ramos et al., 2011) and in Sepetiba Bay, south-eastern Brazil (Pessanha & Araujo, 2012). In the Goiana Estuary, species of the family Gerreidae have mean landings of c. 15⋅1 t year−1 , representing a subsistence resource for the local community (Barletta & Costa, 2009). This information reinforces the importance of estuarine habitats for the life cycle of this family, and their importance for the fishery production from this estuarine region, where their conservation status are under threat (e.g. plastic debris) from human activity (Possatto et al., 2011; Dantas et al., 2012b; Ramos et al., 2012; Lima et al., 2014). Knowledge about the feeding ecology of fishes is extremely important for understanding community dynamics, to improve management of stocks and support estuarine conservation (Blaber, 2000). The diet of the same fish species tends to change with growth (Kerschner et al., 1985; Gning et al., 2008; Pessanha & Araujo, 2012) and hence the resulting seasonal and spatial habitat utilization (Dantas et al., 2013). These changes result from needs to meet nutritional requirements (Gning et al., 2008), variation in morphology (Wainwright & Richard, 1995) and strategies to use partitioned resources, in order to avoid intra and interspecific competition (Wootton, 1998). To understand the role of feeding ecology for fish species, it is important to know how they use resources throughout their life cycle by analysing feeding habits at different ontogenetic phases. Studies of the feeding ecology of Gerreidae species exist for the tropical and sub-tropical waters in Puerto Rico (Austin, 1971), South Africa (Cyrus & Blaber, 1983), north-east Brazil (Teixeira & Helmer, 1997) and southern Brazil (Chaves & Otto, 1998), but studies focusing on Gerreidae diet shifts with ontogenetic development are rare (Kerschner et al., 1985; Gning et al., 2008; Pessanha & Araujo, 2012), and there are no previous reports relating these changes to mouth structure changes for species from the western Atlantic Ocean. With this background, this study examined the feeding habits and the allometric growth of mouth structure of sympatric Gerreidae species [Eugerres brasilianus (Cuvier 1830), Eucinostomus melanopterus (Bleeker 1863) and Diapterus rhombeus (Cuvier 1829)] in the Goiana Estuary. The working hypotheses tested were: (1) there are differences in allometric growth of the mouth among ontogenetic phases of gerreids and they have a relationship with ontogenetic diet shifts, (2) feeding habits of different ontogenetic phases of gerreids change with habitat and season and (3) intra and interspecific competition for food resources occurs among ontogenetic phases of gerreids when using the same estuarine habitats at the same time.
© 2014 The Fisheries Society of the British Isles, Journal of Fish Biology 2014, doi:10.1111/jfb.12499
3
F E E D I N G E C O L O G Y O F S Y M PAT R I C G E R R E I D A E
Equator BRAZIL
07° 31′ 00′′ S
N
Atlantic Ocean 0
1000 2000 km
2
5 km
Main channel
34° 47′ 00′′ W
36°00′ 00′′ W
3
Atlantic Ocean
1
07° 36′ 00′′ S
Fig. 1. Goiana Estuary. Red line marks the Goiana Estuary main channel, and dashed squares limit the upper (1), middle (2) and lower (3) reaches. Image source: Google Earth.
MATERIALS AND METHODS S T U DY A R E A The Goiana Estuary (7∘ 32′ –7∘ 35′ S; 34∘ 54′ –34∘ 58′ W) (Fig. 1) has an area of 4200 ha and multiple habitat types, including the main channel, floodplain, mangrove intertidal creeks and mangrove forest (Barletta & Costa, 2009; Ramos et al., 2011). There are four well-defined seasons, differing not by temperature, but by rainfall patterns: early rainy (March to May), late rainy (June to August), early dry (September to November) and late dry (December to February). The main channel is divided into three different habitats (upper, middle and lower estuary) based on the salinity gradient (Barletta & Costa, 2009). The estuary provides subsistence to some communities through the exploitation of finfishes, crustaceans, shellfish and mangrove forest wood (Barletta & Costa, 2009). FISH SAMPLING Samples were taken monthly from the main channel of the Goiana Estuary from January 2006 to March 2009 with an otter trawl. The net had 35 mm mesh size in the body and 22 mm in the codend (Dantas et al., 2010). In addition, 12 samples were collected in mangrove intertidal creeks at different reaches of the Goiana Estuary from April to May 2008, using a fyke net with 1 cm mesh size (Ramos et al., 2011). For each sample, salinity, dissolved oxygen (mg l−1 ) and water temperature (∘ C) were recorded. After each sample was taken, the fishes captured were preserved on ice and subsequently frozen. After transfer to the laboratory, all individuals were identified to species level (Menezes & Figueiredo, 1980; Carpenter, 2002). The total length (LT , cm) and total body mass (g) of each individual were measured. Stomachs were removed and analysed. To evaluate ontogenetic changes, fishes were allocated to three different size classes (juvenile, sub-adult and adult) (Table I), following the classification proposed by Ramos et al. (2012).
