Humans have modified the natural landscape at global scales through the damming of rivers, the deforestation of natural habitats, and the exploitation of other natural resources. These anthropogenic modifications have ushered in a sixth mass extinction and the modern geological age, the Anthropocene (Dirzo et al. 2014). Over the past 65 million years, the rate of extinction was about 0.1–1 species per million species per year; however, for several vertebrate groups, the extinction rate is markedly higher (Dirzo and Raven 2003; De Vos et al. 2015). One such group of vertebrates, Testudines (turtles), are considered the most imperiled vertebrate order, with 61% of turtle species currently threatened by extinction (Turtle Taxonomy Working Group 2017; Rhodin et al. 2018).

The western alligator snapping turtle (Macrochelys temminckii) is a large freshwater turtle in North America (Ernst and Lovich 2009). This species is highly aquatic and rarely leaves the water except for nesting. Macrochelys temminckii can be found in drainages along the northern Gulf of Mexico from Texas to the Florida panhandle, with records as far north (historically) as Illinois and Indiana (Minton 2001; Kessler et al. 2017). Macrochelys temminckii inhabits rivers, oxbows, and sloughs throughout its distribution (Ernst and Lovich 2009). They consume a wide variety of animals (i.e., crayfish, insects, mollusks, mammals, and birds), but the diet largely consists of fish (Elsey 2006), although they are known to also consume considerable amounts of plant matter (e.g., acorns, Sloan et al. 1996).

Throughout their range in the southeastern United States, alligator snapping turtles have experienced dramatic declines that are directly related to overexploitation for human consumption accompanied by habitat loss (Pritchard 1989; Sloan and Lovich 1995; Jensen and Birkhead 2003; Folt and Godwin 2013; Howey and Dinkelacker 2013). While they are currently not federally protected, the US Fish and Wildlife Service recently released its 90-d findings suggesting that M. temminckii warrants further consideration for federal listing (Mires 2015). This species is also included in Appendix III of the Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES 2006). for protection against overexploitation. In Tennessee, M. temminckii has state protection as threatened, and the commercial and recreational harvesting of this species has been illegal since 1991 under Proclamation 91-30 (Tennessee Fish and Wildlife Commission 1991).

Very little is currently known about the status and distribution of M. temminckii in Tennessee. Unfortunately, the effects of past harvesting pressures and habitat degradation on population abundances and demography cannot be evaluated due to a lack of historical data on harvesting within the state. Relative to other turtle species, M. temminckii has very few verified localities (i.e., 64 total) in Tennessee, with most of these located within the lower Tennessee River of Middle Tennessee (Scott and Redmond 2020). For example, Tennessee Wildlife Resources Agency (TWRA) captured 2 adults in the Beech River in 2019. Other more recent records are of adult carcasses or by-catch from commercial fishermen (Scott and Redmond 2020). A carcass (carapace length = 49.53 cm) was recovered in 2017 from Center Hill Lake, a reservoir created on the Caney Fork River of the Cumberland River drainage. In 2020, an adult was captured on a trout line in Indian Creek. Relatively few verified records are known from Tennessee despite the great potential for habitat (sloughs, slow-moving rivers of the coastal plain such as the Wolf and Hatchie rivers; Scott and Redmond 2020). In general, the lack of records or reported sightings indicates that M. temminckii populations are probably reduced in size (i.e., number of individuals) or extirpated in Tennessee (Pritchard 1989; Ernst and Lovich 2009).

Despite the lack of knowledge on the status and distribution of this species within Tennessee, especially the western portion of the state, TWRA proactively released 425 adults and juveniles at 14 different sites throughout western Tennessee over a 7-yr period (2000–2007) to bolster populations (R.C., unpubl. data, 2020). The release locations were selected based on habitat suitability. Most reintroduced and stocked individuals were acquired from turtle farms in Louisiana. Since the release of these individuals, only one study has investigated postrelease success of these individuals for approximately 1 yr (Ream 2008). Therefore, no assessment of the stocked individuals has since been conducted and the success of the stocking program remains unknown. To date, a systematic study assessing the population status of M. temminckii in western Tennessee and the success of the reintroduction/stocking is needed before the development of an effective conservation and management plan at the state and federal level.

We aimed to assess the distribution of M. temminckii by conducting trapping surveys throughout western Tennessee. We sampled 65 unique locations in 13 counties within the Mississippi River drainages of western Tennessee (Fig. 1). Our sampling efforts covered 11 of the 14 known sites where TWRA released hatchery-reared individuals, allowing us to assess whether stocked individuals had survived. Additionally, we estimate catch-per-unit-effort (CPUE) for M. temminckii to compare with other studies from the surrounding states with known populations.

