INTRODUCTION
Delphinidae is the most successful and diverse among all Cetartiodactyla families which occupy a wide variety of ecosystems (LeDuc, Reference LeDuc, Perrin, Wursig and Thewissen2002), including freshwater (e.g. tucuxi, Sotalia fluviatilis), shallow coastal (e.g. short-beaked common dolphin, Delphinus delphis), and deep pelagic waters (e.g. Fraser's dolphin, Lagenodelphis hosei) in tropical and subpolar environments. Some species have restricted distribution while others are broadly distributed. Others have evolved by adapting to specific niches, generating different ecotypes such as the coastal and offshore bottlenose dolphin (Tursiops truncatus) (Segura et al., Reference Segura, Rocha-Olivares, Flores-Ramirez and Rojas-Bracho2006) and killer whales (Orcinus orca) (Ford, Reference Ford, Perrin, Wursig and Thewissen2002).
A rich diversity in delphinids can be found in waters off the coast of Brazil (Bastida et al., Reference Bastida, Rodríguez, Secchi and da Silva2007). The continental shelf and slope waters in southern Brazil are influenced by the Subtropical Convergence and represent a biogeographical transition zone between Patagonian temperate and Brazilian tropical waters. In the neritic zone, cold and warm water circulation and upwelling processes (Castello et al., Reference Castello, Haimovici, Oderbretch, Vooren, Seeliger, Oderbrech and Castello1997; Garcia, Reference Garcia, Seeliger, Odebrecht and Castello1998, Piola et al., Reference Piola, Campos, Moller, Charo and Martinez2000; Muelbert et al., Reference Muelbert, Acha, Mianzan, Guerrero, Reta, Braga, Garcia, Berasategui, Gomez-Erache and Ramírez2008) influence productivity that can be considered moderate to high (Odebrecht & Garcia, Reference Odebrecht, Garcia, Seeliger, Odebrecht and Castello1997).
Trophic relationships and habitat use are of key importance to the understanding, management and conservation of cetacean populations and their position in complex marine food webs. In southern Brazil, studies on the feeding ecology and habitat use of delphinids have typically focused on two coastal species: the Guiana dolphin (Sotalia guianensis) (e.g. Flores & Bazzalo, Reference Flores and Bazzalo2004; Oshima et al., Reference Oshima, Santos, de, Bazallo, Flores and Pupim2010) and the bottlenose dolphin (e.g. Simões-Lopes & Fabian, Reference Simões-Lopes and Fabian1999; Fruet et al., Reference Fruet, Secchi, Di Tullio and Kinas2011). Information on the trophic ecology of the remaining delphinid species that inhabit the continental shelf or offshore waters is scarce or non-existent, mostly due to their often difficult to work in habitats.
Carcasses washed ashore provide valuable biological material from species from those habitats. Stomach content analysis, for example, can yield direct information on the diet and indirect insight on the feeding environment (Barros & Clarke, Reference Barros, Clarke, Perrin, Würsig and Thewissen2009). However, this information is often biased owing to different digestion rates of prey, overestimating, for example, the importance of prey with chitinous structures (e.g. cephalopod beaks and crustaceans) relative to fish, whose otoliths are rapidly digested by gastric acids (Jobling & Breiby, Reference Jobling and Breiby1986; Santos et al., Reference Santos, Clarke and Pierce2001). Other limitations of analyses of stomach contents are due to the secondary ingestion of prey, which is prey found that was in the digestive tract of the predator's prey, as well as the short feeding time interval integrated by this kind of study, giving only information on recent feeding (Hobson et al., Reference Hobson, Piatt and Pitocchelli1994; Dehn et al., Reference Dehn, Sheffield, Follmann, Duffy, Thomas and O'Hara2006). These latter limitations can be reduced by increasing sample sizes and ensuring samples represent the appropriate temporal and spatial scales.
