Population Demographics of Native Red-Bellied Cooters and Invasive Red-Eared Sliders in a Virginia Lake
Abstract
An 8-wk mark-and-recapture study using baited fyke nets showed that native red-bellied cooters (Pseudemys rubriventris) (estimated population 738–3618) were more abundant than non-native red-eared sliders (Trachemys scripta elegans) (estimated population 256–520) in Lake Matoaka in southeastern Virginia. Red-bellied cooters have maintained a similarly large population since a prior study in 2004 with a balanced female:male sex ratio of 0.91:1, but over that time, capture of red-eared sliders has increased almost 5-fold; their size distribution is broader than that of red-bellied cooters, and their sex ratio is dominated by females (2.02:1). Persistence of native red-bellied cooters in Lake Matoaka may be challenged by the apparent ongoing growth of the invasive red-eared slider population.
Red-eared sliders (Trachemys scripta elegans) are an invasive species of freshwater turtle that has spread across the United States due to its popularity in the pet trade. Native to the Mississippi Basin, red-eared sliders have been imported to more than 70 countries and are considered one of the most invasive species in the world (Lowe et al. 2000; Lambert et al. 2019). The species is known to outcompete native turtles (Lambert et al. 2013), especially when resources are a limiting factor (Pearson et al. 2015), and can quickly become the dominant chelonian where introduced (Snapp et al. 2021). As a generalist species, red-eared sliders often can spread into more stressful habitats that native turtles may not tolerate as well (Tate et al. 2022) and thus may provide ecosystem service benefits in degraded aquatic habitats (Dupuis-Desormeaux et al. 2022), including enhancement of energy flows and nutrient cycling (Lovich et al. 2018).
Red-eared sliders have been present in Lake Matoaka in southeastern Virginia for decades, but a 2-mo turtle survey completed in 2004 captured only 27 invasive red-eared sliders relative to 136 native red-bellied cooters (Pseudemys rubriventris) (E. Thompson, unpublished dat, 2004). In our study completed 18 yrs later, our objective was to document sizes and demographics of populations of native red-bellied cooters and invasive red-eared sliders in Lake Matoaka. We expected to see a growth in the red-eared slider population due to the species’ invasive capabilities, with a concomitant reduction in the size of the red-bellied cooter population. We also documented the occurrence and frequency of melanism in male red-eared sliders, a condition for which individuals gradually become more pigmented. Melanism has been found to coincide with male turtles reaching the size of female turtles at maturity (Hays and McBee 2009). Because melanistic turtles can dominate larger size classes (Lovich et al. 1990), we compared the sizes of melanistic and nonmelanistic male sliders.
Finally, we completed surveys of potential nest sites in selected open space locations adjacent to the lake. Female red-eared sliders may preferentially select open microhabitats as nesting sites (Lloyd and Warner 2019), whereas prior research suggests red-bellied cooters prefer vegetated nesting sites to bare ground nesting sites (Nelson et al. 2009). We use these anecdotal observations along with demographic data to comment on the ongoing differences in population sizes observed between native and nonnative turtles in Lake Matoaka.
METHODS
Our turtle population studies were conducted at 2 trapping sites in Lake Matoaka in Williamsburg, Virginia (lat 37°16′45″N, long 76°42′54″W). The sites of interest were the College Creek and Tinkling Rill deltas, formed where the 2 largest streams discharge to the lake (Fig. 1). Discharge from other tributaries to Lake Matoaka was insufficient to form deltas of consolidated sediment. The 2 trapping sites separated by over 1 km were selected because 1) large numbers of turtles were visually observed there, and 2) water depth was relatively shallow and bottom sediments were sandy so that turtle traps could be set and anchored safely. Other locations around the lake margins were either too deep or the sediments were too unconsolidated to keep traps anchored securely.