© 2014 The Fisheries Society of the British Isles, Journal of Fish Biology 2014, doi:10.1111/jfb.12499
MW MH SL
MW MH SL
MW MH SL
Variables
𝛽1
Juvenile (9⋅54 cm) −2⋅23 1⋅81 ± 0⋅12 −1⋅30 1⋅04 ± 0⋅11 −1⋅26 1⋅26 ± 0⋅08
𝛽0
0⋅85 0⋅72 0⋅87
0⋅80 0⋅37 0⋅31
0⋅93 0⋅89 0⋅86
r2
-
F 2,133 = 7⋅604** F 2,133 = 19⋅913** F 2,133 = 38⋅395**
F 2,354 = 3⋅005NS F 2,354 = 2⋅318NS F 2,354 = 4⋅286*
F-test J×S×A
Table I. Regression coefficients (±s.d.) for mouth variables (y; M W , mouth width; M H , mouth high; SL , snout length) in relation to total length (LT ) in each ontogenetic phase of species studied and summary of F-test results for 𝛽 1 comparison between ontogenetic phases (J, juvenile; S, sub-adult; A, 𝛽 adult). 𝛽 0 is the intercept, 𝛽 1 is the growth coefficient, r2 is the coefficient of determination, model: y = 𝛽0 LT1
4 J . A . A . R A M O S E T A L.
© 2014 The Fisheries Society of the British Isles, Journal of Fish Biology 2014, doi:10.1111/jfb.12499
F E E D I N G E C O L O G Y O F S Y M PAT R I C G E R R E I D A E
(a)
MW
5
(b)
SL
MH
Fig. 2. Measurements of mouth morphology of Gerreidae: (a) mouth height (M H ) and mouth width (M W ), and (b) snout length (SL ).
M O R P H O L O G I C A L M E A S U R E M E N T S A N D A N A LY S I S A Vernier calliper was used to measure each individual (nearest 0⋅1 mm): LT , mouth width (M W ), mouth height (M H ) and snout length (SL ) (i.e. from the anterior edge of the lips of the upper jaws to the anterior edge of the eye with the mouth fully open) (Fig. 2). The allometric growth of each variable was calculated for each size class as a power function of the LT (Huxley, 𝛽 1924) using the following equation: Y = 𝛽0 LT1 + 𝜖, where Y is the response variable (e.g. M W , M H and SL ), 𝛽 0 is the intercept, 𝛽 1 is the growth coefficient and 𝜖 is the error. The growth coefficient was classified as isometric (when 𝛽 1 = 1), allometric positive (𝛽 1 > 1) or allometric negative (𝛽 1 < 1) (van Snik et al., 1997). S T O M A C H C O N T E N T A N A LY S I S Stomach contents of each fish from the different size classes (juvenile, sub-adult and adult), estuary area (upper, middle and lower) and season (early and late dry seasons; early and late rainy seasons) were examined under stereomicroscope (×45 magnification). Each prey item was identified to the lowest possible taxonomic level, with support from the literature (Boltovskoy, 1999; Young & Sewell, 2002). Prey items were washed with distilled water, dried with tissue paper, counted and weighed (±0⋅001 g). To assess the importance of each prey, the index of relative importance (I RI ) proposed by Pinkas et al. (1971) was used after being calculated from the equation: I RI = % F i (% N i + % M i ), where %Fi (frequency of occurrence) is the percentage of stomachs which contained the item i, %Ni (composition in number) is the number of item i expressed as a percentage of the total number of prey in all stomachs examined, %Mi (composition in mass) is the mass of item i expressed as a percentage of the total mass of prey in all stomachs examined (Hyslop, 1980). For algal fragments, a modified I RI proposed by Azzurro et al. (2007) was used with the equation: I RI = % F i % M i . The I RI was expressed as a percentage (%I RI ) of the total I RI contributed by each prey (Cortés, 1997). To analyse the niche breadth, Hill’s index of diversity (I N ) was computed for each size class of each species using the equation: IN = e[−Σ(Pi ln Pi )] , where Pi is the proportion of prey item i among the total number of
© 2014 The Fisheries Society of the British Isles, Journal of Fish Biology 2014, doi:10.1111/jfb.12499
6
J . A . A . R A M O S E T A L.