Fortunately, other complementary methods, such as naturally occurring stable isotopes of key elements and fatty acids (e.g. Herman et al., Reference Herman, Burrows, Wade, Durban, Matkin, LeDuc, Barrett-Lennard and Krahn2005; Krahn et al., Reference Krahn, Herman, Matkin, Durban, Barrett-Lenard, Burrows, Dahlheim, Black, LeDuc and Wade2007), are now available for studies on feeding ecology and habitat preference in aquatic vertebrates. Stable isotope values in animal tissues reflect those in the food webs where they feed (Rubenstein & Hobson, Reference Rubenstein and Hobson2004; Graham et al., Reference Graham, Koch, Newsome, McMahon, Aurioles, West, Bowen, Dawson and Tu2010) and are useful particularly for determining trophic level, identifying major food sources, and assessing foraging habitats (e.g. Das et al., Reference Das, Beans, Holsbeek, Mauger, Berrow, Rogan and Bouquegneau2003; Krahn et al., Reference Krahn, Herman, Matkin, Durban, Barrett-Lenard, Burrows, Dahlheim, Black, LeDuc and Wade2007; Pinela et al., Reference Pinela, Borrell, Cardona and Aguilar2010; Ricchialdelli et al., Reference Riccialdelli, Newsome, Fogel and Goodall2010). Stable isotope ratios of nitrogen (15N/14N) and to a lesser extent carbon (13C/12C) show a stepwise enrichment with increasing trophic level in the marine environment (De Niro & Epstein, Reference De Niro and Epstein1978, Reference De Niro and Epstein1981). Indeed, nitrogen isotopes values change in a predictable fashion between trophic levels, owing to the preferential excretion of the light isotope (Caut et al., Reference Caut, Angulo and Courchamp2009), and so reflect trophic position (De Niro & Epstein, Reference De Niro and Epstein1981; Cabana & Rasmussen, Reference Cabana and Rasmussen1996; McCutchan et al., Reference McCutchan, Lewis, Kendall and McGrath2003). Some studies with captive (Hilderbrand et al., Reference Hilderbrand, Farley, Robbins, Hanley, Titus and Servheen1996; Hobson et al., Reference Hobson, Schell, Renouf and Noseworthy1996; Lesage et al., Reference Lesage, Hammill and Kovacs2002) and free-ranging (Newsome et al., Reference Newsome, Bentall, Tinker, Oftedal, Ralls, Estes and Fogel2010) mammals showed that this increment varies among tissues, species, developmental stage and/or body condition. However, an approximate enrichment of 3.4‰ per trophic level is generally accepted (Post, Reference Post2002). Similar to 15N, 13C content also increases up trophic levels, although an increase of only roughly 1‰ is typically observed (De Niro & Epstein, Reference De Niro and Epstein1978; Peterson & Fry, Reference Peterson and Fry1987). Therefore, predators' carbon isotope values are used as an indicator of the sources at the base of the food web where they feed (Hobson, Reference Hobson1999; Graham et al., Reference Graham, Koch, Newsome, McMahon, Aurioles, West, Bowen, Dawson and Tu2010).
An important aspect to be considered when using stable isotopes is that the turnover rate within a tissue is based on its metabolic rate. Therefore, diet information may be determined over a time frame that varies from a few days (e.g. blood plasma and liver), months (e.g. red blood cells or muscle) to years (e.g. whale baleen or teeth) (Walker & Macko, Reference Walker and Macko1999; Kelly, Reference Kelly2000). Teeth are a particularly informative tissue for tracking the diet of delphinids over their lifetimes because they provide a permanent dietary record for an individual, as, under normal conditions, growth layers in teeth do not resorb or modify (Walker et al., Reference Walker, Potter and Macko1999; Walker & Macko, Reference Walker and Macko1999; Niño-Torres et al., Reference Niño-Torres, Gallo-Reynoso, Galván-Magaña, Escobar-Briones and Macko2006).
In this study a first estimation of carbon and nitrogen isotopes values in teeth of delphinids from southern Brazil, in the subtropical western South Atlantic, is presented. These stable isotope profiles were used to investigate inter- and intra-specific variations in feeding ecology and habitat use of seven species from coastal/estuarine, continental shelf and offshore marine environments.
MATERIALS AND METHODS
Study site and sampling
Fifty specimens of seven species (Table 1) found washed ashore during systematic beach surveys conducted along the southern coast of Rio Grande do Sul State (RS), Brazil, (1993 to 2009) were used for this study (Figure 1). Teeth were extracted from the middle upper or lower jaw, cleaned and stored dry in the Laboratório de Tartarugas e Mamíferos Marinhos (LTMM-IO-FURG) collection.
Fig. 1. Study area, western South Atlantic, southern Brazil. Northern and southern limits of the region surveyed for the collection of stranded delphinids are shown (Mostardas and Chui).