Citation: Chelonian Conservation and Biology: Celebrating 25 Years as the World’s Turtle and Tortoise Journal 23, 1; 10.2744/CCB-1587.1
Turtles were captured using fyke nets with 2.54-cm mesh and 8-m wings, with wooden stakes anchoring the wing tips and cod ends of the nets. By maintaining hoop positions so that a portion of each hoop of the fyke net was above water, we ensured that captured turtles had access to air to breathe. Operationally, turtles swim through 2 internal cones of netting in search of bait placed toward the rear of each trap. Once in a trap, turtles are unable to find their way back out, although Brown et al. (2011a) found that female red-eared sliders may preferentially escape hoop nets. A more recent and comprehensive review of the literature confirmed that turtle trapping with hoop traps can yield escapes that are species-specific, and that overall sampling bias relative to other capture methods such as fyke nets has been untested (Mali et al. 2014). The sampling gear, trap locations, and trapping dates replicated those used in the prior rapid-assessment population study in 2004 (E. Thompson, unpublished data, 2004).
Between 8 June and 29 July 2022, traps were baited and set every weekday and then checked every day Monday–Friday. Traps were left baited over weekends. When checked, all red-bellied cooters and red-eared sliders were separated and retained for processing. All bycatch of other turtle species and fish was released at the site of capture. Most captured fish, however, had been killed and partially consumed by the turtles that were in the trap.
All captured red-bellied cooters and red-eared sliders were sexed using tail length and cloaca location, then measured for total carapace length and width using calipers. Turtles were not weighed, nor were measures made of maturation (e.g., claw length; Cagle 1948). We used a base 2 notching pattern originally described by Cagle (1939) to mark the captured turtles. Recaptured turtles were identified by the unique notching code and then remeasured. After processing, all red-bellied cooters and red-eared sliders captured in fyke nets were released alive near the point of capture in the lake. Mark–recapture data were then used to estimate total population sizes using the Schnabel method (Krebs 2014):
Where
Ci = total number of individuals caught in sample i,
Mi = number of marked individuals at sample i − 1,
Ri = number of marked individuals in sample i, and i is each sampling time.
The Schnabel method assumes that 1) animals do not lose marks between samples, and 2) the population is closed, i.e., there is no dilution (birth or immigration) or loss (death or emigration) between sample periods (Lindeman 1990). This study, completed over the relatively short term of 8 wks, met these assumptions, yielded a high catch per unit effort (Brown et al. 2011b) and provided a rapid assessment of population size. The Schnabel method, however, also assumes that all turtles have an equal probability of capture at each sample time (day). Because we sampled with only 2 traps at 2 locations, that assumption may have been violated if turtles did not mix uniformly throughout the lake. Further, some turtles may exhibit trap-happy or trap-shy behavior, which would skew population estimates. We did not try to account for any of these potential inequalities in probability of turtle capture.
Mean turtle carapace lengths were plotted by species and by sex using size-frequency histograms and compared with 2-sample Kolmogorov-Smirnov (K-S) tests using the “dgof” package in R (R Core Team 2021). Carapace lengths and widths were plotted to further visualize potential interspecific differences in size; χ2 goodness-of-fit tests determined whether the number of females and males of each species was significantly different from a 1:1 sex ratio. To test the assumption that melanism occurs more frequently in older and presumably larger male turtles, mean carapace length of nonmelanistic and melanistic male red-eared sliders was compared using a t-test.
We also collected data from daily surveys of potential turtle nesting sites in the area surrounding the W.M. Keck Environmental Field Lab (Williamsburg, VA) from 8 June to 15 July. Our walks were completed in open upland areas within 400 m of Lake Matoaka where turtles had been observed nesting in prior years (Fig. 1). Turtles found in the process of nesting were marked for use in population estimation. The major nest predators in much of southeastern Virginia are raccoons and crows (Cook et al. 2018), which are known to dig up nests of red-bellied cooters (Nelson et al. 2009) and red-eared sliders (Shipman 2019). Torn eggshells around the opening of a predated nest were considered evidence of raccoon predation, although raccoons seasonally may ingest eggs and shells entirely (Burke et al. 2009). Fully dug but emptied nest cavities with no eggshells present were considered evidence of crow predation because crows tend to fly away with eggs.
