Ecological niche is defined as the multidimensional space comprising the resources used by an organism (Hutchinson 1957). Resource partitioning within species is a strategy used by reptiles to maximize coexistence among individuals (e.g. Marques et al. 2013a, Richards-Dimitrie et al. 2013). Species exhibiting sexual size dimorphism may have sexual differences in nutritional requirements, which can result in the use of different resources (Wearmouth and Sims 2008). In addition, the increase in body size throughout the ontogenetic development is also an important factor that influences the exploitation of food resources (Werner and Gilliam 1984).

Stable isotope methodology has been used to identify overlaps or partitioning of resources within species, since it provides quantitative information about food resource consumption and habitat use (Newsome et al. 2007, Layman et al. 2012). This methodology has advantages over stomach or feces content analysis: it is less traumatic to the animals, it provides a time-integrated measure of diet, and it is not biased toward hard or less digestible food items (Bulté et al. 2008). These characteristics have made it possible to apply isotopic niche approaches toward a better understanding of the interactions among organisms (e.g., Jackson et al. 2011). Isotopic niche is defined as the hypervolume of the isotopic space (n-dimensional), where the axes are the isotopic compositions that represent the bionomic and scenopoetic components of the niche (Newsome et al. 2007). This approach has already been successfully used both to understand the variations in resource exploitation by Chelodina longicollis in a natural-urban gradient in southeastern Australia (Ferronato et al. 2016), and to identify the impact of the invasive species Trachemys scripta elegans (Wied-Neuwied, 1839) on Emys orbicularis (Linnaeus, 1758) in Italy (Balzani et al. 2016).

Mesoclemmys vanderhaegei (Bour, 1973) is a medium-sized chelid turtle with limited geographic distribution in South America: Amazonas, Tocantins, Paraguay, Parana, and Uruguay river basins (Marques et al. 2014). The species has been documented utilizing several different habitat types, including lagoons, small streams and urban/anthropogenic watercourses (Brandão et al. 2002, Brito et al. 2012, Marques et al. 2013b); however, little is known about its biology and ecology (Brito et al. 2009, Marques et al. 2014).

Here we applied stable isotope methodology to ascertain the intraspecific trophic niche variation in M. vanderhaegei at Serra das Araras Ecological Station, state of Mato Grosso, Brazil. First, we compared the isotopic niche width and the degree of overlap between males and females. Additionally, we tested the relationship between isotopic compositions (δ¹⁵N and δ¹³C) and body size.

The data for this study was collected on a reservoir at Serra das Araras Ecological Station (SAES; 15°49'31"S, 57°17'14"W, altitude: 800 m), a protected area in the state of Mato Grosso, Brazil. SAES is located between the cities of Porto Estrela and Cáceres and is dominated by savanna (Total area: 28,700 ha; Fig. 1). The small stream (first order) that originates the dam is oligotrophic, with sandy bottom, aquatic vegetation, periphyton and grass, which are partially submerged in the rainy season.

Mesoclemmys vanderhaegei individuals were captured with funnel traps (1.2 m long × 0.60 m external diameter × 0.30 m entrance diameter; plastic mesh 5.0 × 1.0 mm), between November 2010 and November 2011. The animals were attracted into the funnel traps by a mixture of bovine meat and commercial fish-flavored pellets. Trapped individuals were marked by notching the marginal scutes of their the carapaces (Cagle 1939) and were sexed based on the external examination of secondary sexual characteristics (Brito et al. 2009). All individuals had their straight line carapace length (Stainless Hardenedde caliper, millimeters) and body mass (Pesola, grams) taken. Claw samples (terminal 5 mm) were collected from individuals for the isotopic analyses. The animals were released where they had been captured.

Different types of tissue reflect distinct scales of exploitation of food resources (Dalerum and Angerbjörn 2005). Tissues with high turnover rates reflect the animal’s recent diet, however, inert tissues (e.g., claw) reflect the animal’s diet at the time those tissues were synthetized (Ethier et al. 2010). Therefore, inert tissues are considered good indicators of the general diet of animals (Bowen et al. 2005). Claw tissue can reflect changes in the diet of the freshwater ‘turtle after six months for d¹⁵N and greater than six months δ¹³C (Aresco et al. 2015).

The claw samples were washed with distilled water, dried at 60°C, fragmented to the smallest possible size and placed (0.8–1.0 mg) in small tin capsules, where they were heated in a Carlo Erba elemental analyzer (CHN-1110) coupled to a Delta Plus mass spectrometer in the Laboratório de Ecologia Isotópica, Centro de Energia Nuclear na Agricultura, Universidade de São Paulo, Brazil. The isotopic composition of carbon and nitrogen was expressed in delta (δ) per mil (‰) as folows: δX = (R_sample/R_standard – 1) x 1000; where R are ratios of heavy to light isotopes (¹⁵N/¹⁴N or ¹³C/¹²C) in the sample and standard. The analytical error estimated by repeated measures of internal standard (sugarcane) was 0.5‰ for δ¹⁵N and 0.3‰ for δ¹³C. The international standards used for nitrogen and carbon were atmospheric air (AIR) and Pee Dee Belemnite (PDB), respectively.