Table 1. δ13C and δ15N values±SD (‰) of teeth of delphinids found washed ashore along the southern coast of Brazil.
Min, minimum; Max, maximum.
Analysis of isotope compositions
Stable isotope analysis of teeth was performed following the protocol described in Walker & Macko (Reference Walker and Macko1999). Teeth were dried for 3–4 days in a 60°C oven and cleaned of outer soft tissue with a carbide burr attached to a drill. A low speed saw with a diamond-embedded blade was used to cut through the centre of the tooth in the longitudinal buccal–lingual axis in order to expose the growth layer groups (GLGs; Perrin & Myrick, Reference Perrin and Myrick1980). Exposed dentine was sampled with a small drill bit, taking care that all GLGs were sampled so that the resulting powder would represent the entire life of the individual. The powder obtained was acidified with 30% hydrochloric acid (HCl) to remove biogenic carbonates, which could alter the organic δ13C measurements, and then dried again for 1 hour in a 60°C oven. The goal of preparation is the conversion of the organic samples into gases of suitable purity that can then be analysed by the mass spectrometer. Samples of approximately 5 mg of residual acidified tooth were used for δ13C and δ15N analysis using an elemental analyser (EA) connected to a Micromass Optima Isotope Ratio Mass Spectrometer (IRMS; GV Instruments, Manchester, UK). Natural abundance of stable isotope ratios (13C/12C and 15N/14N) is expressed in a delta notation (δ) as per mil variations (‰) when compared with international standards (e.g. Pee Dee Belemnite (PDB) for carbon and atmospheric N2, for nitrogen). Results were expressed as:
$$\delta X = \lsqb \lpar \hbox{R}_{\rm sample}/\hbox{R}_{\rm standard}\rpar - 1\rsqb ^{\ast} 1000$$where R sample and R standard are the 13C/12C or 15N/14N ratios of the sample and standard, respectively.
Data analysis
All results are presented as the mean ± SD. Data were tested for normality and homogeneity of the variances using the Kolmogorov–Smirnov test and Levene's test, respectively. Comparison of isotopes values among species were conducted using 1-way analysis of variance (ANOVA) techniques followed by a Tukey's honestly significant difference (HSD) test, when a significant difference was found. The null hypothesis of no differences was rejected if P < 0.05. In addition, a cluster analysis (Euclidean distances, complete linkage method) based on δ13C and δ15N mean values of each species was used for the detection of isotope groupings.
RESULTS
Carbon and nitrogen isotope compositions differed significantly among species (ANOVA, F6,41 = 10.78; P < 0.001 and F6,41 = 49.99; P < 0.001, for carbon and nitrogen isotopes, respectively) (Tables 1 & 2; Figure 3). Data from the Fraser's dolphin (N = 1) could not be statistically tested.
Table 2. Results of the Tukey's honestly significant difference post hoc test for multiple comparisons of δ13C and δ15N values from teeth of delphinids found washed ashore along the southern coast of Brazil (see Table 1 for codes). Carbon isotope (δ13C) P values are reported below the diagonal and those for nitrogen isotopes (δ15N) are reported above the diagonal. Significant P values (<0.05) are highlighted in bold.
One killer whale had the lowest δ13C found in our samples (−20‰) and also a low δ15N value (10.5‰) (Figure 2). This 396 cm-long specimen was a very emaciated young female found stranded alive with the stomach full of oceanic salps (Iasis zonaria) and other unidentified planktonic invertebrates. This was considered atypical, thus, this animal was excluded from statistical analyses.
Fig. 2. Teeth δ13C and δ15N values in teeth of delphinids found stranded along the southern coast of Brazil. Dashed line circle indicates 15N-enriched bottlenose dolphins, Tursiops sp.