preys (Ludwig & Reynolds, 1988). In addition, the evenness index (I E ) was calculated to measure how evenly fish rely on food resources using the equation: I E = I N N 0 − 1 , where N 0 is the number of species of prey ingested (Ludwig & Reynolds, 1988). Moreover, Schoener’s index (C) was calculated to evaluate the degree of overlap ∑ in the diets among different size classes of both species, using the equation: C = 1 − 0 ⋅ 5( |N xi − N yi| ), where Nxi is the mean proportion of the number of prey i ingested by a specific size classes of a gerreid species x and Nyi is the mean proportion of the number of prey i ingested by another size class and gerreid species y (Schoener, 1970).
S TAT I S T I C A L A N A LY S I S The F-test was used to check if growth of the mouth structure differs significantly between size classes of each species (Sokal & Rohlf, 1995). Regarding diet analysis, rare prey (%I RI < 4) were excluded to ensure robustness. All diet data were transformed (Box–Cox) to increase normality (Box & Cox, 1964). A factorial analysis of variance (ANOVA) was used to compare prey composition (number and mass of prey), I N and I E among the factors: season, area of estuary and size classes (Zar, 1996). Whenever significant differences were detected by ANOVA, the Bonferroni test was used a posteriori to detect the source of variance (Quinn & Keough, 2002). Canonical correspondence analysis (CCA) was performed to determine the influence of environmental variables (salinity, rainfall, water temperature and dissolved oxygen) on ontogenetic diet shifts (using %I RI of important prey as dependent variable) of E. brasilianus and E. melanopterus. Diapterus rhombeus was not included in this analysis due to the low number of individuals captured resulting from its low occurrence in the habitats studied. Moreover, CCA was used to examine the relationship between mouth growth (using residual values of morphometric data) and prey ingestion (using %I RI ). For this analysis, morphometric data had body size effects removed by extracting size residuals from the linear regression of each variable (M H , M W and SL ) against LT , and using the log10 -transformed values. This procedure avoids computational biases when using morphometric data in multivariate analysis (Davis et al., 2012). A Monte-Carlo permutation test was used to determine which axis and environmental variables were significant to the variability of the dependent variable (ter Braak & Smilauer, 2002). All analyses were carried out using a significance level of 0⋅05.
RESULTS A total of 543 individuals were analysed: 360 E. brasilianus (211 juvenile, 100 sub-adult and 49 adults, LT from 3⋅1 to 40⋅2 cm), 141 E. melanopterus (41 juveniles, 56 sub-adults and 44 adults, LT from 1⋅23 to 15⋅2 cm) and 44 D. rhombeus (four juvenile individuals, eight sub-adults and 37 adults, LT from 5⋅28 to 15⋅4 cm) (Fig. 3). M O U T H G R O W T H R AT E
The allometric growth of M W , M H and SL was described and compared among size classes for each species (Table I). For E. brasilianus, only SL growth differed significantly among size classes (P < 0⋅05), growing faster (𝛽 1 = 1⋅304) during the adult phase [Table I and Fig. 3(c)]. For E. melanopterus, the growth pattern of M W , M H and SL differed significantly among size classes (P < 0⋅01), with the fastest growth observed in juveniles (allometric positive; 𝛽 1 > 1) [Table I and Fig. 3(d)–(f)]. DIET COMPOSITION AND ONTOGENETIC SHIFTS
From 543 Gerreidae specimens analysed, 13 individuals had completely empty stomachs (one juvenile of E. melanopterus; two sub-adults of D. rhombeus; nine juveniles
© 2014 The Fisheries Society of the British Isles, Journal of Fish Biology 2014, doi:10.1111/jfb.12499
7
F E E D I N G E C O L O G Y O F S Y M PAT R I C G E R R E I D A E
(a)
(d) 3
MW
n = 360
2
1
0·6
0·8
n = 141
0
(b) 3 2
0·4
0·2
0·2
0
0
(e)
(h)
0·8 n = 360
1
0·6
0·8 0·6
n = 141
0·4 1 0
(c) 6 n = 360
SL
4 2 0
0 4 8 12 16 20 24 28 32 36 40 44
n = 44
0·4
0·2
0·2
0
0
(f)
(i)
2
2
1·5
1·5
n = 141
1
1
0·5
0·5
0
n = 44
0·6
0·4 1
MH
(g)
0·8
0
2
4
6
8 10 12 14 16 LT (cm)
0
n = 44
0
2
4
6
8 10 12 14 16 18
Fig. 3. Relation of (a, d, g) mouth width (M W ), (b, e, h) mouth height (M H ) and (c, f, i) snout length (SL ) with the total length (LT ) of juveniles ( ), sub-adults ( ) and adults ( ) of (a, b, c) Eugerres brasilianus, (d, e, f) Eucinostomus melanopterus and (g, h, i) Diapterus rhombeus. , phase changes (see Table I).