Normality data and homoscedasticity were tested prior to statistical analyses by Anderson Darling’s test and Levene’s test, respectively. The t-test was used to investigate sexual differences in biometric measurements (carapace length and body mass) and in stable isotope compositions (δ¹⁵N and δ¹³C).

Isotopic niche can be defined as the polygon area formed by the isotopic compositions in δ space (Newsome et al. 2007). Sexual niche widths were evaluated using Bayesian standard ellipse methodology (SEA_B – Jackson et al. 2011). Differences between male and female in niche widths were tested by estimating the number of simulations where one group had a larger SEA than the other. Niche overlap was measured by comparing the extent of male and female overlap of corrected standard ellipses (SEA_c – Jackson et al. 2011). These analyses were performed in the package Stable Isotope Analysis (SIAR – Parnell et al. 2010) of the software R (R Core Team 2013). Linear regression was used to test the relationship to isotopic compositions and biometric measurements (carapace length and body mass). These analyses were performed using Minitab 16 (Minitab Inc., State College, Pennsylvania, USA).

We captured 28 individuals during the study period, 14 males and 14 females. Females had significantly longer carapaces (female = 165.9 ± 12.2 mm; male = 142.1 ± 16.4 mm; t = 4.29; df = 27; p < 0.001) and greater body mass than males (female = 392.9 ± 99.7 g; male = 259.3 ± 73.5 g; t = 4.03; df = 27; p < 0.001).

Niche widths (SEA_B) were not statistically different between the sexes (SEA_B female = 3.76 ‰², 95% Cr.I. = 2.02–5.79 ‰²; SEA_B male = 2.56 ‰², 95% Cr.I. = 1.36-3.93 ‰²; p = 0.145; Fig. 2) and there was a high degree of overlap between the corrected standard ellipses of males and females (M/F: 88.2%; F/M: 50.5%; Fig. 3). The mean values of δ¹⁵N were 8.87 ± 0.54 ‰ (7.89-9.65 ‰) for females and 8.60 ± 0.36 ‰ (8.01–9.19 ‰) for males. The mean values of δ¹³C were -21.19 ± 1.88 ‰ (-25.68–17.58 ‰) for females and -21.30 ± 1.56 ‰ (-24.54–18.22 ‰) for males. There was no difference for both δ¹⁵N (t = 1.45; df = 27; p = 0.162) and δ¹³C (t = 0.15; df = 27; p = 0.879) between sexes.

No significant relationship was found between the values of δ¹⁵N for both carapace length (F = 1.15; df = 27; p = 0.294; r² = 0.04) and body mass (F = 1.83; df = 27; p = 0.188; r² = 0.06). Similar pattern occurred for values of δ¹³C (CL: F = 3.94; df = 27; P = 0.058; r² = 0.13; BM: F = 4.13; df = 27; p = 0.053; r² = 0.13; Fig. 4).

Mesoclemmys vanderhaegei showed consistent sexual dimorphism in body size, corroborating previous studies on the species in others localities (Brito et al. 2012, Marques et al. 2013b), which also found that females are larger than males. However, despite the marked morphological difference between the sexes, their isotopic niche widths did not differ.

Isotopic niche of wildlife species can be influenced by what they consume as well as their habitat (Newsome et al. 2007). The knowledge about the diet of M. vanderhaegei is extremely limited since there have been few studies in natural conditions. This turtle is omnivorous with affinity to carnivory (Cabrera 1998, Rueda-Almonacid et al. 2007). Brito et al. (2016) showed that aquatic insects are the main food item consumed by this species in the Cerrado (frequency of occurrence: Odonata (25%), Diptera (25%), Hemiptera (16%)followed by fish (39%), aquatic plants (9%), fruits (9%), and leaves (6%)). The lack of differences in isotopic niche widths suggests that both sexes can exploit a similar range of food resources available in the study area.

Habitat use can also influence the isotopic niche because animals can move between areas with different isotopic baselines (Hobson 1999, Hobson and Wassenaar 2008). This is the case of crocodilians in the coastal region, which have access to marine and estuarine/freshwater environments (Rosenblatt and Heithaus 2011, Hanson et al. 2015) or felids in agricultural landscapes (forest remnants – C3 x monocultures of sugarcane – C4; Magioli et al. 2014). Males and females of M. vanderhaegei in this study were captured in the same water bodies and have similar δ¹³C values, which suggests that both sexes have the same isotopic baseline. However, future studies should determine isotopic compositions from prey representing the different potential baselines (e.g., benthic, pelagic, terrestrial) in the system, in order to verify their importance in the intersexual niche of M. vanderhaegei.

Several studies indicate ontogenetic differences in the diet of turtles (e.g., Clark and Gibbons 1969, Souza and Abe 1995). Dietary partitioning is a strategy adopted by many species and has the advantage of reducing the intraspecific competition and increasing fitness (Riklefs 2008). However, there was no relationship between isotopic composition and body size in M. vanderhaegei. Nitrogen isotopic composition (δ¹⁵N) indicates the trophic level of an organism, whilecarbon isotopic composition (δ¹³C) indicates the different carbon sources of diet exploited by individuals (Minawaga and Wada 1984, Fry 2006). Future long-term studies should understand how the trophic niche of M. vanderhaegei vary in space (e.g., natural and man-made habitats) and time (e.g., seasonaly).