View Figure

• Download as Powerpoint
• Download Figure

METHODS

From 2016 to 2018, we trapped 65 sites across 4 drainages from 13 counties in western Tennessee (Fig. 1). To standardize trapping effort at each site, we deployed 10 baited hoop nets for 3 trap nights (approximately 72 hrs), for a total of 30 net nights per site. The exceptions to this protocol were limited to 2 sites (Gated Mineral Slough and Miss-Hatchie Hunting Club), which had only 2 trap nights due to mechanical issues with the boat or severe weather. Additional sites (Ballard Slough and Running Reelfoot Bayou) were trapped with more than 10 traps or on multiple occasions due to the known presence of M. temminckii (i.e., pictures from local fisherman). Three sites within the Hatchie National Wildlife Refuge (Bullpen Lake, Hart Lake, and Swan Lake) were trapped with fewer than 10 traps due to the small size of the oxbow.

We used hoop nets of various sizes, including 91-cm hoop nets with 10-cm mesh, 91-cm hoop nets with 4-cm mesh, and 122-cm hoop nets with 4-cm mesh, to reduce bias in trapping of all size classes (Ennen et al. 2021). We set traps near submerged woody vegetation and structures, when available at the sites, and allowed the traps to protrude from the water to allow turtles to surface. Once traps were deployed, they were checked every 24 hrs, usually in the morning. We set on average 10.0 traps (1.27 standard deviation [SD], range 6–18 traps per site) and checked them for roughly 3 days (mean = 2.95 days, SD = 0.21, range = 2–3) for a total of about 30 trap nights per site (mean = 29.46, SD = 2.98, range = 18–36). The variation in these values was due to uncontrollable factors such as severe weather events and stolen traps. We used 5, 91-cm hoop nets with 10-cm mesh at all sites and randomly varied the number (i.e., either 2 or 3) of 91-cm hoop nets with 4-cm mesh and 122-cm hoop nets with 4-cm mesh traps per site. All traps were baited with various types of fish provided by TWRA and local fish markets. All traps were rebaited daily regardless of leftover bait.

We estimated CPUE 2 ways. For comparison to other studies on M. temminckii (e.g., Baxley et al. 2014; Huntzinger et al. 2019), we first estimated CPUE for M. temminckii as total number of individuals captured by the product of number of traps deployed and number of trap nights over the entirety of the survey. Finally, we estimated CPUE by site for each species and calculated the mean and SD across all sites.

For all captured M. temminckii, we measured the following morphological variables: mass (kg), straight-line carapace length (cm), and precloacal tail length (cm). All individuals were then pit tagged and released at the point of capture. All morphological measurements and CPUE calculations were summarized using means and SDs. A tissue sample was taken from each individual and stored in 95% ethanol.

Many of the reintroduced and stocked individuals were released without being permanently marked, thus distinguishing between native and introduced individuals was not possible morphologically. Because of this, we collected tissue samples (i.e., skin) from captured M. temminckii and compared genotypes of trapped individuals from western Tennessee with known Louisiana (Calcasieu and Atchafalaya rivers) and Mississippi (Big Black River) individuals. Genomic DNA was extracted from each tissue sample using a DNeasy Tissue Kit (Qiagen Inc, Valencia, CA) and amplified using microsatellite primers developed and optimized for use with M. temminckii (Mtem109 and Mtem111, Hackler et al. 2007; Mtem007, Mtem013, Mtem017, Mtem019, Mtem021, Mtem025, Mtem026, Mtem028, Mtem029, Mtem030, Mtem033, and Mtem035, L.P., unpubl. data, 2020) using published methods (Selman et al. 2013). Amplifications of samples were conducted in a total volume of 12.5 µl using 8.195 µl of dH₂O, 1.25 µl of 10× standard Taq (Mg-free) buffer (New England BioLabs, Ipswich, MA), 0.25 µl of 2 mM dNTPs, 1 µl of 25 mM MgCl₂, 0.08 µl of Taq polymerase, 0.20 µl of 10-µM concentrations of the M13 tailed forward primer and the reverse primer, 0.075 µl of 0.3 µM of labeled M-13 primer (Eurofins, Louisville, KY), and 1.25 µl of 20–50 ng/µl DNA template. Polymerase chain reaction (PCR) cycling conditions were as follows: initial denaturation at 94°C for 2 min, 35 cycles of 30 sec at 94°C, 30 sec at 56°C, and 1 min at 72°C, with a final elongation of 10 min at 72°C. Microsatellite alleles were visualized on an acrylamide gel using a LICOR 4300 DNA Analyzer and allele sizes were determined using GeneProfiler version 4.05 (LICOR Co., Lincoln, NE). Loci were tested for Hardy-Weinberg equilibrium (HWE) and linkage disequilibrium (LD) using ‘genepop’ (Raymond and Rousset 1995; Rousset 2008) with a sequential Bonferroni correction (Rice 1989) applied to both tests.

