Human-mediated transport of aquatic nonindigenous species (NIS) has been reported extensively in the literature, particularly through transport in the ballast of large, transoceanic ships (Bax et al. 2003; Costello et al. 2007). However, the transport of NIS may also be facilitated when they move from one aquatic system to the next via lock and canal systems (Hulme 2009). Some movements may be innocuous and NIS become naturalized to their new aquatic environments while other NIS become invasive and begin to harm invaded habitats. These species usually cost millions of dollars to control, with examples including sea lampreys (Petromyzon marinus) and zebra mussels (Dreissena polymorpha) in the Great Lakes of the United States and Canada (Smith 1971; Bax et al. 2003).

The Tennessee-Tombigbee Waterway (TTW) in northeastern Mississippi was constructed between 1972 and 1985 to connect the Tennessee and the Tombigbee River systems. This project was completed to facilitate a shipping connection between the Gulf of Mexico and the interior region of the United States (Stine 1991). During and after construction of the TTW, direct impacts occurred to the native fish (Taylor et al. 2008) and mussel communities (Williams et al. 2008; Hamstead 2013). Along with direct impacts to aquatic organisms, the TTW also made a direct path to connect 2 diverse and unique aquatic faunas historically separated by a high ridge of the Tennessee Valley Divide. Indeed, researchers have documented numerous fishes (inland silverside, Menidia beryllina; weed shiner, Notropis texanus; and blacktail shiner, Cyprinella venusta; Etnier and Starnes 1993; Love and Taylor 2004; Shepard et al. 2009; Strongin et al. 2011) and freshwater mussels (flat floater, Anodonta suborbiculata; pink papershell, Potamilus ohiensis; and mapleleaf, Quadrula quadrula; Hartfield and Jones 1989; Hamstead 2013) that have exchanged between the Tennessee and Tombigbee River systems via the TTW. However, to our knowledge, there has been no documentation to support the movement of aquatic turtle fauna across the TTW. Herein, we report the first documented records of 2 riverine turtles, the Ouachita map turtle (Graptemys ouachitensis) and the Alabama map turtle (Graptemys pulchra), both occurring outside of their native ranges, likely facilitated via the TTW.

OBSERVATIONS

Graptemys ouachitensis. — On 17 August 2005, 3 G. ouachitensis specimens were captured from the northern end of Bay Springs Lake of the Tombigbee River system; specimens were captured via basking traps (similar to described in Jones and Hartfield 1995) near Crow's Neck Boat Ramp (Tishomingo County; white star in Fig. 1). These individuals included an adult male (Mississippi Museum of Natural Science [MMNS] 14460; 10.4 cm midline carapace length [CL]), a juvenile female (MMNS 14461; 10.2 cm CL), and an unsexable juvenile (MMNS 14462; 7.0 cm CL; Fig. 2). These are the first recorded individuals of this species in the Tombigbee River system (Lindeman 2013).

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

The 2 nearest specimens of G. ouachitensis were collected from the Tennessee River system of Mississippi (MMNS 5408, 5670; black star in Fig. 1), both of which were collected from Pickwick Lake near J.P. Coleman State Park (Tishomingo County). These native-range specimens were ∼ 55 river kilometers (rkm) from the 3 individuals captured at the northern end of Bay Springs Lake. Therefore, our observations of G. ouachitensis from Bay Springs Lake in the Tombigbee River system represent a movement of ∼ 40 km through the Divide Section Canal from their native range in Pickwick Lake.