and one adult of E. brasilianus). A total of 30 different prey taxa were recorded and the overall diet consisted a variety of benthic prey such as Polychaeta, Copepoda, Ostracoda, Bivalvia and algal fragments (Table II). Moreover, the importance of these prey varied according to fish size. A complete overview of the diet of Gerreidae species is given in Tables SI–SIII (Supporting Information). The preferred prey of juvenile size classes of E. brasilianus were Polychaeta and Asterope sp., Copepoda and Nereis sp. for E. melanopterus and tentacles of Terebellidae for D. rhombeus (Table II). When reaching sub-adult sizes, E. brasilianus reduced the ingestion of Terebellidae tentacles and fed preferentially on Nereis sp. and Asterope sp., E. melanopterus and D. rhombeus decreased their consumption of Polychaeta and increased capture of Pseudodiaptomus spp. and algal fragments (Table II). In the adult size class, Anomalocardia brasiliana and Littorinidae became the most important prey for E. brasilianus, while Nereis sp. followed by algal fragments for E. melanopterus and Longipedia spp. was the predominant prey in the diet for D. rhombeus (Table II). The presence of non-food items (microplastics) was also noted in the stomach contents (Table II). Plastic fragments were observed in the stomach of the three Gerreidae species (Table II). It was absent only in the juvenile size class of D. rhombeus. The ingestion of plastic fragments represented
Trophic niche and habitat shifts of sympatric Gerreidae J. A. A. Ramos* † ‡, M. Barletta† ‡ §, D. V. Dantas†‡, A. R. A. Lima†‡ and M. F. Costa†‡ *Instituto Federal de Educação, Ciência e Tecnologia da Paraíba (IFPB)-Campus Cabedelo, Rua Santa Rita de Cássia, s/n, Jardim Jericó, Cabedelo, Paraíba, CEP 58310-000, Brazil, †Laboratório de Ecologia e Gerenciamento de Ecossistemas Costeiros e Estuarinos, Departamento de Oceanografia, Universidade Federal de Pernambuco, Cidade Universitária, Recife, Pernambuco, CEP 50740-550, Brazil and ‡Instituto de Ecologia e Gerenciamento de Ecossistemas Costeiros e Estuarinos (IEGEA), P. O. Box 8132, Recife, Pernambuco, CEP 51020-970, Brazil (Received 15 January 2014, Accepted 9 July 2014) The diet and mouth growth rates of three Gerreidae species (Eugerres brasilianus, Eucinostomus melanopterus and Diapterus rhombeus) were assessed at different ontogenetic phases (juveniles, sub-adults and adults) in order to detect allometric growth, and whether they are related to habitat and seasonal changes in the Goiana Estuary, north-east Brazil. The importance of each prey for each ontogenetic phase was described using the index of relative importance. The three species showed seasonal ontogenetic shifts in diet and allometric growth of mouth morphology. They also had an exclusively zoobenthic diet, comprising mainly Polychaeta, Copepoda, Ostracoda, Gastropoda and Bivalvia. Mouth development showed a possible influence on diet changes for E. melanopterus. Significant interactions (P < 0⋅01) were detected among seasons, areas and ontogenetic phases for the most important prey for E. brasilianus and E. melanopterus. Diet overlaps are evidence of intra and interspecific competition among gerreids for specific prey. A conceptual model of the competition and seasonal diet shifts among ontogenetic phases of gerreids is given. The sediment ingested due to the feeding mechanisms of Gerreidae species could also partially explain the ingestion of synthetic items observed for all ontogenetic phases, which indicates one of a myriad effects of human activities (e.g. artisanal fishery) in this estuary. © 2014 The Fisheries Society of the British Isles
Key words: allometric growth; Brazil; conceptual model of feeding; diet shift; Goiana Estuary; South America.
