Map turtles and sawbacks (genus Graptemys) represent the most diverse turtle genus in the United States. Fourteen species of Graptemys have been described, most of which are considered riverine (Lindeman 2013). The genus includes species limited to a single river and its drainages (e.g., Pearl map turtle, G. pearlensis) and others with much wider distributions (e.g., northern map turtle, G. geographica,Lindeman 2013). Graptemys spp. fill a broad range of ecological niches and many species are sexually dimorphic, with females having wide heads (i.e., megacephaly) and reaching greater carapace length than males (Lindeman 2013). Increased head width and body size allows for ingestion of larger mollusk prey (Bulté et al. 2008).

Diet shifts of animals, including turtles, in the presence of invasive species are a well-known phenomena (Ernst and Lovich 2009; David et al. 2017). The impact of invasive species on native consumer populations has been understudied; however, there is evidence that long-lived organisms can use and benefit from novel prey sources at least in the short-term (i.e., growth rate, body size; Maerz et al. 2005; King et al. 2006). The extent and impact of introduced bivalves on aquatic ecosystems has been a major focus in freshwater biology (Strayer 1999; Sousa et al. 2008; Carlsson et al. 2011). Among these invaders is the Asian clam (Corbicula fluminea), introduced into the United States in the 1950s and successful because of its explosive hermaphroditic reproduction, multiple feeding strategies, and ability to invade disturbed aquatic habitats (Kraemer 1979; Sousa et al. 2008; Atkinson et al. 2010). In its native range of Southeast Asia, this small bivalve lives in both lentic and lotic systems; however, it is a rampant invader of rivers throughout the southeastern United States (McMahon and Bogan 2001). Diet shifts from native species to invasive mollusks have been reported for several Graptemys species: G. ernsti (Shealy 1976), G. geographica (Moll 1977), G. pseudogeographica kohnii (Shively and Vidrine 1984), G. caglei (Porter 1990), G. versa (Lindeman 2006), G. gibbonsi (Selman and Lindeman 2015), and G. pulchra (Lindeman 2016).

Female Barbour's map turtles (G. barbouri) exhibit the most extreme megacephaly and the species has the greatest sexual size dimorphism among the 5 species in the megacephalic clade in the genus (Lindeman 2013). Graptemys barbouri occurs in Apalachicola drainages including the Flint, Chattahoochee, and Chipola rivers in southwestern Georgia and north Florida and recent records exist for the Ochlockonee, Choctawhatchee, and Wacissa rivers (Lindeman 2013). The species is among the least-studied of the map turtles (Lovich and Ennen 2013; but see Sterrett et al. 2010a, 2010b, 2015) and uncertainty about the status of populations in light of land use change and declining populations of native mussel species in the Apalachicola drainage led to a petition to list the species as Threatened under the US Endangered Species Act in 2011 (Center for Biological Diversity 2011).

Corbicula fluminea was first reported as established in the Flint River drainage in 1967 and by 1991 was reported as widespread across Georgia (Slapcinski 2014). Earlier studies of diet for female G. barbouri in southwest Georgia and northwest Florida had found the dominant prey to be native gastropods and bivalves (Cagle 1952; Sanderson 1974). Herein, we present a more recent description of the dietary components of adult female G. barbouri, demonstrating an apparent shift in the molluscan diet of females of this species from native prey to C. fluminea.

Methods. — This study took place on Ichawaynochaway Creek, a tributary of the Flint River, in Baker County, Georgia. Female map turtles were hand-captured by snorkelers from 2004 to 2009 in coordination with a study on radiotelemetry (Sterrett et al. 2015). Turtles were transported to a laboratory at the Jones Center at Ichauway, where they were measured (maximum carapace length), given a unique identification code by marking the marginal scutes (Cagle 1939), and placed individually in clean plastic containers (55 × 40 × 42 cm). Enough tap water was added to tanks to reach the underside of the carapace and turtles were left in the container overnight or until a fecal sample was present (< 24 hrs). All contents of the fecal sample were strained and sorted using a stainless steel mesh sieve (500 µm) and stored in 70% EtOH. No recaptured individuals were resampled for fecal samples. All turtles were released at their points of capture within 48 hrs.