Some species showed considerable intraspecific variation, either in δ13C or δ15N, however, in those animals presenting extreme values nothing atypical was detected when found washed ashore, as was the case of the killer whale mentioned above. Therefore, their values were considered in the range of the normal distribution of isotopic values for the species. Carbon isotope values for rough-toothed dolphins ranged from −14.3‰ to −11.5‰ and those for false killer whales were from −13.1‰ to −10.8‰ (Table 1; Figures 2 & 3). Furthermore, values of δ15N for the latter species revealed a bimodal distribution, with two specimens with high values (19.4‰ and 19.0‰) and the remainder with a mean δ15N of 12.2‰. The latter included animals from a single mass stranding while the 15N-enriched individuals were found washed ashore alone. There is no stock information available for this species in the region thus we cannot infer if they belong to different ecotypes/stocks. However we treated them as two different groups (A and B), as their significantly different δ15N values (Student's t-test, P < 0.001) suggest different feeding habits. The 15N-enriched group (false killer whale A) was significantly different from the other species (Tukey's HSD test, P < 0.01), with the exception of the bottlenose dolphin (Tukey's HSD test, P = 0.14) and the killer whale (Tukey's HSD test, P = 0.54). δ15N values of the second group (false killer whale B) were significantly different from all dolphin species (Tukey's HSD test, P < 0.001). As for δ13C, group A did not significantly differ from the remainder species (Tukey's HSD test, P > 0.05), while group B was significantly different from bottlenose dolphins and killer whales (Tukey's HSD test, P < 0.01).
Fig. 3. Mean (±SD) δ13C and δ15N values for teeth dentine of delphinids from southern Brazil.
Six bottlenose dolphins were more enriched in 15N (4 males, one female and one individual of unknown sex) (Figure 2). The mean δ15N of these animals was 18.9‰, being 1.6‰ more enriched than the lower δ15N group (17.2‰). Their δ15N values were statistically different (t-test, P < 0.0001) but their carbon stable isotopes values were not different (t-test, P = 0.51), averaging −10.5‰ for the group with enriched nitrogen signatures, and −10.6‰ for the rest of the dolphins. Finally, one bottlenose dolphin showed extremely different and depleted carbon and nitrogen signals (δ13C= −12.0‰ and δ15N = 15.5‰).
Cluster analysis of stable isotope values for carbon and nitrogen defined one group with high trophic level predators leaving the group B of false killer whales (low trophic level predator) on a separate branch. Within the cluster of high trophic level predators, the oceanic Fraser's dolphin was separated from the continental shelf and coastal grouping, with the latter clustered together at a lower distance (Figure 4).
Fig. 4. Tree diagram of delphinids from southern Brazil resulted from cluster analysis of stable isotope ratios of carbon and nitrogen in teeth (see Table 1 for codes).
DISCUSSION
Carbon and nitrogen stable isotope values in teeth dentine of delphinids from southern Brazil reflected different trophic levels and/or feeding environments respectively. An offshore–inshore trend of increasing δ13C values was observed, which is in agreement with the preferred habitat of the species analysed here. Indeed, carbon isotope values varied from highly 13C enriched values found in coastal species, such as bottlenose dolphins (δ13C = −10.6‰) to more depleted signals such as the carbon isotope ratio of the Fraser's dolphin (δ13C = −12.8‰), a typical species from deep pelagic environments. Similarly, previous studies revealed a longitudinal trend in marine environments, where nearshore, benthos linked food webs are more 13C enriched compared to more offshore, pelagic food webs (France, Reference France1995; Burton & Koch, Reference Burton and Koch1999; Takai et al., Reference Takai, Onaka, Ikeda, Yatsu, Kidokoro and Sakamoto2000; Clementz & Koch, Reference Clementz and Koch2001; Lesage et al., Reference Lesage, Hammill and Kovacs2001; Barros et al., Reference Barros, Ostrom, Stricker and Wells2010; Pinela et al., Reference Pinela, Borrell, Cardona and Aguilar2010; Riccialdelli et al., Reference Riccialdelli, Newsome, Fogel and Goodall2010) which is probably a reflection of a gradient of decreasing macrophyte influence (Hill et al., Reference Hill, McQuaid and Kaehler2006). Indeed, phytoplankton have lower δ13C values than many inshore plants (e.g. seagrasses, kelp forests and marsh plants), making inshore carbon sources able to be distinguished from more pelagic sources (Fry & Sherr, Reference Fry and Sherr1984; Hobson et al., Reference Hobson, Piatt and Pitocchelli1994; Clementz & Koch, Reference Clementz and Koch2001).
