The Sabine map turtle (Graptemys sabinensis) is 1 of 14 species that compose the most diverse genus of freshwater turtles in the United States and Canada (Lindeman 2013). The diversification of the 13 Gulf Coastal species of map turtles and sawbacks has occurred over the last 2.5–3.5 million years as a result of the isolation of river drainages (Lamb et al. 1994). The evolutionary success of the map turtles and sawbacks can be attributed largely to the dietary diversity and adaptive trophic morphologies between narrow-headed (microcephalic) and broader-headed (meso- and megacephalic) species of Graptemys (Lindeman 2000, 2013). Microcephalic species feed on algal and invertebrate communities, while broader-headed species primarily consume mollusks (Lindeman 2000, 2013). In many drainages, a microcephalic species occurs in sympatry with a broader-headed species with little competition for food (Lindeman 2000, 2013). On sunny days, map turtles spend much of their time basking on emergent logs and dead branches to increase metabolic rates, promote digestion, dry out algae and parasites on the skin and shell, and accelerate the rate of egg development (Lindeman 2013).

Graptemys sabinensis occurs in the Sabine-Neches, Calcasieu, and Mermentau river systems of Texas and Louisiana (Cagle 1953; Lindeman 2013). The taxon was first regarded as a subspecies of G. pseudogeographica (Cagle 1953) and later as a subspecies of G. ouachitensis (Vogt 1993), but its allopatric distribution, diagnosability without evidence of intergradation, and position in recent phylogenetic analyses support its recognition as a separate species (Lindeman 2013).

The first G. sabinensis specimens were collected in 1893 and 1894 by Gustave Kohn (Cagle 1953; Lindeman 2013), with 4 adult females from the Mermentau River at the town of Mermentau in Acadia Parish (Tulane University Museum, TU 7632, 7634, 7640, and 7690) and 2 adult females from near Vinton, Louisiana (probably from the lower Sabine River WSW of Vinton in Calcasieu Parish; TU 7638 and 7643). The species was not formally described until Cagle (1953) and his students collected a specimen from Calcasieu River in 1948 and a large type series from the Sabine River in 1950. Cagle (1953) noted the Mermentau River specimens in his description; however, in all his extensive Gulf Coast collecting expeditions, the only trapping he ever conducted on the Mermentau drainage was on Lake Arthur, a natural widening of the lower Mermentau River, in 1947 (Cagle and Chaney 1950). This was apparently the only collecting of turtles conducted by anyone on the Mermentau drainage since Kohn's day (see dot-distribution maps for turtles in Dundee and Rossman 1989). Cagle and Chaney (1950) caught no G. sabinensis among 25 turtles taken from Lake Arthur, but their sampling was conducted prior to the development of special collecting techniques that greatly facilitated their later studies of Graptemys, including collection of the type series of G. sabinensis from the Sabine (see Chaney and Smith 1950). The 4 specimens of G. sabinensis collected from the Mermentau River at Mermentau in the 1890s and unpublished sightings by Shively (2001) at that same locality and an upstream locality on a tributary, Bayou Nezpique, are the only records of Sabine map turtles in the Mermentau drainage prior to the present study.

Since the species' description, only a few studies have been conducted on the ecology of Sabine map turtles (Shively and Jackson 1985; Shively 2001; Ewert et al. 2004; Coleman and Gutberlet 2008). Some literature accounts have been inaccurate regarding the range and distribution of the species. For example, Vogt (1993) and Ernst and Lovich (2009) incorrectly restricted the range of the Sabine map turtle to the Sabine drainage, although they cited specimens from the Calcasieu drainage and Shively and Jackson's (1985) study of the species conducted on the Calcasieu drainage, respectively. Recently, Brown et al. (2012) made the same error in spite of the G. sabinensis specimen included in their phylogenetic analyses being from the Calcasieu. In contrast, one recent source has correctly described G. sabinensis as occurring in all 3 river drainages (Buhlmann et al. 2008).