Graptemys pulchra. — On 4 August 2018, W.S. was conducting spotting scope surveys in Tishomingo County, Mississippi, for a related study on Tennessee River drainage map turtles (Graptemys sp.). While surveying the basking structures near the shoreline at Goat Island Recreation Area on Pickwick Lake (white cross in Fig. 1), W.S. observed a female Graptemys basking ∼ 100 m north of the boat ramp (34.956943°N, –88.230961°W). Photographs were taken of the basking female with a Nikon Coolpix p900 and W.S. also took photographs of the individual closer at ∼ 50 m away (Fig. 3). Upon review of the photographs in the camera viewfinder, the visual characteristics did not match any of the native Graptemys species from that portion of the Tennessee River drainage: Graptemys geographica (common map turtle) or G. ouachitensis. Shortly thereafter, L.H. captured the turtle with a dipnet while it was basking on a log. After capture and upon closer examination, we identified the individual as a G. pulchra, a species endemic to the Mobile River system, of which the Tombigbee River is a major tributary. This individual is the first documented occurrence of G. pulchra in Tishomingo County and is also the first record of the species reported for the Tennessee River drainage (Lindeman 2013). The individual was retained as a specimen for the Mississippi Museum of Natural Science (MMNS 19865).

View Figure

• Download as Powerpoint
• Download Figure

For additional verification we collected and extracted a genetic sample from MMNS 19865 and amplified a diagnostic 660-base pair (bp) portion of the mitochondrial control region using primers reported in Spinks and Shaffer (2005). Amplified products were sent to Eurofins for sequencing. The sequence was edited and aligned in Sequencher v. 4.1 (GeneCodes Co., Ann Arbor, MI), and we used NCBI's Basic Local Alignment Search Tool (BLAST) (https://blast.ncbi.nlm.nih.gov/Blast.cgi) feature to compare our sequences with those published in GenBank. Our BLAST search verified that the Tishomingo specimen had the highest percent similarity to G. pulchra (99% similarity).

On 17 September 2018, 2 G. pulchra individuals (1 male, 1 female) were also documented in the TTW of Tishomingo County by G.J.B. (Florida Museum of Natural History [UF] 188673 [photo voucher]). These individuals were both observed in a canal section, ∼ 2 rkm south of the Bay Spring Reservoir spillway (i.e., within the former Tombigbee River system; gray cross in Fig. 1). This partially fills the aforementioned 137-rkm gap between the Goat Island individual and the formerly closest downstream record from Itawamba County (MMNS 5404).

There are 8 G. pulchra specimens collected in the Tombigbee River system of Mississippi (Auburn University Museum of Natural History [AUM] 17131, Illinois Natural History Survey [INHS] 13262, MMNS 2985, 2986, 3235, 5362, 5404, 13790; Fig. 1). Of these 8 specimens, the closest specimen locality was MMNS 5404, which was collected from the East Fork of the Tombigbee River near the town of Smithville (Itawamba County; black cross in Fig. 1). This specimen was ∼ 137 rkm from the individual we captured at Goat Island Recreation Area. This distance includes ∼ 37 rkm through the Divide Section Canal that joins the Tombigbee River to the Tennessee River, 4 lock-and-dam systems (Fulton, John Rankin, G.V. [Sonny] Montgomery, Jamie Whitten), and 1 large reservoir (Bay Springs Lake).

DISCUSSION

One explanation for the occurrences is that these individuals were released as pets via the pet trade. However, this hypothesis does not seem supported for a multitude of reasons. First, neither one of these species is highly desired or commonly observed in the pet trade, unlike other Graptemys species (W.S., pers. obs.). For example, Graptemys pseudogeographica kohnii (Mississippi map turtle) is very popular in the pet trade. Consequently, G. p. kohnii has been observed in numerous aquatic systems outside of its native range, including the Pearl River system of Mississippi (Jones et al. 1991), multiple locations in Florida (Krysko et al. 2011), and other nonnative aquatic systems throughout the United States (Ernst and Lovich 2009). Second, the locations of both G. ouachitensis and G. pulchra are in close proximity to the TTW entrance at Bay Springs Lake and Pickwick Lake, respectively. If an individual moved upstream or downstream through the TTW, locations nearest to the TTW seem to be the most likely place for individuals to be first observed rather than at locations that are more distant. Third, the G. pulchra individual and the 3 G. ouachitensis individuals exhibited external characteristics that seemed to be more indicative of wild individuals (e.g., discoloration of the plastron, shell pits, scute anomalies). Fourth, multiple size classes of G. ouachitensis including unsexable juveniles were observed, which indicates that this species is likely reproducing outside of its native range in Bay Springs Lake.