INTRODUCTION Estuaries are ecosystems with high productivity and biological diversity that provide protection, reproduction and feeding grounds for several species of fishes, which may be resident or spend only part of their life cycles in this ecosystem (Barletta-Bergan et al., 2002a; Barletta et al., 2005, 2008; Dantas et al., 2012a). Within estuaries, mangrove forests have an important ecological role, providing the fauna with a high diversity of habitats (e.g. intertidal mangrove creeks, tidal pools, mangrove roots and crab holes) and other resources (Barletta et al., 2000, 2003; Barletta-Bergan et al., 2002b; §Author to whom correspondence should be addressed. Tel.: +55 81 2126 7223; email: [email protected]
1 © 2014 The Fisheries Society of the British Isles
2
J . A . A . R A M O S E T A L.
Krumme et al., 2005; Ramos et al., 2011). In addition, many species that occur in these areas are economically important, which confers an important argument for the conservation of these areas (Lenanton & Potter, 1987; Barletta & Costa, 2009; Barletta et al., 2010). The fishes of the family Gerreidae are an important marine resource for commercial and artisanal fisheries (Austin, 1971; Barletta & Costa, 2009). Their main characteristics are their highly protrusible mouth and their occurrence in coastal areas, seagrass beds, intertidal mangrove creeks and estuaries (Cyrus & Blaber, 1982; Carpenter, 2002). Many studies of fish ecology have reported their occurrence in coastal areas around the world, such as the Embley Estuary, north Australia (Barletta & Blaber, 2007), their wide distribution and abundance in the Kosi Estuary system, South Africa (Cyrus & Blaber, 1982), in the Sine-Saloum Estuary in Senegal (Gning et al., 2008) and along the western coast of the tropical Atlantic Ocean, as e.g. in the Caeté Estuary, eastern Amazon (Barletta & Blaber, 2007), the Mandaú/Manguaba Lagoon system (Teixeira & Helmer, 1997), the Goiana Estuary, north-east Brazil (Ramos et al., 2011) and in Sepetiba Bay, south-eastern Brazil (Pessanha & Araujo, 2012). In the Goiana Estuary, species of the family Gerreidae have mean landings of c. 15⋅1 t year−1 , representing a subsistence resource for the local community (Barletta & Costa, 2009). This information reinforces the importance of estuarine habitats for the life cycle of this family, and their importance for the fishery production from this estuarine region, where their conservation status are under threat (e.g. plastic debris) from human activity (Possatto et al., 2011; Dantas et al., 2012b; Ramos et al., 2012; Lima et al., 2014). Knowledge about the feeding ecology of fishes is extremely important for understanding community dynamics, to improve management of stocks and support estuarine conservation (Blaber, 2000). The diet of the same fish species tends to change with growth (Kerschner et al., 1985; Gning et al., 2008; Pessanha & Araujo, 2012) and hence the resulting seasonal and spatial habitat utilization (Dantas et al., 2013). These changes result from needs to meet nutritional requirements (Gning et al., 2008), variation in morphology (Wainwright & Richard, 1995) and strategies to use partitioned resources, in order to avoid intra and interspecific competition (Wootton, 1998). To understand the role of feeding ecology for fish species, it is important to know how they use resources throughout their life cycle by analysing feeding habits at different ontogenetic phases. Studies of the feeding ecology of Gerreidae species exist for the tropical and sub-tropical waters in Puerto Rico (Austin, 1971), South Africa (Cyrus & Blaber, 1983), north-east Brazil (Teixeira & Helmer, 1997) and southern Brazil (Chaves & Otto, 1998), but studies focusing on Gerreidae diet shifts with ontogenetic development are rare (Kerschner et al., 1985; Gning et al., 2008; Pessanha & Araujo, 2012), and there are no previous reports relating these changes to mouth structure changes for species from the western Atlantic Ocean. With this background, this study examined the feeding habits and the allometric growth of mouth structure of sympatric Gerreidae species [Eugerres brasilianus (Cuvier 1830), Eucinostomus melanopterus (Bleeker 1863) and Diapterus rhombeus (Cuvier 1829)] in the Goiana Estuary. The working hypotheses tested were: (1) there are differences in allometric growth of the mouth among ontogenetic phases of gerreids and they have a relationship with ontogenetic diet shifts, (2) feeding habits of different ontogenetic phases of gerreids change with habitat and season and (3) intra and interspecific competition for food resources occurs among ontogenetic phases of gerreids when using the same estuarine habitats at the same time.
