The freshwater chelid turtle Phrynops geoffroanus has a wide geographic distribution in South America (Pritchard and Trebbau 1984; Ernst and Barbour 1989; Vanzolini 1994; Souza 2005). The species inhabit streams, large rivers, ponds, and lakes (Souza and Abe 2000) and is also found in urban freshwater environments, including polluted urban rivers and streams (Souza and Abe 2000, 2001; Marques et al. 2008). According some studies, P. geoffroanus may be considered one of the most common components of the neotropical urban fauna (Souza and Abe 2000, 2001).

The ability of P. geoffroanus to live in urban environments makes the species an ideal model to study the effects of freshwater habitat degradation on some aspects of freshwater turtle natural history. According to previous studies, the diet of the species seems to vary across different habitat features (Medem 1960; Molina 1990; Fachin-Teran et al. 1995; Souza and Abe 2000; Dias and Souza 2005).

Despite the high density in urban rivers and the species' wide geographic range, only 1 study described the feeding habits of P. geoffroanus in anthropogenic habitats (Souza and Abe 2000). The aim of present study is to describe the items present in the stomach contents of P. geoffroanus captured at Anhanduizinho River, an urban freshwater water body in Campo Grande, Mato Grosso do Sul.

Methods

We caught 102 P. geoffroanus using gill nets extended between the river banks (Souza and Abe 2001) of a nonchannelized 300-m stretch of the Anhanduizinho River (AR). The study site is a polluted urban water body of the Paraná River basin located in Campo Grande, Mato Grosso do Sul (lat 20°28′S, long 54°40′W), central Brazil. The animals were captured monthly, from September 2004 to April 2005. The typical natural vegetation of the region is the savanna-like vegetation known as cerrado; however, it was totally modified through the AR margins, mainly along the channelized portions. Along the nonchannelized portions, the riverbanks are covered by grass and extend by 10–20 m with a slope of ∼ 20–30 degrees. The AR is one of the main rivers of the city and has significant discharge of human sewage and domestic waste into its waters.

Turtle straight-line carapace length (CL) was measured with metric tapes (0.5 mm) and turtles marked by notching marginal scutes (Cagle 1939) as part of a monitoring project developed in the AR. Each turtle was identified as juvenile, male, or female according to external morphological characteristics (Molina 1989; Souza and Abe 2001). According to Souza and Abe (2001), adult P. geoffronaus shows secondary sexual characteristics at 200–210 mm CL; thus, individuals less than 200 mm may be considered juveniles.

Stomach contents were obtained by stomach flushing (Legler 1977) during 2 hours after capture. Samples were fixed in 10% formalin, preserved in 70% ethyl alcohol, and sorted and identified to the lowest possible taxon under a stereomicroscope. Total individuals of each prey item were counted, and their volume was measured by water displacement (Bjorndal et al. 1997). Variables used for descriptions were frequency of occurrence (%F, percentage of individuals turtles in which a given food item was found), numeric frequency (%N, number of individual prey of a given taxon found as percentage of all animal food items), and volume percentage (%V, volume percentage of individual prey category in relation to all food items detected). The difference between the stomach content volumes of juveniles and adults was tested by Mann-Whitney U-test and considered the total volume of prey items verified in each stomach content.

Results

From 102 captured turtles, we obtained 39 stomach contents (27 adults and 12 juveniles). The temporal distribution of samples and the numeric frequency of the principal prey ingested by P. geoffroanus are shown in Fig. 1. The diet of P. geoffroanus at AR is composed primarily of immature Chironomus sp. (mainly larvae), but the species also ingests naiads, coleopterans, gastropods, fish, plant material, and unidentifiable organic debris (Table 1). The plant material (leaves and branch fragments) had low observed values of %F and %V, and they were probably accidentally ingested. There was no significant difference between the stomach content volumes of adults and juveniles (U  =  112; p > 0.05; Fig. 2).

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Table 1 Composition of stomach contents of Phrynops geoffroanus from the Anhanduizinho River, Mato Grosso do Sul, Brazil. Data refer to all stomach contents sampled (n  =  39) and according to sex or life stages: M, males (n  =  17); F, females (n  =  10); J, juveniles (n  =  12). %F, occurrence frequency; %N, numeric frequency; %V, volume frequency; UM, unidentified material (as described in the text).