Throughout its range, G. sabinensis is sympatric with the false map turtle (Graptemys pseudogeographica; specimens typical of the Mississippi map turtle, G. pseudogeographica kohnii), a species that is much more widely distributed, with additional populations in the Mississippi River and its tributaries as well as 3 additional river drainages in east Texas (Lindeman 2013). Kohn collected a specimen of G. pseudogeographica at Mermentau in 1893 (TU 7631, a juvenile female) and 2 specimens were collected in 1960 at “Rayne,” probably from one of the northeastern tributaries of the drainage, such as Bayou Wikoff (LSUMZ 84350–51), but there are no additional records for the species in the Mermentau drainage.

Graptemys sabinensis is a state-listed species of conservation concern under the Louisiana Comprehensive Wildlife Conservation Strategy (Lester et al. 2005). Graptemys pseudogeographica was found to be “ubiquitous” in Louisiana and was removed from the list of regulated species. The dearth of data from the Mermentau drainage limits our knowledge of the status of these species and is of particular importance for G. sabinensis, considering its much more restricted geographic range. The river banks in the Mermentau drainage lack sandbars and the water is murky and very slow moving because most of the drainage's water level is < 1 m above sea level. The drainage is therefore atypical habitat for southern species of Graptemys, which are more commonly found in rivers with more flow that thus have clearer water and numerous sandbars that serve as nesting beaches (Lindeman 2013).

During the early summers of 2010, 2011, and 2012, we conducted basking surveys on the Mermentau River and 5 of its tributary bayous, as well as 2 smaller bayous that join those tributaries, and Lake Arthur. We determined relative abundance of Sabine map turtles in basking assemblages and basking density per kilometer for stream reaches throughout the drainage. We also report data on G. pseudogeographica and report what appears to be a catastrophic decline in the status of the species in the drainage that is likely due to habitat modification affecting its prey base.

METHODS

Study Area

The Mermentau River is formed by the confluence of bayous Plaquemine Brule and des Cannes in Acadia Parish, Louisiana (Fig. 1). It is joined by 2 additional major tributary bayous, Nezpique from the northwest and Queue de Tortue from the east, before it enters Lake Arthur. Below the lake, Lacassine Bayou and the lower Mermentau River flow together into Grand Lake just below the Gulf Intracoastal Waterway. Boat and point surveys were conducted on all of these sections of the drainage as well as 2 smaller bayous: Bayou Wikoff, a tributary of upper Bayou Plaquemine Brule, and Bayou Chene, a tributary of upper Lacassine Bayou.

View Figure

• Download as Powerpoint
• Download Figure

Throughout the drainage, the riparian forest is dominated by water tupelo (Nyssa aquatica) and bald cypress (Taxodium distichum). The main channel of the Mermentau River is 80 m wide at its origin and 180 m wide at its inflow into Lake Arthur. The tributary bayous are narrower, typically 20–60 m. Deadwood and fallen trees with foliage are ubiquitous throughout the drainage, and turtles use these substrates to bask.

Survey Methods

During 20–27 May 2010, 15 May–7 June 2011, and 17–25 May 2012, we conducted surveys of basking turtles using boat surveys and point counts. Sampling in 2010 was of a preliminary nature, and sampling in 2012 was conducted only to fill in some gaps in drainage coverage (6 point counts and 2 boat surveys); thus, the most intensive sampling was conducted in 2011. We used Canon 18 × 50 image-stabilizer binoculars and Bushnell 20 × 50 binoculars to identify turtles to species and record all turtles seen basking. Species were identified by recognizable head and shell patterns. During boat surveys, we also recorded numbers of “unidentified” turtles that we could not identify before they dove into the water (the great majority of the unidentified category) or for which diagnostic marks were blocked from view. Boat count surveys were taken using a jonboat with an outboard or a canoe with a trolling motor. To avoid startling turtles and to facilitate identification where densities were high, we travelled slowly and stayed 12–18 m from the bank. The mainstem Mermentau River was too wide to survey both banks simultaneously, so we surveyed one bank and then the opposite bank on our return. We surveyed both banks simultaneously on all the narrower tributaries. We conducted point counts at bridge crossings, boat ramps, and other public access points. All surveys were conducted on warm, sunny days between 0900 and 1800 hrs.