© 2014 The Fisheries Society of the British Isles, Journal of Fish Biology 2014, doi:10.1111/jfb.12499
3
F E E D I N G E C O L O G Y O F S Y M PAT R I C G E R R E I D A E
Equator BRAZIL
07° 31′ 00′′ S
N
Atlantic Ocean 0
1000 2000 km
2
5 km
Main channel
34° 47′ 00′′ W
36°00′ 00′′ W
3
Atlantic Ocean
1
07° 36′ 00′′ S
Fig. 1. Goiana Estuary. Red line marks the Goiana Estuary main channel, and dashed squares limit the upper (1), middle (2) and lower (3) reaches. Image source: Google Earth.
MATERIALS AND METHODS S T U DY A R E A The Goiana Estuary (7∘ 32′ –7∘ 35′ S; 34∘ 54′ –34∘ 58′ W) (Fig. 1) has an area of 4200 ha and multiple habitat types, including the main channel, floodplain, mangrove intertidal creeks and mangrove forest (Barletta & Costa, 2009; Ramos et al., 2011). There are four well-defined seasons, differing not by temperature, but by rainfall patterns: early rainy (March to May), late rainy (June to August), early dry (September to November) and late dry (December to February). The main channel is divided into three different habitats (upper, middle and lower estuary) based on the salinity gradient (Barletta & Costa, 2009). The estuary provides subsistence to some communities through the exploitation of finfishes, crustaceans, shellfish and mangrove forest wood (Barletta & Costa, 2009). FISH SAMPLING Samples were taken monthly from the main channel of the Goiana Estuary from January 2006 to March 2009 with an otter trawl. The net had 35 mm mesh size in the body and 22 mm in the codend (Dantas et al., 2010). In addition, 12 samples were collected in mangrove intertidal creeks at different reaches of the Goiana Estuary from April to May 2008, using a fyke net with 1 cm mesh size (Ramos et al., 2011). For each sample, salinity, dissolved oxygen (mg l−1 ) and water temperature (∘ C) were recorded. After each sample was taken, the fishes captured were preserved on ice and subsequently frozen. After transfer to the laboratory, all individuals were identified to species level (Menezes & Figueiredo, 1980; Carpenter, 2002). The total length (LT , cm) and total body mass (g) of each individual were measured. Stomachs were removed and analysed. To evaluate ontogenetic changes, fishes were allocated to three different size classes (juvenile, sub-adult and adult) (Table I), following the classification proposed by Ramos et al. (2012).
© 2014 The Fisheries Society of the British Isles, Journal of Fish Biology 2014, doi:10.1111/jfb.12499
MW MH SL
MW MH SL
MW MH SL
Variables
𝛽1
Juvenile (9⋅54 cm) −2⋅23 1⋅81 ± 0⋅12 −1⋅30 1⋅04 ± 0⋅11 −1⋅26 1⋅26 ± 0⋅08
𝛽0
0⋅85 0⋅72 0⋅87
0⋅80 0⋅37 0⋅31
0⋅93 0⋅89 0⋅86
r2
-
F 2,133 = 7⋅604** F 2,133 = 19⋅913** F 2,133 = 38⋅395**
F 2,354 = 3⋅005NS F 2,354 = 2⋅318NS F 2,354 = 4⋅286*
F-test J×S×A
Table I. Regression coefficients (±s.d.) for mouth variables (y; M W , mouth width; M H , mouth high; SL , snout length) in relation to total length (LT ) in each ontogenetic phase of species studied and summary of F-test results for 𝛽 1 comparison between ontogenetic phases (J, juvenile; S, sub-adult; A, 𝛽 adult). 𝛽 0 is the intercept, 𝛽 1 is the growth coefficient, r2 is the coefficient of determination, model: y = 𝛽0 LT1
4 J . A . A . R A M O S E T A L.
© 2014 The Fisheries Society of the British Isles, Journal of Fish Biology 2014, doi:10.1111/jfb.12499
F E E D I N G E C O L O G Y O F S Y M PAT R I C G E R R E I D A E
(a)
MW
5
(b)
SL
MH
Fig. 2. Measurements of mouth morphology of Gerreidae: (a) mouth height (M H ) and mouth width (M W ), and (b) snout length (SL ).