RESULTS
Population Estimates. —
With 2 traps typically set each day of the study, data were collected over the course of 72 trap days. We captured 130 unique red-bellied cooters and 128 unique red-eared sliders, 15 of which were melanistic (Table 1). We had 6 recaptures of red-bellied cooters and 31 recaptures of red-eared sliders. Using the Schnabel method to calculate population size from mark-recapture data, we estimated 1576 red-bellied cooters (95% confidence interval 738–3618) and 358 red-eared sliders (95% confidence interval 256–520) (Table 1). We also had a variety of bycatch, including common snapping turtles (Chelydra serpentina serpentina), painted turtles (Chrysemys picta), common musk turtles (Sternotherus odoratus), and largemouth bass (Micropterus salmoides). These were not enumerated.
Turtle Size and Sex Data. —
The size range of red-eared sliders (carapace length 10 to 30.2 cm) was more broad than that of red-bellied cooters (16.3 to 31.1 cm), with greater representation in the smaller and presumably younger sizes (Fig. 2). Differences in histograms between species by sex were significant for both females (K-S test, D = 0.63; p < 0.001) and males (K-S test, D = 0.54; p < 0.001). Carapace length–width relationships of the 2 species, however, were similar (Fig. 3). The sex ratio of red-eared sliders was heavily biased toward females (2.02:1) and significantly different from 1:1 (χ2(1,128) = 14.56; p < 0.01); the sex ratio of red-bellied cooters (0.91:1) was not different from 1:1 (χ2(1, 130) = 0.28; p > 0.05). The mean carapace length of the 15 melanistic male red-eared sliders was greater than for the 27 nonmelanistic, male red-eared sliders (20.6 ± 0.3 cm SE vs. 16.7 ± 0.4 cm SE, respectively; t-test, t = 6.98, df = 48, p < 0.001).
Citation: Chelonian Conservation and Biology: Celebrating 25 Years as the World’s Turtle and Tortoise Journal 23, 1; 10.2744/CCB-1587.1
Citation: Chelonian Conservation and Biology: Celebrating 25 Years as the World’s Turtle and Tortoise Journal 23, 1; 10.2744/CCB-1587.1
Nesting and Depredation Observations. —
Eight of 10 female turtles observed on land on nest crawls were red-eared sliders and 2 were red-bellied cooters (Table 1). We also found more than 20 empty nests with evidence of having been dug up, but no shell debris was present, suggesting predation by crows. Indeed, often we were alerted to the presence of turtle nests by the calls and behaviors of crows that frequently maraud nests after turtles finish laying (R.M.C., pers. obs.).
DISCUSSION
Although the native red-bellied cooter remains the more abundant species, our study provides evidence of the ongoing growth of the nonnative red-eared slider population in Lake Matoaka. The estimated population of 1576 red-bellied cooters in 2022 is similar to the population of 2057 turtles estimated in 2004 (E. Thompson, unpublished data, 2004). Relative to the red-bellied cooter population, we estimate a smaller population of 358 nonnative red-eared sliders in 2022, but the population of red-eared sliders could not be estimated in 2004 owing to the small number of captures and recaptures. We captured 128 red-eared sliders in 2022 with 31 recaptures, whereas in 2004 only 27 red-eared sliders were captured, with 2 recaptures. The number of native turtles has remained similar but the number of nonnative turtles has grown substantially, with almost 5 times as many captures in 2022 as in 2004.