We used STRUCTURE 2.3.4 (Pritchard et al. 2000) to determine the number of genetic groups represented by our samples from Tennessee, the Big Black River, and Louisiana (Calcasieu and Atchafalaya rivers). The Big Black River site represented another native population from the Mississippi River system in close proximity to the Tennessee sites, while the Louisiana sites represented areas where the turtle farms might have acquired their broodstock. We excluded Tennessee sites with low sample sizes (Running Reelfoot Bayou and Ghost River section of the Wolf River, each n = 1). We tested values of K from 1–8 using a model of admixed ancestry and assuming correlated allele frequencies between groups (Hubisz et al. 2009). For each value of K, 20 replicates were performed with a burn-in of 50,000 generations followed by 200,000 subsequent generations. The best value of K was determined by comparing the average log-likelihood scores for each K and the ΔK method (Evanno et al. 2005) as calculated by StructureSelector (Li and Liu 2018). We subsequently used CLUMPP 1.1.2 (Jakobsson and Rosenberg 2007) to average the STRUCTURE output for the best value of K, while Distruct 1.1 (Rosenberg 2003) was used to create a graphical presentation of the average ancestry score for each individual.

RESULTS AND DISCUSSION

Our trapping efforts of 1915 net nights and 192 survey nights (mean 2.95 nights per site) resulted in 46.54 ± 54.26 (mean ± SD) individual turtles of all species per site and 1.60 ± 1.83 CPUE per site. By far, Trachemys scripta (pond slider) was the most abundant (1.34 ± 1.70) and widespread (captured in 89% of locations of trap arrays) species of freshwater turtle in west Tennessee followed by Chelydra serpentina (common snapping turtle), Graptemys pseudogeographica (false map turtle), and Apalone spinifera (spiny softshell turtle), respectively (Table 1). These abundance patterns were like those reported in a western Kentucky study focusing effort on M. temminckii (Baxley et al. 2014), where T. scripta was the most abundant species followed by C. serpentina, A. spinifera, and G. pseudogeographica. These findings were similar to turtle assemblages in Illinois (Bluett et al. 2011). In coastal plain drainages of Louisiana and Georgia, T. scripta was still the dominant turtle species captured but other species of freshwater turtles, such as Sternotherus carinatus (razor-backed musk turtle) and Pseudemys concinna (river cooter), had relatively high abundances followed by A. spinifera and M. temminckii (King et al. 2016; Huntzinger et al. 2019).

Table 1. Catch-per-unit-effort (CPUE; mean ± SD) and number of sites found for each of the 11 turtle species captured in west Tennessee from 2016 to 2018 across 65 locations. The drainage column indicates if the species was present in the following drainage: Forked Deer (F), Hatchie (H), Mississippi (M), Obion (O), and Wolf (W).

View Fullscreen

We captured a total of 22 unique M. temminckii from 4 sites (6% of locations of trap arrays) throughout western Tennessee for an overall CPUE of 0.010 ± 0.052 SD (Table 1) and a CPUE of 0.167 ± 0.157 SD for occupied sites. We confirmed the presence of M. temminckii in 2 locations within the Wolf River drainage (Ghost River and Mineral Slough), 1 location within the Obion River drainage (Running Reelfoot Bayou) and 1 location in the flood plain of the Mississippi River (Ballard Slough); none were captured within the Hatchie and Forked Deer drainages (Table 2). In general, M. temminckii are more abundant in southern portion of their range and are considered rare and occurring in low abundances in Tennessee, at the northern periphery of their distribution (Pritchard 1989; Table 3). Besides studies conducted in Kansas (Shipman et al. 1995), Illinois (Bluett et al. 2011), and Kentucky (Baxley et al. 2014), our CPUE is the lowest reported for M. temminckii (Table 3). For comparison, CPUE of M. temminckii reported in the literature ranged from 0.00 to 0.39 with states like Florida (0.25), Arkansas (3 studies: 0.39, 0.28, and 0.23), Oklahoma (0.35), and Georgia (0.20) recording the highest values (Table 3). Only 2 of our sites with M. temminckii, Ballard Slough (0.31) and Mineral Slough (0.30), were comparable to the higher reported CPUE values; whereas the other 2 sites, Ghost River (0.033) and Running Reelfoot Bayou (0.028), were relatively lower (Table 3).

Table 2. Catch-per-unit-effort (CPUE), net nights, and total captures of Macrochelys temminckii by drainage in west Tennessee from 2016 to 2018.

View Fullscreen

Table 3. A comparison of catch-per-unit-effort (CPUE, arranged from highest to lowest) and total captures for Macrochelys temminckii from 17 previous studies and the present study, modified from Baxley et al. (2014). Data for present study appear in bold.