M O R P H O L O G I C A L M E A S U R E M E N T S A N D A N A LY S I S A Vernier calliper was used to measure each individual (nearest 0⋅1 mm): LT , mouth width (M W ), mouth height (M H ) and snout length (SL ) (i.e. from the anterior edge of the lips of the upper jaws to the anterior edge of the eye with the mouth fully open) (Fig. 2). The allometric growth of each variable was calculated for each size class as a power function of the LT (Huxley, 𝛽 1924) using the following equation: Y = 𝛽0 LT1 + 𝜖, where Y is the response variable (e.g. M W , M H and SL ), 𝛽 0 is the intercept, 𝛽 1 is the growth coefficient and 𝜖 is the error. The growth coefficient was classified as isometric (when 𝛽 1 = 1), allometric positive (𝛽 1 > 1) or allometric negative (𝛽 1 < 1) (van Snik et al., 1997). S T O M A C H C O N T E N T A N A LY S I S Stomach contents of each fish from the different size classes (juvenile, sub-adult and adult), estuary area (upper, middle and lower) and season (early and late dry seasons; early and late rainy seasons) were examined under stereomicroscope (×45 magnification). Each prey item was identified to the lowest possible taxonomic level, with support from the literature (Boltovskoy, 1999; Young & Sewell, 2002). Prey items were washed with distilled water, dried with tissue paper, counted and weighed (±0⋅001 g). To assess the importance of each prey, the index of relative importance (I RI ) proposed by Pinkas et al. (1971) was used after being calculated from the equation: I RI = % F i (% N i + % M i ), where %Fi (frequency of occurrence) is the percentage of stomachs which contained the item i, %Ni (composition in number) is the number of item i expressed as a percentage of the total number of prey in all stomachs examined, %Mi (composition in mass) is the mass of item i expressed as a percentage of the total mass of prey in all stomachs examined (Hyslop, 1980). For algal fragments, a modified I RI proposed by Azzurro et al. (2007) was used with the equation: I RI = % F i % M i . The I RI was expressed as a percentage (%I RI ) of the total I RI contributed by each prey (Cortés, 1997). To analyse the niche breadth, Hill’s index of diversity (I N ) was computed for each size class of each species using the equation: IN = e[−Σ(Pi ln Pi )] , where Pi is the proportion of prey item i among the total number of
© 2014 The Fisheries Society of the British Isles, Journal of Fish Biology 2014, doi:10.1111/jfb.12499
6
J . A . A . R A M O S E T A L.
preys (Ludwig & Reynolds, 1988). In addition, the evenness index (I E ) was calculated to measure how evenly fish rely on food resources using the equation: I E = I N N 0 − 1 , where N 0 is the number of species of prey ingested (Ludwig & Reynolds, 1988). Moreover, Schoener’s index (C) was calculated to evaluate the degree of overlap ∑ in the diets among different size classes of both species, using the equation: C = 1 − 0 ⋅ 5( |N xi − N yi| ), where Nxi is the mean proportion of the number of prey i ingested by a specific size classes of a gerreid species x and Nyi is the mean proportion of the number of prey i ingested by another size class and gerreid species y (Schoener, 1970).
S TAT I S T I C A L A N A LY S I S The F-test was used to check if growth of the mouth structure differs significantly between size classes of each species (Sokal & Rohlf, 1995). Regarding diet analysis, rare prey (%I RI < 4) were excluded to ensure robustness. All diet data were transformed (Box–Cox) to increase normality (Box & Cox, 1964). A factorial analysis of variance (ANOVA) was used to compare prey composition (number and mass of prey), I N and I E among the factors: season, area of estuary and size classes (Zar, 1996). Whenever significant differences were detected by ANOVA, the Bonferroni test was used a posteriori to detect the source of variance (Quinn & Keough, 2002). Canonical correspondence analysis (CCA) was performed to determine the influence of environmental variables (salinity, rainfall, water temperature and dissolved oxygen) on ontogenetic diet shifts (using %I RI of important prey as dependent variable) of E. brasilianus and E. melanopterus. Diapterus rhombeus was not included in this analysis due to the low number of individuals captured resulting from its low occurrence in the habitats studied. Moreover, CCA was used to examine the relationship between mouth growth (using residual values of morphometric data) and prey ingestion (using %I RI ). For this analysis, morphometric data had body size effects removed by extracting size residuals from the linear regression of each variable (M H , M W and SL ) against LT , and using the log10 -transformed values. This procedure avoids computational biases when using morphometric data in multivariate analysis (Davis et al., 2012). A Monte-Carlo permutation test was used to determine which axis and environmental variables were significant to the variability of the dependent variable (ter Braak & Smilauer, 2002). All analyses were carried out using a significance level of 0⋅05.
