Methods and Materials

All translocated females (N = 69) were initially captured in late June near the end of the nesting season in 1999 at nesting areas near Swan Lake (Tucker et al. 1998). Some (N = 40) were recaptures of females previously marked and others were first captured in 1999 (N = 29). We recorded carapace length, plastron length, carapace height, carapace width, and gravid mass of captured turtles (Tucker et al. 1998). Holes were drilled in the marginal scutes to uniquely mark each individual. All gravid females had oviposition induced using oxytocin (Tucker et al. 1998, 2007). All were then kept captive for a minimum of 72 hours to reduce the incidence of false nesting (Tucker et al. 1995). All 69 females were then released into a ditch that nesting females from Long Lake routinely traversed to reach their nesting area (see Tucker 2000 and Fig. 1). An additional 33 females first caught in 1999 in Swan Lake were returned to Swan Lake to serve as a control group.

View Figure

• Download as Powerpoint
• Download Figure

Individuals were recaptured in 2 ways. Most were recaptured while they were attempting to nest. Nesting areas (Fig. 1) were visited daily during the nesting season from 2000 to 2006 (see Tucker and Warner [1999] for details on nesting areas). Others were recaptured in aquatic traps set between 2003 and 2006. Baited hoop traps were used and were employed a minimum of 15 days per year at Swan Lake (2003–2006), Gilbert Lake (2003–2006), and Long Lake (2001–2006). We used a minimum of 10 traps in Swan Lake, Gilbert Lake, and Long Lake during each trapping period. Traps were also placed along the channel border of the Illinois River outside Swan Lake and outside Gilbert Lake. We used a minimum of 5 traps on each side of the river. This trapping could only be done occasionally when periods of stable river stages could be predicted. Rising river water levels could submerge channel border traps and cause drowning.

Cross-river movements were recorded during mark and recapture trapping conducted between 2001 and 2006 in order to compare them to homing in females. Each turtle was uniquely marked and then measured and weighed as above. All were released at their original capture location excepting the 69 females translocated from Swan Lake to Long Lake. Cross-river movements were recorded for males and females. The distance between Gilbert Lake and Swan Lake is 1 km compared to 3 km between Long Lake and Swan Lake (Fig. 1). The frequency that each sex made such movements was compared using chi-square analysis. Recapture frequency was also compared using chi-square analysis. Means were compared with the nonparametric Kruskal–Wallis test or the Wilcoxon two-sample test, which ever was appropriate. All statistics were performed using SAS for Windows (SAS Institute 2000).

Results

Between 2000 and 2006, we made 2005 captures of nesting females at the nesting areas associated with Long Lake and 899 captures of nesting females at nesting areas associated with Swan Lake (Fig. 1). Trapping at Swan Lake from 2003 to 2006 resulted in 7035 captures of females; whereas, trapping at Long Lake from 2001 to 2006 yielded 8086 captures of females. The nesting area captures and trap captures resulted in recapture of 40 of the 69 females (58%) that we translocated to Long Lake and in recaptures of 21 of 33 control females (63.6%). The proportion of recaptured females among those translocated and the controls did not differ (χ² = 0.46, df = 1, p = 0.6060). All of the recaptures of translocated females were made at Swan Lake nesting areas (N = 32) or in traps set in Swan Lake (N = 8). No translocated females were recaptured at Long Lake. Recaptured translocated females (N = 40), translocated females that were not recaptured (N = 29), and control females that were not translocated (N = 33) did not vary in any measure of size (Table 1; Kruskal–Wallis tests for each trait, p > 0.05). Size details for these 102 females are in Table 1. During these same years we made a total of 16,294 captures of male turtles (7932 at Long Lake and 9267 at Swan Lake). Males and females caught at Swan Lake were larger than males and females caught at Long Lake (Table 2; Wilcoxon two-sample tests for each trait, p < 0.0001).

Table 1. Descriptive statistics for female red-eared sliders (Trachemys scripta elegans).

View Fullscreen

Table 2. Descriptive statistics for male and female red-eared sliders (Trachemys scripta elegans) caught at Long Lake and Swan Lake.

View Fullscreen

Based on our trapping results, males were less likely than females to make cross-river movements between Gilbert Lake and Swan Lake. Only 31 males (out of 16,294 captures of males) vs. 55 females (out of 18,025 captures of females) were observed crossing the Illinois River between Gilbert Lake and Swan Lake. The female bias was statistically significant (χ² = 6.70, 1 df, p = 0.0097) indicating that more females than males made this trip. In contrast, 22 turtles (9 males and 13 females) that were initially marked in Swan Lake between 2000 and 2006 were later recaptured in Long Lake. No turtles that were initially marked in Long Lake were subsequently recaptured in Swan Lake.