The frequency histograms demonstrated that red-eared sliders exhibit a broader size range of turtles than red-bellied cooters, including many smaller individuals. Recruitment into the invasive slider population appears greater than recruitment into the native red-bellied cooter population based on the higher frequency of observations of these smaller and presumably younger turtles. Beside reduced recruitment, another possible explanation for the absence of small red-bellied cooters may be that these individuals are better at escaping or avoiding the traps than the small red-eared sliders. No studies to date, however, have fully assessed the fidelity of turtle trap captures to the actual demographics of turtle populations (Gamble 2006; Bluett et al. 2011). A prior study found a positive relationship between the hoop size of traps and the average carapace length of the turtles captured, possibly because smaller turtles were better able to escape from larger diameter hoops (Gulette et al. 2019). Another study confirmed that traps with smaller mesh sizes captured smaller turtles (Ennen et al. 2021). Because our bycatch included many small painted turtles (Chrysemys picta) and musk turtles (Sternotherus odoratus) with carapace lengths less than 10 cm, we doubt that only small red-bellied cooters were able to escape the nets when other species were not able to. We believe it more likely that smaller red-bellied cooters are either not present or are active in another part of the lake where our traps were not set.
Based on the similar length–width relationships of the 2 species, red-eared sliders are structurally much the same as red-bellied cooters, suggesting possible competitive overlap for food and other resources. Both species are omnivorous with generalist diets, consuming both plants and animals, but red-bellied cooters tend more toward herbivory (Ernst et al. 1994; Zahn and Hoback 2014). For both species, females dominated the largest size classes; globally, aquatic turtle species tend to show female-biased sexual size dimorphism (Agha et al. 2018). Because its sex ratio was heavily biased toward females, the red-eared slider population appears poised to continue to grow. Still, the size of the red-bellied cooter population currently is about 5 times larger than that of the red-eared slider, so we were surprised by the low number of female red-bellied cooters seen nesting in our study area relative to red-eared sliders. Whereas all 8 female red-eared sliders we encountered in open space on land completed nesting, 1 of the 2 red-bellied cooters abandoned her nest without laying eggs, and the other was not witnessed digging a nest. From these observations, female red-bellied cooters perhaps are more vigilant and less willing to lay eggs in exposed areas compared to the red-eared sliders. Red-bellied cooters may prefer more vegetated and secluded nesting sites as described by Nelson et al. (2009).
The importance of melanism in the red-eared slider population in Lake Matoaka is unclear. Melanism can be induced by thermal and substrate cues in laboratory-reared red-eared sliders (Rowe et al. 2016). Results from experimental enclosures, however, do not suggest any reproductive advantage, as females do not show preference for melanistic males, which behave largely the same as nonmelanistic males (Stone et al. 2015). Melanistic male sliders are almost 4 cm larger, on average, than nonmelanistic males in Lake Matoaka. With no reports of melanistic males in the 2004 study and assuming that red-eared sliders live for decades (Gibbons 1987), the melanistic males identified in 2022 may be older on average than the nonmelanistic males, confirming prior research indicating melanism is a factor of age and/or sexual maturity in the male red-eared sliders (Lovich et al. 1990; Tucker et al. 1995).
To sum, our study documented a sustaining, large population of native red-bellied cooters in Lake Matoaka, but the population of nonnative red-eared sliders appears to be growing. The number of red-eared slider captures in 2022 was almost 5 times greater than in 2004, and the population now appears to include a broad range of size classes. Melanistic male red-eared sliders are not uncommon, and female red-eared sliders far outnumber males and are nesting in open areas surrounding the lake. With the addition of a growing population of nonnative, red-eared sliders, food and other habitat resources for native turtles may become limiting, creating conditions that favor further population increase of this invasive species.
Sampling site locations (marked by triangles) in the deltas of the 2 largest tributaries to Lake Matoaka in southeastern Virginia (starred, inset). Hydrography data courtesy of National Hydrography. Dotted line indicates the path surrounding the W.M. Keck Environmental Field Lab walked daily to observe nesting activity.
Frequency histogram of carapace length for male and female (a) red-eared sliders and (b) red-bellied cooters.
Carapace length vs. carapace width in red-bellied cooters and red-eared sliders. Points represent unique turtle observations. For red-bellied cooters n = 130; for red-eared sliders n = 128.
Contributor Notes
Handling Editor: Jeffrey A. Seminoff