Fecal samples were identified and sorted by prey type categories (e.g., native or nonnative bivalve, gastropod, plant material, other macroinvertebrate, and nonfood items such as rocks and twigs). The volume of each prey type for each individual turtle was determined using water displacement (0.01 ml). An index of relative importance for each prey category i (IRI_i) was calculated according to Bjorndal et al. (1997):

where F_i = frequency of occurrence and V_i = mean volumetric percentage. An IRI is a standardized, composite measure of diet that takes into account the frequency and volume of diet items among individuals of a species and ultimately represents the importance of each diet item to the diet of species in a population, which is also a measure that is comparable among diet studies (Bjorndal et al. 1997; Hart et al. 2002).

Results and Discussion. — We collected fecal samples from 35 female G. barbouri. The mean size of 34 individuals was 256.31 ± 49.29 mm maximum straight-lined carapace length (range = 115.9–308.0 mm); 1 individual was not measured. Fecal material of female G. barbouri in Ichawaynochaway Creek was predominantly composed of shells of C. fluminea (Table 1). The IRI for C. fluminea was 98 versus 2 for gastropods (Elimia spp.; Table 1). Corbicula fluminea was found in 94% of samples and composed on average 87% of the volume of samples across individuals (Table 1). In contrast, native gastropods (Elimia spp.) were found in 40% of samples and constituted only 4% of the volume of fecal samples (Table 1). Other items in the fecal material of females included native bivalves (a single valve of Villosa vibex), in addition to leeches, plant material (e.g., small twigs), and other materials such as rocks, sediment, and interestingly, a shed snake skin. Studies that examine fecal contents are biased toward detecting hard-bodied prey items and it is unclear whether soft-bodied prey items were not detected in this study as a result of methodology; however, there is little evidence that female G barbouri actively feed on soft-bodied prey items (e.g., insects) and insect exoskeletons do readily pass in the feces in studies of other species. We found no relationship between body size and C. fluminea volume in samples (r² = 0.012, p = 0.54).

Table 1. Diet data from fecal material samples of 35 female Barbour's map turtle (Graptemys barbouri) from Ichawaynochaway Creek in southwestern Georgia, including the percent occurrence, mean percent volume, and the index of relative importance.

View Fullscreen

In the absence of data on changes in prey abundance and availability, we cannot explain the apparent dietary shift in female G. barbouri from gastropods to C. fluminea. However, the overwhelming density of C. fluminea in aquatic ecosystems, including rivers, has likely made C. fluminea a locally abundant and widely distributed novel food source (Sousa et al. 2008). Corbicula fluminea can reach extremely high densities (up to 483/m² in Piedmont rivers of Georgia; McDowell and Byers 2019) and, while density estimates are not available for Ichawaynochaway Creek, they have been estimated at 55/m² in a reservoir south of Ichawaynochaway Creek (Patrick et al. 2017). Also, C. fluminea are relatively small compared with native bivalves, which allows female G. barbouri to easily crush even the largest individuals with their muscular jaws (S.C.S., pers. obs.). In contrast, gastropods were found to make up only 4% of the volume of diets of female G. barbouri. Surprisingly, only one native bivalve was found in the samples, despite reportedly high diversity of mollusks in Ichawaynochaway Creek (Smith et al. 2015). Corbicula fluminea likely competes for the same resources and has a more flexible diet than some native mollusks in Ichawaynochaway Creek (Atkinson et al. 2010). However, it is also possible that competition for space may result in high densities of C. fluminea (i.e., burrowing activity and displacement), which may influence native mollusk populations (Sousa et al. 2008), resulting in the use of C. fluminea by G. barbouri. Further studies will be required to understand how other factors such as prey density, handling time, and nutritional quality play into the switch to this relatively novel and currently abundant diet source.

Many aquatic predators have responded positively to C. fluminea invasions and some may even play a role in controlling this invasive species. Fish community predation had a 29-fold negative effect on abundance of C. fluminea in a lake system in the southern United States (Robinson and Wellborn 1988). Furthermore, myriad fish species are known to consume C. fluminea (Garcia and Protogino 2005; Cantanhêde et al. 2008). However, freshwater turtles (Graptemys spp., Sternotherus spp.) have received less recognition for their role as predators of Asian clams (but see Atkinson 2013 and Lindeman 2013). Many female Graptemys spp. have strictly molluscivorous diets as adults and have shifted their diet to consuming exclusively C. fluminea since their introduction in the southeastern United States (Lindeman 2006, 2013). Understanding how freshwater turtles affect invasive bivalves is vital to conservation of freshwater systems (Moll 1977; Lovich et al. 2018). Additionally, it is unknown how novel food resources like exotic mollusks affect native turtle populations, which presents an important opportunity for future studies.