Bottlenose dolphins and killer whales had the highest δ13C values, reflecting their coastal feeding habits (Table 1). Furthermore, their δ15N values also overlap, suggesting a similar trophic level for these two coastal predators. In southern Brazil, coastal bottlenose dolphins form small resident populations usually associated with estuaries and river mouths (Castello & Pinedo, Reference Castello and Pinedo1977; Simões-Lopes & Fabian, Reference Simões-Lopes and Fabian1999). A resident population of 84–86 animals (Dalla Rosa, Reference Dalla Rosa1999; Fruet et al., Reference Fruet, Secchi, Di Tullio and Kinas2011) inhabits the Patos Lagoon estuary and its adjacent coastal areas. Studies of the diet of bottlenose dolphins from RS based on stomach content analysis confirmed a coastal feeding habit with the white croaker (Micropogonias furnieri), the cutlass fish (Trichiurus lepturus) and the drum (Paralonchurus brasiliensis) being the most important prey (Pinedo, Reference Pinedo1982; Mehsen et al., Reference Mehsen, Secchi, Fruet and Di Tullio2005). Nitrogen isotope values were also high. As noted above, due to the trophic enrichment in 15N through the food chain, a high δ15N is expected for this high-trophic-level predator. Two groups of bottlenose dolphins differing in their δ15N values were identified (Figure 2), which can be interpreted as a resource partitioning that may be occurring among bottlenose dolphins from southern Brazil. The observed mean values of the 15N-enriched group (δ15N = 18.9‰) and the 15N-depleted group (17.2‰) are higher and similar to, respectively, than values found for teeth of bottlenose dolphins from the western North Atlantic coast (δ15N = 16.8‰, Walker et al., Reference Walker, Potter and Macko1999; δ15N = 17.6 and16.8‰ for the inner and outer part of the tooth dentine, Knoff et al., Reference Knoff, Hohn and Macko2008) and ~5–7‰ higher than values found by Barros et al. (Reference Barros, Ostrom, Stricker and Wells2010) in Sarasota Bay (δ15N = 11.9‰) and the Gulf of Mexico (δ15N = 12.7‰). The latter used a different methodology for preparing the samples, centrifuging the dentine powder after demineralization, to separate collagen from non-collagenous proteins, and performing a lipid extraction. However, lipid extraction has a small influence in δ15N by introducing an average fractionation of about 0.25‰ (Post et al., Reference Post, Layman, Arrington, Takimoto, Quattrochi and Montaña2007). Therefore, regional variations in diet/food web structure and/or nitrogen isotopes at the base of the food webs are likely to be the cause of the observed differences. Moreover, Abreu et al. (Reference Abreu, Costa, Bemvenuti, Odebrecht, Granéli and Anesio2006) reported high values of δ15N inside the Patos Lagoon estuary and argued that this is probably an effect of nutrient input from domestic and industrial sewage. Nitrogen isotope content of wastewater has higher values due to ammonium volatilization and denitrification processes during sewage treatment that removes the lighter 14N faster than the 15N (Macko & Ostrom, Reference Macko, Ostrom, Lajtha and Michener1994; McClelland et al., Reference McClelland, Valiela and Michener1997). In addition, stormwater may also be considered as an enrichment factor, due to the thermodynamically favoured volatilization of isotopically depleted 14NH3 from stormwater as it flows across hot surfaces (Dillon & Chanton, Reference Dillon and Chanton2007). Estuarine 15N-enriched waters could influence the isotopic composition of adjacent waters; however, no data on stable isotopes of the food chain of the coastal adjacent waters of the estuary are available at this time, thus precluding further interpretations.
Additionally, a bottlenose dolphin with a clearly different isotope signal was identified, which suggests that it may have fed in another region (Figure 2). In a preliminary isotope analysis comparing bottlenose dolphins from two areas, this animal was clustered together with a northern group of dolphins found stranded along the São Paulo State, Brazil (25°00′S 47°50′W) (Botta et al., Reference Botta, Hohn, Macko, Santos and Secchi2010a). Three hypotheses exist for the origin of this animal: (a) this individual could be part of the northern form, as proposed by Barreto (Reference Barreto2000); (b) it could be a disperser from a southern population of bottlenose dolphins from Uruguay, which are known to move to coastal adjacent waters of the Patos Lagoon estuary (Laporta et al., Reference Laporta, Fruet, Di Tullio and Secchi2008); or (c) it could belong to an offshore ecotype. However, based on the cranial characters proposed by Barreto (Reference Barreto2000) to distinguish northern from southern forms (e.g. shape of the pterigoyds and their separation) we could infer that this animal was a northern form individual, which could explain its distinct isotope compositions. As stated above, carbon and nitrogen signals of the southern São Paulo population were similar to the δ13C and δ15N values found for this animal (Botta et al., Reference Botta, Hohn, Macko, Santos and Secchi2010a). Nevertheless, the possibility of this animal belonging to an offshore group cannot be discarded, as no information on cranial morphometry/shape and/or isotopic signatures for this group are available so far. Isotope compositions in teeth of offshore bottlenose dolphins from other areas are similar for nitrogen, but lower for carbon to those presented by this animal (e.g. δ15N = 14.8‰ and δ13C = −13.9‰, western North Atlantic; Walker & Macko, Reference Walker and Macko1999). Indeed, the carbon isotope value for this bottlenose dolphin (−12‰) was similar to those of continental shelf species (e.g. Atlantic spotted dolphin) and higher than that presented by the offshore Fraser's dolphin.