View Fullscreen

Of the 4 sites where we detected M. temminckii, only 2 possessed sample sizes sufficiently large for the STRUCTURE analysis; the remaining 2 sites (Running Reelfoot Bayou and Ghost River portion of the Wolf River) only produced one individual each. Although the ΔK analysis suggested a K of 3, the log-likelihood scores plateaued at K of 4 (mean Ln K = –2469.9). These 4 genetic groups represented the Big Black River, Ballard Slough, Calcasieu, and Atchafalaya sites (Fig. 2). The Mineral Slough site was composed mostly of individuals with ancestry in the Atchafalaya group, although one individual had a much higher ancestry in the Ballard Slough group. The presence of individuals in Tennessee with genetic ancestry in 2 different groups suggests that these sites represent a native population as well as individuals derived from introductions. In particular, the shared ancestry of the Mineral Slough and Atchafalaya sites suggests that some individuals from Mineral Slough were translocated from Louisiana.

View Figure

• Download as Powerpoint
• Download Figure

Macrochelys temminckii individuals in west Tennessee averaged 27.8 ± 6.52 cm in carapace length and 7.02 ± 3.86 kg in mass (Table 4). All individuals were considered female or juvenile based on mature individual carapace and precloacal length using criteria provided by Dobie (1971). The largest individual captured (SCL = 39.40 cm, mass = 14.60 kg) was a female from Ballard Slough. In comparison with other studies, the carapace length and weight were similar for the collective population of M. temminckii in western Tennessee to other locations within their range (Trauth et al. 1998; Jensen and Birkhead 2003; Boundy and Kennedy 2006; Riedle et al. 2008; Folt and Godwin 2013; Folt et al. 2016). Individuals from native populations were larger (SCL = 32.05 ± 4.92 cm, mass = 9.53 ± 2.98 kg) compared with the individuals from the reintroduced population (i.e., Mineral Slough: SCL = 21.66 ± 1.82 cm, mass = 3.39 ± 0.77 kg). The individuals at Mineral Slough were released in 2007 as juveniles (SCL = 6.40 ± 0.8 cm, mass = 0.074 ± 0.027 kg).

Table 4. A morphological comparison of Macrochelys temminckii among the 4 sites in western Tennessee. Morphological measurements are summarized using mean ± SD (range) for samples > 1.

View Fullscreen

Our trapping efforts deliberately targeted reintroduction sites (11 of the 14) from TWRA's efforts between 1992 and 2005. All released individuals were from Louisiana—either hatchery-raised individuals or native Louisiana populations. Of the 11 reintroduction sites, we captured M. temminckii (n = 9) from only one reintroduction site, Mineral Slough. In general, reintroduction and translocation for reptiles are rarely successful (Dodd and Seigel 1991; Germano and Bishop 2009); however, the persistence of reintroduced individuals at Mineral Slough 15 yrs postrelease is encouraging for future reintroductions within Tennessee. The absence of postrelease monitoring at most reintroduction sites in Tennessee inhibits our ability to identify sources of failure and success. We recommend monitoring reintroduced individuals for at least a 4-yr period postrelease to elucidate sources of success and failure because most failures occur within that duration (Bubac et al. 2019). Additionally, we recommend correcting larger issues, such as habitat loss and degradation, before initiating another reintroduction program. How these larger issues, which are related to the initial extirpation or species declines, are addressed is ultimately the best predictor of success or failure for reintroductions (Bubac et al. 2019).

Our study provides evidence that M. temminckii is not widely distributed or abundant in western Tennessee, but research is still needed in middle Tennessee to determine the state status of this rare species. Unfortunately, there are no historical trapping data for M. temminckii available to accurately estimate population declines or trends in any part of Tennessee. It is entirely possible that M. temminckii were never common in Tennessee or further north (Minton 2001; Baxley et al. 2014; Kessler et al. 2017). Nonetheless, it is likely that abundances have decreased through time due to habitat modifications (i.e., channelization for agriculture) and potentially to fishing pressures of limb lines and bycatch (Folt and Godwin 2013; Huntzinger et al. 2019). Several of the major drainages in western Tennessee—Obion, Forked Deer, and Loosahatchie—are channelized. In general, channelization has a negative impact upon turtle species richness (Vandewalle and Christiansen 1996), and the abundance of several turtle species can be decreased (Moll 1980; Reese and Welsh 1998). These reductions in species richness and abundance via channelization are attributed to reductions in food resources and critical habitats (e.g., nesting beaches and sand bars, overwintering habitat; Bodie 2001). However, data on the impacts of stream and landscape modification on M. temminckii in western Tennessee are not currently available. Finally, we find that past reintroduction efforts have likely established one population, which suggests that successful reintroduction is possible, with more research needed to develop an effective reintroduction strategy for Tennessee.