We measured the length in kilometers of each section of river surveyed by boat by using the path measure function in Google Earth's satellite map imagery. To account for many sections being surveyed more than once, sometimes for differing lengths, we calculated weighted averages of Sabine map turtles seen per kilometer for each section with repeated surveys.

RESULTS

We conducted 61 boat surveys covering 154.6 total river km in 15 distinct sections of the drainage as well as 61 total point counts at 24 separate access points (Fig. 1). Graptemys sabinensis accounted for 73% of all turtles identified in boat surveys and point counts (Table 1). We observed 8 other turtle species, most prominently 2 other emydids, Trachemys scripta (slider turtle, 18%) and Pseudemys concinna (river cooter, 6%); each of the other 6 species accounted for < 1.5% of all turtles seen (Table 1).

Table 1. Summary of relative abundance of turtle species in basking surveys on the Mermentau drainage, 2010–2012.

View Fullscreen

Bayou Plaquemine Brule had notably high basking densities of G. sabinensis, ranging from 17.7 to 34.5/km, with the highest point counts being at a drawbridge where substantial deadwood accumulated near the bridge and its flood control structures (Fig. 1). Basking densities on the Mermentau above Lake Arthur were also high, ranging from 14.0 to 28.5/km. The numbers of G. sabinensis per kilometer on lower reaches of bayous Nezpique and des Cannes were intermediate, ranging from 5.0 to 6.5/km. Bayou Queue de Tortue and Lacassine Bayou had very low densities of G. sabinensis, 1.3/km and 0.4–0.7/km, respectively. Lake Arthur had a moderately high density of 9.7/km, while the Mermentau below Lake Arthur had a low density of 1.4/km. On Bayou Plaquemine Brule above its confluence with Bayou Wikoff and on upper reaches of bayous Nezpique and des Cannes, densities dropped off dramatically. G. sabinensis was not seen in point counts at bridges on the 2 smaller bayous, Wikoff and Chene. Numbers seen during point counts, although not expressed as densities, tended to be higher near reaches that had higher densities in boat surveys (Fig. 1).

During boat surveys, 15% of turtles evacuated into the water before they could be identified or did not present diagnostic features; this percentage ranged from 8% to 56% for the 15 survey segments. Under the assumption that relative abundance of unidentified turtles matched that of identified turtles, corrected basking densities would thus typically be 15% higher, with a range from 8% higher in Lake Arthur (i.e., 10.5/km) to 56% higher on the uppermost reach of Bayou Nezpique (i.e., 0.3/km).

Relative abundance of Sabine map turtles was greatest along the Mermentau above Lake Arthur, in the lake, and in the lower sections of the 3 northernmost tributaries of the river (Fig. 2). Their relative abundance declined greatly at more upstream localities of the northern tributaries, on the 2 southern bayous that had the lowest densities of the species, and on the Mermentau below Lake Arthur. These results are similar to the pattern for basking density and there was a highly significant positive correlation of relative abundance with ln-transformed basking density (r  =  0.91, p < 0.0001; Fig. 3). The 2 sites that departed farthest from the regression line—both with high basking density but lower-than-expected relative abundance of G. sabinensis—were the uppermost section of the Mermentau (with 31% T. scripta and 5.1% painted turtles, Chrysemys picta) and the Mermentau immediately above Lake Arthur (with 16% P. concinna and 15% T. scripta). These 2 sites had the highest and third-highest basking densities for combined species exclusive of G. sabinensis. Combined basking density of other turtle species was not significantly correlated with basking density of G. sabinensis (r  =  0.12, p  =  0.68).