The presence of killer whales in coastal waters of southern Brazil is seasonal, with the records from winter and spring months being more common (Dalla Rosa et al., Reference Dalla Rosa, Secchi, Lailson-Brito and Azevedo2005). The weakfish (Cynoscion guatucupa), the eagle stingray (Myliobatis sp.) and cephalopods have been recorded as prey for killer whales in Brazilian waters (Dalla Rosa et al., Reference Dalla Rosa, Secchi, Lailson-Brito and Azevedo2005). Although no remains of cetaceans were found in the stomach of the killer whales analysed, a franciscana dolphin (Pontoporia blainvillei) was reported in the stomach content of a killer whale stranded in this region (Ott & Danilewicz, Reference Ott and Danilewicz1997). Owing to differences in the period of time integrated by stomach contents and stable isotopes, these two animals may have been also eating small cetaceans, as denoted by their high δ15N and δ13C, which are similar to values found in franciscana teeth from this area (Botta et al., Reference Botta, Hohn, Macko, Santos, Di Beneditto, Ramos, Bertozzi, Fraco-Trecu, Barbosa, Cremer, Failla, Iñiguez and Secchi2010b).
The killer whale with the lowest δ13C found in our samples (−20‰) and also with a low δ15N value (10.5‰) was considered atypical. The stomach contents of this individual (oceanic salps (Iasis zonaria) and other unidentified planktonic invertebrates) could explain the low observed stable carbon and nitrogen isotope values found. Indeed, filter-feeder salps occupy low trophic levels in the oceanic environments (Madin, Reference Madin1974). Elsewhere, salps and other components of the zooplankton have δ13C values around −20‰ (Hatase et al., Reference Hatase, Takai, Matsuzawa, Sakamoto, Omuta, Goto, Arai and Fujiwara2002; Bode et al., Reference Bode, Alvarez-Ossorio, Carrera and Lorenzo2004).
In the western South Atlantic, short-beaked common dolphins and Atlantic spotted dolphins inhabit shallow waters over the continental shelf and upper slope (Zerbini et al., Reference Zerbini, Secchi, Bassoi, Dalla Rosa, Higa, de Sousa, Moreno, Möller and Caon2004; Moreno et al., Reference Moreno, Zerbini, Danilewicz, Santos, de, Simões-Lopes, Lailson-Brito and Azevedo2005; Tavares et al., Reference Tavares, Moreno, Siciliano, Rodríguez, Santos, Lailson-Brito and Fabián2010), where they feed on small meso/epipelagic fish and squids (Santos & Haimovici, Reference Santos and Haimovici2002; Melo et al., Reference Melo, Santos, Bassoi, Araujo, Lailson-Brito, Dornelles and Azevedo2010; E.R. Secchi, personal observation). The intermediate δ13C values found in teeth of these species' individuals are presumably reflecting this pelagic phytoplankton-dependent food web. The δ13C value found for common dolphins is consistent with values found in other studies (~− 16‰ in muscle, Das et al., Reference Das, Beans, Holsbeek, Mauger, Berrow, Rogan and Bouquegneau2003; ~ −12‰ in bone, Pinela et al., Reference Pinela, Borrell, Cardona and Aguilar2010).
