DIET COMPOSITION OF PELAGIC FISH IN THE SOUTHWESTERN ATLANTIC , BRAZIL : AN ISOTOPIC MIXTURES

The purpose of this study is to characterize, through isotopic mixture models (13C and 15N), the diet of pelagic top predators of the Southwestern Atlantic Ocean. Thus, samples of muscular tissue were collected from individuals of the species: Xiphias gladius, Thunnus obesus, Thunnus alalunga, Thunnus albacares, Prionace glauca, Alopias superciliosus and Isurus oxyrinchus. Results indicate the diet of X. gladius is primarily composed by cephalopods (Ommastrephidae squids). A similar pattern is presented by T. alalunga, whose food is also mainly composed by Ommastrephidae squids. T. obesus and T. albacares present similar compositions, but smaller pelagic fish species are important to their diet compositions. A. superciliosus feeds mainly on Scombridae species (T. albacares and T. alalunga). A. superciliosus diet is composed mainly by Scombridae species, differing from that one from P. glauca, whose diet consists of smaller pelagic fish. The use of information in a conjugated way of the predators’ stomach contents as well as the isotopic tissue analyses may be a important step towards a more complete understanding of pelagic food web composition in Southwestern Atlantic Ocean.


INTRODUCTION
A trophic web consists in representation of the feeding relationships among predators and their prey in an ecologic community.The interactions may indicate the energetic flow within the referred ecosystem (Pimm, 1982).By meaning as direct interactions inside the trophic web, competition and predation may be assessed using traditional approaches (stomach contents analysis, fecal composition), but they may also be analysed using relatively recent techniques, including stable isotopes (Caut et al., 2006).As observed by Hildebrand et al. (1996), the studies based only on stomach contents may present some limitations.As the digestion rates may interfere, estimations about the consumed bulk may become more problematic and even more difficult to estimate what is assimilated by the predator.In this situation the use of stable isotopes (mainly 13 C, 15 N and 34 S) should be an important, though complementary but useful tool to reconstruct the consumer's diet, as such data may provide a continuous measurement about the trophic position of the components in the community.They also help to integrate the energy assimilation through the mass flow at distinct compartments until they reach the individual in the community (Peterson & Fry, 1987;Post, 2002).
According to Phillips (2001), the isotopic signature of a consumer may be interpreted as pondered mixture, which reflects the relative contribution distinct items represent to the diet of the consumer.Its proportion may be determined by means of mathematic mixture models.These models consist of a series of lineal equations, which maintain the mass balance to estimate the potential contribution the sources furnish to the system (details in Phillips, 2001.;Phillips & Gregg, 2003).The literature registers studies, which succeeded by using models of isotopic mixtures to describe the predators` diet composition as is the case of researches conducted by Ben-David et al. (1997a, b), Melville and Connolly (2003) and Caut et al. (2006).

Data collection
Samples were obtained in collaboration from the Santos (SP, Brazil) tuna fishery fleet.Such vessel operations occur at the Southwestern Atlantic Ocean in an area comprised by 17 -35° S and 27 -52° W out of the continental shelf (as observed by Amorim et al., 1998) (Figure 1).Based on stomach contents observations, samples of the most representative prey were collected in order to complement the analyses.They were represented by: Ommastrephidae squids, pelagic fishes (Gempylus serpens, Trichiurus lepturus, Brama brama, Cubiceps sp.), and Decapoda crustaceans (infraorder Caridea).Such samples were directly collected from the stomach contents of the fish species under study, and identified using taxonomic criteria based on Figueiredo & Menezes (1980;2000), Menezes & Figueiredo (1980;1985), Clarke (1986), Nesis (1987), andFigueiredo et al. (2002).

Diet characterization
The isotopic signature ( 13 C/ 12 C and 15 N/ 14 N) of each sampled species was then used to determine the diet characterization.The potential contribution of the sources to each consumer was estimated by means of the resolution expressed by the equations seen below: Where f a and f m represent the proportions each potential m food item contributes to the consumer`s diet.The isotopic traces from each source are represented by δ a , δ b , δ c , and δ m ; and δ corresponds to the isotopic signature (mixture) of the consumer.The calculations were made using the software IsoSource (Phillips & Gregg, 2003), following the methods of aggregation indicated by Phillips et al. (2005).

RESULTS
The prey proportions for the diets of X. gladius, A. superciliosus, P. glauca, T. albacares, T. obesus and T. alalunga are shown in Table I.The equations system resolution corresponding to I. oxyrinchus did not get a viable solution.In consequence, it was impossible to obtain an access to the potential food sources to this species.At Figure 2, presents the pelagic trophic web diagram from Southwestern Atlantic Ocean.
The calculations indicate that the X. gladius diet may be primarily composed by cephalopods (Ommastrephidae squids), followed by scombrid fishes (mainly T. albacares) and other smaller pelagic fish.Among the analyzed sharks, A. superciliosus may compose its diet with scombrid species mainly; (T.albacares and T. alalunga) in a distinct way when compared to P. glauca, whose diet is primarily composed by smaller pelagic fish, followed by Ommastrephidae squids.
The T. albacares and T. obesus present similar food compositions for their diets.Both species feed on small pelagic fish, and other kinds of prey are secondary to their diets.It may be important to emphasize that Decapoda crustaceans play an important role to the diet of T. albacares, in which they perform approximately 20% of the total assimilated food.The T. alalunga, on the other hand, composes its diet with Ommastrephidae squids in first place.Fish and crustaceans play a secondary role to its diet.