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Of the 2 Graptemys species that coexist in the Mermentau drainage, 99.8% observed were G. sabinensis, while only 0.2% were G. pseudogeographica. Graptemys pseudogeographica was seen only during point surveys and at only 2 localities, 3 times at a site on upper Bayou Plaquemine Brule (on one occasion, 2 individuals were seen) and once on a site on a middle reach of Bayou Nezpique (asterisks in Fig. 2). Photo vouchers were taken for both localities (FLMNH 166405 and 166403, respectively).

DISCUSSION

Our surveys showed G. sabinensis to be widespread and abundant throughout most of the Mermentau drainage. The Mermentau above Lake Arthur and Bayou Plaquemine Brule had the highest basking densities. Lake Arthur and the other 2 northern (upstream) tributaries, bayous des Cannes and Nezpique, had intermediate basking densities. The Mermentau below Lake Arthur and the 2 southern tributaries, Bayou Queue de Tortue and Lacassine Bayou, had the lowest basking densities.

On all of the tributaries, the basking density of G. sabinensis declined toward zero at more upstream localities, where relative abundance was also low. The strong correlation between basking density and relative abundance of G. sabinensis and the lack of a correlation of their density to that of other basking turtles suggest the decline is a species-specific effect in G. sabinensis, rather than being indicative of a general decline in the suitability of turtle habitat at more upstream localities. On the Whiskey Chitto River in the neighboring Calcasieu River drainage, populations of G. pseudogeographica and T. scripta were found to extend much farther upstream than those of G. sabinensis (Shively and Jackson 1985). Amount of sunlight may be a major factor in determining the upstream limit of G. sabinensis: as the width of the river narrows and the tree cover creates a thicker canopy, there is less sunlight and subsequently less algal aufwuchs supporting prey populations, as well as fewer sunlit places suitable for basking (Shively and Jackson 1985). Low densities of G. sabinensis on the Mermentau below Lake Arthur, Lacassine Bayou, and Bayou Queue de Tortue cannot be attributed to stream width, however. High salinity in closer proximity to the Gulf of Mexico may limit these turtles in the southern (downstream) portions of the drainage.

Two previous studies report basking densities for G. sabinensis on the Sabine and upper Calcasieu drainages that are comparable to the highest densities observed in the present study on the upper Mermentau and Bayou Plaquemine Brule (Table 2). Shively (2001) saw an average of 31/km on the lower Sabine River and densities as high as 27/km and 24/km on the upper Calcasieu and its largest tributary, the Whiskey Chitto River, respectively. Mean basking densities of G. sabinensis observed by Coleman and Gutberlet (2008) on a 1.5-km stretch of the upper Sabine River in Texas that they surveyed every other week ranged from 13.8/km in winter to 47.4/km in summer. Shively (2001) also reported lower densities on 2 tributary bayous of the Sabine drainage, similar to our findings for the other 4 smaller tributary bayous of the Mermentau system.

Table 2. Basking densities reported for narrow-headed species of Graptemys from rivers. Indented drainages are tributaries of the rivers listed above them. No corrections are made for unidentified turtles observed. NR  =  not reported.

View Fullscreen

The status of Mermentau populations of G. sabinensis can also be assessed by comparison with population surveys of the other 4 species of microcephalic (narrow-headed) Graptemys (Table 2). While many of the densities for the other 4 species are comparable to densities we observed for G. sabinensis in the Mermentau drainage, all 4 other species have also been seen at substantially higher densities, in the range of 50–294 turtles/km. Differences in survey methodology may affect densities observed in 2 ways. First, most of the highest numbers in Table 2 were reported from surveys conducted on foot, either without use of a boat from bridge crossings or concealed spots on shorelines (Lindeman 1998, 1999; Shively 2001; Coleman and Gutberlet 2008) or after landing a boat and walking along a sandbar (Carr 2001; Selman and Qualls 2009). Compared with boat surveys, such surveys are less likely to startle turtles into the water before they can be recorded, and they provide more time to scan to ensure that the maximum numbers of turtles that are basking are seen. Second, lengths of stream surveyed and repetition of surveys both vary greatly among studies and may greatly affect the variability observed in basking densities. Averaging density over either more frequent surveys or longer stream reaches would tend to decrease the highest densities reported. Nevertheless, it appears Mermentau populations of G. sabinensis are relatively substantial. The high relative abundance we observed—G. sabinensis outnumbered 8 other species observed by almost 3∶1 overall, and by better than 10∶1 in some reaches—is an oft-reported phenomenon for microcephalic species of Graptemys (Table 2), further supporting the conclusion that the Mermentau populations of the species are relatively healthy.