The rough-toothed dolphin is generally found in deep-offshore waters (Miyazaki & Perrin, Reference Miyazaki, Perrin, Ridgway and Harrison1994; Jefferson, Reference Jefferson, Perrin, Wursig and Thewissen2002); however coastal sightings of this species are relatively common along Brazilian waters, mainly for the south-eastern coast (Lodi, Reference Lodi1992; Ott & Danilewicz, Reference Ott and Danilewicz1996; Flores & Ximenes, Reference Flores and Ximenes1997; Lodi & Hetzel, Reference Lodi and Hetzel1998). The carbon isotope signal found for the species was similar to that of continental shelf species, indicating that a shallower water habitat is also used in the subtropical western South Atlantic. Diet reported for this species included fish and squid (Miyazaki & Perrin, Reference Miyazaki, Perrin, Ridgway and Harrison1994). Nitrogen stable isotopes for this species did not differ from the rest of the continental shelf species, thus suggesting feeding at similar trophic positions (Figure 4). Finally, a rough-toothed dolphin with a 13C depleted value was identified (Figure 2). This animal had δ13C even lower than the oceanic Fraser's dolphin which can reflect a more oceanic feeding habitat used by this specimen, which could belong to an offshore group of rough-toothed dolphins.
False killer whale habitats are primarily oceanic and their main prey are deep-sea cephalopods and fish (Odell & McClune, Reference Odell, McClune, Ridgway and Harrison1999; Baird, Reference Baird, Perrin, Wursig and Thewissen2002). Isotopically, two groups of divergent trophic level and/or habitats could be identified. One group of 6 false killer whales presented low carbon and nitrogen isotopes signals (false killer whale B: Figures 2 &3; Table 1). These animals were part of a mass stranding of 14 individuals that occurred in winter of 1995. The stomach contents of four of these animals (the remainder were empty) revealed only cephalopod prey, mainly Ommastrephes bartramii (Andrade et al., Reference Andrade, Pinedo and Barreto2001). This squid is a member of the oceanic Ommastrephiidae family and together with Ilex argentinus are common prey found in mass stranded false killer whales from Argentina too (Koen-Alonso et al., Reference Koen-Alonso, Pedraza, Schiavini, Goodall and Crespo1999) and are also important in the diet of other upper slope and oceanic adjacent water predators (Santos & Haimovici, Reference Santos and Haimovici2001, Reference Santos and Haimovici2002). The only isotope information published for these cephalopods off RS area, revealed low carbon and nitrogen stable isotope content (δ13C = −16.7‰ and δ15N= 9.3‰, for a combined sample of I. argentinus and O. bartramii; Bugoni et al., Reference Bugoni, McGill and Furness2010). After accounting for a combination of tissue-dependent Δ13C collagen-muscle (~4‰) and a trophic discrimination factor (1‰) by subtracting a total of 5‰ from false killer whales' dentine δ13C value (Koch, Reference Koch, Michener and Lajtha2007), comparison with omastrephids data confirmed the teuthophagic feeding habit of this group of false killer whales. On the other hand, two individuals presented different isotopic values from this mass stranded group (false killer whale A: Figures 2 &3). One of these specimens, a 333 cm-long false killer whale washed ashore in 2004, had a high 13C content (−10.8‰), indicating a coastal habitat, and 15N enriched dentine (19‰), indicating it was feeding at a high trophic level. Besides cephalopod prey, Sciaenidae and Serranidae fish were reported in the diet of this species in southern Brazil (Pinedo & Rosas, Reference Pinedo and Rosas1989). Other authors indicate that this species could also prey upon small cetaceans (Odell & McClune, Reference Odell, McClune, Ridgway and Harrison1999; Baird, Reference Baird, Perrin, Wursig and Thewissen2002). Nevertheless, the stable isotope proxy, together with previously-reported stomach content information suggest that at least some false killer whales in southern Brazil have a more coastal piscivorous feeding habit. Indeed, this species has been observed in coastal areas off southern Brazil (LTMM-IO-FURG, unpublished data). Finally, one young specimen (total length = 165 cm) had a low δ13C (−13.1‰) but a high δ15N (19.4‰). This animal was found stranded alone in 2006. Higher δ15N relative to the mother's signal is expected for lactate-feeding calves because mothers are catabolizing their own tissues for producing milk, which leads to a higher ‘trophic level’ of their offspring (Hobson & Sease, Reference Hobson and Sease1998; Walker & Macko, Reference Walker and Macko1999; Newsome et al., Reference Newsome, Koch, Etnier and Aurioles-Gamboa2006, Reference Newsome, Etnier, Monson and Fogel2009; Knoff et al., Reference Knoff, Hohn and Macko2008). The calf's lower δ13C value is probably influenced by the high lipid content in the 13C depleted milk (Hobson & Sease, Reference Hobson and Sease1998; Newsome et al., Reference Newsome, Koch, Etnier and Aurioles-Gamboa2006, Reference Newsome, Etnier, Monson and Fogel2009). Based on its total length, this animal was probably still nursing, as false killer whales usually lactate for 18–24 months (Odell & McClune, Reference Odell, McClune, Ridgway and Harrison1999), which could be a plausible explanation for the isotopes signals found.