Discussion

Homing ability is clearly well developed in adult female red-eared sliders (Trachemys scripta elegans). Not only did they return to their original capture sites, but statistically, essentially all of them did. The lack of difference in recapture rates between translocated females and the control group strongly suggests that nearly all translocated turtles returned to Swan Lake. We note that in previous studies of homing, no reference group or control was used. Consequently, without a control group, it is not possible to estimate the proportion of translocated turtles that homed. Instead, only the proportion of those recaptured can be determined. No inferences can be made about those turtles not recaptured without a control group. Ours is the first homing study of turtles to use a control group.

This high rate of return is likely not due to turtles including both lakes in their home ranges. During the course of the study, we found only 22 turtles that had moved from Swan Lake to Long Lake out of 19,018 captures that we made. Turtles originally marked in Long Lake were never recaptured in Swan Lake. Management practices could explain this one-way movement. In Long Lake, water is not drawn down, but portions of Swan Lake are dewatered each year. We frequently see turtles crossing from Swan Lake to the Illinois River during such events. Apparently most of these turtles move to the nearest available backwater (Gilbert Lake, Fig. 1). Thus it seems unlikely that our Swan Lake females would have been in Long Lake before. It should also be noted that 32 of these females not only returned to Swan Lake but they also returned to the same nesting area from which they had been removed. Nest site fidelity has been previously demonstrated for slider females from Long Lake (Tucker 2001).

These same management practices may also cause the size differences between turtles from Swan Lake and Long Lake. The basis for the difference is that young turtles (2–4 years old) are much less common among turtles trapped in Swan Lake; whereas, these young turtles make up a large portion of the turtles trapped in Long Lake. Underlying differences in recruitment rates may explain the apparent difference in age structure, and we continue to study recruitment at both lakes. Habitat differences between the 2 lakes are many. Swan Lake does not have submersed aquatic vegetation, and a levee borders one side of the lake. The levee has little cover for juvenile turtles, and the lake has large numbers of predatory birds that line the banks and moist-soil units when portions of the lake are drawn down. In contrast, Long Lake has extensive cover in the form of emergent vegetation and shrubs that extend into the lake. Moreover, submersed aquatic vegetation is present in parts of the lake. Finally, Long Lake is not dewatered on a regular basis.

Similar to map turtles (Freedberg et al. 2005), female sliders have well developed homing ability. Nest site fidelity is also well developed in slider females (Tucker 2001). Well developed homing ability and nest site fidelity were used as evidence of natal homing by Freedberg et al. (2005). It could be possible that that natal homing is part of the life history strategy of the slider because sliders and map turtles have similar patterns of homing and nest site fidelity.

Our experiment concentrated only on adult females, and we did not attempt to translocate juveniles or males. Males and young turtles may or may not be as adept at homing as adult females. Smar and Chambers (2005) found that male stinkpots (Sternotherus odoratus) were more likely than females to home after translocation. Adult sliders are mobile, and the Illinois River is not a barrier as we recaptured turtles that moved from one side of the river to the other. Interestingly, we found females more likely to make this journey than males.

Generally, male sliders are known to be more mobile than females possibly because increased mobility may enhance mating opportunities (Morreale et al. 1984; Parker 1984). Moreover, male sliders have larger home ranges than do female sliders (Schubauer et al. 1990). Our observations do not necessarily contradict these findings. We did not determine home ranges and males may have larger home ranges in our study area as well. Nonetheless, one other study of sliders found that males make fewer long-distance moves (Bodie and Semlitsch 2000).

Moreover, the cross-river movements we observed are extrapopulational (Gibbons et al. 1990). We suspect that females move in response to low water levels. The water level in Gilbert Lake is often lowered during summer months to encourage growth of moist-soil vegetation. Females seem to respond to draw downs by seeking deeper water sites and may be more likely to leave than males. Females return each year during the nesting season due to their nest site fidelity (Tucker 2001) making them more likely to be captured on both sides of the river (Tucker 2001).

The frequency of cross-river movements in sliders should be kept in perspective. They represent only 86 recaptures (0.8%) out of a total of 10,373 recaptures made from 19,018 trapped turtles marked and released at our study areas between 1994 and 2006. Sliders seem not to be nearly as mobile as riverine turtles such as map turtles (Freedberg et al. 2005). These map turtles identified as Graptemys kohnii returned from as far away as 6 km within 1–2 days of translocation. The female sliders that we studied also were capable of rapid movements. One individual was found making a false nesting attempt (see Tucker et al. 1995) at the Swan Lake nesting area within 48 hours of its release some 3 km away. This recapture was not counted in our totals, but it demonstrated that females could rapidly return from Long Lake.

The ease with which sliders can cross the Illinois River stands in contrast to the barrier presented by other rivers to turtle dispersal. Bodie and Semlitsch (2000) found that turtles, including sliders, were apparently unable to cross the Missouri River. Turtles, however, regardless of species inhabiting backwater lakes or riverside lentic habitats, cannot be assumed to be limited in their mobility by lotic habitats. Only comprehensive trapping surveys on both sides of rivers can determine their impact on movements of aquatic turtles.