View Fullscreen

Discussion

Our study verified the predominance of insects in P. geoffroanus stomach contents. The plant material was probably ingested accidentally. Brazilian chelids seem to be omnivores that feed on a wide range of prey, from vegetables (fruits and seeds) to vertebrates (fish, amphibians, and lizards), with a clear predominance of insects with an aquatic larval phase (Souza 2004). The studied population of P. geoffroanus fits such a dietary characterization.

With respect to the main ingested prey, the cosmopolitan chironomid flies are holometabolic insects with a larval aquatic phase. These immature insects may reach high densities in sediments rich in organic matter (Marques et al. 2008) and play an important role in freshwater food chains, representing the main connection between producers and primary consumers (Tokeshi 1995). In fact, chironomids are one of the most abundant aquatic insects in benthic arthropod communities (Roque et al. 2003), acting as an important food resource for many species around the world, such as fish (Mackey 1979; Teixeira and Bennemann 2007; Rolla et al. 2009), turtles (Georges et al. 1986; Souza and Abe 2000), and salamanders (Muenz et al. 2008). Although prey availability was not studied by us at AR, we note that chironomids should be abundant at this site since it had obvious pollution caused by human sewage and domestic waste (Souza et al. 2008).

Souza and Abe (2000) observed high values of Chironomus (C. cf. plumosus) in the diet of P. geoffroanus captured at Ribeirão Preto stream, a polluted urban river in São Paulo state. These authors hypothesized that P. geoffroanus may take some advantage of the urban environment, given the absence of predators and food abundance. The predominance of Chironomus sp. in the diet of P. geoffroanus at AR also suggests a high abundance of flies in the studied environment, which may be advantageous for the turtles, given the high values of this animal prey in the analyzed stomach contents. However, at the AR, the real impacts of pollution and other human interferences on P. geoffroanus feeding habits remains unknown. Studies on the influences of habitat degradation on the feeding habits and other biological aspects of P. geoffroanus deserve future attention.

Besides the predominance of aquatic stages of Chironomus sp. in the diet of P. geoffroanus, the present study also described the presence of other invertebrates, fishes, and plant material in the stomach contents of turtles. Studies on the feeding habits of P. geoffronaus indicate a high plasticity in diet (Medem 1960; Molina 1990; Fachin-Teran et al. 1995; Souza and Abe 2000; Dias and Souza 2005). In captivity, P. geoffronaus showed an exclusively carnivorous diet, with no differences between adults and juveniles. In captivity, both juveniles and adults feed manly on meat (bovine), fish, and insects, with no ontogenetic diet variation (Molina 1990). A study on the food habits of an assemblage of 5 species of freshwater turtles in Rondonia, a region dominated by the Amazon forest, showed a high diverse diet for P. geoffroanus, which includes algae, seeds and fruits, plant material, insects, fish, crabs, and shrimp (Fachin-Teran et al. 1995). Dias and Souza (2005) described a diet composed of crustaceans, fish, and aquatic insects for P. geoffroanus captured at an impoundment on the Paraná River, São Paulo. In an urban environment at Ribeirão Preto, São Paulo, P. geoffroanus fed on large quantities of Chironomus cf. plumosus and also on terrestrial insects (like cockroaches), mollusks, carrion, and organic debris (Souza and Abe 2000). The presence of organic debris was also observed in the present work as unidentified material, which, as proposed by Souza and Abe (2000), may have originated from carcasses and domestic waste, materials frequently observed in the AR.

In the present work there was no statistical difference between the stomach content volumes of juveniles and adults. Studies on the feeding habits of freshwater turtles have shown a tendency for juveniles to be more carnivorous than adults, with clear ontogenetic changes in diet (Georges et al. 1993; Souza and Abe 1998; Souza 2004). In carnivorous species, juveniles ingest a higher prey number than adults (Chessman 1983; Souza and Abe 1998). This pattern is explained by the hypothesis that juveniles must grow faster than adults, given their higher susceptibility to predators due their small size and thin carapace (Georges 1993; Souza 2004; Martins and Souza 2008). Probably the absence of observed difference between the stomach contents of adults and juveniles in the present work was due to the inefficiency of stomach-flushing technique in acquiring all items presents within the stomachs of studied turtles. However, if the main item ingested by P. geoffroanus in present study (Chironomus sp.) produce a satisfactory energetic return, then limited amounts of this resource would be adequate for fast growth. Studies on the nutrient composition of larvae and pupae of Chironomus genus deserves future attention since this insect group is a common resource used by many freshwater predators (Mackey 1979; Georges et al. 1986; Souza and Abe 2000; Teixeira and Bennemann 2007; Muenz et al. 2008; Rolla et al. 2009).