The native range of G. pulchra extends throughout the Mobile River system of Alabama, northwestern Georgia, and northeastern Mississippi inclusive of the Alabama, Black Warrior, Cahaba, Coosa, Tallapoosa, and Tombigbee rivers (Fig. 1). Therefore, the record from Pickwick Lake is outside of its native range. Even though map turtles can move through lock-and-dam systems (Bennett et al. 2010), the excessively long pathway for G. pulchra from the previously closest known specimen to the Pickwick Lake location seems unlikely for this individual or its parents. A more likely explanation is that prior to the construction of the TTW, G. pulchra occurred farther upstream than the existing specimen record suggests. Recent studies of similar broad-headed map turtles (Graptemys gibbonsi and Graptemys pearlensis) have found them occurring farther upstream in smaller tributaries than documented in the specimen record (Selman and Qualls 2009; Lindeman et al. 2020). If we apply these recent upstream findings of other Graptemys species to G. pulchra, it seems probable that G. pulchra historically extended much farther upstream in the Tombigbee River system. For example, they may have occurred upstream into Mackey's Creek in southern Tishomingo County before it was impounded as Bay Springs Lake (near the town of Moore's Mill). There is no way to determine this, but regardless of the exact scenario, the minimum movement would still be ∼ 56 rkm moving through the reservoir and the Divide Section Canal. Our observations increase the credibility that G. pulchra can exist in both canal and reservoir segments. However, it is unknown at this point if G. pulchra populations within reservoir and canal sections are viable populations or merely remnant populations persisting in unsuitable habitats following the construction of the TTW.

Graptemys ouachitensis is commonly found to occur and even thrive in lentic habitats (Lindeman 2013) throughout much of the Mississippi River system (Fig. 1). Thus, the movement downstream from Pickwick Lake, mostly through a lentic canal (i.e., the TTW divide section) and into another reservoir is not surprising. For the genus Graptemys, it is not typical for species to be successful in reservoirs, which indicates that G. ouachitensis is a generalist in the genus that can readily establish populations in connected reservoirs and waterways (Lindeman 2013).

With all the available evidence, we suggest that our observations were indeed associated with movements of G. pulchra and G. ouachitensis via the TTW from their native ranges into nonnative ranges. At this point, it is unknown how many individuals have made the drainage exchange and to what extent the previously allopatric species are interacting. One problematic issue with this exchange is that many Graptemys species will readily hybridize (Godwin et al. 2014), which is possible because of the relatively recent evolutionary radiation within the genus (Thomson et al. 2018). Therefore, a looming threat is the hybridization of previously allopatric species that are now syntopic in distribution, particularly the microcephalic species (G. ouachitensis and G. nigrinoda, the blackknobbed sawback). Several records for G. nigrinoda occur in the upstream Tombigbee River system (north to Itawamba County; MMNS 1055, 3690, 7733), so it seems possible that this species may also move upstream and encounter G. ouachitensis. Thereafter, it seems plausible that hybridization of these 2 species could occur and, consequently, genetic introgression across these species may compromise these distinct evolutionary lineages.

In conclusion, we suggest that continued monitoring is needed to document the movement of these Graptemys species (and potentially other turtles including different lineages of Pseudemys spp.) into nonnative drainages, while also capturing individuals to determine levels of introgression (Harrison and Larson 2014). For the former, monitoring the range expansions of G. pulchra and G. ouachitensis should be undertaken in the Tennessee River system above the Yellow Creek embayment of Pickwick Lake and below Bay Springs Lake, respectively. For the latter, it would be of interest to determine if introgression is localized (i.e., introgressed alleles occur only where the 2 parental species occur) or dispersed (i.e., introgressed alleles of 1 species flow into “pure” populations of the other species; Heiser 1973). Also, any Graptemys individuals encountered outside of their native range in this region should be collected and removed from the environment until such genetic studies are completed.