To what extent the high relative abundance of microcephalic species stems from true numerical dominance in rivers vs. their being much more heliophilic than sympatric species is unknown. The answer could be determined using a combination of approaches, including multispecies mark–resight population estimates (e.g., by extending the approach of Selman and Qualls 2009 to a larger set of species) and focused basking observations of marked or telemetered individuals of sympatric species.

The Louisiana Comprehensive Wildlife Conservation Strategy (Lester et al. 2005) stated that G. sabinensis may have been extirpated from parts of its range (although no source is given for that unlikely assertion). Graptemys pseudogeographica is a much more widespread species in Louisiana, being found almost statewide and in diverse habitats (Dundee and Rossman 1989; Carr 2001; Lindeman 2013). In the Mermentau drainage, however, G. sabinensis outnumbered its congener by a ratio of 357∶1. Kohn's collecting at Mermentau in 1893–1894 is noteworthy in that it included a specimen of G. pseudogeographica (TU 7631) in addition to 4 specimens of G. sabinensis, suggesting that the ratio of the 2 species was likely more equitable in the past.

While it is typical that microcephalic Graptemys species outnumber sympatric meso- or megacephalic congeners, in our surveys the ratio of the 2 species exceeded by more than an order of magnitude all previous ratios reported for sympatric pairs of Graptemys (reviewed in Lindeman 2013). The previous greatest discrepancy was the 9.6∶1 ratio of G. oculifera to G. pearlensis in the Pearl River drainage (Lindeman 1998, 1999). In addition, while Godwin (2001, 2003) reported the ratio of G. nigrinoda to G. pulchra throughout the Mobile Bay drainage to be 3.7∶1, it was highly variable in different portions of the drainage, reaching as high as 34∶1 in the Black Warrior River, 23∶1 in the Tensaw delta, and 10∶1 in the Tombigbee River. For G. sabinensis and G. pseudogeographica in sympatry in other drainages, Coleman and Gutberlet (2008) reported a ratio of 8.4∶1 in the upper Sabine River and Shively (2001) reported ratios of 1.5∶1 in the Calcasieu drainage and 0.8∶1 in the lower Sabine drainage.

We found no evidence that G. pseudogeographica is found upstream of the range limits of G. sabinensis, as has been reported for tributaries of the Calcasieu and Sabine rivers (Shively and Jackson 1985; Shively 2001), despite our surveys extending upstream of the main populations of G. sabinensis on all 5 major tributaries (Figs. 1 and 2). Hence it is clear that in the Mermentau drainage, G. pseudogeographica is an extraordinarily rare species clinging to a tenuous existence (Fig. 4).

View Figure

• Download as Powerpoint
• Download Figure

Decimated mussel populations may explain the exceptionally low numbers of G. pseudogeographica that we observed, because mussels are important prey for female G. pseudogeographica (Vogt 1981; Shively and Vidrine 1984; Shively 2001); none of the 5 observations we made were of adult females. Extensive dredging conducted in nearly the entire Mermentau drainage has greatly altered the habitat, transforming streams from clear and sandy-bottomed to silty and mud-bottomed and eliminating almost the entire native mussel fauna (Vidrine 1993, 2010). The Mermentau drainage is thus marked by a strong contrast in the status of its 2 native Graptemys species, the narrower-headed G. sabinensis being strongly predominant in assemblages of basking turtles, while the broader-headed G. pseudogeographica has been nearly extirpated. Future studies of map turtles in the Mermentau drainage should emphasize searches for relictual mussel populations that may support small, persistent populations of G. pseudogeographica.