The carbon isotope content for the Fraser's dolphin was lower than that of typical continental shelf delphinids, which closely represents its oceanic, deep water habitat. This is a tropical species, but with unusual strandings reported for subtropical areas (e.g. Praderi et al., Reference Praderi, Praderi and Garcia1992; Pinedo et al., Reference Pinedo, Lammardo and Barreto2001; Laporta et al., Reference Laporta, Praderi, Le Bas and Crespo2002; Moreno et al., Reference Moreno, Danilewicz, Borges-Martins, Ott, Caon and Oliveira2003). This individual was a male with a total length of 236 cm. Length at sexual maturity is about 220–230 cm (Dolar, Reference Dolar, Perrin, Wursig and Thewissen2002) indicating that this animal was probably sexually mature. Mesopelagic fish, crustaceans and cephalopods are among the preferential prey for this species, which are captured mainly from near surface to at least 600 m (Robison & Craddock, Reference Robison and Craddock1983). However, the diet reported for individuals stranded in southern Brazil included demersal fish, the cutlass fish (Trichiurus lepturus), the epipelagic squid Loligo sanpaulensis and the demersal/pelagic cephalopod, Argonauta nodosa, besides some penneidae shrimp (Pinedo et al., Reference Pinedo, Lammardo and Barreto2001; Santos & Haimovici, Reference Santos and Haimovici2001; Moreno et al., Reference Moreno, Danilewicz, Borges-Martins, Ott, Caon and Oliveira2003; Melo et al., Reference Melo, Santos, Bassoi, Araujo, Lailson-Brito, Dornelles and Azevedo2010). This coastal feeding habit can be considered as unusual and likely occurred because the animals were outside their home range (Moreno et al., Reference Moreno, Danilewicz, Borges-Martins, Ott, Caon and Oliveira2003). Therefore, stomach contents may occasionally represent unusual local feeding while stable isotope values of teeth reflect the typical long-term feeding habits for this individual.
This first study on stable isotope values of delphinids from southern Brazil revealed some important information about trophic ecology and feeding environments of delphinids inhabiting coastal, shelf and oceanic environments in the subtropical western South Atlantic. Substantial information about the trophic ecology and feeding environment can be derived from biochemical analyses in teeth. Furthermore, its combination with traditional methods, such as stomach content analysis, direct observation and telemetry, and biochemical methods could be a powerful tool to investigate the natural history of marine mammals (Pauly et al., Reference Pauly, Trites, Capuli and Christensen1998). However, the almost non-existent environmental isotope data for this southern region calls for additional sampling efforts and analyses involving the main components of these marine food webs which include these high trophic level predators.
ACKNOWLEDGEMENTS
We are indebted to all of the researchers and volunteers from the LTMM-IO-FURG and LMM-MORG and to Lilia Fidelix for helping to collect data on stranded delphinids along Rio Grande do Sul. We would like to thank the Yaqu Pacha Foundation (Germany), the Society of Marine Mammalogy (United States of America), Cetacean Society International (United States of America) and Museu Oceanográfico/Fundação Universidade Federal do Rio Grande (MO-FURG, Brazil) for financial or logistical support to this project. The Conselho Nacional de Desenvolvimento Científico e Tecnológico—CNPq (Brazil) provided scholarships to E.R. Secchi (PQ 305219/2008-1) and to S.B. (141668/2007-5 and SWE 201247/2009-7). This article is part of Silvina Botta's PhD thesis in Biological Oceanography (Post-Graduate Course in Biological Oceanography—IO–FURG, RS, Brazil) and is a contribution of the Research Group Ecologia e Conservação da Megafauna Marinha-EcoMega/CNPq.