RESULTS A total of 543 individuals were analysed: 360 E. brasilianus (211 juvenile, 100 sub-adult and 49 adults, LT from 3⋅1 to 40⋅2 cm), 141 E. melanopterus (41 juveniles, 56 sub-adults and 44 adults, LT from 1⋅23 to 15⋅2 cm) and 44 D. rhombeus (four juvenile individuals, eight sub-adults and 37 adults, LT from 5⋅28 to 15⋅4 cm) (Fig. 3). M O U T H G R O W T H R AT E
The allometric growth of M W , M H and SL was described and compared among size classes for each species (Table I). For E. brasilianus, only SL growth differed significantly among size classes (P < 0⋅05), growing faster (𝛽 1 = 1⋅304) during the adult phase [Table I and Fig. 3(c)]. For E. melanopterus, the growth pattern of M W , M H and SL differed significantly among size classes (P < 0⋅01), with the fastest growth observed in juveniles (allometric positive; 𝛽 1 > 1) [Table I and Fig. 3(d)–(f)]. DIET COMPOSITION AND ONTOGENETIC SHIFTS
From 543 Gerreidae specimens analysed, 13 individuals had completely empty stomachs (one juvenile of E. melanopterus; two sub-adults of D. rhombeus; nine juveniles
© 2014 The Fisheries Society of the British Isles, Journal of Fish Biology 2014, doi:10.1111/jfb.12499
7
F E E D I N G E C O L O G Y O F S Y M PAT R I C G E R R E I D A E
(a)
(d) 3
MW
n = 360
2
1
0·6
0·8
n = 141
0
(b) 3 2
0·4
0·2
0·2
0
0
(e)
(h)
0·8 n = 360
1
0·6
0·8 0·6
n = 141
0·4 1 0
(c) 6 n = 360
SL
4 2 0
0 4 8 12 16 20 24 28 32 36 40 44
n = 44
0·4
0·2
0·2
0
0
(f)
(i)
2
2
1·5
1·5
n = 141
1
1
0·5
0·5
0
n = 44
0·6
0·4 1
MH
(g)
0·8
0
2
4
6
8 10 12 14 16 LT (cm)
0
n = 44
0
2
4
6
8 10 12 14 16 18
Fig. 3. Relation of (a, d, g) mouth width (M W ), (b, e, h) mouth height (M H ) and (c, f, i) snout length (SL ) with the total length (LT ) of juveniles ( ), sub-adults ( ) and adults ( ) of (a, b, c) Eugerres brasilianus, (d, e, f) Eucinostomus melanopterus and (g, h, i) Diapterus rhombeus. , phase changes (see Table I).
and one adult of E. brasilianus). A total of 30 different prey taxa were recorded and the overall diet consisted a variety of benthic prey such as Polychaeta, Copepoda, Ostracoda, Bivalvia and algal fragments (Table II). Moreover, the importance of these prey varied according to fish size. A complete overview of the diet of Gerreidae species is given in Tables SI–SIII (Supporting Information). The preferred prey of juvenile size classes of E. brasilianus were Polychaeta and Asterope sp., Copepoda and Nereis sp. for E. melanopterus and tentacles of Terebellidae for D. rhombeus (Table II). When reaching sub-adult sizes, E. brasilianus reduced the ingestion of Terebellidae tentacles and fed preferentially on Nereis sp. and Asterope sp., E. melanopterus and D. rhombeus decreased their consumption of Polychaeta and increased capture of Pseudodiaptomus spp. and algal fragments (Table II). In the adult size class, Anomalocardia brasiliana and Littorinidae became the most important prey for E. brasilianus, while Nereis sp. followed by algal fragments for E. melanopterus and Longipedia spp. was the predominant prey in the diet for D. rhombeus (Table II). The presence of non-food items (microplastics) was also noted in the stomach contents (Table II). Plastic fragments were observed in the stomach of the three Gerreidae species (Table II). It was absent only in the juvenile size class of D. rhombeus. The ingestion of plastic fragments represented