DISCUSSION
According to the stable isotopes 13 C and 15 N data, recorded from the muscular tissues of X. gladius, one may infer that its diet may in great part be composed by cephalopods (Ommastrephidae).These cephalopods constitute a common reference to other researches, concerning analyses about stomach contents of such fish species (Gorni et al., 2013).Some authors identify more frequently species, which belong to the genera Illex, Ommastrephes, Dosidicus and Todarodes as components of the stomach contents of X. gladius (Zavala-Camin, 1981;Stillwell & Kohler, 1985;Simões & Andrade, 2000;Vaske-Júnior, 2000;Ibañez et al., 2004;Castillo et al., 2007;Letelier et al., 2009).The exception cited by Letelier et al. (2009) in a mention related to occurrence of fish pieces (Thunnus) found in stomach.The detection about the assimilation of mesopelagic prey as some species of Ommastrephidae squids (Nesis, 1987) and epipelagic ones (Figueiredo & Menezes, 2000) may corroborate the attributed daily vertical displacement pattern of X. gladius (Carey & Robison, 1981;Takahashi et al., 2003).
The tissue isotopic concentration of A. superciliosus revealed a clear dominance of teleosts, in an agreement to the observations made by Polo-Silva (2007) and Preti et al. (2008) whose studies were performed at the Pacific Ocean.In more detail, the most representative fish families were Paralepididae, Merluccidae and Sciaenidae.Gorni et al. (2013) report great contribution of Teuthida squids and Trichiuridae fishes in A. superciliosus stomach content.Meanwhile, in this present study based on isotopic mixture results, there happens to occur a disagreement to those observations, as Scombrid fishes should compose more than 80% of the assimilated prey.So, the lack of research of food composition and isotopic information for A. superciliosus at the Southwestern Atlantic may, for now, be considered, at least, a precocious conclusion.The P. glauca, in turn, feeds mainly on teleosts, and the squids should only constitute a complementary diet.This pattern confirms results obtained by Bornatowski & Schwingel (2008) and Vaske-Junior et al. (2009).Notwithstanding this, a result based on the analyses of 68 stomachs of such sharks caught at oceanic waters off the southern Brazilian coast (Vaske-Junior & Rincón-Filho, 1998) indicate a diet to be mainly based on squids, represented most frequently by the batipelagic Chiroteuthis veranyi.
Following the isotopic analysis, T. albacares may be considered as food item to all the top predators of the region, as it signalizes to this direction in this study.The tissues of this tuna sustain a signal that this species feeds on smaller teleosts, in accordance to observations made by Zavala-Camin (1981), Vaske-Júnior (2000), Vaske-Junior et al. (2003) and Satoh et al. (2004).These fish assimilate mainly specimens of Bramidae and Trichiuridae.These observations reinforce the idea that T. albacares uses to prey over monospecific fish concentrations, which come to more superficial layers of the water (Dragovich & Potthoff, 1972;Bard et al., 2002).In a similar way, the isotopic signature indicated T. obesus feed primarily on pelagic fish, and cephalopods should only constitute secondary prey.By analyzing 44 stomach contents of T. obesus, Vaske-Júnior (2000), registered fish to compose more than 86% of the ingested prey there observed.Liming et al. (2005), in an analysis based on 272 stomach contents of T. obesus, caught at the Central Atlantic Ocean, also characterize pelagic fish as their most important prey, as they perform almost 50% of the total food within their stomachs.Cephalopod mollusks constitute, therefore, a secondary food item to this species, in which they constitute 30% of the whole food they consume.
The T. alalunga, should be, though presenting a differentiated feeding habit, the tuna species to which cephalopods represent the most important food component.More than 80% of the food items are based on cephalopods.This tendency to a teutophagic habit may corroborate the observations made by Zavala-Camin (1981), who, analysing 741 stomachs, registered predominance for mollusks (59.9%), to which fish and crustaceans were secondary food items.In any case, Sabatié et al. (2003), Satoh et al. (2004) and Gorni et al. (2013) show distinct results, in which fish and crustaceans are predominant.
It should be noted that research on the food web of top predators presents a great challenge.Access to large numbers of samples is limited by the high costs involved in capturing these animals on open seas.Therefore, the food patterns recorded in the present study are restricted to the period of this work.
Divergences by comparing results originated from isotopic mixture models and those obtained by stomach contents analyses may be related to particular issues related to each used method.Following Schindler & Lubetkin (2004), the use of stable isotopes may follow the destinies of distinct energy sources and materials through the trophic webs, becoming, therefore, important complementary tools to analyses of food debris.Though more accurate in determining the preyed species, the evaluation of gastric contents may provide a more restrict sight about the feeding habits of the consumers, as such methods limit their analyses to shorter temporal scales (Estrada et al., 2005;Kojadinovic et al., 2008).
Thus, the utilization in a conjugated way of data obtained by investigating food stomach contents and stable isotopes shown in their tissues may be an important way to improve a direction to a better and more complete knowledge about the ecologic relationships presented by the pelagic species at the Southwestern Atlantic Ocean.

Figure 2 -
Figure 2 -Pelagic trophic web diagram from Southwestern Atlantic Ocean.Thinner lines represents contribution of less than 25% in food source; medium thickness lines indicate the contribution of 25 to 50% of the diet, and thicker lines represent more than 50% of the diet.

Table I -
Mean and standard deviations of the prey proportions to the diets of Xiphias gladius, Alopias superciliosus, Prionace glauca, Thunnus albacares, T. obesus and T. alalunga inferred by the use of models of isotopic mixture (δ 13 C and δ 15 N).